Faculty Profiles - MORI, Kensaku

Artificial intelligence-based diagnostic imaging system with virtual enteroscopy and virtual unfolded views to evaluate small bowel lesions in Crohn's disease. Invited Reviewed

Furukawa K, Oda M, Watanabe O, Nakamura M, Yamamura T, Maeda K, Mori K, Kawashima H

Revista espanola de enfermedades digestivas 2024.3

Development of real-time navigation system for laparoscopic hepatectomy using magnetic micro sensor. Invited Reviewed

Igami T, Hayashi Y, Yokyama Y, Mori K, Ebata T

Minimally invasive therapy & allied technologies : MITAT : official journal of the Society for Minimally Invasive Therapy page： 1 - 11 2024.1

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Minimally Invasive Therapy and Allied Technologies

Background: We report a new real-time navigation system for laparoscopic hepatectomy (LH), which resembles a car navigation system. Material and methods: Virtual three-dimensional liver and body images were reconstructed using the “New-VES” system, which worked as roadmap during surgery. Several points of the patient’s body were registered in virtual images using a magnetic position sensor (MPS). A magnetic transmitter, corresponding to an artificial satellite, was placed about 40 cm above the patient’s body. Another MPS, corresponding to a GPS antenna, was fixed on the handling part of the laparoscope. Fiducial registration error (FRE, an error between real and virtual lengths) was utilized to evaluate the accuracy of this system. Results: Twenty-one patients underwent LH with this system. Mean FRE of the initial five patients was 17.7 mm. Mean FRE of eight patients in whom MDCT was taken using radiological markers for registration of body parts as first improvement, was reduced to 10.2 mm (p =.014). As second improvement, a new MPS as an intraoperative body position sensor was fixed on the right-sided chest wall for automatic correction of postural gap. The preoperative and postoperative mean FREs of 8 patients with both improvements were 11.1 mm and 10.1 mm (p =.250). Conclusions: Our system may provide a promising option that virtually guides LH.

DOI： 10.1080/13645706.2023.2301594

Deep learning model for extensive smartphone-based diagnosis and triage of cataracts and multiple corneal diseases. Invited Reviewed

Ueno Y, Oda M, Yamaguchi T, Fukuoka H, Nejima R, Kitaguchi Y, Miyake M, Akiyama M, Miyata K, Kashiwagi K, Maeda N, Shimazaki J, Noma H, Mori K, Oshika T

The British journal of ophthalmology 2024.1

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：British Journal of Ophthalmology

Aim To develop an artificial intelligence (AI) algorithm that diagnoses cataracts/corneal diseases from multiple conditions using smartphone images. Methods This study included 6442 images that were captured using a slit-lamp microscope (6106 images) and smartphone (336 images). An AI algorithm was developed based on slit-lamp images to differentiate 36 major diseases (cataracts and corneal diseases) into 9 categories. To validate the AI model, smartphone images were used for the testing dataset. We evaluated AI performance that included sensitivity, specificity and receiver operating characteristic (ROC) curve for the diagnosis and triage of the diseases. Results The AI algorithm achieved an area under the ROC curve of 0.998 (95% CI, 0.992 to 0.999) for normal eyes, 0.986 (95% CI, 0.978 to 0.997) for infectious keratitis, 0.960 (95% CI, 0.925 to 0.994) for immunological keratitis, 0.987 (95% CI, 0.978 to 0.996) for cornea scars, 0.997 (95% CI, 0.992 to 1.000) for ocular surface tumours, 0.993 (95% CI, 0.984 to 1.000) for corneal deposits, 1.000 (95% CI, 1.000 to 1.000) for acute angle-closure glaucoma, 0.992 (95% CI, 0.985 to 0.999) for cataracts and 0.993 (95% CI, 0.985 to 1.000) for bullous keratopathy. The triage of referral suggestion using the smartphone images exhibited high performance, in which the sensitivity and specificity were 1.00 (95% CI, 0.478 to 1.00) and 1.00 (95% CI, 0.976 to 1.000) for’urgent’, 0.867 (95% CI, 0.683 to 0.962) and 1.00 (95% CI, 0.971 to 1.000) for’semi-urgent’, 0.853 (95% CI, 0.689 to 0.950) and 0.983 (95% CI, 0.942 to 0.998) for’routine’ and 1.00 (95% CI, 0.958 to 1.00) and 0.896 (95% CI, 0.797 to 0.957) for’observation’, respectively. Conclusions The AI system achieved promising performance in the diagnosis of cataracts and corneal diseases.

DOI： 10.1136/bjo-2023-324488

Endoscope Automation Framework with Hierarchical Control and Interactive Perception for Multi-Tool Tracking in Minimally Invasive Surgery. Invited Reviewed

Fozilov K, Colan J, Davila A, Misawa K, Qiu J, Hayashi Y, Mori K, Hasegawa Y

Sensors (Basel, Switzerland) Vol. 23 ( 24 ) 2023.12

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Sensors

In the context of Minimally Invasive Surgery, surgeons mainly rely on visual feedback during medical operations. In common procedures such as tissue resection, the automation of endoscopic control is crucial yet challenging, particularly due to the interactive dynamics of multi-agent operations and the necessity for real-time adaptation. This paper introduces a novel framework that unites a Hierarchical Quadratic Programming controller with an advanced interactive perception module. This integration addresses the need for adaptive visual field control and robust tool tracking in the operating scene, ensuring that surgeons and assistants have optimal viewpoint throughout the surgical task. The proposed framework handles multiple objectives within predefined thresholds, ensuring efficient tracking even amidst changes in operating backgrounds, varying lighting conditions, and partial occlusions. Empirical validations in scenarios involving single, double, and quadruple tool tracking during tissue resection tasks have underscored the system’s robustness and adaptability. The positive feedback from user studies, coupled with the low cognitive and physical strain reported by surgeons and assistants, highlight the system’s potential for real-world application.

DOI： 10.3390/s23249865

Artificial intelligence for evaluating the risk of gastric cancer: reliable detection and scoring of intestinal metaplasia with deep learning algorithms. Invited Reviewed

Iwaya M, Hayashi Y, Sakai Y, Yoshizawa A, Iwaya Y, Uehara T, Kitagawa M, Fukayama M, Mori K, Ota H

Gastrointestinal endoscopy Vol. 98 ( 6 ) page： 925 - 933.e1 2023.12

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Gastrointestinal Endoscopy

Background and Aims: Gastric cancer (GC) is associated with chronic gastritis. To evaluate the risk, the Operative Link on Gastric Intestinal Metaplasia Assessment (OLGIM) system was constructed and showed a higher GC risk in stage III or IV patients, determined by the degree of intestinal metaplasia (IM). Although the OLGIM system is useful, evaluating the degree of IM requires substantial experience to produce precise scoring. Whole-slide imaging is becoming routine, but most artificial intelligence (AI) systems in pathology are focused on neoplastic lesions. Methods: Hematoxylin and eosin–stained slides were scanned. Images were divided into each gastric biopsy tissue sample and labeled with an IM score. IM was scored as follows: 0 (no IM), 1 (mild IM), 2 (moderate IM), and 3 (severe IM). Overall, 5753 images were prepared. A deep convolutional neural network (DCNN) model, ResNet50, was used for classification. Results: ResNet50 classified images with and without IM with a sensitivity of 97.7% and specificity of 94.6%. IM scores 2 and 3, involved as criteria of stage III or IV in the OLGIM system, were classified by ResNet50 in 18%. The respective sensitivity and specificity values of classifying IM between scores 0 and 1 and 2 and 3 were 98.5% and 94.9%, respectively. The IM scores classified by pathologists and the AI system were different in only 438 images (7.6%), and we found that ResNet50 tended to miss small foci of IM but successfully identified minimal IM areas that pathologists missed during the review. Conclusions: Our findings suggested that this AI system would contribute to evaluating the risk of GC accuracy, reliability, and repeatability with worldwide standardization.

DOI： 10.1016/j.gie.2023.06.056

Deep learning-based prediction model for postoperative complications of cervical posterior longitudinal ligament ossification Invited Reviewed

Sadayuki Ito, Hiroaki Nakashima, Toshitaka Yoshii, Satoru Egawa, Kenichiro Sakai, Kazuo Kusano, Shinji Tsutui, Takashi Hirai, Yu Matsukura, Kanichiro Wada, Keiichi Katsumi, Masao Koda, Atsushi Kimura, Takeo Furuya, Satoshi Maki, Narihito Nagoshi, Norihiro Nishida, Yukitaka Nagamoto, Yasushi Oshima, Kei Ando, Masahiko Takahata, Kanji Mori, Hideaki Nakajima, Kazuma Murata, Masayuki Miyagi, Takashi Kaito, Kei Yamada, Tomohiro Banno, Satoshi Kato, Tetsuro Ohba, Satoshi Inami, Shunsuke Fujibayashi, Hiroyuki Katoh, Haruo Kanno, Masahiro Oda, Kensaku Mori, Hiroshi Taneichi, Yoshiharu Kawaguchi, Katsushi Takeshita, Morio Matsumoto, Masashi Yamazaki, Atsushi Okawa, Shiro Imagama

Europian Spine Journal Vol. 32 ( 11 ) page： 3797 - 3806 2023.11

Social Relation Atmosphere Recognition with Relevant Visual Concepts Invited

JI Ying, WANG Yu, MORI Kensaku, KATO Jien

IEICE Transactions on Information and Systems Vol. E106.D ( 10 ) page： 1638 - 1649 2023.10

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：The Institute of Electronics, Information and Communication Engineers

Social relationships (e.g., couples, opponents) are the foundational part of society. Social relation atmosphere describes the overall interaction environment between social relationships. Discovering social relation atmosphere can help machines better comprehend human behaviors and improve the performance of social intelligent applications. Most existing research mainly focuses on investigating social relationships, while ignoring the social relation atmosphere. Due to the complexity of the expressions in video data and the uncertainty of the social relation atmosphere, it is even difficult to define and evaluate. In this paper, we innovatively analyze the social relation atmosphere in video data. We introduce a Relevant Visual Concept (RVC) from the social relationship recognition task to facilitate social relation atmosphere recognition, because social relationships contain useful information about human interactions and surrounding environments, which are crucial clues for social relation atmosphere recognition. Our approach consists of two main steps: (1) we first generate a group of visual concepts that preserve the inherent social relationship information by utilizing a 3D explanation module; (2) the extracted relevant visual concepts are used to supplement the social relation atmosphere recognition. In addition, we present a new dataset based on the existing Video Social Relation Dataset. Each video is annotated with four kinds of social relation atmosphere attributes and one social relationship. We evaluate the proposed method on our dataset. Experiments with various 3D ConvNets and fusion methods demonstrate that the proposed method can effectively improve recognition accuracy compared to end-to-end ConvNets. The visualization results also indicate that essential information in social relationships can be discovered and used to enhance social relation atmosphere recognition.

DOI： 10.1587/transinf.2023PCP0008

Automated Detection and Diagnosis of Spinal Schwannomas and Meningiomas Using Deep Learning and Magnetic Resonance Imaging. Invited Reviewed

Ito S, Nakashima H, Segi N, Ouchida J, Oda M, Yamauchi I, Oishi R, Miyairi Y, Mori K, Imagama S

Journal of clinical medicine Vol. 12 ( 15 ) 2023.8

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Journal of Clinical Medicine

Spinal cord tumors are infrequently identified spinal diseases that are often difficult to diagnose even with magnetic resonance imaging (MRI) findings. To minimize the probability of overlooking these tumors and improve diagnostic accuracy, an automatic diagnostic system is needed. We aimed to develop an automated system for detecting and diagnosing spinal schwannomas and meningiomas based on deep learning using You Only Look Once (YOLO) version 4 and MRI. In this retrospective diagnostic accuracy study, the data of 50 patients with spinal schwannomas, 45 patients with meningiomas, and 100 control cases were reviewed, respectively. Sagittal T1-weighted (T1W) and T2-weighted (T2W) images were used for object detection, classification, training, and validation. The object detection and diagnosis system was developed using YOLO version 4. The accuracies of the proposed object detections based on T1W, T2W, and T1W + T2W images were 84.8%, 90.3%, and 93.8%, respectively. The accuracies of the object detection for two spine surgeons were 88.9% and 90.1%, respectively. The accuracies of the proposed diagnoses based on T1W, T2W, and T1W + T2W images were 76.4%, 83.3%, and 84.1%, respectively. The accuracies of the diagnosis for two spine surgeons were 77.4% and 76.1%, respectively. We demonstrated an accurate, automated detection and diagnosis of spinal schwannomas and meningiomas using the developed deep learning-based method based on MRI. This system could be valuable in supporting radiological diagnosis of spinal schwannomas and meningioma, with a potential of reducing the radiologist’s overall workload.

DOI： 10.3390/jcm12155075

Activities of National Institute of Informatics in Japan Invited Reviewed

Kitsuregawa M., Urushidani S., Yamaji K., Takakura H., Hasuo I., Sato I., Ishikawa F., Echizen I., Mori K.

Communications of the ACM Vol. 66 ( 7 ) page： 58 - 63 2023.6

Speedometer for withdrawal time monitoring during colonoscopy: a clinical implementation trial. Invited Reviewed

Barua I, Misawa M, Glissen Brown JR, Walradt T, Kudo SE, Sheth SG, Nee J, Iturrino J, Mukherjee R, Cheney CP, Sawhney MS, Pleskow DK, Mori K, Løberg M, Kalager M, Wieszczy P, Bretthauer M, Berzin TM, Mori Y

Scandinavian journal of gastroenterology Vol. 58 ( 6 ) page： 664 - 670 2023.6

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Scandinavian Journal of Gastroenterology

Objectives: Meticulous inspection of the mucosa during colonoscopy, represents a lengthier withdrawal time, but has been shown to increase adenoma detection rate (ADR). We investigated if artificial intelligence-aided speed monitoring can improve suboptimal withdrawal time. Methods: We evaluated the implementation of a computer-aided speed monitoring device during colonoscopy at a large academic endoscopy center. After informed consent, patients ≥18 years undergoing colonoscopy between 5 March and 29 April 2021 were examined without the use of the speedometer, and with the speedometer between 29 April and 30 June 2021. All colonoscopies were recorded, and withdrawal time was assessed based on the recordings in a blinded fashion. We compared mean withdrawal time, percentage of withdrawal time ≥6 min, and ADR with and without the speedometer. Results: One hundred sixty-six patients in each group were eligible for analyses. Mean withdrawal time was 9 min and 6.6 s (95% CI: 8 min and 34.8 s to 9 min and 39 s) without the use of the speedometer, and 9 min and 9 s (95% CI: 8 min and 45 s to 9 min and 33.6 s) with the speedometer; difference 2.3 s (95% CI: −42.3–37.7, p = 0.91). The ADRs were 45.2% (95% CI: 37.6–52.8) without the speedometer as compared to 45.8% (95% CI: 38.2–53.4) with the speedometer (p = 0.91). The proportion of colonoscopies with withdrawal time ≥6 min without the speedometer was 85.5% (95% CI: 80.2–90.9) versus 86.7% (95% CI: 81.6–91.9) with the speedometer (p = 0.75). Conclusions: Use of speed monitoring during withdrawal did not increase withdrawal time or ADR in colonoscopy. ClinicalTrials.gov Identifier: NCT04710251.

DOI： 10.1080/00365521.2022.2154616

Correction to: Gaussian affinity and GIoU-based loss for perforation detection and localization from colonoscopy videos. Invited Reviewed

Jiang K, Itoh H, Oda M, Okumura T, Mori Y, Misawa M, Hayashi T, Kudo SE, Mori K

International journal of computer assisted radiology and surgery Vol. 18 ( 5 ) page： 807 2023.5

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：International Journal of Computer Assisted Radiology and Surgery

The original version of this article unfortunately contained a mistake. The incorrect notations were given in the author’s affiliations.

DOI： 10.1007/s11548-023-02899-x

Gaussian affinity and GIoU-based loss for perforation detection and localization from colonoscopy videos. Invited Reviewed

Jiang K, Itoh H, Oda M, Okumura T, Mori Y, Misawa M, Hayashi T, Kudo SE, Mori K

International journal of computer assisted radiology and surgery Vol. 18 ( 5 ) page： 795 - 805 2023.5

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：International Journal of Computer Assisted Radiology and Surgery

Purpose: Endoscopic submucosal dissection (ESD) is a minimally invasive treatment for early gastric cancer. However, perforations may happen and cause peritonitis during ESD. Thus, there is a potential demand for a computer-aided diagnosis system to support physicians in ESD. This paper presents a method to detect and localize perforations from colonoscopy videos to avoid perforation ignoring or enlarging by ESD physicians. Method: We proposed a training method for YOLOv3 by using GIoU and Gaussian affinity losses for perforation detection and localization in colonoscopic images. In this method, the object functional contains the generalized intersection over Union loss and Gaussian affinity loss. We propose a training method for the architecture of YOLOv3 with the presented loss functional to detect and localize perforations precisely. Results: To qualitatively and quantitatively evaluate the presented method, we created a dataset from 49 ESD videos. The results of the presented method on our dataset revealed a state-of-the-art performance of perforation detection and localization, which achieved 0.881 accuracy, 0.869 AUC, and 0.879 mean average precision. Furthermore, the presented method is able to detect a newly appeared perforation in 0.1 s. Conclusions: The experimental results demonstrated that YOLOv3 trained by the presented loss functional were very effective in perforation detection and localization. The presented method can quickly and precisely remind physicians of perforation happening in ESD. We believe a future CAD system can be constructed for clinical applications with the proposed method.

DOI： 10.1007/s11548-022-02821-x

Class-wise confidence-aware active learning for laparoscopic images segmentation Invited Reviewed

Jie Qiu, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 18 ( 3 ) page： 473 - 482 2023.3

A skeleton context-aware 3D fully convolutional network for abdominal artery segmentation Invited Reviewed

Ruiyun Zhu, Masahiro Oda, Yuichiro Hayashi, Takayuki Kitasaka, Kazunari Misawa, Michitaka Fujiwara & Kensaku Mori Vol. 18 ( 3 ) page： 461 - 472 2023.3

Deep Learning-Based Seminal Vesicle and Vas Deferens Recognition in the Posterior Approach of Robot-Assisted Radical Prostatectomy. Invited Reviewed

Takeshita N, Sakamoto S, Kitaguchi D, Takeshita N, Yajima S, Koike T, Ishikawa Y, Matsuzaki H, Mori K, Masuda H, Ichikawa T, Ito M

Urology Vol. 173 page： 98 - 103 2023.3

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Urology

Objective: To develop a convolutional neural network to recognize the seminal vesicle and vas deferens (SV-VD) in the posterior approach of robot-assisted radical prostatectomy (RARP) and assess the performance of the convolutional neural network model under clinically relevant conditions. Methods: Intraoperative videos of robot-assisted radical prostatectomy performed by the posterior approach from 3 institutions were obtained between 2019 and 2020. Using SV-VD dissection videos, semantic segmentation of the seminal vesicle-vas deferens area was performed using a convolutional neural network-based approach. The dataset was split into training and test data in a 10:3 ratio. The average time required by 6 novice urologists to correctly recognize the SV-VD was compared using intraoperative videos with and without segmentation masks generated by the convolutional neural network model, which was evaluated with the test data using the Dice similarity coefficient. Training and test datasets were compared using the Mann–Whitney U-test and chi-square test. Time required to recognize the SV-VD was evaluated using the Mann–Whitney U-test. Results: From 26 patient videos, 1 040 images were created (520 SV-VD annotated images and 520 SV-VD non-displayed images). The convolutional neural network model had a Dice similarity coefficient value of 0.73 in the test data. Compared with original videos, videos with the generated segmentation mask promoted significantly faster seminal vesicle and vas deferens recognition (P < .001). Conclusion: The convolutional neural network model provides accurate recognition of the SV-VD in the posterior approach RARP, which may be helpful, especially for novice urologists.

DOI： 10.1016/j.urology.2022.12.006

Database-driven patient-specific registration error compensation method for image-guided laparoscopic surgery Invited Reviewed

Yuichiro Hayashi, Kazunari Misawa, Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 18 ( 1 ) page： 63 - 69 2023.1

Computer-Aided Size Estimation of Colorectal Polyps Invited Reviewed

Hotta K., Itoh H., Mori Y., Misawa M., Mori K., Kudo S.e.

Techniques and Innovations in Gastrointestinal Endoscopy Vol. 25 ( 2 ) page： 186 - 188 2023.1

Surgical area recognition from laparoscopic images in laparoscopic gastrectomy for gastric cancer using label smoothing and uncertainty Invited Reviewed

Yuichiro Hayashi,Kazunari Misawa, Kensaku Mori

Proc. SPIE 12466 Vol. 12466 2023

A semantic segmentation method for laparoscopic images using semantically similar groups Invited Reviewed

Uramoto L., Hayashi Y., Oda M., Kitasaka T., Misawa K., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12466 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

In this paper, we present a segmentation method for laparoscopic images using semantically similar groups for multi-class semantic segmentation. Accurate semantic segmentation is a key problem for computer assisted surgeries. Common segmentation models do not explicitly learn similarities between classes. We propose a model that, in addition to learning to segment an image into classes, also learns to segment it into human-defined semantically similar groups. We modify the LinkNet34 architecture by adding a second decoder with an auxiliary task of segmenting the image into these groups. The feature maps of the second decoder are merged into the final decoder. We validate our method against our base model LinkNet34 and a larger LinkNet50. We find that our proposed modification increased the performance both with mean Dice (average +1.5%) and mean Intersection over Union metrics (average +2.8%) on two laparoscopic datasets.

