BACKGROUND: The aim of this study was to investigate the efficacy of diagnosing depth of wall invasion of gastric cancer on endoscopic images using computer-aided pattern recognition. METHODS: The back propagation algorithm was used for computer training. Data of 344 patients who underwent gastrectomy or endoscopic tumor resection between 2001 and 2010 and their 902 endoscopic images were collected. The images were divided into ten groups among which the number of patients and images were almost equally distributed according to T staging. The computer learning was performed using about 800 images from all but one group, and the accuracy rate of diagnosing the depth of wall invasion of gastric cancer was calculated using the remaining group of about 90 images. The various numbers of input layers, hidden layers, and learning counts were updated, and the ideal setting was decided. Similar learning and diagnostic procedures were repeated ten times using every group and all 902 images were tested. The accuracy rate was calculated based on the ideal setting. RESULTS: The most appropriate setting was a resolution of 16 × 16, a hidden layer of 240, and a learning count of 50. In the next step, using all the images on the ideal setting, the overall accuracy rate was 64.7%. The diagnostic accuracy was 77.2, 49.1, 51.0, and 55.3% in the T1, T2, T3, and T4 stagings, respectively. The accuracy was 68.9% in T1a(M) staging and 63.6% in T1b(SM) staging. The positive predictive values were 80.1, 41.6, 51.4, and 55.8% in the T1, T2, T3, and T4 staging, respectively. It was 69.2% in T1a(M) staging and 68.3% in T1b(SM) staging. CONCLUSION: Computer-aided diagnosis is useful for diagnosing depth of wall invasion of gastric cancer on endoscopic images.
BACKGROUND: The aim of this study was to investigate the efficacy of diagnosing depth of wall invasion of gastric cancer on endoscopic images using computer-aided pattern recognition. METHODS: The back propagation algorithm was used for computer training. Data of 344 patients who underwent gastrectomy or endoscopic tumor resection between 2001 and 2010 and their 902 endoscopic images were collected. The images were divided into ten groups among which the number of patients and images were almost equally distributed according to T staging. The computer learning was performed using about 800 images from all but one group, and the accuracy rate of diagnosing the depth of wall invasion of gastric cancer was calculated using the remaining group of about 90 images. The various numbers of input layers, hidden layers, and learning counts were updated, and the ideal setting was decided. Similar learning and diagnostic procedures were repeated ten times using every group and all 902 images were tested. The accuracy rate was calculated based on the ideal setting. RESULTS: The most appropriate setting was a resolution of 16 × 16, a hidden layer of 240, and a learning count of 50. In the next step, using all the images on the ideal setting, the overall accuracy rate was 64.7%. The diagnostic accuracy was 77.2, 49.1, 51.0, and 55.3% in the T1, T2, T3, and T4 stagings, respectively. The accuracy was 68.9% in T1a(M) staging and 63.6% in T1b(SM) staging. The positive predictive values were 80.1, 41.6, 51.4, and 55.8% in the T1, T2, T3, and T4 staging, respectively. It was 69.2% in T1a(M) staging and 68.3% in T1b(SM) staging. CONCLUSION: Computer-aided diagnosis is useful for diagnosing depth of wall invasion of gastric cancer on endoscopic images.
Authors: Sonia J Wakelin; Chris Deans; Trevor J Crofts; Paul L Allan; John N Plevris; Simon Paterson-Brown Journal: Eur J Radiol Date: 2002-02 Impact factor: 3.528
Authors: K Yamashita; T Yoshiura; H Arimura; F Mihara; T Noguchi; A Hiwatashi; O Togao; Y Yamashita; T Shono; S Kumazawa; Y Higashida; H Honda Journal: AJNR Am J Neuroradiol Date: 2008-04-03 Impact factor: 3.825
Authors: Andre Esteva; Katherine Chou; Serena Yeung; Nikhil Naik; Ali Madani; Ali Mottaghi; Yun Liu; Eric Topol; Jeff Dean; Richard Socher Journal: NPJ Digit Med Date: 2021-01-08