Machine-Learning-Based Multiple Abnormality Prediction with Large-Scale Chest Computed Tomography Volumes

Abstract

Machine learning models for radiology benefit from large-scale data sets with high quality labels for abnormalities. We curated and analyzed a chest computed tomography (CT) data set of 36,316 volumes from 19,993 unique patients. This is the largest multiply-annotated volumetric medical imaging data set reported. To annotate this data set, we developed a rule-based method for automatically extracting abnormality labels from free-text radiology reports with an average F-score of 0.976 (min 0.941, max 1.0). We also developed a model for multi-organ, multi-disease classification of chest CT volumes that uses a deep convolutional neural network (CNN). This model reached a classification performance of AUROC greater than 0.90 for 18 abnormalities, with an average AUROC of 0.773 for all 83 abnormalities, demonstrating the feasibility of learning from unfiltered whole volume CT data. We show that training on more labels improves performance significantly: for a subset of 9 labels - nodule, opacity, atelectasis, pleural effusion, consolidation, mass, pericardial effusion, cardiomegaly, and pneumothorax - the model's average AUROC increased by 10% when the number of training labels was increased from 9 to all 83. All code for volume preprocessing, automated label extraction, and the volume abnormality prediction model will be made publicly available. The 36,316 CT volumes and labels will also be made publicly available pending institutional approval.

arXiv comment20 pages, 3 figures, 5 tables (appendices additional). Published in Medical Image Analysis (October 2020)

References in corpus (17)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Distributed Representations of Words and Phrases and their Compositionality
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• The NumPy array: a structure for efficient numerical computation
• TensorFlow: A system for large-scale machine learning
• Transfusion: Understanding Transfer Learning for Medical Imaging
• Visualizing the Loss Landscape of Neural Nets
• Multi-source Transfer Learning with Convolutional Neural Networks for Lung Pattern Analysis
• Diagnose like a Radiologist: Attention Guided Convolutional Neural Network for Thorax Disease Classification
• Learning to diagnose from scratch by exploiting dependencies among labels
• NegBio: a high-performance tool for negation and uncertainty detection in radiology reports
• Weakly Supervised Deep Learning for Thoracic Disease Classification and Localization on Chest X-rays
• Abnormality Detection and Localization in Chest X-Rays using Deep Convolutional Neural Networks
• Weakly Supervised Medical Diagnosis and Localization from Multiple Resolutions
• TieNet: Text-Image Embedding Network for Common Thorax Disease Classification and Reporting in Chest X-rays
• Learning to recognize Abnormalities in Chest X-Rays with Location-Aware Dense Networks
• Multi-task Learning for Chest X-ray Abnormality Classification on Noisy Labels

Cited by in corpus (11)

• Ontology-driven weak supervision for clinical entity classification in electronic health records
• Deep learning-driven pulmonary artery and vein segmentation reveals demography-associated vasculature anatomical differences
• Personalized and privacy-preserving federated heterogeneous medical image analysis with PPPML-HMI
• Explainable multiple abnormality classification of chest CT volumes
• Use HiResCAM instead of Grad-CAM for faithful explanations of convolutional neural networks
• CT-AGRG: Automated Abnormality-Guided Report Generation from 3D Chest CT Volumes
• Enhancing chest X-ray datasets with privacy-preserving large language models and multi-type annotations: a data-driven approach for improved classification
• MvKeTR: Chest CT Report Generation with Multi-View Perception and Knowledge Enhancement
• Classification of Multiple Diseases on Body CT Scans using Weakly Supervised Deep Learning
• Detecting Spurious Correlations with Sanity Tests for Artificial Intelligence Guided Radiology Systems
• Structure Observation Driven Image-Text Contrastive Learning for Computed Tomography Report Generation