DOI： 10.1117/12.2654636

膀胱鏡画像におけるtiny-YOLOを用いた腫瘍検出 Invited Reviewed

牟田口淳, 小田昌宏, 猪口淳一, 森健策, 江藤正俊

生体医工学 Vol. Annual61 ( Abstract ) page： 255_2 - 255_2 2023

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：公益社団法人日本生体医工学会

【背景】膀胱癌は経尿道手術後に再発が多い腫瘍であり、膀胱鏡での腫瘍の見落としが原因とされている。内視鏡での観察は、従来の白色光(WLI)の他に、NBIを使用するが、いずれの腫瘍検出精度は検者の技量・経験に依存するため、検査の再現性・客観性が少ないことが課題である。近年、人工知能(AI)が多くの医療分野で活用されており、AIによる検査は、客観性・再現性を持った上で、エキスパートレベルと同程度の診断能を持つ可能性があるとされている。今回、WLI/NBI膀胱鏡画像を用いて、AIによる腫瘍検出の精度を検証した。【方法】2019年から2021年まで、経尿道的膀胱腫瘍切除術(TURBT)の際に、WLI/NBIを用いて観察を行った症例の手術動画から膀胱鏡画像を作成し、腫瘍を含む画像を腫瘍画像、腫瘍を含まない画像を正常画像と定義した。腫瘍画像内の膀胱腫瘍を矩形でアノテーションを行い、テストデータ用の画像を用いてAIによる感度、特異度、陽性的中率を評価した。AIでの物体検出はtiny-YOLOを用い、腫瘍検出精度の検証を行った。【結果】WLIとNBIから、それぞれ腫瘍画像をそれぞれtiny-YOLOで学習を行い、腫瘍画像(WLI: 525枚、NBI:219枚)と正常画像(WLI:98枚、NBI:108枚)で精度検証を行った。AIによる物体検出の感度/特異度/陽性的中率は、WLIで87.8%/88.8%/97.7%、NBIで82.2%/81.4%/90.0%であった。【結論】膀胱鏡画像において、AIにより比較的良好に腫瘍検出が可能であった。更なる精度改善、リアルタイム検出への課題について、文献的考察を加え報告する。

DOI： 10.11239/jsmbe.annual61.255_2

看護師の職域プレゼンティズム（心身不調）をIoTセンシングで早期検出する Invited Reviewed

山下佳子, 大山慎太郎, 鈴木輝彦, 坂本祐二, 出野義則, 山下暁士, 赤川里美, 藤井晃子, 白鳥義宗, 森健策

生体医工学 Vol. Annual61 ( Abstract ) page： 225_1 - 225_1 2023

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：公益社団法人日本生体医工学会

看護職は心身不調に由来する顕在化していない労働生産性低下状態（プレゼンティズム）が他職種より多く、不規則な勤務形態や強い心身ストレスにより、離職に至る率も高いのが現状である。類似研究は腰痛に注目してウェアラブルデバイスで筋硬度測定によりプレゼンティズム発生を予測するが、性別や運動経験の有無など個人の特性を大きく反映するため予測が難しいと報告がある。そこで本研究では、プレゼンティズムの主な要因で筋骨格系疼痛の原因となる看護行動を、Internet of Things (IoT)センシングで抽出・判別することを目的とする。既設のBLE 屋内測位(AoA法）アンテナを利用した加速度センサデバイスによる看護行動データの収集と、観察者の直接観察による看護行動に対するラベルデータ（看護行動、実施場所、行動の開始・終了時刻を入力）作成を実施した。看護行動は身体プレゼンティズムリスク(腰痛・膝痛・頚肩腕痛)に分類し定義した。1分毎の行動データ（位置情報の平均値、加速度の最大・最小・平均値・標準偏差）、及び時刻を入力データとし、ラベルデータを教師データとして機械学習を行い、行動認識モデルを構築した。看護行動割合は、腰痛43.6%、膝痛3.3%、頚肩腕痛7.4%であった。IoTセンシングで個人ごとのプレゼンティズムに発展する行動をモニタリングすることにより、プレゼンティズムを予防することが可能と示唆された。

DOI： 10.11239/jsmbe.annual61.225_1

医用画像とAI Invited

森健策

カレントテラピー Vol. 41 page： 79 - 79 2023

TriMix: A General Framework for Medical Image Segmentation from Limited Supervision Invited Reviewed

Zheng Z., Hayashi Y., Oda M., Kitasaka T., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 13846 LNCS page： 185 - 202 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

We present a general framework for medical image segmentation from limited supervision, reducing the reliance on fully and densely labeled data. Our method is simple, jointly trains triple diverse models, and adopts a mix augmentation scheme, and thus is called TriMix. TriMix imposes consistency under a more challenging perturbation, i.e., combining data augmentation and model diversity on the tri-training framework. This straightforward strategy enables TriMix to serve as a strong and general learner learning from limited supervision using different kinds of imperfect labels. We conduct extensive experiments to show TriMix’s generic purpose for semi- and weakly-supervised segmentation tasks. Compared to task-specific state-of-the-arts, TriMix achieves competitive performance and sometimes surpasses them by a large margin. The code is available at https://github.com/MoriLabNU/TriMix.

DOI： 10.1007/978-3-031-26351-4_12

Thrombosis region extraction and quantitative analysis in confocal laser scanning microscopic image sequence in in-vivo imaging Invited Reviewed

Wu Y., Oda M., Hayashi Y., Kawamura S., Takebe T., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12468 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

In this paper, we propose a scheme that includes automated extraction of thrombus regions and quantitative analysis of thrombosis in confocal laser scanning microscope (CLSM) blood flow image sequence. Making thrombosis model in animal models play an important role in the development of antithrombotic drugs and ascertaining thrombosis mechanisms. Making thrombosis model in cerebral cortex of mice is usually observed using a CLSM in the fluorescence mode. However, some small changes of thrombus regions are not easily observed in CLSM blood flow image sequences. In addition, it is not easy for researchers to quantitatively analyze the degree of thrombosis. Therefore, we propose a scheme to achieve automatic thrombosis region extraction and quantitative analysis. In which, our thrombosis region extraction method uses analysis of changing pattern of thrombosis regions in CLSM blood flow image sequence. Experimental results showed that our scheme can help biological researchers observe and analyze the changes of thrombosis in animal models and reduced the use of fluorescent thrombus markers.

DOI： 10.1117/12.2654632

Real Bronchoscopic Images-based Bronchial Nomenclature: a Preliminary Study Invited Reviewed

Wang C., Hayashi Y., Oda M., Kitasaka T., Takabatake H., Mori M., Honma H., Natori H., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12466 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This article describes a method for bronchial nomenclature using real bronchoscopic (RB) images and pre-built knowledge base of branches. The bronchus has a complex tree-like structure, which increases the difficulty of bronchoscopy. Therefore, a bronchoscopic navigation system is used to help physicians during examination. Conventional navigation system used preoperative CT images and real bronchoscopic images to obtain the camera pose for navigation, whose accuracy is influenced by organ deformation. We propose a bronchial nomenclature method to estimate branch names for bronchoscopic navigation. This method consists of a bronchus knowledge base construction model, a camera motion estimation module, an anatomical structure tracking module, and a branch name estimation module. The knowledge base construction module is used to find the relationship of each branch. The anatomical tracking module is used to track the bronchial orifice (BO) extracted in each RB frame. The camera motion estimation module is used to estimate the camera motion between two frames. The branch name estimation module uses the pre-built bronchus knowledge base and BO tracking results to find the name of each branch. Experimental results showed that it is possible to estimate branch names using only RB images and the pre-built knowledge base of branches.

DOI： 10.1117/12.2654508

Priority attention network with Bayesian learning for fully automatic segmentation of substantia nigra from neuromelanin MRI Invited Reviewed

Hu T., Itoh H., Oda M., Saiki S., Hattori N., Kamagata K., Aoki S., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12464 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

Neuromelanin magnetic resonance imaging (NM-MRI) has been widely used in the diagnosis of Parkinson’s disease (PD) for its significantly enhanced contrast between the PD-related structure, the substantia nigra (SN) and surrounding tissues. To develop the computer-aided diagnosis (CAD) system of PD and reduce the labor burden of clinicians, precise and automatic segmentation of SN is becoming more and more desired. This paper proposes a novel network combining the priority gating attention and Bayesian learning for improving the accuracy of fully automatic SN segmentation from NM-MRI. Different from the conventional gated attention model, the proposed network uses the prior SN probability map for guiding the attention computation and reducing the potential disruptions introduced by the background. Additionally, to lower the risks of over-fitting and estimate the confidence scores for the segmentation results, Bayesian learning with Monte Carlo dropout is applied in the training and testing phases. The quantitative results showed that the proposed network acquired the averaged Dice score of 79.46% in comparison with the baseline model 77.93%.

DOI： 10.1117/12.2655112

Octree Cube Constraints in PBD Method for High Resolution Surgical Simulation Invited Reviewed

Miyazaki R., Hayashi Y., Oda M., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12464 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper proposes a deformable tissue model that introduces octree lattice vertex layout and cubic constraints to the orthodox PBD (Position Based Dynamics) method. Surgical simulation is expected to provide a safe method for training in surgery, which is especially useful for preoperative education of inexperienced surgeons and/or for the case a prior attempt is required. To build a surgical simulator, it is necessary to develop organ models with deformations and interaction algorithms between surgical instruments and organ models, all of which must be performed in real time. Since existing surgical simulators focus on real-time performance, the resolution of organ models is limited. The proposed method restricts the vertex locations of the PBD method to the vertices of the octree lattice to save computation time while maintaining a high deformation resolution. To obtain appropriate results even for large deformations, three-dimensional constraints are applied to each octree cube as the constraints of the PBD method. In the simulations, we tested the overall deformation by dropping a liver model and the local deformation scene by laparoscopic clipping. As a result, we achieved deformation simulations at 26.5 fps for the model with approximately 2,672 cube elements and 20,659 vertices.

DOI： 10.1117/12.2654092

Multi-view Guidance for Self-supervised Monocular Depth Estimation on Laparoscopic Images via Spatio-Temporal Correspondence Invited Reviewed

Li W., Hayashi Y., Oda M., Kitasaka T., Misawa K., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 14228 LNCS page： 429 - 439 2023

This work proposes an innovative self-supervised approach to monocular depth estimation in laparoscopic scenarios. Previous methods independently predicted depth maps ignoring spatial coherence in local regions and temporal correlation between adjacent images. The proposed approach leverages spatio-temporal coherence to address the challenges of textureless areas and homogeneous colors in such scenes. This approach utilizes a multi-view depth estimation model to guide monocular depth estimation when predicting depth maps. Moreover, the minimum reprojection error is extended to construct a cost volume for the multi-view model using adjacent images. Additionally, a 3D consistency of the point cloud back-projected from predicted depth maps is optimized for the monocular depth estimation model. To benefit from spatial coherence, deformable patch-matching is introduced to the monocular and multi-view models to smooth depth maps in local regions. Finally, a cycled prediction learning for view synthesis and relative poses is designed to exploit the temporal correlation between adjacent images fully. Experimental results show that the proposed method outperforms existing methods in both qualitative and quantitative evaluations. Our code is available at https://github.com/MoriLabNU/MGMDepthL.

DOI： 10.1007/978-3-031-43996-4_41

Masked Frequency Consistency for Domain-Adaptive Semantic Segmentation of Laparoscopic Images Invited Reviewed

Zhao X., Hayashi Y., Oda M., Kitasaka T., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 14220 LNCS page： 663 - 673 2023

Semantic segmentation of laparoscopic images is an important issue for intraoperative guidance in laparoscopic surgery. However, acquiring and annotating laparoscopic datasets is labor-intensive, which limits the research on this topic. In this paper, we tackle the Domain-Adaptive Semantic Segmentation (DASS) task, which aims to train a segmentation network using only computer-generated simulated images and unlabeled real images. To bridge the large domain gap between generated and real images, we propose a Masked Frequency Consistency (MFC) module that encourages the network to learn frequency-related information of the target domain as additional cues for robust recognition. Specifically, MFC randomly masks some high-frequency information of the image to improve the consistency of the network’s predictions for low-frequency images and real images. We conduct extensive experiments on existing DASS frameworks with our MFC module and show performance improvements. Our approach achieves comparable results to fully supervised learning method on the CholecSeg8K dataset without using any manual annotation. The code is available at github.com/MoriLabNU/MFC.

DOI： 10.1007/978-3-031-43907-0_63

Improved method for COVID-19 Classification of Complex-Architecture CNN from Chest CT volumes using Orthogonal Ensemble Networks Invited Reviewed

Toda R., Oda M., Hayashi Y., Otake Y., Hashimoto M., Akashi T., Aoki S., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12465 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper introduces the improved method for the COVID-19 classification based on computed tomography (CT) volumes using a combination of a complex-architecture convolutional neural network (CNN) and orthogonal ensemble networks (OEN). The novel coronavirus disease reported in 2019 (COVID-19) is still spreading worldwide. Early and accurate diagnosis of COVID-19 is required in such a situation, and the CT scan is an essential examination. Various computer-aided diagnosis (CAD) methods have been developed to assist and accelerate doctors' diagnoses. Although one of the effective methods is ensemble learning, existing methods combine some major models which do not specialize in COVID-19. In this study, we attempted to improve the performance of a CNN for the COVID-19 classification based on chest CT volumes. The CNN model specializes in feature extraction from anisotropic chest CT volumes. We adopt the OEN, an ensemble learning method considering inter-model diversity, to boost its feature extraction ability. For the experiment, We used chest CT volumes of 1283 cases acquired in multiple medical institutions in Japan. The classification result on 257 test cases indicated that the combination could improve the classification performance.

DOI： 10.1117/12.2653792

Classification of COVID-19 cases from chest CT volumes using hybrid model of 3D CNN and 3D MLP-Mixer Invited Reviewed

Oda M., Zheng T., Hayashi Y., Otake Y., Hashimoto M., Akashi T., Aoki S., Mori K.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE Vol. 12465 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper proposes an automated classification method of COVID-19 chest CT volumes using improved 3D MLP-Mixer. Novel coronavirus disease 2019 (COVID-19) spreads over the world, causing a large number of infected patients and deaths. Sudden increase in the number of COVID-19 patients causes a manpower shortage in medical institutions. Computer-aided diagnosis (CAD) system provides quick and quantitative diagnosis results. CAD system for COVID-19 enables efficient diagnosis workflow and contributes to reduce such manpower shortage. In image-based diagnosis of viral pneumonia cases including COVID-19, both local and global image features are important because viral pneumonia cause many ground glass opacities and consolidations in large areas in the lung. This paper proposes an automated classification method of chest CT volumes for COVID-19 diagnosis assistance. MLP-Mixer is a recent method of image classification using Vision Transformer-like architecture. It performs classification using both local and global image features. To classify 3D CT volumes, we developed a hybrid classification model that consists of both a 3D convolutional neural network (CNN) and a 3D version of the MLP-Mixer. Classification accuracy of the proposed method was evaluated using a dataset that contains 1205 CT volumes and obtained 79.5% of classification accuracy. The accuracy was higher than that of conventional 3D CNN models consists of 3D CNN layers and simple MLP layers.

DOI： 10.1117/12.2654706

Automated Detection of the Thoracic Ossification of the Posterior Longitudinal Ligament Using Deep Learning and Plain Radiographs. Invited Reviewed

Ito S, Nakashima H, Segi N, Ouchida J, Oda M, Yamauchi I, Oishi R, Miyairi Y, Mori K, Imagama S

BioMed research international Vol. 2023 page： 8495937 2023

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：BioMed Research International

Ossification of the ligaments progresses slowly in the initial stages, and most patients are unaware of the disease until obvious myelopathy symptoms appear. Consequently, treatment and clinical outcomes are not satisfactory. This study is aimed at developing an automated system for the detection of the thoracic ossification of the posterior longitudinal ligament (OPLL) using deep learning and plain radiography. We retrospectively reviewed the data of 146 patients with thoracic OPLL and 150 control cases without thoracic OPLL. Plain lateral thoracic radiographs were used for object detection, training, and validation. Thereafter, an object detection system was developed, and its accuracy was calculated. The performance of the proposed system was compared with that of two spine surgeons. The accuracy of the proposed object detection model based on plain lateral thoracic radiographs was 83.4%, whereas the accuracies of spine surgeons 1 and 2 were 80.4% and 77.4%, respectively. Our findings indicate that our automated system, which uses a deep learning-based method based on plain radiographs, can accurately detect thoracic OPLL. This system has the potential to improve the diagnostic accuracy of thoracic OPLL.

DOI： 10.1155/2023/8495937

AI画像解析による内視鏡外科手術手技のビデオ評価及び手術支援システムの構築 Invited Reviewed

安井昭洋, 内田広夫, 森健策, 石田昇平, 出家亨一, 檜顕成, 城田千代栄, 小田昌宏, 林雄一郎

生体医工学 Vol. Annual61 ( Abstract ) page： 127_2 - 127_2 2023

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：公益社団法人日本生体医工学会

【はじめに】術後成長発達する小児患者にとって、低侵襲手術は非常に重要である。しかし患者数は限られているため、しっかりとした手術を行うためにoff the job-training(OJT)が重要である。さらにOJTでの効率的な手技獲得には、手技を客観的に評価しfeed backを行うシステムが必須である。また安全で効率的な内視鏡手術を行うためには、臓器の位置関係の把握が必要であるため、術中ナビゲーションは重要な要件となる。これらの課題に対して、AIを用いた内視鏡手技評価および手術支援システムの構築に着手しており現状の成果を報告する。【方法と結果】食道閉鎖症モデルを用いた吻合手技を被験者に課し、各被験者の手技を最初に人の目で「check 表」「エラー項目」「時間」を用いて評価した。次にビデオから検出した鉗子の動きと人が判定した手技優劣の関係性をAIで学習させ、上位88％・下位95％の精度で手技優劣が自動判定可能となった。この結果を解析することで今まで必要だった50項目以上の肉眼チェックが、わずか７項目チェックするだけで手技の優劣を判断できることが明らかになった。現在食道閉鎖症の手術画像を用いて、食道・迷走神経・気管を深層学習させ、各種構造物の自動認識を進めている。【まとめ】AI画像解析により内視鏡手技の優劣をビデオで判定可能となった。この結果から新たに効率的な手技判断基準を定めることができた。術中ナビゲーションは現在精度のさらなる向上を目指している。

DOI： 10.11239/jsmbe.annual61.127_2

Anatomy Aware-based 2.5D Bronchoscope Tracking for Image-guided Bronchoscopic Navigation Invited Reviewed

Cheng Wang, Masahiro Oda, Yuichiro Hayashi, Takayuki Kitasaka, Hayato Itto, Hirotoshi honma,Hirotsugu takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2022.12

Surgical Assistance in Laparoscopic Surgery Using AI Invited Reviewed

HAYASHI Yuichiro, MORI Kensaku

Medical Imaging Technology Vol. 40 ( 4 ) page： 164 - 169 2022.9

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society of Medical Imaging Technology

Laparoscopic surgery is currently performed as one of surgical procedures for various organs. Laparoscopic surgery requires highly skill of surgeon. Therefore, computer aided surgery which is used computer technology for assisting surgery has been researched. These assistances are sometimes referred as surgical navigation. In recent year, as the development of AI (Artificial Intelligence) technology, there are many researches on laparoscopic video analysis using deep learning for assisting laparoscopic surgery. In this paper, we introduce our researches about laparoscopic video analysis using AI.

DOI： 10.11409/mit.40.164

AI for VR Invited Reviewed

森健策

日本VR医学会学術大会プログラム・抄録集 Vol. 2022 ( 0 ) page： 9 - 9 2022.8

BONE MARROW EDEMA SCORE IN HAND X-RAY FILM BY AI DEEP LEARNING ASSOCIATE WITH MRI BONE EDEMA IN RHEUMATOID ARTHRITIS. Invited Reviewed

Katayama, K; Pan, D; Oda, M; Okubo, T; Mori, K

ANNALS OF THE RHEUMATIC DISEASES Vol. 81 page： 1773 - 1774 2022.6

30年間の医用画像研究経験を振り返り未来を考える Invited Reviewed

森健策

情報・システムソサイエティ誌 Vol. 27 ( 1 ) page： 16 - 17 2022.5

Depth Estimation from Single-shot Monocular Endoscope Image Using Image Domain Adaptation And Edge-Aware Depth Estimation Reviewed

Masahiro Oda, Hayato Itoh, Kiyohito Tanaka, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization Vol. 10 ( 3 ) page： 266 - 273 2022.5

Uncertainty meets 3D-spatial feature in colonoscopic polyp-size determination Invited Reviewed

Hayato Itoh, Masahiro Oda, Kai Jiang, Yuichi Mori, Masashi Misawa, Shin-Ei Kudo, Kenichiro Imai, Sayo Ito, Kinichi Hotta, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization Vol. 10 ( 3 ) page： 289 - 298 2022.5

Spatially variant biases considered self-supervised depth estimation based on laparoscopic videos Reviewed

Wenda Li, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kazunari Misawa, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization Vol. 10 ( 3 ) page： 274 - 282 2022.5

SR-CycleGAN: super-resolution of clinical CT to micro-CT level with multi-modality super-resolution loss Reviewed

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Takayasu Moriya, Shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

Journal of Medical Imaging Vol. 9 ( 2 ) page： 024003-1 - 28 2022.4

Artificial Intelligence for Computer Vision in Surgery: A Call for Developing Reporting Guidelines. Invited Reviewed

Kitaguchi D, Watanabe Y, Madani A, Hashimoto DA, Meireles OR, Takeshita N, Mori K, Ito M, Computer Vision in Surgery International Collaborative.

Annals of surgery Vol. 275 ( 4 ) page： e609 - e611 2022.4

Evaluation in real-time use of artificial intelligence during colonoscopy to predict relapse of ulcerative colitis: a prospective study Invited Reviewed

Yasuharu Maeda, Shin-ei Kudo, Noriyuki Ogata, Masashi Misawa, Marietta Iacucci, Mayumi Homma, Tetsuo Nemoto, Kazumi Takishima, Kentaro Mochida, Hideyuki Miyachi, Toshiyuki Baba, Kensaku Mori, Kazuo Ohtsuka, Yuichi Mori

Gastrointestinal Endoscopy 2022.4

Artificial Intelligence-Based Total Mesorectal Excision Plane Navigation in Laparoscopic Colorectal Surgery. Invited Reviewed

Igaki T, Kitaguchi D, Kojima S, Hasegawa H, Takeshita N, Mori K, Kinugasa Y, Ito M

Diseases of the colon and rectum 2022.2

A cascaded fully convolutional network framework for dilated pancreatic duct segmentation Invited Reviewed

Chen Shen, Holger R. Roth, Yuichiro Hayashi, Masahiro Oda, Tadaaki Miyamoto, Gen Sato, Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 17 ( 2 ) page： 343 - 354 2022.2

impact of the clinical use of artificial intelligence-assisted neoplasia detection for colonoscopy: a large-scale prospective, propensity score-matched study (with video) Reviewed

Misaki Ishiyama, Shin-ei Kudo, Masashi Misawa, Yuichi Mori, Yasuharu Maeda, Katsuro Ichimasa, Toyoki Kudo, Takemasa Hayashi, Kunihiko Wakamura, Hideyuki Miyachi, Fumio Ishida, Hayato Itoh, Masahiro Oda, Kensaku Mori

Vol. 95 ( 1 ) page： 155 - 163 2022.1

Joint multi organ and tumor segmentation from partial labels using federated learning Invited Reviewed

Chen Shen, Pochuan Wang, Dong Yang, Daguang Xu, Masahiro Oda, Po-Ting Chen, Kao-Lang Liu, Wei-Chin Liao, Chiou-Shann Fuh, Kensaku Mori , Weichung Wang, Holger R. Roth

,LNCS 13573 Vol. 13573 page： 58 - 67 2022

Depth-based branching level estimation for bronchoscopic navigation Invited Reviewed

Cheng Wang, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Hirotsugu Takabatake, Masaki Mori, Hirotoshi Honma, Hiroshi Natori , Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( 10 ) page： 1795 - 1804 2021.10

Artificial intelligence and computer-aided diagnosis for colonoscopy: where do we stand now? Invited Reviewed

Kudo Shin-Ei, Mori Yuichi, Abdel-aal Usama M., Misawa Masashi, Itoh Hayato, Oda Masahiro, Mori Kensaku

TRANSLATIONAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6 page： 64 2021.10

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Translational Gastroenterology and Hepatology

Computer-aided diagnosis (CAD) for colonoscopy with use of artificial intelligence (AI) is catching increased attention of endoscopists. CAD allows automated detection and pathological prediction, namely optical biopsy, of colorectal polyps during real-time endoscopy, which help endoscopists avoid missing and/or misdiagnosing colorectal lesions. With the increased number of publications in this field and emergence of the AI medical device that have already secured regulatory approval, CAD in colonoscopy is now being implemented into clinical practice. On the other side, drawbacks and weak points of CAD in colonoscopy have not been thoroughly discussed. In this review, we provide an overview of CAD for optical biopsy of colorectal lesions with a particular focus on its clinical applications and limitations.

