Yuichi Mori, MD, PhD; Shin-ei Kudo, MD, PhD; Masashi Misawa, MD, PhD; Yutaka Saito, MD, PhD; Hiroaki Ikematsu, MD, PhD; Kinichi Hotta, MD; Kazuo Ohtsuka, MD, PhD; Fumihiko Urushibara, MD, PhD; Shinichi Kataoka, MD; Yushi Ogawa, MD; Yasuharu Maeda, MD, PhD; Kenichi Takeda, MD, PhD; Hiroki Nakamura, MD; Katsuro Ichimasa, MD, PhD; Toyoki Kudo, MD, PhD; Takemasa Hayashi, MD, PhD; Kunihiko Wakamura, MD, PhD; Fumio Ishida, MD, PhD; Haruhiro Inoue, MD, PhD; Hayato Itoh, PhD; Masahiro Oda, PhD; Kensaku Mori, PhD
Acknowledgment: The authors thank Olympus and Cybernet Systems for their instrumental support of the present study. They also thank Ms. Nancy Schatken of Edanz for her medical writing and editorial assistance, which was funded by the Japan Society for the Promotion of Science (grant 17H05305).
Grant Support: By Grants-in-Aid for Scientific Research (grant 17H05305) from the Japan Society for the Promotion of Science.
Disclosures: Dr. Y. Mori reports personal fees from Olympus and grants from The Japan Society for the Promotion of Science during the conduct of the study, grants from Japan Agency for Medical Research and Development outside the submitted work, and a patent (JP6059271) licensed to Showa University and Cybernet Systems. Drs. S. Kudo and Misawa report personal fees from Olympus during the conduct of the study, grants from Japan Agency for Medical Research and Development outside the submitted work, and a patent (JP6059271) licensed to Showa University and Cybernet Systems. Dr. Ohtsuka reports personal fees and nonfinancial support from Olympus outside the submitted work. Dr. Inoue reports personal fees from Olympus during the conduct of the study. Dr. K. Mori reports grants from Cybernet Systems during the conduct of the study; reports grants from Olympus, Ziosoft, and NTT outside the submitted work; and has been issued a patent for the Endoscope Observation Assistance System. Authors not named here have disclosed no conflicts of interest. Disclosures can also be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M18-0249.
Editors' Disclosures: Christine Laine, MD, MPH, Editor in Chief, reports that her spouse has stock options/holdings with Targeted Diagnostics and Therapeutics. Darren B. Taichman, MD, PhD, Executive Editor, reports that he has no financial relationships or interests to disclose. Cynthia D. Mulrow, MD, MSc, Senior Deputy Editor, reports that she has no relationships or interests to disclose. Deborah Cotton, MD, MPH, Deputy Editor, reports that she has no financial relationships or interest to disclose. Jaya K. Rao, MD, MHS, Deputy Editor, reports that she has stock holdings/options in Eli Lilly and Pfizer. Sankey V. Williams, MD, Deputy Editor, reports that he has no financial relationships or interests to disclose. Catharine B. Stack, PhD, MS, Deputy Editor for Statistics, reports that she has stock holdings in Pfizer and Johnson & Johnson.
Reproducible Research Statement:Study protocol: See the . Statistical code: Available by request from, and written agreement with, Dr. Y. Mori (e-mail, firstname.lastname@example.org). Data set: Not available.
Corresponding Author: Yuichi Mori, MD, PhD, Digestive Disease Center, Showa University Northern Yokohama Hospital, 35-1 Chigasaki-chuo, Tsuzuki-ku, Yokohama 224-8503, Japan; e-mail, email@example.com.
Current Author Addresses: Drs. Y. Mori, S. Kudo, Misawa, Urushibara, Kataoka, Ogawa, Maeda, Takeda, Nakamura, Ichimasa, T. Kudo, Hayashi, Wakamura, and Ishida: Digestive Disease Center, Showa University Northern Yokohama Hospital, 35-1 Chigasaki-chuo, Tsuzuki-ku, Yokohama 224-8503, Japan.
Dr. Saito: Endoscopy Division, National Cancer Center Hospital, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan.
Dr. Ikematsu: Department of Gastroenterology and Endoscopy, National Cancer Center Hospital East, 6-5-1 Kashiwanoha, Kashiwa 227-8577, Japan.
