Eladio Rodriguez-Diaz1, György Baffy2, Wai-Kit Lo2, Hiroshi Mashimo2, Gitanjali Vidyarthi3, Shyam S Mohapatra4, Satish K Singh5. 1. Research Service, VA Boston Healthcare System, Boston, MA; Department of Biomedical Engineering, Boston University College of Engineering, Boston, MA. 2. Department of Medicine, Section of Gastroenterology, VA Boston Healthcare System, Boston, MA; Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA. 3. Section of Gastroenterology, James A. Haley Veterans Hospital, Tampa, FL; Department of Medicine, Morsani College of Medicine, University of South Florida, Tampa, FL. 4. Research Service, James A. Haley Veterans Hospital, Tampa, FL; Department of Medicine, Morsani College of Medicine, University of South Florida, Tampa, FL. 5. Research Service, VA Boston Healthcare System, Boston, MA; Department of Biomedical Engineering, Boston University College of Engineering, Boston, MA; Department of Medicine, Section of Gastroenterology, VA Boston Healthcare System, Boston, MA; Department of Medicine, Boston University School of Medicine, Boston, MA; Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA.
Abstract
BACKGROUND AND AIMS: Artificial intelligence (AI)-based computer-aided diagnostic (CADx) algorithms are a promising approach for real-time histology (RTH) of colonic polyps. Our aim is to present a novel in situ CADx approach that seeks to increase transparency and interpretability of results by generating an intuitive augmented visualization of the model's predicted histology over the polyp surface. METHODS: We developed a deep learning model using semantic segmentation to delineate polyp boundaries and a deep learning model to classify subregions within the segmented polyp. These subregions were classified independently and were subsequently aggregated to generate a histology map of the polyp's surface. We used 740 high-magnification narrow-band images from 607 polyps in 286 patients and over 65,000 subregions to train and validate the model. RESULTS: The model achieved a sensitivity of .96, specificity of .84, negative predictive value (NPV) of .91, and high-confidence rate (HCR) of .88, distinguishing 171 neoplastic polyps from 83 non-neoplastic polyps of all sizes. Among 93 neoplastic and 75 non-neoplastic polyps ≤5 mm, the model achieved a sensitivity of .95, specificity of .84, NPV of .91, and HCR of .86. CONCLUSIONS: The CADx model is capable of accurately distinguishing neoplastic from non-neoplastic polyps and provides a histology map of the spatial distribution of localized histologic predictions along the delineated polyp surface. This capability may improve interpretability and transparency of AI-based RTH and offer intuitive, accurate, and user-friendly guidance in real time for the clinical management and documentation of optical histology results. Published by Elsevier Inc.
BACKGROUND AND AIMS: Artificial intelligence (AI)-based computer-aided diagnostic (CADx) algorithms are a promising approach for real-time histology (RTH) of colonic polyps. Our aim is to present a novel in situ CADx approach that seeks to increase transparency and interpretability of results by generating an intuitive augmented visualization of the model's predicted histology over the polyp surface. METHODS: We developed a deep learning model using semantic segmentation to delineate polyp boundaries and a deep learning model to classify subregions within the segmented polyp. These subregions were classified independently and were subsequently aggregated to generate a histology map of the polyp's surface. We used 740 high-magnification narrow-band images from 607 polyps in 286 patients and over 65,000 subregions to train and validate the model. RESULTS: The model achieved a sensitivity of .96, specificity of .84, negative predictive value (NPV) of .91, and high-confidence rate (HCR) of .88, distinguishing 171 neoplastic polyps from 83 non-neoplastic polyps of all sizes. Among 93 neoplastic and 75 non-neoplastic polyps ≤5 mm, the model achieved a sensitivity of .95, specificity of .84, NPV of .91, and HCR of .86. CONCLUSIONS: The CADx model is capable of accurately distinguishing neoplastic from non-neoplastic polyps and provides a histology map of the spatial distribution of localized histologic predictions along the delineated polyp surface. This capability may improve interpretability and transparency of AI-based RTH and offer intuitive, accurate, and user-friendly guidance in real time for the clinical management and documentation of optical histology results. Published by Elsevier Inc.
Authors: Ana García-Rodríguez; Yael Tudela; Henry Córdova; Sabela Carballal; Ingrid Ordás; Leticia Moreira; Eva Vaquero; Oswaldo Ortiz; Liseth Rivero; F Javier Sánchez; Miriam Cuatrecasas; Maria Pellisé; Jorge Bernal; Glòria Fernández-Esparrach Journal: Endosc Int Open Date: 2022-09-14