DOI： 10.21037/tgh.2019.12.14

Deep learning system for automatic detection of bladder tumors in cystoscopic images Invited Reviewed

Mutaguchi J., Oda M., Ueda S., Kinoshita F., Naganuma H., Matsumoto T., Lee K., Monji K., Kashiwagi K., Takeuchi A., Shiota M., Inokuchi J., Mori K., Eto M.

EUROPEAN UROLOGY Vol. 79 page： S1022 - S1023 2021.6

Development of Diagnosis Assistant AI for COVID-19 Patients Invited Reviewed

ODA Masahiro, ZHENG Tong, HAYASHI Yuichiro, MORI Kensaku

Medical Imaging Technology Vol. 39 ( 1 ) page： 13 - 19 2021.1

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society of Medical Imaging Technology

We introduce AIs for diagnosis assistance of COVID-19 patients from CT volumes that were developed in Nagoya University. Novel coronavirus disease 2019（COVID-19）spreads over the world causing the large number of infected patients and deaths. Diagnosis assistance by AI is effective for diagnosing the large number of patients that are caused by infective diseases. We developed diagnosis assistant AIs for COVID-19 cases that evaluate the typical-ness of COVID-19 case based on image appearances from a CT volume. We developed three essential methods for the AIs including the lung region segmentation method, the clustering method of lung region, and the COVID-19 typical-ness evaluation method. To develop the AIs, we utilized huge number of medical images stored in the cloud platform of medical bigdata. We confirmed our AI has high performance for diagnosis assistance in the evaluations using CT volumes of real COVID-19 patients.

DOI： 10.11409/mit.39.13

Automated Detection of Spinal Schwannomas Utilizing Deep Learning Based on Object Detection from MRI Invited Reviewed International coauthorship

Sadayuki Ito, Kei Ando, Kazuyoshi Kobayashi, Hiroaki Nakashima Masahiro Oda, Masaaki Machino, Shunsuke Kanbara, Taro Inoue, Hidetoshi Yamaguchi, Hiroyuki Koshimizu, Kensaku Mori, Naoki Ishiguro, Shiro Imagama

Spine Vol. 46 ( 2 ) page： 95 - 100 2021.1

Depth Estimation from Single-shot Monocular Endoscope Image Using Image Domain Adaptation And Edge-Aware Depth Estimation Invited Reviewed

Masahiro Oda, Hayato Itoh, Kiyohito Tanaka, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2021

Context encoder guided self-supervised siamese depth estimation based on stereo laparoscopic images Invited Reviewed

Wenda Li, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kazunari Misawa, Kensaku Mori

Proc. SPIE 11598, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling Vol. 11598 2021

Bronchial orifice segmentation on bronchoscopic video frames based on generative adversarial depth estimation Invited Reviewed

Cheng Wang, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Hirotsugu Takabatake, Masaki Mori, Hirotoshi Honma, Hiroshi Natori, Kensaku Mori

Proc. SPIE 11598, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling Vol. 11598 page： 115980N-1 - 7 2021

COVID-19 Infection Segmentation from Chest CT Images Based on Scale Uncertainty Invited Reviewed

Masahiro Oda, Tong Zheng, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori

LNCS12969 Vol. 12969 LNCS page： 88 - 97 2021

Attention-guided pancreatic duct segmentation from abdominal CT volumes Invited Reviewed

Chen Shen, Holger Roth, Yuichiro Hayashi, Masahiro Oda, Takaaki Miyamoto, Gen Sato, Kensaku Mori

LNCS12969 Vol. 12969 LNCS page： 46 - 55 2021

CURRENT STATUS AND FUTURE PERSPECTIVE ON ARTIFICIAL INTELLIGENCE FOR LOWER ENDOSCOPY Invited Reviewed

MISAWA Masashi, KUDO Shin-ei, MORI Yuichi, MAEDA Yasuharu, OGAWA Yushi, ICHIMASA Katsuro, KUDO Toyoki, WAKAMURA Kunihiko, HAYASHI Takemasa, MIYACHI Hideyuki, BABA Toshiyuki, ISHIDA Fumio, ITOH Hayato, ODA Masahiro, MORI Kensaku

GASTROENTEROLOGICAL ENDOSCOPY Vol. 63 ( 7 ) page： 1402 - 1416 2021

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Japan Gastroenterological Endoscopy Society

The global incidence and mortality rate of colorectal cancer remains high. Colonoscopy is regarded as the gold standard examination for detecting and eradicating neoplastic lesions. However, there are some uncertainties in colonoscopy practice that are related to limitations in human performance. First, approximately one-fourth of colorectal neoplasms are missed on a single colonoscopy. Second, it is still difficult for nonexperts to perform adequately regarding optical biopsy. Third, recording of some quality indicators (e.g. cecal intubation, bowel preparation, and withdrawal speed） which are related to adenoma detection rate, is sometimes incomplete. With recent improvements in machine learning techniques and advances in computer performance, artificial intelligence-assisted computer-aided diagnosis is being increasingly utilized by endoscopists. In particular, the emergence of deep-learning, data-driven machine learning techniques have made the development of computer-aided systems easier than that of conventional machine learning techniques, the former currently being considered the standard artificial intelligence engine of computer-aided diagnosis by colonoscopy. To date, computer-aided detection systems seem to have improved the rate of detection of neoplasms. Additionally, computer-aided characterization systems may have the potential to improve diagnostic accuracy in real-time clinical practice. Furthermore, some artificial intelligence-assisted systems that aim to improve the quality of colonoscopy have been reported. The implementation of computer-aided system clinical practice may provide additional benefits such as helping in educational poorly performing endoscopists and supporting real-time clinical decision-making. In this review, we have focused on computer-aided diagnosis during colonoscopy reported by gastroenterologists and discussed its status, limitations, and future prospects.

DOI： 10.11280/gee.63.1402

Clinical and Genetic Characteristics of 18 Patients from 13 Japanese Families with CRX-associated retinal disorder: Identification of Genotype-phenotype Association Invited Reviewed

Fujinami-Yokokawa Y., Fujinami K., Kuniyoshi K., Hayashi T., Ueno S., Mizota A., Shinoda K., Arno G., Pontikos N., Yang L., Liu X., Sakuramoto H., Katagiri S., Mizobuchi K., Kominami T., Terasaki H., Nakamura N., Kameya S., Yoshitake K., Miyake Y., Kurihara T., Tsubota K., Miyata H., Iwata T., Tsunoda K., Nishimura T., Hayashizaki Y., Kondo M., Shimozawa N., Horiguchi M., Yamamoto S., Kuze M., Naoi N., Machida S., Shimada Y., Nakamura M., Fujikado T., Hotta Y., Takahashi M., Mochizuki K., Murakami A., Kondo H., Ishida S., Nakazawa M., Hatase T., Matsunaga T., Maeda A., Noda K., Tanikawa A., Yamamoto S., Yamamoto H., Araie M., Aihara M., Nakazawa T., Sekiryu T., Kashiwagi K., Kosaki K., Piero C., Fukuchi T., Hayashi A., Hosono K., Mori K., Tanaka K., Furuya K., Suzuki K., Kohata R., Yanagi Y., Minegishi Y., Iejima D., Suga A., Rossmiller B.P., Pan Y., Oshima T., Nakayama M., Teruyama Y., Yamamoto M., Minematsu N., Sanbe H., Mori D., Kijima Y., Mawatari G., Kurata K., Yamada N., Itoh M., Kawaji H., Murakawa Y.

Scientific Reports Vol. 10 ( 1 ) 2020.12

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Scientific Reports

Inherited retinal disorder (IRD) is a leading cause of blindness, and CRX is one of a number of genes reported to harbour autosomal dominant (AD) and recessive (AR) causative variants. Eighteen patients from 13 families with CRX-associated retinal disorder (CRX-RD) were identified from 730 Japanese families with IRD. Ophthalmological examinations and phenotype subgroup classification were performed. The median age of onset/latest examination was 45.0/62.5 years (range, 15–77/25–94). The median visual acuity in the right/left eye was 0.52/0.40 (range, −0.08–2.00/−0.18–1.70) logarithm of the minimum angle of resolution (LogMAR) units. There was one family with macular dystrophy, nine with cone-rod dystrophy (CORD), and three with retinitis pigmentosa. In silico analysis of CRX variants was conducted for genotype subgroup classification based on inheritance and the presence of truncating variants. Eight pathogenic CRX variants were identified, including three novel heterozygous variants (p.R43H, p.P145Lfs*42, and p.P197Afs*22). A trend of a genotype-phenotype association was revealed between the phenotype and genotype subgroups. A considerably high proportion of CRX-RD in ADCORD was determined in the Japanese cohort (39.1%), often showing the mild phenotype (CORD) with late-onset disease (sixth decade). Frequently found heterozygous missense variants located within the homeodomain underlie this mild phenotype. This large cohort study delineates the disease spectrum of CRX-RD in the Japanese population.

DOI： 10.1038/s41598-020-65737-z

Clinical characteristics in patients with ossification of the posterior longitudinal ligament: A prospective multi-institutional cross-sectional study Invited Reviewed

Hirai T., Yoshii T., Ushio S., Mori K., Maki S., Katsumi K., Nagoshi N., Takeuchi K., Furuya T., Watanabe K., Nishida N., Watanabe K., Kaito T., Kato S., Nagashima K., Koda M., Ito K., Imagama S., Matsuoka Y., Wada K., Kimura A., Ohba T., Katoh H., Matsuyama Y., Ozawa H., Haro H., Takeshita K., Watanabe M., Matsumoto M., Nakamura M., Yamazaki M., Okawa A., Kawaguchi Y.

Scientific Reports Vol. 10 ( 1 ) 2020.12

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Scientific Reports

Ossification of the posterior longitudinal ligament (OPLL) can occur throughout the entire spine and can sometimes lead to spinal disorder. Although patients with OPLL sometimes develop physical limitations because of pain, the characteristics of pain and effects on activities of daily living (ADL) have not been precisely evaluated in OPLL patients. Therefore, we conducted a multi-center prospective study to assess whether the symptoms of cervical OPLL are different from those of cervical spondylosis (CS). A total of 263 patients with a diagnosis of cervical OPLL and 50 patients with a diagnosis of CS were enrolled and provided self-reported outcomes, including responses to the Japanese Orthopaedic Association (JOA) Cervical Myelopathy Evaluation Questionnaire (JOACMEQ), JOA Back Pain Evaluation Questionnaire (JOABPEQ), visual analog scale (VAS), and SF-36 scores. The severity of myelopathy was significantly correlated with each domain of the JOACMEQ and JOABPEQ. There was a negative correlation between the VAS score for each domain and the JOA score. There were significantly positive correlations between the JOA score and the Mental Health, Bodily Pain, Physical Functioning, Role Emotional, and Role Physical domains of the SF-36. One-to-one matching resulted in 50 pairs of patients with OPLL and CS. Although there was no significant between-group difference in scores in any of the domains of the JOACMEQ or JOABPEQ, the VAS scores for pain or numbness in the buttocks or limbs were significantly higher in the CS group; however, there was no marked difference in low back pain, chest tightness, or numbness below the chest between the two study groups. The scores for the Role Physical and Body Pain domains of the SF-36 were significantly higher in the OPLL group than in the CS group, and the mean scores for the other domains was similar between the two groups. The results of this study revealed that patients with OPLL were likely to have neck and low back pain and restriction in ADL. No specific type of pain was found in patients with OPLL when compared with those who had CS.

DOI： 10.1038/s41598-020-62278-3

Synthetic laparoscopic video generation for machine learning-based surgical instrument segmentation from real laparoscopic video and virtual surgical instruments Reviewed International coauthorship

Takuya Ozawa,Yuichiro Hayashi,Hirohisa Oda, Masahiro Oda, Takayuki Kitasaka, Nobuyoshi Takeshita, Masaaki Ito, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2020.11

Clinical impact of Endoscopic Surgical Skill Qualification System (ESSQS) by Japan Society for Endoscopic Surgery (JSES) for laparoscopic distal gastrectomy and low anterior resection based on the National Clinical Database (NCD) registry. Invited Reviewed

Akagi T, Endo H, Inomata M, Yamamoto H, Mori T, Kojima K, Kuroyanagi H, Sakai Y, Nakajima K, Shiroshita H, Etoh T, Saida Y, Yamamoto S, Hasegawa H, Ueno H, Kakeji Y, Miyata H, Kitagawa Y, Watanabe M

Annals of gastroenterological surgery Vol. 4 ( 6 ) page： 721 - 734 2020.11

A visual SLAM-based bronchoscope tracking scheme for bronchoscopic navigation Invited Reviewed

Cheng Wang, Masahiro Oda, Yuichiro Hayashi, Benjamin Villard, Takayuki Kitasaka, Hirotsugu Takabatake, Masaki Mori, Hirotoshi Honma, Hiroshi Natori, Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 10 ) page： 1619 - 1630 2020.10

Station number assignment to abdominal lymph node for assisting gastric cancer surgery Reviewed

Yuichiro Hayashi, Kazunari Misawa, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2020.10

A deformable model for navigated laparoscopic gastrectomy based on finite elemental method Invited Reviewed

Chen Tao, Wei Guodong, Xu Lili, Shi Weili, Xu Yikai, Zhu Yongyi, Hayashi Yuichiro, Oda Hirohisa, Oda Masahiro, Hu Yanfeng, Yu Jiang, Jiang Zhengang, Li Guoxin, Mori Kensaku

MINIMALLY INVASIVE THERAPY & ALLIED TECHNOLOGIES Vol. 29 ( 4 ) page： 210 - 216 2020.7

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Minimally Invasive Therapy and Allied Technologies

Background: Accurate registration for surgical navigation of laparoscopic surgery is highly challenging due to vessel deformation. Here, we describe the design of a deformable model with improved matching accuracy by applying the finite element method (FEM). Material and methods: ANSYS software was used to simulate an FEM model of the vessel after pull-up based on laparoscopic gastrectomy requirements. The central line of the FEM model and the central line of the ground truth were drawn and compared. Based on the material and parameters determined from the animal experiment, a perigastric vessel FEM model of a gastric cancer patient was created, and its accuracy in a laparoscopic gastrectomy surgical scene was evaluated. Results: In the animal experiment, the FEM model created with Ogden foam material exhibited better results. The average distance between the two central lines was 6.5mm, and the average distance between their closest points was 3.8 mm. In the laparoscopic gastrectomy surgical scene, the FEM model and the true artery deformation demonstrated good coincidence. Conclusion: In this study, a deformable vessel model based on FEM was constructed using preoperative CT images to improve matching accuracy and to supply a reference for further research on deformation matching to facilitate laparoscopic gastrectomy navigation.

DOI： 10.1080/13645706.2019.1625926

Automated laparoscopic colorectal surgery workflow recognition using artificial intelligence: Experimental research Invited Reviewed

Kitaguchi Daichi, Takeshita Nobuyoshi, Matsuzaki Hiroki, Oda Tatsuya, Watanabe Masahiko, Mori Kensaku, Kobayashi Etsuko, Ito Masaaki

INTERNATIONAL JOURNAL OF SURGERY Vol. 79 page： 88 - 94 2020.7

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：International Journal of Surgery

Background: Identifying laparoscopic surgical videos using artificial intelligence (AI) facilitates the automation of several currently time-consuming manual processes, including video analysis, indexing, and video-based skill assessment. This study aimed to construct a large annotated dataset comprising laparoscopic colorectal surgery (LCRS) videos from multiple institutions and evaluate the accuracy of automatic recognition for surgical phase, action, and tool by combining this dataset with AI. Materials and methods: A total of 300 intraoperative videos were collected from 19 high-volume centers. A series of surgical workflows were classified into 9 phases and 3 actions, and the area of 5 tools were assigned by painting. More than 82 million frames were annotated for a phase and action classification task, and 4000 frames were annotated for a tool segmentation task. Of these frames, 80% were used for the training dataset and 20% for the test dataset. A convolutional neural network (CNN) was used to analyze the videos. Intersection over union (IoU) was used as the evaluation metric for tool recognition. Results: The overall accuracies for the automatic surgical phase and action classification task were 81.0% and 83.2%, respectively. The mean IoU for the automatic tool segmentation task for 5 tools was 51.2%. Conclusions: A large annotated dataset of LCRS videos was constructed, and the phase, action, and tool were recognized with high accuracy using AI. Our dataset has potential uses in medical applications such as automatic video indexing and surgical skill assessments. Open research will assist in improving CNN models by making our dataset available in the field of computer vision.

DOI： 10.1016/j.ijsu.2020.05.015

Artificial intelligence for magnifying endoscopy, endocytoscopy, and confocal laser endomicroscopy of the colorectum Invited Reviewed

Mori, Y; Kudo, SE; Misawa, M; Itoh, H; Oda, M; Mori, K

TECHNIQUES AND INNOVATIONS IN GASTROINTESTINAL ENDOSCOPY Vol. 22 ( 2 ) page： 56 - 60 2020.4

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Techniques and Innovations in Gastrointestinal Endoscopy

Because magnifying endoscopy is considered to be more accurate at predicting the histology of colorectal polyps than nonmagnifying endoscopy, it has been attracting a lot of attention, especially in Japan. However, use of magnifying endoscopy is not yet widespread because of its limited availability and the difficulty in interpreting the acquired images. Application of artificial intelligence (AI) is now changing this situation because it helps less-skilled endoscopists to accurately interpret magnified images. Research in this field initially focused on magnifying endoscopy with narrow-band imaging as the target of AI. Most previously published retrospective studies have reported over 90% sensitivity in differentiation of neoplastic lesions; however, automatically indicating the region of interest (ROI) of the polyps that AI should analyze has been found to be challenging. To address this practical problem, some researchers have started to adopt contact endomicroscopy as a target for AI. Contact endomicroscopy includes endocytoscopy (520-fold magnification, Olympus, Tokyo, Japan) and confocal laser endomicroscopy (1000-fold magnification, Mauna Kea, Paris, France). These forms of contact endomicroscopy provide ultramagnified images that make it unnecessary to manually select the ROI because the entire image acquired by contact endomicroscopy is the ROI of the targeted polyps. This strength of contact endomicroscopy has contributed to early implementation of this technology into clinical practice, which may change the utility of magnifying endoscopy in clinical settings and help increase its use globally in the near future.

DOI： 10.1016/j.tgie.2019.150632

Clinical application of a surgical navigation system based on virtual thoracoscopy for lung cancer patients: real time visualization of area of lung cancer before induction therapy and optimal resection line for obtaining a safe surgical margin during surgery Invited Reviewed

Nakamura Shota, Hayashi Yuichiro, Kawaguchi Koji, Fukui Takayuki, Hakiri Shuhei, Ozeki Naoki, Mori Shunsuke, Goto Masaki, Mori Kensaku, Yokoi Kohei

JOURNAL OF THORACIC DISEASE Vol. 12 ( 3 ) page： 672 - 679 2020.3

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Journal of Thoracic Disease

Background: We have developed a surgical navigation system that presents virtual thoracoscopic images using computed tomography (CT) image data, as if you are observing intra-thoracic cavity in synchronization with the real thoracoscopic view. Using this system, we made it possible to simultaneously visualize the ‘area of lung cancer before induction therapy’ and the ‘optimal resection line for obtaining a safe surgical margin’ as a virtual thoracoscopic view. We applied this navigation system in the clinical setting in operations for lung cancer patients with chest wall invasion after induction chemoradiotherapy. Methods: The proposed surgical navigation system consisted of a three-dimensional (3D) positional tracker and a virtual thoracoscopy system. The 3D positional tracker was used to recognize the positional information of the real thoracoscope. The virtual thoracoscopy system generated virtual thoracoscopic views based on CT image data. Combined with these two technologies, patient-to-image registration was performed in two patients, and the results generated a virtual thoracoscopic view that was synchronized with the real thoracoscopic view. Results: The operations were started with video-assisted thoracic surgery (VATS), and the navigation system was activated at the same time. The virtual thoracoscopic view was synchronized with the real thoracoscopic view, which also simultaneously indicated the ‘area of lung cancer before induction therapy’ and the ‘optimal resection lines for obtaining a safe surgical margin’. We marked the optimal lines using an electric scalpel, and then performed lobectomy and chest wall resection with a sufficient surgical margin using these landmarks. Pathological examinations confirmed that the surgical margin was negative. No complications related to the navigation system were encountered during or after the procedures. Conclusions: Using this proposed navigation system, we could obtain a ‘CT-derived virtual intrathoracic 3D view of the patient’ that was aligned with the thoracoscopic view during surgery. The accurate identification of areas of cancer invasion before induction therapy using this system might be a useful for determining optimal surgical resection lines.

DOI： 10.21037/jtd.2019.12.108

Tensor-cut: A tensor-based graph-cut blood vessel segmentation method and its application to renal artery segmentation Reviewed International coauthorship

Chenglong Wang, Masahiro Oda, Yuichiro Hayashi, Yasushi Yoshino, Tokunori Yamamoto, Alejandro F. Frangi, Kensaku Mori

Medical Image Analysis Vol. 60 page： 101623 2020.2

Automated eye disease classification method from anterior eye image using anatomical structure focused image classification technique Invited Reviewed

Oda Masahiro, Yamaguchi Takefumi, Fukuoka Hideki, Ueno Yuta, Mori Kensaku

MEDICAL IMAGING 2020: COMPUTER-AIDED DIAGNOSIS Vol. 11314 2020

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper presents an automated classification method of infective and non-infective diseases from anterior eye images. Treatments for cases of infective and non-infective diseases are different. Distinguishing them from anterior eye images is important to decide a treatment plan. Ophthalmologists distinguish them empirically. Quantitative classification of them based on computer assistance is necessary. We propose an automated classification method of anterior eye images into cases of infective or non-infective disease. Anterior eye images have large variations of the eye position and brightness of illumination. This makes the classification difficult. If we focus on the cornea, positions of opacified areas in the corneas are different between cases of the infective and non-infective diseases. Therefore, we solve the anterior eye image classification task by using an object detection approach targeting the cornea. This approach can be said as "anatomical structure focused image classification". We use the YOLOv3 object detection method to detect corneas of infective disease and corneas of non-infective disease. The detection result is used to define a classification result of an image. In our experiments using anterior eye images, 88.3% of images were correctly classified by the proposed method.

DOI： 10.1117/12.2549951

[Pharmacological action and clinical effect of tedizolid phosphate (SIVEXTRO® Tablets 200 mg, for iv infusion 200 mg), a novel oxazolidinone-class antibacterial drug]. Invited Reviewed

Mori M, Takase A

Nihon yakurigaku zasshi. Folia pharmacologica Japonica Vol. 155 ( 5 ) page： 332 - 339 2020

3Dプリンティングの最新動向 Invited

森健策

インナービジョン Vol. 35 page： 36 - 37 2020

Automated Pancreas Segmentation Using Multi-institutional Collaborative Deep Learning Invited Reviewed

Wang P., Shen C., Roth H.R., Yang D., Xu D., Oda M., Misawa K., Chen P.T., Liu K.L., Liao W.C., Wang W., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 12444 LNCS page： 192 - 200 2020

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

The performance of deep learning based methods strongly relies on the number of datasets used for training. Many efforts have been made to increase the data in the medical image analysis field. However, unlike photography images, it is hard to generate centralized databases to collect medical images because of numerous technical, legal, and privacy issues. In this work, we study the use of federated learning between two institutions in a real-world setting to collaboratively train a model without sharing the raw data across national boundaries. We quantitatively compare the segmentation models obtained with federated learning and local training alone. Our experimental results show that federated learning models have higher generalizability than standalone training.