Dr. Hotta: Division of Endoscopy, Shizuoka Cancer Center, 1007 Shimonagakubo, Nagaizumi-tyo, Shunto-gun, Shizuoka-ken 411-8777, Japan.
Dr. Ohtsuka: Department of Endoscopy, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan.
Dr. Inoue: Digestive Disease Center, Showa University Koto-Toyosu Hospital, 5-1-38 Toyosu, Koto-ku, Tokyo 135-8577, Japan.
Drs. Itoh, Oda, and K. Mori: Graduate School of Informatics, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-0814, Japan.
Author Contributions: Conception and design: Y. Mori, S. Kudo, H. Inoue, K. Mori.
Analysis and interpretation of the data: Y. Mori.
Drafting of the article: Y. Mori.
Critical revision for important intellectual content: Y. Mori, M. Misawa, Y. Saito, H. Ikematsu, K. Hotta, K. Ohtsuka, F. Urushibara, Y. Ogawa, Y. Maeda, K. Takeda, H. Nakamura, K. Ichimasa, T. Kudo, T. Hayashi.
Final approval of the article: Y. Mori, S. Kudo, M. Misawa, Y. Saito, H. Ikematsu, K. Hotta, K. Ohtsuka, F. Urushibara, S. Kataoka, Y. Ogawa, Y. Maeda, K. Takeda, H. Nakamura, K. Ichimasa, T. Kudo, T. Hayashi, K. Wakamura, F. Ishida, H. Inoue, H. Itoh, M. Oda, K. Mori.
Provision of study materials or patients: Y. Mori, M. Misawa, Y. Saito, H. Ikematsu, K. Hotta, K. Ohtsuka, Y. Ogawa, H. Itoh, M. Oda.
Obtaining of funding: Y. Mori, S. Kudo.
Administrative, technical, or logistic support: S. Kudo, F. Ishida, K. Mori.
Collection and assembly of data: Y. Mori, M. Misawa, Y. Saito, F. Urushibara, S. Kataoka, Y. Ogawa, Y. Maeda, K. Takeda, H. Nakamura, K. Ichimasa, T. Kudo, T. Hayashi, K. Wakamura, K. Mori.
Computer-aided diagnosis (CAD) for colonoscopy may help endoscopists distinguish neoplastic polyps (adenomas) requiring resection from nonneoplastic polyps not requiring resection, potentially reducing cost.
To evaluate the performance of real-time CAD with endocytoscopes (×520 ultramagnifying colonoscopes providing microvascular and cellular visualization of colorectal polyps after application of the narrow-band imaging [NBI] and methylene blue staining modes, respectively).
Single-group, open-label, prospective study. (UMIN [University hospital Medical Information Network] Clinical Trial Registry: UMIN000027360).
791 consecutive patients undergoing colonoscopy and 23 endoscopists.
Real-time use of CAD during colonoscopy.
CAD-predicted pathology (neoplastic or nonneoplastic) of detected diminutive polyps (≤5 mm) on the basis of real-time outputs compared with pathologic diagnosis of the resected specimen (gold standard). The primary end point was whether CAD with the stained mode produced a negative predictive value (NPV) of 90% or greater for identifying diminutive rectosigmoid adenomas, the threshold required to “diagnose-and-leave” nonneoplastic polyps. Best- and worst-case scenarios assumed that polyps lacking either CAD diagnosis or pathology were true- or false-positive or true- or false-negative, respectively.
Overall, 466 diminutive (including 250 rectosigmoid) polyps from 325 patients were assessed by CAD, with a pathologic prediction rate of 98.1% (457 of 466). The NPVs of CAD for diminutive rectosigmoid adenomas were 96.4% (95% CI, 91.8% to 98.8%) (best-case scenario) and 93.7% (CI, 88.3% to 97.1%) (worst-case scenario) with stained mode and 96.5% (CI, 92.1% to 98.9%) (best-case scenario) and 95.2% (CI, 90.3% to 98.0%) (worst-case scenario) with NBI.
Two thirds of the colonoscopies were conducted by experts who had each experienced more than 200 endocytoscopies; 186 polyps not assessed by CAD were excluded.
Real-time CAD can achieve the performance level required for a diagnose-and-leave strategy for diminutive, nonneoplastic rectosigmoid polyps.
Japan Society for the Promotion of Science.
