PENet-a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging

• Type of Material: Article
• Authors: Shih-Cheng Huang, Stanford University; Tanay Kothari, Stanford University; Imon Banerjee, Emory University; Chris Chute, Stanford University; Robyn L. Ball, Stanford University; Norah Borus, Stanford University; Andrew Huang, Stanford University; Bhavik N. Patel, Stanford University; Pranav Rajpurkar, Stanford University; Jeremy Irvin, Stanford University; Jared Dunnmon, Stanford University; Joseph Bledsoe, Intermountain Medical Center, Salt Lake Valley; Katie Shpanskaya, Stanford University; Abhay Dhaliwal, Michigan State University; Roham Zamanian, Stanford University; Andrew Y. Ng, Stanford University; Matthew P. Lungren, Stanford University
• Language: English
• Date: 2020-04-24
• Publisher: NATURE RESEARCH
• Publication Version: Final Published Version
• Copyright Statement: © The Author(s) 2020
• License: Creative Commons Attribution 4.0 International
• Final Published Version (URL): http://dx.doi.org/10.1038/s41746-020-0266-y
• Title of Journal or Parent Work: NPJ DIGITAL MEDICINE
• Volume: 3
• Issue: 1
• Start Page: 61
• End Page: 61
• Grant/Funding Information: Research reported in this publication was supported by the National Library of Medicine of the National Institutes of Health under Award number R01LM012966, Stanford Child Health Research Institute (Stanford NIH-NCATS-CTSA Grant #UL1 TR001085).
• Keywords: Life Sciences & Biomedicine; Medical Informatics; Science & Technology; PRACTICE GUIDELINE-EVALUATION; ANGIOGRAPHY; ERROR; RESIDENTS; Health Care Sciences & Services; RADIOLOGISTS; PERFORMANCE; COMPUTER-AIDED DETECTION; AMERICAN THORACIC SOCIETY/SOCIETY; ARTIFICIAL NEURAL-NETWORK
• Research Categories: Health Sciences, Radiology; Health Sciences, Health Care Management

Abstract

Pulmonary embolism (PE) is a life-threatening clinical problem and computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosis. Prompt diagnosis and immediate treatment are critical to avoid high morbidity and mortality rates, yet PE remains among the diagnoses most frequently missed or delayed. In this study, we developed a deep learning model—PENet, to automatically detect PE on volumetric CTPA scans as an end-to-end solution for this purpose. The PENet is a 77-layer 3D convolutional neural network (CNN) pretrained on the Kinetics-600 dataset and fine-tuned on a retrospective CTPA dataset collected from a single academic institution. The PENet model performance was evaluated in detecting PE on data from two different institutions: one as a hold-out dataset from the same institution as the training data and a second collected from an external institution to evaluate model generalizability to an unrelated population dataset. PENet achieved an AUROC of 0.84 [0.82–0.87] on detecting PE on the hold out internal test set and 0.85 [0.81–0.88] on external dataset. PENet also outperformed current state-of-the-art 3D CNN models. The results represent successful application of an end-to-end 3D CNN model for the complex task of PE diagnosis without requiring computationally intensive and time consuming preprocessing and demonstrates sustained performance on data from an external institution. Our model could be applied as a triage tool to automatically identify clinically important PEs allowing for prioritization for diagnostic radiology interpretation and improved care pathways via more efficient diagnosis.