DOI： 10.1007/978-3-030-60548-3_19

Abdominal artery segmentation method from CT volumes using fully convolutional neural network Invited Reviewed

Masahiro Oda, Holger R. Roth, Takayuki Kitasaka, Kazunari Misawa, Michitaka Fujiwara, Kensaku Mori,

International Journal of Computer Assisted Radiology and Surgery Vol. 14 ( 12 ) page： 2069 - 2081 2019.12

A view of three dimensional unit structures of alveoli in peripheral lung Invited Reviewed

Natori Hiroshi, Takabatake Hirotsugu, Mori Masaki, Oda Masahiro, Mori Kensaku, Koba Hiroyuki, Takahashi Hiroki

EUROPEAN RESPIRATORY JOURNAL Vol. 54 2019.9

ARTIFICIAL INTELLIGENCE-ASSISTED POLYP DETECTION SYSTEM FOR COLONOSCOPY, BASED ON THE LARGEST AVAILABLE COLLECTION OF CLINICAL VIDEO DATA FOR MACHINE LEARNING Invited Reviewed

Misawa Masashi, Kudo Shinei, Mori Yuichi, Cho Tomonari, Kataoka Shinichi, Maeda Yasuharu, Ogawa Yushi, Takeda Kenichi, Nakamura Hiroki, Ichimasa Katsuro, Toyoshima Naoya, Ogata Noriyuki, Kudo Toyoki, Hisayuki Tomokazu, Hayashi Takemasa, Wakamura Kunihiko, Baba Toshiyuki, Ishida Fumio, Itoh Hayato, Oda Masahiro, Mori Kensaku

GASTROINTESTINAL ENDOSCOPY Vol. 89 ( 6 ) page： AB646 - AB647 2019.6

Development of a new laparoscopic detection system for gastric cancer using near-infrared light-emitting clips with glass phosphor Invited Reviewed

Inada S., Nakanishi H., Oda M., Mori K., Ito A., Hasegawa J., Misawa K., Fuchi S.

Micromachines Vol. 10 ( 2 ) 2019.1

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Micromachines

Laparoscopic surgery is now a standard treatment for gastric cancer. Currently, the location of the gastric cancer is identified during laparoscopic surgery via the preoperative endoscopic injection of charcoal ink around the primary tumor; however, the wide spread of injected charcoal ink can make it difficult to accurately visualize the specific site of the tumor. To precisely identify the locations of gastric tumors, we developed a fluorescent detection system comprising clips with glass phosphor (Yb 3+ , Nd 3+ doped to Bi 2 O 3 -B 2 O 3 -based glasses, size: 2 mm × 1 mm × 3 mm) fixed in the stomach and a laparoscopic fluorescent detection system for clip-derived near-infrared (NIR) light (976 nm). We conducted two ex vivo experiments to evaluate the performance of this fluorescent detection system in an extirpated pig stomach and a freshly resected human stomach and were able to successfully detect NIR fluorescence emitted from the clip in the stomach through the stomach wall by the irradiation of excitation light (λ: 808 nm). These results suggest that the proposed combined NIR light-emitting clip and laparoscopic fluorescent detection system could be very useful in clinical practice for accurately identifying the location of a primary gastric tumor during laparoscopic surgery.

DOI： 10.3390/mi10020081

Automated Hand Eye Calibration in Laparoscope Holding Robot for Robot Assisted Surgery Invited Reviewed

Jiang Shuai, Hayashi Yuichiro, Wang Cheng, Oda Masahiro, Kitasaka Takayuki, Misawa Kazunari, Mori Kensaku

INTERNATIONAL WORKSHOP ON ADVANCED IMAGE TECHNOLOGY (IWAIT) 2019 Vol. 11049 2019

Colonoscope tracking method based on shape estimation network Invited Reviewed

Oda Masahiro, Roth Holger R., Kitasaka Takayuki, Furukawa Kazuhiro, Miyahara Ryoji, Hirooka Yoshiki, Navab Nassir, Mori Kensaku

MEDICAL IMAGING 2019: IMAGE-GUIDED PROCEDURES, ROBOTIC INTERVENTIONS, AND MODELING Vol. 10951 2019

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper presents a colonoscope tracking method utilizing a colon shape estimation method. CT colonography is used as a less-invasive colon diagnosis method. If colonic polyps or early-stage cancers are found, they are removed in a colonoscopic examination. In the colonoscopic examination, understanding where the colonoscope running in the colon is difficult. A colonoscope navigation system is necessary to reduce overlooking of polyps. We propose a colonoscope tracking method for navigation systems. Previous colonoscope tracking methods caused large tracking errors because they do not consider deformations of the colon during colonoscope insertions. We utilize the shape estimation network (SEN), which estimates deformed colon shape during colonoscope insertions. The SEN is a neural network containing long short-term memory (LSTM) layer. To perform colon shape estimation suitable to the real clinical situation, we trained the SEN using data obtained during colonoscope operations of physicians. The proposed tracking method performs mapping of the colonoscope tip position to a position in the colon using estimation results of the SEN. We evaluated the proposed method in a phantom study. We confirmed that tracking errors of the proposed method was enough small to perform navigation in the ascending, transverse, and descending colons.

DOI： 10.1117/12.2512729

Automatic segmentation of eyeball structures from micro-CT images based on sparse annotation Invited Reviewed

Takaaki Sugino, Holger Roth, Masahiro Oda, Seiji Omata, Shinya Sakuma, Fumihito Arai, Kensaku Mori

Proc. SPIE 10578, Medical Imaging 2018 Vol. 10578 page： 105780V-1-105780V-6 2018

Cascade classification of lesions for automated pathological diagnosis Invited Reviewed

Hayato Ito, Yuichi Mori, Masashi Misawa, Masahiro Oda, Shin-ei Kudo, Kensaku Mori

Proc. SPIE 10575, Medical Imaging 2018 Vol. 10575 page： 1057516-1-1057516-6 2018

3Dプリンタの最新動向 Invited

森健策

インナービジョン Vol. 33 page： 35 - 36 2018

BESNet: Boundary-Enhanced Segmentation of Cells in Histopathological Images Invited Reviewed

Oda Hirohisa, Roth Holger R., Chiba Kosuke, Sokolic Jure, Kitasaka Takayuki, Oda Masahiro, Hinoki Akinari, Uchida Hiroo, Schnabel Julia A., Mori Kensaku

MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2018, PT II Vol. 11071 page： 228 - 236 2018

Develop and Validate a Finite Element Method Model for Deformation Matching of Laparoscopic Gastrectomy Navigation Invited Reviewed

Chen Tao, Wei Guodong, Shi Weili, Hayashi Yuichiro, Oda Masahiro, Jiang Zhengang, Li Guoxin, Mori Kensaku

MEDICAL IMAGING 2018: IMAGE-GUIDED PROCEDURES, ROBOTIC INTERVENTIONS, AND MODELING Vol. 10576 2018

Deep Learning and Its Application to Medical Image Segmentation Invited Reviewed

ROTH Holger R., SHEN Chen, ODA Hirohisa, ODA Masahiro, HAYASHI Yuichiro, MISAWA Kazunari, MORI Kensaku

Medical Imaging Technology Vol. 36 ( 2 ) page： 63 - 71 2018

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society of Medical Imaging Technology

One of the most common tasks in medical imaging is semantic segmentation. Achieving this segmentation automatically has been an active area of research, but the task has been proven very challenging due to the large variation of anatomy across different patients. However, recent advances in deep learning have made it possible to significantly improve the performance of image recognition and semantic segmentation methods in the field of computer vision. Due to the data driven approaches of hierarchical feature learning in deep learning frameworks, these advances can be translated to medical images without much difficulty. Several variations of deep convolutional neural networks have been successfully applied to medical images. Especially fully convolutional architectures have been proven efficient for segmentation of 3D medical images. In this article, we describe how to build a 3D fully convolutional network (FCN) that can process 3D images in order to produce automatic semantic segmentations. The model is trained and evaluated on a clinical computed tomography (CT) dataset and shows stateof-the-art performance in multi-organ segmentation.

DOI： 10.11409/mit.36.63

Colon Shape Estimation Method for Colonoscope Tracking Using Recurrent Neural Networks Invited Reviewed

Oda Masahiro, Roth Holger R., Kitasaka Takayuki, Furukawa Kasuhiro, Miyahara Ryoji, Hirooka Yoshiki, Goto Hidemi, Navab Nassir, Mori Kensaku

MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2018, PT IV Vol. 11073 page： 176 - 184 2018

Airway Segmentation from 3D Chest CT Volumes Based on Volume of Interest Using Gradient Vector Flow Invited Reviewed

MENG Qier, KITASAKA Takayuki, ODA Masahiro, UENO Junji, MORI Kensaku

Medical Imaging Technology Vol. 36 ( 3 ) page： 133 - 146 2018

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society of Medical Imaging Technology

In this paper, we propose a new airway segmentation algorithm from 3D chest CT volumes based on the volume of interest (VOI). The algorithm segments each bronchial branch by recognizing the airway regions from the trachea using the VOIs to segment each branch. A VOI is placed to envelop the branch currently being processed. Then a cavity enhancement filter is performed only inside the current VOI so that each branch is extracted. At the same time, we perform a leakage detection scheme to avoid any leakage regions inside the VOI. Next the gradient vector flow magnitude map and a tubular-likeness function are computed in each VOI. This assists the predictions of both the position and direction of the next child VOIs to detect the next child branches to continue the tracking algorithm. Finally, we unify all of the extracted airway regions to form a complete airway tree. We used a dataset that includes 50 standard-dose human chest CT volumes to evaluate our proposed algorithm. The average extraction rate was approximately 78.1% with a significantly decreased false positive rate compared to the previous method.

DOI： 10.11409/mit.36.133

3Dプリンターの基礎と医療応用 Invited Reviewed

森健策

心臓 Vol. 49 ( 11 ) page： 1104 - 1113 2017.11

Automated mediastinal lymph node detection from CT volumes based on intensity targeted radial structure tensor analysis Invited

Hirohisa Oda, Kanwal K. Bhatia, Masahiro Oda, Takayuki Kitasaka, Shingo Iwano, Hirotoshi Homma, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Julia A. Schnabel, Kensaku Mori

Journal of Medical Imaging Vol. 4 ( 04 ) page： 1 2017.11

Automatic segmentation of head anatomical structures from sparsely-Annotated images Invited Reviewed

Sugino T., Roth H.R., Eshghi M., Oda M., Chung M.S., Mori K.

2017 IEEE International Conference on Cyborg and Bionic Systems, CBS 2017 Vol. 2018-January page： 145 - 149 2017.7

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：2017 IEEE International Conference on Cyborg and Bionic Systems, CBS 2017

Bionic humanoid systems, which are elaborate human models with sensors, have been developed as a tool for quantitative evaluation of doctors' psychomotor skills and medical device performances. For creation of the elaborate human models, this study presents automated segmentation of head sectioned images using sparsely-Annotated data based on deep convolutional neural network. We applied the following fully convolutional networks (FCNs) to the sparse-Annotation-based segmentation: a standard FCN and a dilated convolution based FCN. To validate the availability of FCNs for segmentation of head structures from sparse annotation, we performed 8- and 243-label segmentation experiments using different two sets of head sectioned images in the Visible Korean Human project. In the segmentation experiments, only 10% of all images in each data set were used for training data. Both of the FCNs could achieve the mean segmentation accuracy of more than 85% in the 8-label segmentation. In the 243-label segmentation, though the mean segmentation accuracy was about 50%, the results suggested that the FCNs, especially the dilated convolution based FCNs, had potential to achieve accurate segmentation of anatomical structures, except for small-sized and complex-shaped tissues, even from sparse annotation.

DOI： 10.1109/CBS.2017.8266085

Artificial Intelligence for Endocytoscopy Provides Fully Automated Diagnosis of Histological Remission in Ulcerativ E. Coli Tis Invited Reviewed

Yasuharu Maeda, Kudo Shinei, Mori Yuichi, Misawa Masashi, Wakamura Kunihiko, Hayashi Seiko, Ogata Noriyuki, Takeda Kenichi, Kudo Toyoki, Hayashi Takemasa, Katagiri Atsushi, Ishida Fumio, Ohtsuka Kazuo, Oda Masahiro, Mori Kensaku

GASTROINTESTINAL ENDOSCOPY Vol. 85 ( 5 ) page： AB248 - AB248 2017.5

Can Artificial Intelligence Correctly Diagnose Sessile Serrated Adenomas/Polyps? Invited Reviewed

Mori Yuichi, Kudo Shinei, Ogawa Yushi, Misawa Masashi, Takeda Kenichi, Kudo Toyoki, Wakamura Kunihiko, Hayashi Takemasa, Ichimasa Katsuro, Maeda Yasuharu, Toyoshima Naoya, Nakamura Hiroki, Katagiri Atsushi, Baba Toshiyuki, Ishida Fumio, Oda Masahiro, Mori Kensaku, Inoue Haruhiro

GASTROINTESTINAL ENDOSCOPY Vol. 85 ( 5 ) page： AB510 - AB510 2017.5

Diagnostic Ability of Automated Diagnosis System Using Endocytoscopy for Invasive Colorectal Cancer Invited Reviewed

Takeda Kenichi, Kudo Shinei, Mori Yuichi, Kataoka Shinichi, Yasuharu Maeda, Ogawa Yushi, Nakamura Hiroki, Misawa Masashi, Kudo Toyoki, Wakamura Kunihiko, Hayashi Takemasa, Katagiri Atsushi, Baba Toshiyuki, Hidaka Eiji, Ishida Fumio, Inoue Haruhiro, Oda Masahiro, Mori Kensaku

GASTROINTESTINAL ENDOSCOPY Vol. 85 ( 5 ) page： AB408 - AB408 2017.5

Computer-Aided Diagnosis Based on Endocytoscopy With Narrow-Band Imaging Allows Accurate Diagnosis of Diminutive Colorectal Lesions Invited Reviewed

Misawa Masashi, Kudo Shinei, Mori Yuichi, Takeda Kenichi, Kataoka Shinichi, Nakamura Hiroki, Maeda Yasuharu, Ogawa Yushi, Yamauchi Akihiro, Igarashi Kenta, Hayashi Takemasa, Kudo Toyoki, Wakamura Kunihiko, Katagiri Atsushi, Baba Toshiyuki, Ishida Fumio, Oda Masahiro, Mori Kensaku

GASTROINTESTINAL ENDOSCOPY Vol. 85 ( 5 ) page： AB57 - AB57 2017.5

Influence of using 3D images and 3D-printed objects on the formation of spatial mental models of experts and novices Invited Reviewed

MAEHIGASHI Akihiro, MIWA Kazuhisa, ODA Masahiro, NAKAMURA Yoshihiko, MORI Kensaku, IGAMI Tsuyoshi

JSAI Technical Report, SIG-ALST Vol. 79 ( 0 ) page： 08 2017.3

3D FCN Feature Driven Regression Forest-Based Pancreas Localization and Segmentation Invited Reviewed

Oda Masahiro, Shimizu Natsuki, Roth Holger R., Karasawa Ken'ichi, Kitasaka Takayuki, Misawa Kazunari, Fujiwara Michitaka, Rueckert Daniel, Mori Kensaku

DEEP LEARNING IN MEDICAL IMAGE ANALYSIS AND MULTIMODAL LEARNING FOR CLINICAL DECISION SUPPORT Vol. 10553 page： 222 - 230 2017

This paper presents a fully automated atlas-based pancreas segmentation method from CT volumes utilizing 3D fully convolutional network (FCN) feature-based pancreas localization. Segmentation of the pancreas is difficult because it has larger inter-patient spatial variations than other organs. Previous pancreas segmentation methods failed to deal with such variations. We propose a fully automated pancreas segmentation method that contains novel localization and segmentation. Since the pancreas neighbors many other organs, its position and size are strongly related to the positions of the surrounding organs. We estimate the position and the size of the pancreas (localization) from global features by regression forests. As global features, we use intensity differences and 3D FCN deep learned features, which include automatically extracted essential features for segmentation. We chose 3D FCN features from a trained 3D U-Net, which is trained to perform multi-organ segmentation. The global features include both the pancreas and surrounding organ information. After localization, a patient-specific probabilistic atlas-based pancreas segmentation is performed. In evaluation results with 146 CT volumes, we achieved 60.6% of the Jaccard index and 73.9% of the Dice overlap.

DOI： 10.1007/978-3-319-67558-9_26

3Dプリンタ・ユーザーインターフェイス等の最新動向 Invited

森健策

インナービジョン Vol. 32 page： 44 - 45 2017

3D Printing for Medical Application Invited

MORI Kensaku

Medical Imaging and Information Sciences Vol. 34 ( 1 ) page： 1 - 6 2017

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：MEDICAL IMAGING AND INFORMATION SCIENCES

This paper briefly introduces 3D printing technology for medical applications. 3D printing technology is obtaining a lot of attentions from various fields including rapid prototyping and manufacturing, home use and medical applications. Medical applications of 3D printing are including:（a）diagnostic aid,（b）surgical aid,（c）medical education,（d）medical training, and（e）re-generative medicine. In this short summary, we will briefly explain the various mechanism of 3D printing and several printing techniques for reproducing organ models by 3D printing techniques. Also the recent topics in international conferences are introduced here.

DOI： 10.11318/mii.34.1

3Dプリンタの医療応用 Invited

森健策

医用画像情報学会雑誌 Vol. 34 page： 1 - 6 2017

3Dプリンタ-の基礎と医療応用 Invited

森健策

月刊心臓 Vol. 49 page： 1104 - 1113 2017

Automatic Segmentation of Head Anatomical Structures from Sparsely-annotated Images Invited Reviewed

Sugino Takaaki, Roth Holger R., Eshghi Mohammad, Oda Masahiro, Chung Min Suk, Mori Kensaku

2017 IEEE INTERNATIONAL CONFERENCE ON CYBORG AND BIONIC SYSTEMS (CBS) page： 145 - 149 2017

Computer-Aided Diagnosis of Mammographic Masses Using Geometric Verification-Based Image Retrieval Invited Reviewed

Li Qingliang, Shi Weili, Yang Huamin, Zhang Huimao, Li Guoxin, Chen Tao, Mori Kensaku, Jiang Zhengang

MEDICAL IMAGING 2017: COMPUTER-AIDED DIAGNOSIS Vol. 10134 2017

Comparison of the Deep-Learning-Based Automated Segmentation Methods for the Head Sectioned Images of the Virtual Korean Human Project Invited Reviewed

Eshghi Mohammad, Roth Holger R., Oda Masahiro, Chung Min Suk, Mori Kensaku

PROCEEDINGS OF THE FIFTEENTH IAPR INTERNATIONAL CONFERENCE ON MACHINE VISION APPLICATIONS - MVA2017 page： 290 - 293 2017

A study on improvement of airway segmentation using Hybrid method Invited Reviewed

Qier M., Kitasaka T., Nimura Y., Oda M., Mori K.

Proceedings - 3rd IAPR Asian Conference on Pattern Recognition, ACPR 2015 page： 549 - 553 2016.6

Authorship：Corresponding author Language：English Publishing type：Research paper (scientific journal) Publisher：Proceedings - 3rd IAPR Asian Conference on Pattern Recognition, ACPR 2015

This paper presents a method for extracting an airway region from 3D chest CT volumes that uses a combination of tube enhancement filters, voxel classification based on machine learning methods and graph-cut algorithm. Lots of previous methods utilize region growing or level set algorithms without any prior knowledge of bronchi, which always fail when they reach to the peripheral bronchi. In this paper, a method of extraction based on airway shape and machine learning is proposed. The proposed method detects candidate voxels of bronchial regions by using two types of enhancement filters, and a classifier model is built for selecting the proper candidates regions based on intensity and shape features and finally the selected candidate voxels are connected by graph-cut algorithm. We applied this method on six cases of 3D chest CT volumes. The results show that this method can extract the smaller airway branches without leaking into the lung parenchyma areas.

DOI： 10.1109/ACPR.2015.7486563

Characterization of Colorectal Lesions Using a Computer-Aided Diagnostic System for Narrow-Band Imaging Endocytoscopy Invited Reviewed

Misawa Masashi, Kudo Shin-ei, Mori Yuichi, Nakamura Hiroki, Kataoka Shinichi, Maeda Yasuharu, Kudo Toyoki, Hayashi Takemasa, Wakamura Kunihiko, Miyachi Hideyuki, Katagiri Atsushi, Baba Toshiyuki, Ishida Fumio, Inoue Haruhiro, Nimura Yukitaka, Mori Kensaku

GASTROENTEROLOGY Vol. 150 ( 7 ) page： 1531 - + 2016.6

Influences of 3D images and 3D-printed objects on mental model construction of liver structure Invited Reviewed

MAEHIGASHI Akihiro, MIWA Kazuhisa, ODA Masahiro, NAKAMURA Yoshihiko, MORI Kensaku, IGAMI Tsuyoshi

JSAI Technical Report, SIG-ALST Vol. 76 ( 0 ) page： 14 2016.3

Authorship：Corresponding author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society for Artificial Intelligence

In this study, we experimentally investigated the influence of a three-dimensional (3D) graphic image and a 3D-printed object on a spatial reasoning task in a situation where liver resection surgery was presupposed. The results of the study indicated that using a 3D-printed object produced more accurate and faster mental model construction of a liver structure than a 3D image. Using a 3D-printed object was assumed to reduce cognitive load and information accessing cost more than using a 3D image.

DOI： 10.11517/jsaialst.76.0_14

Cascade registration of micro CT volumes taken in multiple resolutions Invited Reviewed

Nagara K., Oda H., Nakamura S., Oda M., Homma H., Takabatake H., Mori M., Natori H., Rueckert D., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 9805 LNCS page： 269 - 280 2016

In this paper, we present a preliminary report of a multiscale registration method between micro-focus X-ray CT (micro CT) volumes taken in different scales. 3D fine structures of target objects can be observed on micro CT volumes, which are difficult to observe on clinical CT volumes. Micro CT scanners can scan specimens in various resolutions. In their high resolution volumes, ultra fine structures of specimens can be observed, while scanned areas are limited to very small. On the other hand, in low resolution volumes, large areas can be captured, while fine structures of specimens are difficult to observe. The fusion volume of the high and low resolution volumes will have benefits of both. Because the difference of resolutions between the high and low resolution volumes may vary greatly, an intermediate resolution volume is required for successful fusion of volumes. To perform such volume fusion, a cascade multi-resolution registration technique is required. To register micro CT volumes that have quite different resolutions, we employ a cascade co-registration technique. In the cascade co-registration process, intermediate resolution volumes are used in a registration process of the high and low resolution volumes. In the registration between two volumes, we apply two steps registration techniques. In the first step, a block division is used to register two resolution volumes. Afterward, we estimate the fine spatial positions relating the registered two volumes using the Powell method. The registration result can be used to generate a fusion volume of the high and low resolution volumes.