A diminutive adenoma (A to C) and hyperplastic polyp (D to F) are seen in the rectum on white light (A and D) and endocytoscopic images (magnification, ×520) in NBI (B and E) and stained mode (C and F). Endocytoscopic images of the adenoma reveal networking microvessels in NBI mode and slit-like lumens (arrowheads) with slightly swollen nuclei (arrows) in stained mode. In contrast, endocytoscopic images of the hyperplastic polyp reveal obscure microvessels in NBI mode and serrated lumens (arrowheads) and small intracellular granules (arrows) in stained mode. NBI = narrow-band imaging.
Real-time CAD for use with an endocytoscope (CF-Y0058-I/CF-H290ECI [Olympus]) and its algorithm.
The computer on which CAD is installed is connected directly to an endoscopy unit (Evis Lucera Elite CV 290 [Olympus]), allowing fully automated diagnosis. Endoscopists can initiate CAD output simply by pushing the capture button on the endoscope. The algorithm for CAD in both NBI and stained modes comprises 3 steps: texture analysis, classification, and pathologic prediction. These analyses take 0.4 seconds, immediately outputting the pathologic prediction of the target polyp as either neoplastic or nonneoplastic, along with the probability of the diagnosis (0% to 100%). CAD = computer-aided diagnosis; NBI = narrow-band imaging.
CAD = computer-aided diagnosis; SSA/P = sessile serrated adenoma or polyp.
Appendix Table 1. Comparison Between Polyps Assessed by CAD and Those Not Assessed by CAD*
Table 1. Patient and Colonoscopy Characteristics
Table 2. Characteristics of Diminutive Polyps*
Table 3. Pathologic Prediction of CAD for Diminutive Polyps in the Rectosigmoid Colon*
Appendix Table 2. Pathologic Prediction of CAD for Diminutive Polyps in the Proximal–Rectosigmoid Colon*
Table 4. Performance of CAD for Diminutive Polyps*
Distribution of the number of images captured with endocytoscopy in NBI mode for 466 diminutive polyps.
NBI = narrow-band imaging.
Distribution of the number of images captured with endocytoscopy in stained mode for the 466 diminutive polyps.
Appendix Table 3. Image-Level Analysis
Appendix Table 4. Analysis That Included Histologically Excluded Polyps*
Distribution of time required to capture the first endocytoscopic image with NBI mode for the 466 diminutive polyps.
Distribution of time required to capture the first endocytoscopic image in stained mode for the 466 diminutive polyps.
Appendix Table 5. Time Required to Capture a First Endocytoscopic Image, by Examiner Expertise
Relationship between the number of images per polyp captured with endocytoscopy in NBI mode and the NPV for diminutive adenomas.
This post hoc analysis was conducted with 215 diminutive polyps, for which ≥15 images were captured by using the NBI mode in the whole colon. NBI = narrow-band imaging; NPV = negative predictive value.
Relationship between the number of images per polyp captured with endocytoscopy in stained mode and the NPV for diminutive adenomas.
This post hoc analysis was conducted with 290 diminutive polyps, for which ≥15 images were captured by using the stained mode in the whole colon. NPV = negative predictive value.
Appendix Table 6. Performance of CAD, Experts, and Nonexperts on a Test Set of 450 Diminutive Polyps*
Appendix Table 7. CAD Performance During the First Half (1 June to 15 September 2017) and Second Half (16 September to 31 December 2017) of the Study
Appendix Table 8. Performance of CAD, by Experts Versus Nonexperts
Appendix Table 9. CAD Performance, by Morphology
Appendix Table 10. Number of Nonanalyzable Images per Polyp, by First Half Versus Second Half and by Examiner Expertise*
The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.
Real-time use of CAD, which was trained with 61 925 images of colorectal polyps, correctly predicted the pathology of 2 types of diminutive polyp: an adenoma (neoplastic) and a juvenile polyp (nonneoplastic).
Mori Y, Kudo S, Misawa M, Saito Y, Ikematsu H, Hotta K, et al. Real-Time Use of Artificial Intelligence in Identification of Diminutive Polyps During Colonoscopy: A Prospective Study. Ann Intern Med. ;169:357–366. doi: 10.7326/M18-0249
Download citation file:
Published: Ann Intern Med. 2018;169(6):357-366.
Published at www.annals.org on 14 August 2018
Colonoscopy/Sigmoidoscopy, Gastroenterology/Hepatology, Hematology/Oncology.
Results provided by:
Copyright © 2019 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use