DOI： 10.1007/978-3-319-43775-0_24

3Dプリンタ・ユーザーインターフェイス等の最新動向 Invited

森健策

インナービジョン Vol. - page： 44 - 45 2016

3Dプリンティングのハンドリングのノウハウ Invited

森健策

インナービジョン Vol. 31 page： 20 - 24 2016

Automated anatomical labeling of abdominal arteries and hepatic portal system extracted from abdominal CT volumes Invited Reviewed

Matsuzaki Tetsuro, Oda Masahiro, Kitasaka Takayuki, Hayashi Yuichiro, Misawa Kazunari, Mori Kensaku

MEDICAL IMAGE ANALYSIS Vol. 20 ( 1 ) page： 152 - 161 2015.2

Automated torso organ segmentation from 3D CT images using conditional random field Invited Reviewed

Yukitaka Nimura, Yuichiro Hayashi, Takayuki Kitasaka, Kazunari Misawa, and Kensaku Mori

Proceedings of SPIE Medical Imaging 2016 Vol. 9785 2015

3Dプリンティングの現状と将来展望:医用画像処理と3Dプリンタによる臓器実体モデル作成とその利用 Invited

森健策

光技術コンタクト Vol. 53 page： 20 - 27 2015

Automated branching pattern report generation for laparoscopic surgery assistance Invited Reviewed

Oda Masahiro, Matsuzaki Tetsuro, Hayashi Yuichiro, Kitasaka Takayuki, Misawa Kazunari, Mori Kensaku

MEDICAL IMAGING 2015: COMPUTER-AIDED DIAGNOSIS Vol. 9414 2015

A study on improvement of airway segmentation using Hybrid method Invited Reviewed

Qier Meng, Kitasaka Takayuki, Nimura Yukitaka, Oda Masahiro, Mori Kensaku

PROCEEDINGS 3RD IAPR ASIAN CONFERENCE ON PATTERN RECOGNITION ACPR 2015 page： 549 - 553 2015

Automated Torso Organ Segmentation from 3D CT Images using Structured Perceptron and Dual Decompostion Invited Reviewed

Nimura Yukitaka, Hayashi Yuichiro, Kitasaka Takayuki, Mori Kensaku

MEDICAL IMAGING 2015: COMPUTER-AIDED DIAGNOSIS Vol. 9414 2015

Development of a new detection device using a glass clip emitting infrared fluorescence for laparoscopic surgery of gastric cancer Invited Reviewed

Inada Shunko Albano, Fuchi Shingo, Mori Kensaku, Hasegawa Junichi, Misawa Kazunari, Nakanishi Hayao

6TH INTERNATIONAL CONFERENCE ON OPTICAL, OPTOELECTRONIC AND PHOTONIC MATERIALS AND APPLICATIONS (ICOOPMA) 2014 Vol. 619 2015

Development and clinical application of surgical navigation system for laparoscopic hepatectomy Invited Reviewed

Hayashi Yuichiro, Igami Tsuyoshi, Hirose Tomoaki, Nagino Masato, Mori Kensaku

MEDICAL IMAGING 2015: IMAGE-GUIDED PROCEDURES, ROBOTIC INTERVENTIONS, AND MODELING Vol. 9415 2015

Connection method of separated luminal regions of intestine from CT volumes Invited Reviewed

Oda Masahiro, Kitasaka Takayuki, Furukawa Kazuhiro, Watanabe Osamu, Ando Takafumi, Hirooka Yoshiki, Goto Hidemi, Mori Kensaku

MEDICAL IMAGING 2015: COMPUTER-AIDED DIAGNOSIS Vol. 9414 2015

SGSR: style-subnets-assisted generative latent bank for large-factor super-resolution with registered medical image dataset. Invited Reviewed International journal

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

International journal of computer assisted radiology and surgery Vol. 19 ( 3 ) page： 493 - 506 2024.3

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: We propose a large-factor super-resolution (SR) method for performing SR on registered medical image datasets. Conventional SR approaches use low-resolution (LR) and high-resolution (HR) image pairs to train a deep convolutional neural network (DCN). However, LR-HR images in medical imaging are commonly acquired from different imaging devices, and acquiring LR-HR image pairs needs registration. Registered LR-HR images have registration errors inevitably. Using LR-HR images with registration error for training an SR DCN causes collapsed SR results. To address these challenges, we introduce a novel SR approach designed specifically for registered LR-HR medical images. METHODS: We propose style-subnets-assisted generative latent bank for large-factor super-resolution (SGSR) trained with registered medical image datasets. Pre-trained generative models named generative latent bank (GLB), which stores rich image priors, can be applied in SR to generate realistic and faithful images. We improve GLB by newly introducing style-subnets-assisted GLB (S-GLB). We also propose a novel inter-uncertainty loss to boost our method's performance. Introducing more spatial information by inputting adjacent slices further improved the results. RESULTS: SGSR outperforms state-of-the-art (SOTA) supervised SR methods qualitatively and quantitatively on multiple datasets. SGSR achieved higher reconstruction accuracy than recently supervised baselines by increasing peak signal-to-noise ratio from 32.628 to 34.206 dB. CONCLUSION: SGSR performs large-factor SR while given a registered LR-HR medical image dataset with registration error for training. SGSR's results have both realistic textures and accurate anatomical structures due to favorable quantitative and qualitative results. Experiments on multiple datasets demonstrated SGSR's superiority over other SOTA methods. SR medical images generated by SGSR are expected to improve the accuracy of pre-surgery diagnosis and reduce patient burden.

DOI： 10.1007/s11548-023-03037-3

YOLOv7-RepFPN: Improving real-time performance of laparoscopic tool detection on embedded systems Invited Reviewed

Yuzhang Liu, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kensaku Mori

Healthcare Technology Letters 2024

Language：English Publishing type：Research paper (scientific journal)

This study focuses on enhancing the inference speed of laparoscopic tool detection on embedded devices. Laparoscopy, a minimally invasive surgery technique, markedly reduces patient recovery times and postoperative complications. Real-time laparoscopic tool detection helps assisting laparoscopy by providing information for surgical navigation, and its implementation on embedded devices is gaining interest due to the portability, network independence and scalability of the devices. However, embedded devices often face computation resource limitations, potentially hindering inference speed. To mitigate this concern, the work introduces a two-fold modification to the YOLOv7 model: the feature channels and integrate RepBlock is halved, yielding the YOLOv7-RepFPN model. This configuration leads to a significant reduction in computational complexity. Additionally, the focal EIoU (efficient intersection of union) loss function is employed for bounding box regression. Experimental results on an embedded device demonstrate that for frame-by-frame laparoscopic tool detection, the proposed YOLOv7-RepFPN achieved an mAP of 88.2% (with IoU set to 0.5) on a custom dataset based on EndoVis17, and an inference speed of 62.9 FPS. Contrasting with the original YOLOv7, which garnered an 89.3% mAP and 41.8 FPS under identical conditions, the methodology enhances the speed by 21.1 FPS while maintaining detection accuracy. This emphasizes the effectiveness of the work.

DOI： 10.1049/htl2.12072

Towards better laparoscopic video segmentation: A class-wise contrastive learning approach with multi-scale feature extraction Invited Reviewed

Luyang Zhang, Yuichiro Hayashi, Masahiro Oda, Kensaku Mori

Healthcare Technology Letters 2024

Language：English Publishing type：Research paper (scientific journal)

The task of segmentation is integral to computer-aided surgery systems. Given the privacy concerns associated with medical data, collecting a large amount of annotated data for training is challenging. Unsupervised learning techniques, such as contrastive learning, have shown powerful capabilities in learning image-level representations from unlabelled data. This study leverages classification labels to enhance the accuracy of the segmentation model trained on limited annotated data. The method uses a multi-scale projection head to extract image features at various scales. The partitioning method for positive sample pairs is then improved to perform contrastive learning on the extracted features at each scale to effectively represent the differences between positive and negative samples in contrastive learning. Furthermore, the model is trained simultaneously with both segmentation labels and classification labels. This enables the model to extract features more effectively from each segmentation target class and further accelerates the convergence speed. The method was validated using the publicly available CholecSeg8k dataset for comprehensive abdominal cavity surgical segmentation. Compared to select existing methods, the proposed approach significantly enhances segmentation performance, even with a small labelled subset (1–10%) of the dataset, showcasing a superior intersection over union (IoU) score.

DOI： 10.1049/htl2.12069

Revisiting instrument segmentation: Learning from decentralized surgical sequences with various imperfect annotations Invited Reviewed

Zhou Zheng, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kensaku Mori

Healthcare Technology Letters 2024

Language：English Publishing type：Research paper (scientific journal)

This paper focuses on a new and challenging problem related to instrument segmentation. This paper aims to learn a generalizable model from distributed datasets with various imperfect annotations. Collecting a large-scale dataset for centralized learning is usually impeded due to data silos and privacy issues. Besides, local clients, such as hospitals or medical institutes, may hold datasets with diverse and imperfect annotations. These datasets can include scarce annotations (many samples are unlabelled), noisy labels prone to errors, and scribble annotations with less precision. Federated learning (FL) has emerged as an attractive paradigm for developing global models with these locally distributed datasets. However, its potential in instrument segmentation has yet to be fully investigated. Moreover, the problem of learning from various imperfect annotations in an FL setup is rarely studied, even though it presents a more practical and beneficial scenario. This work rethinks instrument segmentation in such a setting and propose a practical FL framework for this issue. Notably, this approach surpassed centralized learning under various imperfect annotation settings. This method established a foundational benchmark, and future work can build upon it by considering each client owning various annotations and aligning closer with real-world complexities.

DOI： 10.1049/htl2.12068

Artificial intelligence in a prediction model for post-ERCP pancreatitis Reviewed

Takahashi Hidekazu, Eizaburo Ohno, Taiki Furukawa, Kentaro Yamao, Takuya Ishikawa, Yasuyuki Mizutani, Tadashi Iida, Yoshimune Shiratori, Shintaro Oyama, Junji Koyama, Kensaku Mori, Yuichiro Hayashi, Masahiro Oda, Takahisa Suzuki, Hiroki Kawashima

Digestive Endoscopy 2023.7

DEVELOPMENT OF A MACHINE-LEARNING MODEL FOR PREDICTING POST-ERCP PANCREATITIS Invited Reviewed

Gastrointestinal Endoscopy Vol. 97 ( 6 ) page： AB656 - AB656 2023.6

Development of panorama vision ring for thoracoscopy. Invited Reviewed International journal

Takayuki Kitasaka, Shota Nakamura, Yuichiro Hayashi, Tsuyoshi Nakai, Yasuhiro Nakai, Kensaku Mori, Toyofumi Fengshi Chen-Yoshikawa

International journal of computer assisted radiology and surgery Vol. 18 ( 5 ) page： 945 - 952 2023.5

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: Minimally invasive surgery (MIS) using a thoraco- or laparoscope is becoming a more common surgical technique. In MIS, a magnified view from a thoracoscope helps surgeons conduct precise operations. However, there is a risk of the visible area becoming narrow. To confirm that the operation field is safe, the surgeon will draw the thoracoscope back to check the marginal area of the target and insert it again many times during MIS. To reduce the surgeon's load, we aim to visualize the entire thoracic cavity using a newly developed device called "panorama vision ring" (PVR). METHOD: The PVR is used instead of a wound retractor or a trocar. It is a ring-type socket with one big hole for the thoracoscope and four small holes for tiny cameras placed around the big hole. The views from the tiny cameras are fused into one wider view that visualizes the entire thoracic cavity. A surgeon can proceed with an operation by checking what exists outside of the thoracoscopic view. Also, she/he can check whether or not bleeding has occurred from the image of the entire cavity. RESULTS: We evaluated the view-expansion ability of the PVR by using a three-dimensional full-scale thoracic model. The experimental results showed that the entire thoracic cavity could be visible in a panoramic view generated by the PVR. We also demonstrated pulmonary lobectomy in virtual MIS using the PVR. Surgeons could perform a pulmonary lobectomy while checking the entire cavity. CONCLUSION: We developed the PVR, which uses tiny auxiliary cameras to create a panoramic view of the entire thoracic cavity during MIS. We aim to make MIS safer for patients and more comfortable for surgeons through the development of the PVR.

DOI： 10.1007/s11548-023-02859-5

Oesophagus Achalasia Diagnosis from Esophagoscopy Based on a Serial Multi-scale Network Invited

CT像からの腸閉塞検出における中心線を用いた後処理手法

陳思睿, 小田紘久, 安芹, 林雄一郎, 北坂孝幸, 滝本愛太朗, 檜顕成, 内田広夫, 鈴木耕次郎, 小田昌宏, 森健策

電子情報通信学会技術研究報告(MI)信学技報 Vol. 122 ( 417 ) page： 223 - 228 2023.3

医用画像とAI Reviewed

カレントテラピー Vol. 41 ( 39 ) page： 79 - 79 2023.3

Comparative study of the Small Intestine segmentation based on 2D and 3D U-Nets

Qin An, Hirohisa Oda, Sirui Chen, Yuichiro Hayashi, Takayuki Kitasaka, Hiroo Uchida, Akinari Hinoki, Kojiro Suzuki, Aitaro Takimoto, Masahiro Oda, Kensaku Mori

Vol. 122 ( 417 ) page： 46 - 51 2023.3

Average Templateを使ったCT像からのCOVID-19異常陰影領域セグメンテーション手法

柳凱, 小田昌宏, 鄭通, 林雄一郎, 大竹義人, 橋本正弘, 明石敏昭, 青木茂樹, 森　健策

電子情報通信学会技術研究報告(MI) 信学技報 Vol. 122 ( 47 ) page： 40 - 45 2023.3

術前画像情報を用いた腹腔鏡映像からの血管位置予測の検討

榎本圭吾, 林雄一郎, 北坂孝幸, 小田昌宏, 三澤一成, 森健策

電子情報通信学会技術研究報告(MI) 信学技報 Vol. 122 ( 417 ) page： 63 - 68 2023.3

グラフニューラルネットワークを用いた血管名自動命名における臓器特徴の有効性の調査

出口智也, 林雄一郎, 北坂孝幸, 小田昌宏, 三澤一成, 森健策

電子情報通信学会技術研究報告(MI), 信学技報 Vol. 122 ( 417 ) page： 105 - 110 2023.3

L-former : a lightweight transformer for realistic medical image generation and its application to super-resolution Reviewed

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi,shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

proc. SPIE 2023.2

Priority attention network with Bayesian learning for fully automatic segmentation of substantia nigra from neuromelanin MRI Invited Reviewed

Tao Hu, Hayato Itoh, Masahiro Oda, Shinji Saiki, Nobutaka Hattori, Koji Kamagata, Shigeki Aoki, Kensaku Mori

Proc.SPIE12464 2023.2

`Thrombosis region extraction and quantitative analysis in confocal laser scanning microscopic image sequence in in-vivo imaging Invited Reviewed

Yunheng Wu, Masahiro Oda,Yuichiro Hayashi, Shuntaro Kawamura, Takanori Takebe, Kensaku Mori

Proc.SPIE 2023.2

Real bronchoscopic images-based bronchial nomenclature: a preliminary study Invited Reviewed

Cheng Wang, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Proc.SPIE 2023.2

Improved method for COVID-19 classification of complex-architecture CNN from chest CT volumes using orthogonal ensemble networks Invited Reviewed

Ryo Toda, Masahiro Oda, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori

Proc.SPIE 2023.2

Octree cube constraints in PBD method for high resolution surgical simulation Reviewed

Rintaro Miyazaki, Yuichiro Hayashi, Masahiro Oda, Kensaku Mori

Proc.SPIE 2023.2

Classification of COVID-19 cases from chest CT volumes using hybrid model of 3D CNN and 3D MLP-mixer Invited Reviewed

Masahiro Oda, Tong Zheng, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori

Proc.SPIE 2023.2

A semantic segmentation method for laparoscopic images using semantically similar groups Reviewed

Leo Uramoto, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kazunari Misawa, Kensaku Mori

Proc.SPIE 2023.2

Kai Jiang, Masahiro Oda, Yuichiro Hayashi, Hironari Shiwaku, Masashi Misawa, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization Vol. 11 ( 4 ) page： 1 - 10 2023.2

KST-Mixer: kinematic spatio-temporal data mixer for colon shape estimation Invited

Masahiro Oda, Kazuhiro Furukawa, Nassir Navab, Kensaku Mori

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization Vol. 11 ( 4 ) page： 1 - 7 2023.1

人工知能（AI）の最新動向　RSNA2022におけるAI関連セッション Reviewed

インナービジョン Vol. 38 ( 2 ) page： 33 - 34 2023.1

L-former : A Lightweight Transformer for Realistic Medical Image Generation and its Application to Super-resolution Invited Reviewed

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

MEDICAL IMAGING 2023 Vol. 12464 2023

Language：English Publishing type：Research paper (international conference proceedings) Publisher：SPIE-INT SOC OPTICAL ENGINEERING

Medical image analysis approaches such as data augmentation and domain adaption need huge amounts of realistic medical images. Generating realistic medical images by machine learning is a feasible approach. We propose L-former, a lightweight Transformer for realistic medical image generation. L-former can generate more reliable and realistic medical images than recent generative adversarial networks (GANs). Meanwhile, L-former does not consume as high computational cost as conventional Transformer-based generative models. L-former uses Transformers to generate low-resolution feature vectors at shallow layers, and uses convolutional neural networks to generate high-resolution realistic medical images at deep layers. Experimental results showed that L-former outperformed conventional GANs by FID scores 33.79 and 76.85 on two datasets, respectively. We further conducted a downstream study by using the images generated by L-former to perform a super-resolution task. A high PSNR score of 27.87 proved L-former's ability to generate reliable images for super-resolution and showed its potential for applications in medical diagnosis.

DOI： 10.1117/12.2653776

Boundary-aware Feature and Prediction Refinement for Polyp Segmentation Reviewed

ie Qiu, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kensaku Mori, ``Boundary-aware Feature and Prediction Refinement for Polyp Segmentation

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2022.12

Classification and visual explanation for COVID-19 pneumonia from CT images using triple learning Invited Reviewed

Sota Kato, Masahiro Oda, Kensaku Mori, Akinobu Shimizu, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Kazuhiro Hotta

Scientific Reports Vol. 12 ( 1 ) page： 20840 2022.12

Language：English Publishing type：Research paper (scientific journal) Publisher：Springer Science and Business Media LLC

Abstract

This study presents a novel framework for classifying and visualizing pneumonia induced by COVID-19 from CT images. Although many image classification methods using deep learning have been proposed, in the case of medical image fields, standard classification methods are unable to be used in some cases because the medical images that belong to the same category vary depending on the progression of the symptoms and the size of the inflamed area. In addition, it is essential that the models used be transparent and explainable, allowing health care providers to trust the models and avoid mistakes. In this study, we propose a classification method using contrastive learning and an attention mechanism. Contrastive learning is able to close the distance for images of the same category and generate a better feature space for classification. An attention mechanism is able to emphasize an important area in the image and visualize the location related to classification. Through experiments conducted on two-types of classification using a three-fold cross validation, we confirmed that the classification accuracy was significantly improved; in addition, a detailed visual explanation was achieved comparison with conventional methods.

DOI： 10.1038/s41598-022-24936-6

Other Link： https://www.nature.com/articles/s41598-022-24936-6

Positive-gradient-weighted object activation mapping: visual explanation of object detector towards precise colorectal-polyp localisation. Invited Reviewed International journal

Hayato Itoh, Masashi Misawa, Yuichi Mori, Shin-Ei Kudo, Masahiro Oda, Kensaku Mori

International journal of computer assisted radiology and surgery Vol. 17 ( 11 ) page： 2051 - 2063 2022.11

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: Precise polyp detection and localisation are essential for colonoscopy diagnosis. Statistical machine learning with a large-scale data set can contribute to the construction of a computer-aided diagnosis system for the prevention of overlooking and miss-localisation of a polyp in colonoscopy. We propose new visual explaining methods for a well-trained object detector, which achieves fast and accurate polyp detection with a bounding box towards a precise automated polyp localisation. METHOD: We refine gradient-weighted class activation mapping for more accurate highlighting of important patterns in processing a convolutional neural network. Extending the refined mapping into multiscaled processing, we define object activation mapping that highlights important object patterns in an image for a detection task. Finally, we define polyp activation mapping to achieve precise polyp localisation by integrating adaptive local thresholding into object activation mapping. We experimentally evaluate the proposed visual explaining methods with four publicly available databases. RESULTS: The refined mapping visualises important patterns in each convolutional layer more accurately than the original gradient-weighted class activation mapping. The object activation mapping clearly visualises important patterns in colonoscopic images for polyp detection. The polyp activation mapping localises the detected polyps in ETIS-Larib, CVC-Clinic and Kvasir-SEG database with mean Dice scores of 0.76, 0.72 and 0.72, respectively. CONCLUSIONS: We developed new visual explaining methods for a convolutional neural network by refining and extending gradient-weighted class activation mapping. Experimental results demonstrated the validity of the proposed methods by showing that accurate visualisation of important patterns and localisation of polyps in a colonoscopic image. The proposed visual explaining methods are useful for the interpreting and applying a trained polyp detector.

DOI： 10.1007/s11548-022-02696-y

Pattern Analysis of Substantia Nigra in Parkinson Disease by Fifth-Order Tensor Decomposition and Multi-sequence MRI Reviewed

Hayato Itoh, Tao Hu, Masahiro Oda, Shinji Saiki, Koji Kamagata, Nobutaka Hattori, Shigeki Aoki,Kensaku Mori

LNCS13594 2022.10

Enhancing Model Generalization for Substantia Nigra Segmentation Using a Test-time Normalization-Based Method Reviewed

Tao Hu, Hayato Itoh, Masahiro Oda, Yuichiro Hayashi, Zhongyang Lu, Shinji Saiki, Nobutaka Hattori, Koji Kamagata, Shigeki Aoki, Kanako K. Kumamaru, Toshiaki Akashi, Kensaku Mori

LNCS13437 page： 736 - 744 2022.9

Geometric Constraints for Self-supervised Monocular Depth Estimation on Laparoscopic Images with Dual-task Consistency Reviewed

Wenda Li, Yuichiro Hayashi, Masahiro Oda, akayuki Kitasaka, Kazunari Misawa, Kensaku Mori

LNCS13434 page： 467 - 477 2022.9

Confocal Laser Scanning Microscope Image Super Resolution for Biomedical Research Based on Two-Stage Generative Adversarial Network Invited Reviewed

Yunheng WU, Masahiro ODA, Yuichiro HAYASHI, Takanori TAKEBE, Shogo NAGATA, Shuntaro KAWAMURA, Kensaku MORI

page： 138 - 139 2022.7

Co-Training for Semi-Supervised CT Segmentation of COVID-19 Invited Reviewed

Kai LIU, Masahiro ODA, Tong ZHENG, Yuichiro HAYASHI, Yoshito OTAKE, Masahiro HASHIMOTO, Toshiaki AKASHI, Shigeki AOKI, Kensaku MORI

page： 114 - 115 2022.7

A Novel Centroid-attention based Hybrid Model for Subarachnoid Hemorrhage Classification on Imbalanced Data Invited Reviewed

Zhongyang LU, Masahiro ODA,1, Yuichiro HAYASHI, Tao Hu, Hayato ITOH, Takeyuki WATADANI, Osamu ABE，Kensaku MORI

page： 104 - 105 2022.7

テンソル分解を用いた黒質緻密部の3 次元パターン表現に関する初期的検討 Invited Reviewed

伊東隼人, 小田昌宏, 斉木臣二, 服部信考, 鎌形康司, 青木茂樹, 森健策

第41回日本医用画像工学会大会予稿集 page： 124 - 125 2022.7

境界情報を考慮する損失関数を用いたFCN による腹部 CT 像からの臓器領域抽出に関する研究 Invited Reviewed

大野真奈, 申忱, Holger R. Roth, 小田昌宏, 林雄一郎, 三澤一成, 森健策

第41回日本医用画像工学会大会予稿集 page： 106 - 107 2022.7

GNN による血管名自動命名手法における臓器特徴の利用に関する検討 Invited Reviewed

出口智也, 林雄一郎, 北坂孝幸, 小田昌宏, 三澤一成, 森健策

　第41回日本医用画像工学会大会予稿集 2022.7

大腸外科領域における情報支援内視鏡外科手術システムの開発 Invited Reviewed

]長谷川寛, 北口大地, 小島成浩, 竹下修由, 森健策, 伊藤雅昭

日本コンピュータ外科学会誌第31回日本コンピュータ外科学会大会特集号 Vol. 24 ( 2 ) 2022.6

腹腔鏡映像からの血管領域自動抽出におけるDilated U‒Netの段数が抽出精度に与える影響 Invited Reviewed

榎本圭吾，林雄一郎，北坂孝幸，小田昌宏, 三澤一成, 森　健策

日本コンピュータ外科学会誌第31回日本コンピュータ外科学会大会特集号 Vol. 24 ( 2 ) page： 22(5)-3 2022.6

腹腔鏡下胃切除術支援のための腹腔鏡映像からの膵臓領域抽出の検討 Invited Reviewed

林雄一郎, 辻真治, 丘杰, 小田昌宏, 三澤一成，森健策

日本コンピュータ外科学会誌第31回日本コンピュータ外科学会大会特集号 Vol. 24 ( 2 ) page： 22(5)-4 2022.6

nnU‒Netによる肺マイクロCT像からの小葉間隔壁抽出 Invited Reviewed

深井大輔，小田紘久，椎名健，林雄一郎，鄭通，中村彰太，小田昌宏，森健策

日本コンピュータ外科学会誌第31回日本コンピュータ外科学会大会特集号 Vol. 24 ( 2 ) page： 22(5)-6 2022.6

コンピュータ外科におけるAIとVisionのドッキング―知能と知覚の結合による新たなコンピュータ外科 Invited Reviewed

森健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 24 ( 2 ) 2022.6

3D bronchus anatomical structure measurement on real bronchoscopic images based on depth images estimated by deep neural network Reviewed

C. Wang, Y. Hayashi, M. Oda, T. Kitasaka, H. Takabatake, M. Mori, H. Honma, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 17 page： s48 - s49 2022.6

Extraction of respiratory bronchioles and alveolar ducts from micro-CT volumes with distance-based tubular structure filter Invited Reviewed

T. Shiina, H. Oda, T. Zheng, S. Nakamura, Y. Hayashi, M. Oda, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 17 page： s117 - s118 2022.6

2D+3D registration in deformation-adaptive super-resolution for medical images Invited Reviewed

T. Zheng, H. Oda, T. Hu, Y. Hayashi, S. Nakamura, M. Mori, H. Takabatake, H. Natori, M. Oda, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 17 page： s105 - s106 2022.6

Automatic Detection of Bladder Tumors in Narrow-Band Imaging Cystoscopic Images by tiny-YOLO Invited Reviewed

J. Mutaguchi, M. Oda, E. Kashiwagi, J. Inokuchi, K. Mori, M. Eto

International Journal of Computer Assisted Radiology and Surgery Vol. 17 page： s86 - s86 2022.6

Laparoscopic image classification based on surgical areas in laparoscopic gastrectomy, Invited Reviewed

Y. Hayashi, K. Misawa, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 17 page： s57 - 58 2022.6

30年間の医用画像研究経験を振り返り未来を考える Reviewed

情報・システムソサイエティ誌 Vol. 27 ( 1 ) 2022.5

Automated classification method of COVID-19 cases from chest CT volumes using 2D and 3D hybrid CNN for anisotropic volumes Reviewed

Masahiro Oda, Tong Zheng, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori

Proc. SPIE 12033 2022.3

Effective hyperparameter optimization with proxy data for multi-organ segmentation Reviewed

Chen Shen, Holger R. Roth, Vishwesh Nath, Yuichiro Hayashi, Masahiro Oda, Kazunari Misawa, Kensaku Mori

Proc. SPIE 12032 2022.3

Coarse-to-fine cascade framework for cross-modality super-resolution on clinical/micro CT dataset, Reviewed

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

Proc. SPIE 12032 2022.3

Bronchial orifice tracking-based branch level estimation for bronchoscopic navigation Reviewed

Cheng Wang, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Hitotsugu Takabatake, Masaki Mori, Hirotoshi Honma, Hiroshi Natori, Kensaku Mori

Proc. SPIE 12034 2022.3

Taking full advantage of uncertainty estimation: an uncertainty-assisted two-stage pipeline for multi-organ segmentation Reviewed

Zhou Zheng, Masahiro Oda, Kazunari Misawa, Kensaku Mori

Proc. SPIE 12033 2022.3

Multiclass prediction for improving intestine segmentation on non-fecal-tagged CT volume Reviewed

Hirohisa Oda, Yuichiro Hayashi, Takayuki Kitasaka, Aitaro Takimoto, Akinari,Hinoki, Hiroo Uchida, Kojiro Suzuki, Masahiro Oda, Kensaku Mori

Proc. SPIE 12033 2022.3

Spatial label smoothing via aleatoric uncertainty for bleeding region segmentation from laparoscopic videos Reviewed

Jie Qiu, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Nobuyoshi Takeshita, Masaaki Ito, Kensaku Mori

Proc. SPIE 12032 2022.3

Size-reweighted cascaded fully convolutional network for substantia nigra segmentation from T2 MRI Reviewed

Tao Hu, Hayato Itoh, Masahiro Oda, Shinji Saiki, Nobutaka Hattori, Koji Kamagata, Shigeki Aoki, Kensaku Mori

Proc. SPIE 12032 2022.3

Substantia nigra analysis by tensor decomposition of T2-weighted images for Parkinson’s disease diagnosis Invited Reviewed

Hayato Itoh, Masahiro Oda, Shinji Saiki, Nobutaka Hattori, Koji Kamagata, Shigeki Aoki, Kensaku Mori

Proc. SPIE 12032 2022.3

Self-supervised depth estimation with uncertainty-weight joint loss function based on laparoscopic videos Reviewed

Wenda Li, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kazunari Misawa, Kensaku Mori

Proc. SPIE 12034 2022.3

人工知能(AI)最新動向学会発表を中心に(2)-Digital Poster発表を中心に

森　健策

インナービジョン Vol. 37 ( 2 ) page： 30 - 31 2022.2

Aorta-aware GAN for non-contrast to artery contrasted CT translation and its application to abdominal aortic aneurysm detection. Invited Reviewed International journal

Tao Hu, Masahiro Oda, Yuichiro Hayashi, Zhongyang Lu, Kanako Kunishima Kumamaru, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori

International journal of computer assisted radiology and surgery Vol. 17 ( 1 ) page： 97 - 105 2022.1

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: Artery contrasted computed tomography (CT) enables accurate observations of the arteries and surrounding structures, thus being widely used for the diagnosis of diseases such as aneurysm. To avoid the complications caused by contrast agent, this paper proposes an aorta-aware deep learning method to synthesize artery contrasted CT volume form non-contrast CT volume. METHODS: By introducing auxiliary multi-resolution segmentation tasks in the generator, we force the proposed network to focus on the regions of aorta and the other vascular structures. Then, the segmentation results produced by the auxiliary tasks were used to extract aorta. The detection of abnormal CT images containing aneurysm was implemented by estimating the maximum axial radius of aorta. RESULTS: In comparison with the baseline models, the proposed network with auxiliary tasks achieved better performances with higher peak signal-noise ratio value. In aorta regions which are supposed to be the main region of interest in many clinic scenarios, the average improvement can be up to 0.33dB. Using the synthesized artery contrasted CT, the F score of aneurysm detection achieved 0.58 at slice level and 0.85 at case level. CONCLUSION: This study tries to address the problem of non-contrast to artery contrasted CT modality translation by employing a deep learning model with aorta awareness. The auxiliary tasks help the proposed model focus on aorta regions and synthesize results with clearer boundaries. Additionally, the synthesized artery contrasted CT shows potential in identifying slices with abdominal aortic aneurysm, and may provide an option for patients with contrast agent allergy.

DOI： 10.1007/s11548-021-02492-0

医用画像処理による人体構造の解析とその診断治療への応用～ 30年間の医用画像研究経験を振り返り未来を考える～

森健策

電子情報通信学会技術研究報告(MI), MI2021-74 Vol. 121 ( 347 ) page： 127 - 132 2022.1

胸部CT像からのCOVID-19に関連した所見文の自動生成の検討

岡崎真治, 林雄一郎, 小田昌宏, 橋本正弘, 陣崎雅弘, 明石敏昭, 青木茂樹, 森健策

電子情報通信学会技術研究報告(MI), MI2021-57 Vol. 121 ( 347 ) page： 49 - 54 2022.1

高精度な大腸ポリープ検出に向けた物体検出モデルの解析

伊東隼人, 三澤将史, 森悠一, 工藤進英, 小田昌宏, 森健策

電子情報通信学会技術研究報告(MI), MI2021-63 Vol. 121 ( 347 ) page： 86 - 87 2022.1

大規模腹腔鏡動画像データベース構築に向けたオンラインアノテーションツールの開発

伊東隼人, 潘冬平, 小澤卓也, 小田昌宏, 竹下修由, 伊藤雅昭, 森健策

電子情報通信学会技術研究報告(MI), MI2021-65 Vol. 121 ( 347 ) page： 86 - 87 2022.1

深層学習に基づくマウスのクラニアルウィンドウ画像における血管セグメンテーションの考察 Invited Reviewed

呉運恒, 小田昌宏, 林雄一郎, 武部貴則, 森健策

電子情報通信学会技術研究報告(MI) Vol. 121 ( 347 ) page： 174 - 179 2022.1

Performance improvement of weakly supervised fully convolutional networks by skip connections for brain structure segmentation Reviewed

Takaaki Sugino, Holger R. Roth, Mashiro Oda, Taichi Kin, Nobuhito Saito, Yoshikazu Nakajima, Kensaku Mori,

Medical Physics Vol. 48 ( 11 ) page： 7215 - 7227 2021.11

胸腔鏡下手術におけるパノラマビジョンリングの基礎開発 Reviewed

北坂孝幸，林　雄一郎，中村彰太，芳川豊史，森　健策，中井　剛，中井康博

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 318 2021.11

自己教師あり学習による腹腔鏡動画像の手術器具セグメンテーション Reviewed

丘　傑，林　雄一郎，小澤卓也，小田昌宏, 北坂孝幸，三澤一成, 竹下修由, 伊藤雅昭,　森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 217 - 218 2021.11

距離変換と管状構造フィルタによる肺マイクロCT画像からの細気管支・肺胞管抽出手法の検討 Reviewed

椎名　健, 小田紘久, 鄭　通, 中村彰太, 林　雄一郎, 小田昌宏, 森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 219 - 220 2021.11

大規模腹腔鏡動画像データベース構築に向けたアノテーションツール開発 Reviewed

伊東隼人，潘　冬平，小澤卓也，小田昌宏, 竹下修由, 伊藤雅昭, 森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 243 - 244 2021.11

腹腔鏡下胃切除術支援のための腹腔鏡映像からの術中操作の予測に関する初期検討 Reviewed

林雄一郎, 三澤一成, 森健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 248 2021.11

CT像の非等方性を考慮した3D CNNによるCOVID‒19症例の自動分類手法 Reviewed

小田昌宏, 鄭通, 林　雄一郎, 大竹義人, 橋本正弘, 明石敏明, 青木茂樹, 森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 265 - 266 2021.11

MRI 画像からの大脳基底核のAIセグメンテーション―Skip Connection による抽出精度向上の検討 Reviewed

杉野貴明，金　太一，斎藤　季，川瀬利弘，小野木真哉，齊藤延人，森　健策, 中島義和

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 286 2021.11

CT 像からの腸管領域抽出改善に関する基礎的検討 Reviewed

小田紘久，林　雄一郎，北坂孝幸，滝本愛太朗，檜　顕成，内田広夫，鈴木耕次郎，　小田昌宏, 森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 287 - 288 2021.11

複数の畳み込み範囲を持つグラフニューラルネットワークによる血管名自動命名手法の検討 Reviewed

出口智也，林　雄一郎，北坂孝幸，小田昌宏, 三澤一成，森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 299 - 300 2021.11

腹腔鏡下胃切除術の手術ナビゲーションにおける位置合わせ誤差の補正に関する検討 Reviewed

林　雄一郎, 三澤一成，森　健策

日本コンピュータ外科学会誌第30回日本コンピュータ外科学会大会特集号 Vol. 23 ( 4 ) page： 301 - 302 2021.11

Binary polyp-size classification based on deep-learned spatial information Invited Reviewed

Hayato Itoh, Masahiro Oda, Kai Jiang, Yuichi Mori, Masashi Misawa, Shin-Ei Kudo, Kenichiro Imai, Sayo Ito, Kinichi Hotta, Kensaku Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( 10 ) page： 1817 - 1828 2021.10

胸部 CT 像からの COVID-19 症例の自動分類手法

小田　昌宏, 鄭通, 林雄一郎, 大竹　義人, 橋本正弘, 明石敏昭, 森健策

第40回日本医用画像工学会大会予稿集 page： 65 - 67 2021.10

VR Organ Puzzle: A Virtual Reality Application for the Education of Human Anatomy

Siqi LI, Yuichiro HAYASHI，Michitaka FUJIWARA, Masahiro ODA, Kensaku MORI

page： 489 - 491 2021.10

Non-contrast to Artery Contrast CT Translation Via Representation-Aligned Generative Model

Tao HU, Masahiro ODA, Yuichiro HAYASHI, Zhongyang LU, Toshiaki AKASHI, Shigeki AOKI, Kensaku MORI

2021.10

Clinical CT Super-resolution Utilizing Registered Clinical – Micro CT Database

Tong ZHENG, Hirohisa ODA, Yuichiro HAYASHI, Shota NAKAMURA, Masaki MORI, Hirotsugu TAKABATAKE, Hiroshi NATORI, Masahiro ODA, Kensaku MORI

page： 394 - 400 2021.10

距離マップを利用した肺マイクロ CT 像からの肺胞抽出

椎名健, 小田紘久, 鄭通, 中村彰太, 林雄一郎, 小田昌宏, 森健策

第40回日本医用画像工学会大会予稿集 page： 374 - 377 2021.10

ピットパターン特徴量の解析に向けた超拡大内視鏡画像の再構成法に関する初期的検討

伊東隼人, 小田昌宏, 森悠一, 三澤将史, 工藤進英, 森健策

第40回日本医用画像工学会大会予稿集 page： 309 - 317 2021.10

深層学習とディジタルファントムを用いた骨陰影低減技術の開発

五島風汰, 田中利恵, 小田昌宏, 森健策, 高田宗尚, 田村昌也, 松本勲

第40回日本医用画像工学会大会予稿集 page： 270 - 272 2021.10

3D Kidney Tumor Semantic Segmentation using Cascaded Convolutional Networks Invited Reviewed

第40回日本医用画像工学会大会予稿集 page： 243 - 248 2021.10

Attention 機構を導入したグラフニューラルネットワークによる,

出口智也, 林雄一郎, 北坂孝幸, 小田昌宏,1 三澤一成, 森健策

第40回日本医用画像工学会大会予稿集 page： 239 - 241 2021.10

深度情報を利用した FCN による腹腔鏡映像からの血管領域自動抽出の検討

榎本圭吾, 林雄一郎, 北坂孝幸, 小田昌宏,1 伊藤雅昭, 竹下修由, 三澤一成, 森健策

第40回日本医用画像工学会大会予稿集 page： 235 - 241 2021.10

Vascular Structure Segmentation in Stereomicroscope Image

Yunheng WU, Masahiro ODA, Yuichiro HAYASHI, Takanori TAKEBE, Kensaku MORI

page： 229 - 234 2021.10

Synthesized Perforation Detection from Endoscopy Videos Using Model Training with Synthesized Images by GAN

Kai Jiang, Hayato Itoh, Masahiro Oda, Taishi Okumura, Yuichi Mori, Masashi Misawa, Takemasa Hayashi, Shin-Ei Kudo, Kensaku Mori

page： 199 - 201 2021.10

Improving Classification Accuracy of Hands' Bone Marrow Edema by Transfer Learning

Dongping PAN, Masahiro ODA, Kou KATAYAMA, Takanobu Okubo, Kensaku MORI

page： 150 - 157 2021.10

腸閉塞・イレウスの病変箇所特定における診断支援システムの精度評価

小田紘久, 林雄一郎, 北坂孝幸, 玉田雄大, 滝本愛太朗, 檜顕成, 内田広夫, 鈴木耕次郎, 小田昌宏, 森健策

第40回日本医用画像工学会大会予稿集 page： 129 - 131 2021.10

Self-attention Class Balanced DenseNet_LSTM framework for Subarachnoid Hemorrhage CT image Classification on Extremely Imbalanced Brain CT Dataset

第40回日本医用画像工学会大会予稿集 page： 69 - 75 2021.10

VR Organ Puzzle: A Virtual Reality Application for the Education of Human Anatomy

Siqi LI, Yuichiro HAYASHI，Michitaka FUJIWARA, Masahiro ODA, Kensaku MORI

page： 493 - 498 2021.10

Multi-task Federated Learning for Heterogeneous Pancreas Segmentation Invited Reviewed International coauthorship

Chen Shen, Pochuan Wang, Holger R. Roth, Dong Yang, Daguang Xu, Masahiro Oda, Weichung Wang, Chiou-Shann Fuh, Po-Ting Chen, Kao-Lang Liu, Wei-Chih Liao, Kensaku Mori

LNCS12969 page： 101 - 110 2021.9

Super-Resolution by Latent Space Exploration: Training with Poorly-Aligned Clinical and Micro CT Image Dataset Invited Reviewed

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Shota Nakamura, Masahiro Oda, Kensaku Mori

LNCS12965 page： 24 - 33 2021.9

Intestine segmentation with small computational cost for diagnosis assistance of ileus and intestinal obstruction Invited Reviewed

Hirohisa Oda, Yuichiro Hayashi, Takayuki Kitasaka, Aitaro Takimoto, Akinari Hinoki, Hiroo Uchida, Kojiro Suzuki, Masahiro Oda, Kensaku Mori

LNCS 12969 page： 3 - 12 2021.9

泌尿器科領域における画像処理

森健策

泌尿器科 Vol. 14 ( 2 ) page： 213 - 220 2021.8

Can artificial intelligence help to detect dysplasia in patients with ulcerative colitis? Invited International journal

Yasuharu Maeda, Shin-Ei Kudo, Noriyuki Ogata, Masashi Misawa, Yuichi Mori, Kensaku Mori, Kazuo Ohtsuka

Endoscopy Vol. 53 ( 07 ) page： E273 - E274 2021.7

Micro-CT-assisted cross-modality super-resolution of clinical CT: utilization of synthesized training dataset Reviewed

T. Zheng, H. Oda, S. Nakamura, M. Mori, H. Takabatake, H. Natori, M. Oda, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 12 - 14 2021.6

Unsupervised colonoscopic depth estimation by domain translations with a Lambertian-reflection keeping auxiliary task Invited Reviewed

Hayato Itoh, Masahiro Oda, Yuichi Mori, Masashi Misawa, Shin-Ei Kudo, Kenichiro Imai, Sayo Ito, Kinichi Hotta, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY Vol. 16 ( 6 ) page： 989 - 1001 2021.6

Language：English Publishing type：Research paper (scientific journal) Publisher：SPRINGER HEIDELBERG

Purpose A three-dimensional (3D) structure extraction technique viewed from a two-dimensional image is essential for the development of a computer-aided diagnosis (CAD) system for colonoscopy. However, a straightforward application of existing depth-estimation methods to colonoscopic images is impossible or inappropriate due to several limitations of colonoscopes. In particular, the absence of ground-truth depth for colonoscopic images hinders the application of supervised machine learning methods. To circumvent these difficulties, we developed an unsupervised and accurate depth-estimation method. Method We propose a novel unsupervised depth-estimation method by introducing a Lambertian-reflection model as an auxiliary task to domain translation between real and virtual colonoscopic images. This auxiliary task contributes to accurate depth estimation by maintaining the Lambertian-reflection assumption. In our experiments, we qualitatively evaluate the proposed method by comparing it with state-of-the-art unsupervised methods. Furthermore, we present two quantitative evaluations of the proposed method using a measuring device, as well as a new 3D reconstruction technique and measured polyp sizes. Results Our proposed method achieved accurate depth estimation with an average estimation error of less than 1 mm for regions close to the colonoscope in both of two types of quantitative evaluations. Qualitative evaluation showed that the introduced auxiliary task reduces the effects of specular reflections and colon wall textures on depth estimation and our proposed method achieved smooth depth estimation without noise, thus validating the proposed method. Conclusions We developed an accurate depth-estimation method with a new type of unsupervised domain translation with the auxiliary task. This method is useful for analysis of colonoscopic images and for the development of a CAD system since it can extract accurate 3D information.

DOI： 10.1007/s11548-021-02398-x

AIによる内視鏡外科手術支援と開発基盤としての手術動画データベース構築 Invited Reviewed

竹下修由, 森健策, 伊藤正昭

消化器外科 Vol. 44 ( 7 ) page： 1159 - 1166 2021.6

Three-dimensional surgical plan printing for assisting liver surgery, Reviewed

Y. Hayashi, T. Igami, Y. Nakamura, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 104 - 106 2021.6

COVID-19 lung infection and normal region segmentation from CT volumes using FCN with local and global spatial feature encoder Reviewed

M. Oda, Y. Hayashi, Y. Otake, M. Hashimoto, T. Akashi, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 19 - 20 2021.6

Experimental evaluation of loss functions in YOLO-v3 training for the perforation detection and localization in colonoscopic videos Reviewed

K. Jiang, H. Itoh, M. Oda, T. Okumura, Y. Mori, M. Misawa, T. Hayashi, S. E. Kudo, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 74 - 75 2021.6

Blood vessel regions segmentation from laparoscopic videos using fully convolutional networks with multi field of view input Reviewed

K. Mori, S. Morimitsu, S. Yamamoto, T. Ozawa, T. Kitasaka, Y. Hayashi, M. Oda, M. Ito, N. Takeshita, K. Misawa

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 56 - 57 2021.6

`Intestine segmentation combining Watershed transformation and machine learning-based distance map estimation Reviewed

H. Oda, Y. Hayashi, T. Kitasaka, Y. Tamada, A. Takimoto, A. Hinoki, H. Uchida, K. Suzuki, H. Itoh, M. Oda, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 89 - 90 2021.6

Real-time deformation simulation of hollow organs based on XPBD with small time steps and air mesh for surgical simulation Reviewed

S. Li, Y. Hayashi, M. Oda, K. Misawa, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 21 - 25 2021.6

Spatial Information Considered Module based on Attention Mechanism for Self-Supervised Depth Estimation from Laparoscopic Image Pairs Reviewed

W. Li, Y. Hayashi, M. Oda, T. Kitasaka, K. Misawa, K. Mori

nternational Journal of Computer Assisted Radiology and Surgery Vol. 16 ( sup.1 ) page： 45 - 46 2021.6

機械学習によるCOVID-19症例CT画像の診断支援 Reviewed

森　健策

映像情報メディア学会誌 Vol. 75 ( 3 ) page： 326 - 329 2021.5

Development of a computer-aided detection system for colonoscopy and a publicly accessible large colonoscopy video database (with video). Invited Reviewed International journal

Masashi Misawa, Shin-Ei Kudo, Yuichi Mori, Kinichi Hotta, Kazuo Ohtsuka, Takahisa Matsuda, Shoichi Saito, Toyoki Kudo, Toshiyuki Baba, Fumio Ishida, Hayato Itoh, Masahiro Oda, Kensaku Mori

Gastrointestinal endoscopy Vol. 93 ( 4 ) page： 960 - 967 2021.4

Language：English Publishing type：Research paper (scientific journal)

BACKGROUND AND AIMS: Artificial intelligence (AI)-assisted polyp detection systems for colonoscopic use are currently attracting attention because they may reduce the possibility of missed adenomas. However, few systems have the necessary regulatory approval for use in clinical practice. We aimed to develop an AI-assisted polyp detection system and to validate its performance using a large colonoscopy video database designed to be publicly accessible. METHODS: To develop the deep learning-based AI system, 56,668 independent colonoscopy images were obtained from 5 centers for use as training images. To validate the trained AI system, consecutive colonoscopy videos taken at a university hospital between October 2018 and January 2019 were searched to construct a database containing polyps with unbiased variance. All images were annotated by endoscopists according to the presence or absence of polyps and the polyps' locations with bounding boxes. RESULTS: A total of 1405 videos acquired during the study period were identified for the validation database, 797 of which contained at least 1 polyp. Of these, 100 videos containing 100 independent polyps and 13 videos negative for polyps were randomly extracted, resulting in 152,560 frames (49,799 positive frames and 102,761 negative frames) for the database. The AI showed 90.5% sensitivity and 93.7% specificity for frame-based analysis. The per-polyp sensitivities for all, diminutive, protruded, and flat polyps were 98.0%, 98.3%, 98.5%, and 97.0%, respectively. CONCLUSIONS: Our trained AI system was validated with a new large publicly accessible colonoscopy database and could identify colorectal lesions with high sensitivity and specificity. (Clinical trial registration number: UMIN 000037064.).

DOI： 10.1016/j.gie.2020.07.060

Contrastive Learningを用いた肺野CT画像からCOVID-19の自動判定 Invited Reviewed

加藤聡太, 堀田一弘, 小田昌宏, 森健策, 大竹義人, 橋本正弘, 明石敏昭

電子情報通信学会技術研究報告(MI) Vol. 120 ( 432 ) page： 82 - 86 2021.3

カスケードCNNによる腹腔鏡動画からの出血領域自動抽出 Invited Reviewed

山本翔太, 林雄一郎, 盛満慎太郎, 北坂孝幸, 小田昌宏, 竹下修由, 伊藤雅昭, 森健策

電子情報通信学会技術研究報告(MI) Vol. 120 ( 431 ) page： 172 - 175 2021.3

Spectral-based Convolutional Graph Neural Networksを用いた腹部動脈領域の血管名自動命名に関する研究 Invited Reviewed

日比裕太, 林雄一郎, 北坂孝幸, 伊東隼人, 小田昌宏, 三澤一成, 森健策

電子情報通信学会技術研究報告(MI) Vol. 120 ( 431 ) page： 176 - 181 2021.3

Artificial Intelligence System to Determine Risk of T1 Colorectal Cancer Metastasis to Lymph Node. Invited Reviewed International journal

Shin-Ei Kudo, Katsuro Ichimasa, Benjamin Villard, Yuichi Mori, Masashi Misawa, Shoichi Saito, Kinichi Hotta, Yutaka Saito, Takahisa Matsuda, Kazutaka Yamada, Toshifumi Mitani, Kazuo Ohtsuka, Akiko Chino, Daisuke Ide, Kenichiro Imai, Yoshihiro Kishida, Keiko Nakamura, Yasumitsu Saiki, Masafumi Tanaka, Shu Hoteya, Satoshi Yamashita, Yusuke Kinugasa, Masayoshi Fukuda, Toyoki Kudo, Hideyuki Miyachi, Fumio Ishida, Hayato Itoh, Masahiro Oda, Kensaku Mori

Gastroenterology Vol. 160 ( 4 ) page： 1075 - 1084.e2 2021.3

Language：English Publishing type：Research paper (scientific journal)

BACKGROUND & AIMS: In accordance with guidelines, most patients with T1 colorectal cancers (CRC) undergo surgical resection with lymph node dissection, despite the low incidence (∼10%) of metastasis to lymph nodes. To reduce unnecessary surgical resections, we used artificial intelligence to build a model to identify T1 colorectal tumors at risk for metastasis to lymph node and validated the model in a separate set of patients. METHODS: We collected data from 3134 patients with T1 CRC treated at 6 hospitals in Japan from April 1997 through September 2017 (training cohort). We developed a machine-learning artificial neural network (ANN) using data on patients' age and sex, as well as tumor size, location, morphology, lymphatic and vascular invasion, and histologic grade. We then conducted the external validation on the ANN model using independent 939 patients at another hospital during the same period (validation cohort). We calculated areas under the receiver operator characteristics curves (AUCs) for the ability of the model and US guidelines to identify patients with lymph node metastases. RESULTS: Lymph node metastases were found in 319 (10.2%) of 3134 patients in the training cohort and 79 (8.4%) of /939 patients in the validation cohort. In the validation cohort, the ANN model identified patients with lymph node metastases with an AUC of 0.83, whereas the guidelines identified patients with lymph node metastases with an AUC of 0.73 (P < .001). When the analysis was limited to patients with initial endoscopic resection (n = 517), the ANN model identified patients with lymph node metastases with an AUC of 0.84 and the guidelines identified these patients with an AUC of 0.77 (P = .005). CONCLUSIONS: The ANN model outperformed guidelines in identifying patients with T1 CRCs who had lymph node metastases. This model might be used to determine which patients require additional surgery after endoscopic resection of T1 CRCs. UMIN Clinical Trials Registry no: UMIN000038609.

DOI： 10.1053/j.gastro.2020.09.027

Label cleaning and propagation for improved segmentation performance using fully convolutional networks. Invited Reviewed International journal

Takaaki Sugino, Yutaro Suzuki, Taichi Kin, Nobuhito Saito, Shinya Onogi, Toshihiro Kawase, Kensaku Mori, Yoshikazu Nakajima

International journal of computer assisted radiology and surgery Vol. 16 ( 3 ) page： 349 - 361 2021.3

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: In recent years, fully convolutional networks (FCNs) have been applied to various medical image segmentation tasks. However, it is difficult to generate a large amount of high-quality annotation data to train FCNs for medical image segmentation. Thus, it is desired to achieve high segmentation performances even from incomplete training data. We aim to evaluate performance of FCNs to clean noises and interpolate labels from noisy and sparsely given label images. METHODS: To evaluate the label cleaning and propagation performance of FCNs, we used 2D and 3D FCNs to perform volumetric brain segmentation from magnetic resonance image volumes, based on network training on incomplete training datasets from noisy and sparse annotation. RESULTS: The experimental results using pseudo-incomplete training data showed that both 2D and 3D FCNs could provide improved segmentation results from the incomplete training data, especially by using three orthogonal annotation images for network training. CONCLUSION: This paper presented a validation for label cleaning and propagation based on FCNs. FCNs might have the potential to achieve improved segmentation performances even from sparse annotation data including possible noises by manual annotation, which can be an important clue to more efficient annotation.

DOI： 10.1007/s11548-021-02312-5

Unsupervised segmentation of COVID-19 infected lung clinical CT volumes using image inpainting and representation learning Reviewed

Tong Zheng, Masahiro Oda, Chenglong Wang, Takayasu Moriya, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Kensaku Mori

Proc. SPIE 11596, Medical Imaging, 2021: Image Processing page： 115963F-1 - 6 2021.2

Extremely imbalanced subarachnoid hemorrhage detection based on enseNet-LSTM network with class-balanced loss and transfer learning Reviewed

Zhongyang Lu, Masahiro Oda, Yuichiro Hayashi, Tao Hu, Hayato Itoh, Takeyuki Watadani, Osamu Abe, Masahiro Hashimoto, Masahiro Jinzaki, Kensaku Mori

Proceedings Volume 11597, Medical Imaging 2021: Computer-Aided Diagnosis page： 115971Z 2021.2

Single-shot three-dimensional reconstruction for colonoscopic image analysis Reviewed

Hayato Itoh, Masahiro Oda, Yuichi Mori, Masashi Misawa, Shin-ei Kudo, Kinnichi Hotta, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Proc. SPIE 11598, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling page： 115980E-1 - 6 2021.2

Intestinal region reconstruction of ileus cases from 3D CT images based on graphical representation and its visualization Reviewed

Hirohisa Oda, Yuichiro Hayashi, Takayuki Kitasaka, Yudai Tamada, Aitaro Takimoto, Akinari Hinoki, Hiroo Uchida, Kojiro Suzuki, Hayato Itoh, Masahiro Oda, Kensaku Mori

Proc.SPIE 11597, Medical Imaging 2021: Computer-Aided Diagnosis 2021.2

Lung infection and normal region segmentation from CT volumes of COVID-19 cases Reviewed

Masahiro Oda, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Kensaku Mori

Proc. SPIE 11597, Medical Imaging 2021: Computer-Aided Diagnosis page： 115972X-1 - 6 2021.2

Extraction of lung and lesion regions from COVID-19 CT volumes using 3D fully convolutional networks Reviewed

Yuichiro Hayashi, Masahiro Oda, Chen Shen, Masahiro Hashimoto, Yoshito Otake, Toshiaki Akashi, Kensaku Mori

Proc.SPIE 11597, Medical Imaging 2021: Computer-Aided Diagnosis page： 115972A-1 - 6 2021.2

人工知能（AI）最新動向　AI研究から見たRSNA-AIの広がりを感じる大会 Invited

森健策

インナービジョン Vol. 36 ( 2 ) page： 25 - 26 2021.2

CT画像XAI技術で「新型コロナウィルス肺炎」を85%精度で識別

小田昌宏, 森健策

C-press Vol. 120 page： 5 - 6 2021.2

COVID-19診断支援AI開発における名古屋大学の取り組み

小田昌宏, 鄭通, 林雄一郎, 森健策

Vol. 39 ( 1 ) page： 13 - 19 2021.2

New method for the assessment of perineural invasion from perihilar cholangiocarcinoma Reviewed

Hiroshi Tanaka, Tsuyoshi Igami, Yoshie Shimoyama, Tomoki Ebata, Yukihiro Yokoyama, Kensaku Mori, Masato Nagino,

Surgery Today Vol. 51 ( 2 ) page： 136 - 143 2021.1

Current status and future perspective on artificial intelligence for lower endoscopy. Invited International journal

Masashi Misawa, Shin-Ei Kudo, Yuichi Mori, Yasuharu Maeda, Yushi Ogawa, Katsuro Ichimasa, Toyoki Kudo, Kunihiko Wakamura, Takemasa Hayashi, Hideyuki Miyachi, Toshiyuki Baba, Fumio Ishida, Hayato Itoh, Masahiro Oda, Kensaku Mori

Digestive endoscopy : official journal of the Japan Gastroenterological Endoscopy Society Vol. 33 ( 2 ) page： 273 - 284 2021.1

Language：Japanese Publishing type：Research paper (scientific journal)

The global incidence and mortality rate of colorectal cancer remains high. Colonoscopy is regarded as the gold standard examination for detecting and eradicating neoplastic lesions. However, there are some uncertainties in colonoscopy practice that are related to limitations in human performance. First, approximately one-fourth of colorectal neoplasms are missed on a single colonoscopy. Second, it is still difficult for non-experts to perform adequately regarding optical biopsy. Third, recording of some quality indicators (e.g. cecal intubation, bowel preparation, and withdrawal speed) which are related to adenoma detection rate, is sometimes incomplete. With recent improvements in machine learning techniques and advances in computer performance, artificial intelligence-assisted computer-aided diagnosis is being increasingly utilized by endoscopists. In particular, the emergence of deep-learning, data-driven machine learning techniques have made the development of computer-aided systems easier than that of conventional machine learning techniques, the former currently being considered the standard artificial intelligence engine of computer-aided diagnosis by colonoscopy. To date, computer-aided detection systems seem to have improved the rate of detection of neoplasms. Additionally, computer-aided characterization systems may have the potential to improve diagnostic accuracy in real-time clinical practice. Furthermore, some artificial intelligence-assisted systems that aim to improve the quality of colonoscopy have been reported. The implementation of computer-aided system clinical practice may provide additional benefits such as helping in educational poorly performing endoscopists and supporting real-time clinical decision-making. In this review, we have focused on computer-aided diagnosis during colonoscopy reported by gastroenterologists and discussed its status, limitations, and future prospects.

DOI： 10.1111/den.13847

Dense-layer-based YOLO-v3 for detection and localization of colon perforations Invited Reviewed

Kai Jiang, Hayato Itoh, Masahiro Oda, Taishi Okumura, Yuichi Mori, Masashi Misawa, Takemasa Hayashi, Shin-Ei Kudo, Kensaku Mori

Medical Imaging 2021: Computer-Aided Diagnosis Vol. 11597 page： 115971A-1 - 6 2021

Robust endocytoscopic image classification based on higher-order symmetric tensor analysis and multi-scale topological statistics. Invited Reviewed International journal

Hayato Itoh, Yukitaka Nimura, Yuichi Mori, Masashi Misawa, Shin-Ei Kudo, Kinichi Hotta, Kazuo Ohtsuka, Shoichi Saito, Yutaka Saito, Hiroaki Ikematsu, Yuichiro Hayashi, Masahiro Oda, Kensaku Mori

International journal of computer assisted radiology and surgery Vol. 15 ( 12 ) page： 2049 - 2059 2020.12

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: An endocytoscope is a new type of endoscope that enables users to perform conventional endoscopic observation and ultramagnified observation at the cell level. Although endocytoscopy is expected to improve the cost-effectiveness of colonoscopy, endocytoscopic image diagnosis requires much knowledge and high-level experience for physicians. To circumvent this difficulty, we developed a robust endocytoscopic (EC) image classification method for the construction of a computer-aided diagnosis (CAD) system, since real-time CAD can resolve accuracy issues and reduce interobserver variability. METHOD: We propose a novel feature extraction method by introducing higher-order symmetric tensor analysis to the computation of multi-scale topological statistics on an image, and we integrate this feature extraction with EC image classification. We experimentally evaluate the classification accuracy of our proposed method by comparing it with three deep learning methods. We conducted this comparison by using our large-scale multi-hospital dataset of about 55,000 images of over 3800 patients. RESULTS: Our proposed method achieved an average 90% classification accuracy for all the images in four hospitals even though the best deep learning method achieved 95% classification accuracy for images in only one hospital. In the case with a rejection option, the proposed method achieved expert-level accurate classification. These results demonstrate the robustness of our proposed method against pit pattern variations, including differences of colours, contrasts, shapes, and hospitals. CONCLUSIONS: We developed a robust EC image classification method with novel feature extraction. This method is useful for the construction of a practical CAD system, since it has sufficient generalisation ability.

DOI： 10.1007/s11548-020-02255-3

[総論]AI時代を見据えた消化器外科手術 AIによる大腸T2癌リンパ節転移予測 Invited Reviewed

中原健太, 石田文生, 一政克朗, 森悠一, 三澤将史, 澤田成彦, 工藤進英, Villard Ben, 伊東隼人, 森健策

日本消化器外科学会総会 Vol. 75回 page： WS15 - 6 2020.12

多元計算解剖モデルと人工知能に基づく診断治療支援 Invited Reviewed

森　健策

映像情報メディア学会誌 Vol. 74 ( 6 ) page： 909-915 2020.11

Improving contrast and spatial resolution in crystal analyzer based x ray dark field imaging Reviewed

Masami Ando Yuki Nakao Ge Jin Hiroshi Sugiyama Naoki Sunaguchi Yongjin Sung Yoshifumi Suzuki Yong Sun Michio Tanimoto Katsuhiro Kawashima Tetsuya Yuasa Kensaku Mori Shu Ichihara Rajiv Gupta,

Medical Physics Vol. 47 ( 11 ) page： 5505 - 5513 2020.11

Preliminary study of Loss-Function Design for Detection and Localization of Perforations with YOLO-v3 in Colonoscopic Images

Vol. 22 ( 4 ) page： 348 - 349 2020.11

腸閉塞およびイレウスの診断支援システムにおける距離マップの導入

小田紘久, 林雄一郎, 北坂孝幸，玉田雄大,滝本愛太朗，檜顕成, 内田広夫，鈴木耕次郎, 伊東隼人，小田昌宏，森　健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 282 - 284 2020.11

局所情報に注目した腹腔鏡動画像からの出血領域抽出

山本翔太, 盛満慎太郎，林　雄一郎，北坂　孝幸，小田　昌宏，伊藤　雅昭，竹下　修由,　森　健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 285 - 286 2020.11

Dilated convolution を用いた腹腔鏡動画像からの血管領域抽出における空間情報利用に関する検討

盛満慎太郎, 山本翔太, 小澤卓也, 北坂孝幸, 林　雄一郎, 小田昌宏, 伊藤雅昭, 竹下修由，三澤一成, 森　健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 287 - 288 2020.11

表現学習に基づくクラスタリングによるCOVID-19 肺CT像からの病変部抽出手法

鄭通, 小田昌宏, 王成龍，林　雄一郎，橋本正弘, 大竹義人，明石敏昭, 森　健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 294 - 295 2020.11

深層学習によるMRI画像からの神経鞘腫の自動位置検出

小田昌宏, 伊藤定之, 今釜史郎, 森健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 296 - 297 2020.11

腹腔鏡動画像用オンラインアノテーションツールの開発

屠芸豪, 伊東隼人，小澤卓也，小田昌宏, 竹下修由，伊藤雅昭，森健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 306 - 307 2020.11

呼吸器外科における仮想胸腔鏡像による手術ナビゲーションシステムを用いた手術支援の検討

林雄一郎, 中村彰太，森健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 337 2020.11

気管支鏡ナビゲーションのための敵対的生成による内視鏡画像深度推定の評価

王成, 小田昌宏，林雄一郎，北坂孝幸，本間裕敏，高畠博嗣，森雅樹，名取博, 森　健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 338 - 339 2020.11

SUN database : 大腸ポリープ自動検出器の精度評価に向けた試験用画像

伊東隼人, 三澤将史，森悠一，小田　昌宏，工藤　進英, 森健策

日本コンピュータ外科学会誌第29回日本コンピュータ外科学会大会特集号 Vol. 22 ( 4 ) page： 346 - 347 2020.11

Cost savings in colonoscopy with artificial intelligence-aided polyp diagnosis: an add-on analysis of a clinical trial (with video). Invited Reviewed International journal

Yuichi Mori, Shin-Ei Kudo, James E East, Amit Rastogi, Michael Bretthauer, Masashi Misawa, Masau Sekiguchi, Takahisa Matsuda, Yutaka Saito, Hiroaki Ikematsu, Kinichi Hotta, Kazuo Ohtsuka, Toyoki Kudo, Kensaku Mori

Gastrointestinal endoscopy Vol. 92 ( 4 ) page： 905 - 911 2020.10

Language：English Publishing type：Research paper (scientific journal)

BACKGROUND AND AIMS: Artificial intelligence (AI) is being implemented in colonoscopy practice, but no study has investigated whether AI is cost saving. We aimed to quantify the cost reduction using AI as an aid in the optical diagnosis of colorectal polyps. METHODS: This study is an add-on analysis of a clinical trial that investigated the performance of AI for differentiating colorectal polyps (ie, neoplastic versus non-neoplastic). We included all patients with diminutive (≤5 mm) rectosigmoid polyps in the analyses. The average colonoscopy cost was compared for 2 scenarios: (1) a diagnose-and-leave strategy supported by the AI prediction (ie, diminutive rectosigmoid polyps were not removed when predicted as non-neoplastic), and (2) a resect-all-polyps strategy. Gross annual costs for colonoscopies were also calculated based on the number and reimbursement of colonoscopies conducted under public health insurances in 4 countries. RESULTS: Overall, 207 patients with 250 diminutive rectosigmoid polyps (104 neoplastic, 144 non-neoplastic, and 2 indeterminate) were included. AI correctly differentiated neoplastic polyps with 93.3% sensitivity, 95.2% specificity, and 95.2% negative predictive value. Thus, 105 polyps were removed and 145 were left under the diagnose-and-leave strategy, which was estimated to reduce the average colonoscopy cost and the gross annual reimbursement for colonoscopies by 18.9% and US$149.2 million in Japan, 6.9% and US$12.3 million in England, 7.6% and US$1.1 million in Norway, and 10.9% and US$85.2 million in the United States, respectively, compared with the resect-all-polyps strategy. CONCLUSIONS: The use of AI to enable the diagnose-and-leave strategy results in substantial cost reductions for colonoscopy.

DOI： 10.1016/j.gie.2020.03.3759

Realistic Endoscopic Image Generation Method Using Virtual-to-real Image-domain Translation

Prediction of dose distribution from luminescence image of water using a deep convolutional neural network for particle therapy Reviewed

Takuya Yabe, Seiichi Yamamoto, Masahiro Oda, Kensaku Mori, Toshiyuki Toshito, Takashi Akagi

Medical Physics Vol. 47 ( 9 ) page： 3882-3891 2020.9

名古屋大学スーパーコンピュータ「不老」における医用画像処理

大島聡史, 小田昌宏, 片桐孝洋, 森健策

電子情報通信学会技術研究報告(MI) Vol. 120 ( 156 ) page： 69 - 74 2020.9

ニューラルネットワークとSpherical K-meansを用いた胃壁マイクロCT像からの層構造および腫瘍抽出の検討

御手洗翠, 小田紘久, 杉野貴明, 守谷享泰, 伊東隼人, 小田昌宏, 小宮山琢真, 古川和宏, 宮原良二, 藤城光弘, 森雅樹, 高畠博嗣, 名取博, 森健策

電子情報通信学会技術研究報告(MI) Vol. 120 ( 156 ) page： 1 - 6 2020.9

腹腔鏡手術動画像データベース構築に向けたリモートアノテーションツールのプロトタイプ開発

屠芸豪, 伊東隼人, 小澤卓也, 小田昌宏, 竹下修由, 伊藤雅昭, 森健策

第39回日本医用画像工学会大会予稿集 page： 611 - 615 2020.9

Preliminary Study of Perforation Detection and Localization for Colonoscopy Video

Kai Jiang, Hayato Itoh, Masahiro Oda, Taishi Okumura, Yuichi Mori, Masashi Misawa, Takemasa Hayashi, Shin-Ei Kudo, Kensaku Mori

page： 142 - 147 2020.9

読影レポート解析を利用した医用画像データベースからのアノテーション付きデータセット作成に関する初期検討

林雄一郎, 鈴村悠輝, 岡崎真治, 小田昌宏, 森健策

第39回日本医用画像工学会大会予稿集 page： 163 - 167 2020.9

A study on Subarachnoid Hemorrhage automatic detection utilized Transfer Learning on extremely imbalanced brain CT datasets

Zhongyang Lu, Masahiro Oda, Yuichiro Hayashi, Tao Hu, Hayato Ito，Takeyuki Watadani，Osamu Abe，Masahiro Hashimoto，Masahiro Jinzaki，Kensaku Mori

page： 168 - 172 2020.9

COVID-19 症例の定量評価のためのCT 像からの肺野自動セグメンテーション

小田昌宏, 林雄一郎, 大竹義人, 橋本正弘, 明石敏昭, 森健策

第39回日本医用画像工学会大会予稿集 page： 181 - 184 2020.9

Preliminary Study on Classification of Interstitial Cystitis Using Cystoscopy Images

Tao Chu, Masahiro Oda, Akira Furuta, Tokunori Yamamoto, Kensaku Mori

page： 186 - 191 2020.9

Dilated convolution を用いた FCN による腹腔鏡動画像からの血管領域抽出

盛満慎太郎, 山本翔太, 北坂孝幸, 林雄一郎, 小田昌宏, 竹下修由, 伊藤雅昭, 三澤一成, 森健策

第39回日本医用画像工学会大会予稿集 page： 230 - 233 2020.9

大域及び局所情報を用いた深層学習による出血領域自動セグメンテーション

山本翔太, 盛満慎太郎, 林雄一郎, 北坂孝幸, 小田昌宏, 伊藤雅昭, 竹下修由, 森健策

第39回日本医用画像工学会大会予稿集 page： 246 - 249 2020.9

広範囲の隣接関係を考慮したグラフニューラルネットワークを用いた腹部動脈血管名自動命名の検討

日比裕太, 林雄一郎, 北坂孝幸, 伊東隼人, 小田昌宏, 三澤一成, 森健策

第39回日本医用画像工学会大会予稿集 page： 268 - 271 2020.9

Cross-phase CT Image Registration Using Convolutional Neural Network

Tao Hu, Masahiro Oda, Yuichiro Hayashi, Zhongyang Lu, Kanako Kunishishima Kumamaru, Shigeki Aoki, Kensaku Mori

page： 276 - 280 2020.9

Unsupervised 3D Super-resolution of Clinical CT Volumes by Utilizing Multi-axis 2D Super-resolution

第39回日本医用画像工学会大会予稿集 page： 377 - 384 2020.9

Preliminary Study on Classification of Hands' Bone Marrow Edema Using X-ray Images

Dongping Pan, Masahiro Oda, Kou Katayama, Takanobu Okubo, Kensaku Mori

page： 488 - 493 2020.9

大腸内視鏡のための教師なし深度画像推定法における補助タスク検討

伊東隼人, 小田昌宏, 森悠一, 三澤将史, 工藤進英, 堀田欣一, 高畠博嗣, 森雅樹, 名取博, 森健策

第39回日本医用画像工学会大会予稿集 page： 563 - 568 2020.9

Spatial Information Considered Self-Supervised Depth Estimation Based on Image Pairs from Stereo Laparoscope

Wenda Li, Yuichiro Hayashi, Masahiro Oda, Takayuki Kitasaka, Kazunari Misawa, Kensaku Mori

page： 602 - 606 2020.9

泌尿器科画像診断に用いられるAI技術とその応用

森　健策

泌尿器外科 Vol. 33 ( 6 ) page： 557-561 2020.6

Detecting ganglion cells on virtual slide images: Macroscopic masking by superpixel Reviewed

H. Oda, Y. Tamada, K. Nishio, T. Kitasaka, H. Amano, K. Chiba, A. Hinoki, H. Uchida, M. Oda, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S169 - S170 2020.6

An Application of Multi-organ Segmentation from Thick-slice Abdominal CT Volumes using Transfer Learning Reviewed

C. Shen, M. Oda, H. Roth, H. Oda, Y. Hayashi, K. Misawa, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S17 - S18 2020.6

Virtual cleansing by unpaired image translation of intestines for detecting obstruction Reviewed

K. Nishio, H. Oda, T. Kitasaka, Y. Tamada, H. Amano, A. Takimoto, K. Chiba, Y. Hayashi, H. Itoh, M. Oda, A. Hinoki, H. Uchida, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S21 - S22 2020.6

TinyLoss: loss function for tiny image difference evaluation and its application to unpaired non-contrast to contrast abdominal CT estimation Reviewed

M. Oda, T. Hu, K. K. Kumamaru, T. Akashi, S. Aoki, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S25 - S26 2020.6

SR-CycleGAN V2: CycleGAN-based unsupervised superresolution with pixel-shuffling Reviewed

T. Zheng, H. Oda, T. Moriya, T. Sugino, S. Nakamura, M. Oda, M. Mori, H. Takabatake, H. Natori, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S27 - S28 2020.6

Blood vessel segmentation from laparoscopic video using ConvLSTM U-Net

S. Morimitsu, S. Yamamoto, T. Ozawa, T. Kitasaka, Y. Hayashi, M. Oda, M. Ito, N. Takeshita, K. Misawa, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S63 - S64 2020.6

Extraction of blood vessel regions in liver from CT volumes using fully convolutional networks for computer assisted liver surgery Reviewed

Y. Hayashi, C. Shen, T. Igami, M. Nagino, K. Mori

International Journal of Computer Assisted Radiology and Surgery Vol. 15 ( 1 ) page： S152 - S153 2020.6

AIによるSSA/Pの超拡大内視鏡診断 Invited Reviewed

小川悠史, 工藤進英, 森悠一, 三澤将史, 片岡伸一, 前田康晴, 一政克朗, 石垣智之, 工藤豊樹, 若村邦彦, 林武雅, 馬場俊之, 石田文生, 伊東隼人, 小田昌宏, 森健策

日本大腸検査学会雑誌 Vol. 36 ( 2 ) page： 125 - 125 2020.5

バーチャル・リアリティ手術シュミレーター（VRS）の意義と今後の展望

藤原道隆, 林雄一郎, 高見秀樹, 田中千恵, 森健策, 小寺泰弘

臨床外科 Vol. 75 ( 4 ) page： 476 - 482 2020.4

Cardiac fiber tracking on super high-resolution CT images: a comparative study. Invited International journal

Hirohisa Oda, Holger R Roth, Takaaki Sugino, Naoki Sunaguchi, Noriko Usami, Masahiro Oda, Daisuke Shimao, Shu Ichihara, Tetsuya Yuasa, Masami Ando, Toshiaki Akita, Yuji Narita, Kensaku Mori

Journal of medical imaging (Bellingham, Wash.) Vol. 7 ( 2 ) page： 026001 - 026001 2020.3

Language：Japanese Publishing type：Research paper (scientific journal)

Purpose: High-resolution cardiac imaging and fiber analysis methods are required to understand cardiac anatomy. Although refraction-contrast x-ray CT (RCT) has high soft tissue contrast, it cannot be commonly used because it requires a synchrotron system. Microfocus x-ray CT ( μ CT ) is another commercially available imaging modality. Approach: We evaluate the usefulness of μ CT for analyzing fibers by quantitatively and objectively comparing the results with RCT. To do so, we scanned a rabbit heart by both modalities with our original protocol of prepared materials and compared their image-based analysis results, including fiber orientation estimation and fiber tracking. Results: Fiber orientations estimated by two modalities were closely resembled under the correlation coefficient of 0.63. Tracked fibers from both modalities matched well the anatomical knowledge that fiber orientations are different inside and outside of the left ventricle. However, the μ CT volume caused incorrect tracking around the boundaries caused by stitching scanning. Conclusions: Our experimental results demonstrated that μ CT scanning can be used for cardiac fiber analysis, although further investigation is required in the differences of fiber analysis results on RCT and μ CT .

DOI： 10.1117/1.JMI.7.2.026001

3Dプリンティングの最新動向 Invited

森　健策

インナービジョン Vol. 35 ( 2 ) page： 36-37 2020.2

Spatial information-embedded fully convolutional networks for multi-organ segmentation with improved data augmentation and instance normalization Reviewed

Chen Shen, Chenglong Wang, Holger R. Roth, Masahiro Oda, Yuichiro Hayashi, Kazunari Misawa, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11313 2020.2

Organ segmentation from full-size CT images using memory-efficient FCN Reviewed

Chenglong Wang, Masahiro Oda, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11314 2020.2

Multi-modality super-resolution loss for GAN-based super-resolution of clinical CT images using micro CT image database Reviewed

Tong Zheng, Hirohisa Oda, Takayasu Moriya, Takaaki Sugino, Shota Nakamura, Masahiro Oda, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11313 2020.2

Visualizing intestines for diagnostic assistance of ileus based on intestinal region segmentation from 3D CT images Reviewed

Hirohisa Oda, Kohei Nishio, Takayuki Kitasaka, Hizuru Amano, Aitaro Takimoto, Akinari Hinoki, Hiroo Uchida, Kojiro Suzuki, Hayato Itoh, Masahiro Oda, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11314 2020.2

Visualising decision-reasoning regions in computer-aided pathological pattern diagnosis of endoscytoscopic images based on CNN weights analysis Reviewed

Hayato Itoh, Zhongyang Lu, Yuichi Mori, Masashi Misawa, Masahiro Oda, Shin-ei Kudo, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11314 2020.2

Usefulness of fine-tuning for deep learning based multi-organ regions segmentation method from non-contrast CT volumes using small training dataset Reviewed

Yuichiro Hayashi, Chen Shen, Holger R. Roth, Masahiro Oda, Kazunari Misawa, Masahiro Jinzaki, Masahiro Hashimoto, Kanako K. Kumamaru, Shigeki Aoki, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11314 2020.2

Automated eye disease classification method from anterior eye image using anatomical structure focused image classification technique Reviewed

Masahiro Oda, Naoyuki Maeda, Takefumi Yamaguchi, Hideki Fukuoka, Yuta Ueno, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11314 2020.2

Improved visual SLAM for bronchoscope tracking and registration with pre-operative CT images

Cheng Wang, Masahiro Oda, Yuichiro Hayashi, Takayuki Kitasaka, Hirotoshi Honma, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Medical Imaging 2020: Image Processing Vol. 11315 2020.2

CycleGANによる腸管電子洗浄とその腸管閉塞部位検出への応用

西尾光平, 小田紘久, 千馬耕亮, 北坂孝幸, 林雄一郎, 伊東隼人, 小田昌宏, 檜顕成, 内田広夫, 森健策

電子情報通信学会技術研究報告(MI) Vol. 119 ( 399 ) page： 243 - 248 2020.1

臨床肺CT画像と切除肺マイクロCT画像の非剛体位置合わせ手法の検討

波多腰慎矢, 小田紘久, 林雄一郎, Holger R. Roth, 中村彰太, 小田昌宏, 森雅樹, 高畠博嗣, 名取博, 森健策

電子情報通信学会技術研究報告(MI) Vol. 119 ( 399 ) page： 249 - 254 2020.1

MICCAI 2019参加報告

小田昌宏, 伊東隼人, 宮内翔子, 諸岡健一, 松崎博貴, 花岡昇平, 古川亮, 増谷佳孝, 森健策

電子情報通信学会技術研究報告(MI) Vol. 119 ( 399 ) page： 219 - 226 2020.1

仮想腹腔鏡画像生成と深層学習による腹腔鏡画像からの術具領域セグメンテーション

小澤卓也, 林雄一郎, 小田紘久, 小田昌宏, 北坂孝幸, 竹下修由, 伊藤雅昭, 森健策

電子情報通信学会技術研究報告(MI) Vol. 119 ( 399 ) page： 129 - 134 2020.1

腹腔鏡下手術支援のためのU-Netに基づく腹腔鏡動画像からの出血領域の推定

山本翔太, 林雄一郎, 盛満慎太郎, 小澤卓也, 北坂孝幸, 小田昌宏, 竹下修由, 伊藤雅昭, 森健策

電子情報通信学会技術研究報告(MI) Vol. 119 ( 399 ) page： 209 - 214 2020.1

Stable polyp-scene classification via subsampling and residual learning from an imbalanced large dataset. International journal

Hayato Itoh, Holger Roth, Masahiro Oda, Masashi Misawa, Yuichi Mori, Shin-Ei Kudo, Kensaku Mori

Healthcare technology letters Vol. 6 ( 6 ) page： 237 - 242 2019.12

Language：English Publishing type：Research paper (scientific journal)

This Letter presents a stable polyp-scene classification method with low false positive (FP) detection. Precise automated polyp detection during colonoscopies is essential for preventing colon-cancer deaths. There is, therefore, a demand for a computer-assisted diagnosis (CAD) system for colonoscopies to assist colonoscopists. A high-performance CAD system with spatiotemporal feature extraction via a three-dimensional convolutional neural network (3D CNN) with a limited dataset achieved about 80% detection accuracy in actual colonoscopic videos. Consequently, further improvement of a 3D CNN with larger training data is feasible. However, the ratio between polyp and non-polyp scenes is quite imbalanced in a large colonoscopic video dataset. This imbalance leads to unstable polyp detection. To circumvent this, the authors propose an efficient and balanced learning technique for deep residual learning. The authors' method randomly selects a subset of non-polyp scenes whose number is the same number of still images of polyp scenes at the beginning of each epoch of learning. Furthermore, they introduce post-processing for stable polyp-scene classification. This post-processing reduces the FPs that occur in the practical application of polyp-scene classification. They evaluate several residual networks with a large polyp-detection dataset consisting of 1027 colonoscopic videos. In the scene-level evaluation, their proposed method achieves stable polyp-scene classification with 0.86 sensitivity and 0.97 specificity.

DOI： 10.1049/htl.2019.0079

Masahiro Oda, Kiyohito Tanaka, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

MICCAI 2019 Healthcare Technology Letters Vol. 6 ( 6 ) page： 214 - 219 2019.12

Realistic Endoscopic Image Generation Method Using Virtual-to-real Image-domain Translation

Masahiro Oda, Kiyohito Tanaka, Hirotsugu Takabatake, Masaki Mori, Hiroshi Natori, Kensaku Mori

Healthcare Technology Letters Vol. 6 ( 6 ) page： 214-219 2019.12

Generative Adversarial Networks Showcase: Their Mechanisms and Radiological Applications

Masahiro Oda, Hirohisa Oda, Kanako K. Kumamaru, Shigeki Aoki, Hiroshi Natori, Kensaku Mori, Masaki Mori, Hirotsugu Takabatake

RSNA2019 page： AI020-EB-X 2019.12

Automatic Quantitative Analysis of Kidney Tumor Using 3D Fully Convolutional Network

Chenglong Wang, Masahiro Oda, Yuichiro Hayashi, Naoto Sassa, Tokunori Yamamoto, Kensaku Mori

RSNA2019 page： UR002-EB-X 2019.12

Technique for Improving Accuracy of Deep Learning-based Multi-Organ Segmentation from CT Volumes

Chen Shen, Hirohisa Oda, MENG, Masahiro Oda, Holger R. Roth, Yuichiro Hayashi, Kensaku Mori, Takayuki Kitasaka, Kazunari Misawa

RSNA2019 page： N021-EC-X 2019.12

Micro Lung Cancer Analysis Based on Micro CT Imaging Using Generative Adversarial Network

Kensaku Mori, Takayasu Moriya, Hirohisa Oda, MENG , Midori Mitarai, Masahiro Oda, Shota Nakamura, Takaaki Sugino, Holger R. Roth

RSNA2019 page： CH007-EC-X 2019.12

大腸肛門病理学におけるAI利用の将来像

森健策

日本大腸肛門病学学術集会　抄録号 page： A17 2019.10

機械学習を用いた医療支援

森健策

第84回日本泌尿器気学会東部総会, プログラム抄録集 page： 123 2019.10

胸部領域AIの歴史と今後-歴史的研究を振り返りながら今後を展望する-

森健策

臨床画像 Vol. 35 ( 10 ) page： 1139-1149 2019.10

Unsupervised Segmentation of Micro-CT Images of Lung Cancer Specimen Using Deep Generative Models

Takayasu Moriya, Hirohisa Oda, Midori Mitarai, Shota Nakamura, Holger R. Roth, Masahiro Oda, Kensaku Mori

MICCAI 2019 Vol. LNCS 11769 page： 240-248 2019.10

Tubular Structure Segmentation Using Spatial Fully Connected Network With Radial Distance Loss for 3D Medical images

Chenglong Wang, Yuichiro Hayashi, Masahiro Oda, Hayato Itoh, Takayuki Kitasaka, Alejandro Frangi, Kensaku Mori

MICCAI 2019 Vol. LNCS 11769 page： 348-356 2019.10

Intelligent Image Synthesis to Attack a egmentation CNN Using Adversarial Learning

Liang Chen, Paul Bentley, Kensaku Mori, Kazunari Misawa, Michitaka Fujiwara, Daniel Rueckert

MICCAI 2019 Vol. LNCS 11827 page： 90-99 2019.10

Spaciousness Filters for Non-contrast CT Volume Segmentation of the Intestine Region for Emergency Ileus Diagnosis

Hirohisa Oda, Kohei Nishio, Takayuki Kitasaka, Benjamin Villard, Hizuru Amano, Kosuke Chiba, Akinari Hinoki, Hiroo Uchida, Kojiro Suzuki, Hayato Itoh, Masahiro Oda, Kensaku Mori

MICCAI 2019 Vol. LNCS 11840 page： 104-114 2019.10

人工知能時代の医療を考える

森　健策

EAJ NEWS「AI×医療」特集号 ( 181 ) page： 6-8 2019.10

Precise estimation of renal vascular dominant regions using spatially aware fully convolutional networks, tensor-cut and Voronoi diagrams

Chenglong Wang, Holger R. Roth, Takayuki Kitasaka, Masahiro Oda, Yuichiro Hayashi, Yasushi Yoshino, Tokunori Yamamoto, Naoto Sassa, Momokazu Goto, Kensaku Mori

Computerized Medical Imaging and Graphics Vol. 77 ( 10642 ) page： 1-13 2019.10

Spaciousness Filters for Non-contrast CT Volume Segmentation of the Intestine Region for Emergency Ileus Diagnosis

Hirohisa Oda, Kohei Nishio, Takayuki Kitasaka, Benjamin Villard, Hizuru Amano, Kosuke Chiba, Akinari Hinoki, Hiroo Uchida, Kojiro Suzuki, Hayato Itoh, Masahiro Oda, Kensaku Mori

LNCS 11840 page： 104-114 2019.10

Unsupervised Segmentation of Micro-CT Images of Lung Cancer Specimen Using Deep Generative Models

Takayasu Moriya, Hirohisa Oda, Midori Mitarai, Shota Nakamura, Holger R. Roth, Masahiro Oda, Kensaku Mori,

LNCS11769 page： 240-248 2019.10

Tubular Structure Segmentation Using Spatial Fully Connected Network With Radial Distance Loss for 3D Medical images

Chenglong Wang, Yuichiro Hayashi, Masahiro Oda, Hayato Itoh, Takayuki Kitasaka, Alejandro Frangi, Kensaku Mori

LNCS11769 page： 348-356 2019.10

Intelligent Image Synthesis to Attack a segmentation CNN Using Adversarial Learning

Liang Chen, Paul Bentley, Kensaku Mori, Kazunari Misawa, Michitaka Fujiwara, Daniel Rueckert,

LNCS11827 page： 90-99 2019.10

Artificial Intelligence-assisted System Improves Endoscopic Identification of Colorectal Neoplasms. Reviewed International journal

Shin-Ei Kudo, Masashi Misawa, Yuichi Mori, Kinichi Hotta, Kazuo Ohtsuka, Hiroaki Ikematsu, Yutaka Saito, Kenichi Takeda, Hiroki Nakamura, Katsuro Ichimasa, Tomoyuki Ishigaki, Naoya Toyoshima, Toyoki Kudo, Takemasa Hayashi, Kunihiko Wakamura, Toshiyuki Baba, Ishida Fumio, Haruhiro Inoue, Hayato Itoh, Masahiro Oda, Kensaku Mori

Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association Vol. 18 ( 8 ) page： 1874 - 1881 2019.9

Language：English Publishing type：Research paper (scientific journal)

BACKGROUND & AIMS: Precise optical diagnosis of colorectal polyps could improve the cost-effectiveness of colonoscopy and reduce polypectomy-related complications. However, it is difficult for community-based non-experts to obtain sufficient diagnostic performance. Artificial intelligence-based systems have been developed to analyze endoscopic images; they identify neoplasms with high accuracy and low interobserver variation. We performed a multi-center study to determine the diagnostic accuracy of EndoBRAIN, an artificial intelligence-based system that analyzes cell nuclei, crypt structure, and microvessels in endoscopic images, in identification of colon neoplasms. METHODS: The EndoBRAIN system was initially trained using 69,142 endocytoscopic images, taken at 520-fold magnification, from patients with colorectal polyps who underwent endoscopy at 5 academic centers in Japan from October 2017 through March 2018. We performed a retrospective comparative analysis of the diagnostic performance of EndoBRAIN vs that of 30 endoscopists (20 trainees and 10 experts); the endoscopists assessed images from 100 cases produced via white-light microscopy, endocytoscopy with methylene blue staining, and endocytoscopy with narrow-band imaging. EndoBRAIN was used to assess endocytoscopic, but not white-light, images. The primary outcome was the accuracy of EndoBrain in distinguishing neoplasms from non-neoplasms, compared with that of endoscopists, using findings from pathology analysis as the reference standard. RESULTS: In analysis of stained endocytoscopic images, EndoBRAIN identified colon lesions with 96.9% sensitivity (95% CI, 95.8%-97.8%), 100% specificity (95% CI, 99.6%-100%), 98% accuracy (95% CI, 97.3%-98.6%), a 100% positive-predictive value (95% CI, 99.8%-100%), and a 94.6% negative-predictive (95% CI, 92.7%-96.1%); these values were all significantly greater than those of the endoscopy trainees and experts. In analysis of narrow-band images, EndoBRAIN distinguished neoplastic from non-neoplastic lesions with 96.9% sensitivity (95% CI, 95.8-97.8), 94.3% specificity (95% CI, 92.3-95.9), 96.0% accuracy (95% CI, 95.1-96.8), a 96.9% positive-predictive value, (95% CI, 95.8-97.8), and a 94.3% negative-predictive value (95% CI, 92.3-95.9); these values were all significantly higher than those of the endoscopy trainees, sensitivity and negative-predictive value were significantly higher but the other values are comparable to those of the experts. CONCLUSIONS: EndoBRAIN accurately differentiated neoplastic from non-neoplastic lesions in stained endocytoscopic images and endocytoscopic narrow-band images, when pathology findings were used as the standard. This technology has been authorized for clinical use by the Japanese regulatory agency and should be used in endoscopic evaluation of small polyps more widespread clinical settings. UMIN clinical trial no: UMIN000028843.

DOI： 10.1016/j.cgh.2019.09.009