Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automated triage of COVID-19 from various lung abnormalities using chest CT features

Dor Amran, Maayan Frid-Adar, Nimrod Sagie, Jannette Nassar, Asher Kabakovitch, Hayit Greenspan

TL;DR

A fully automated AI based system that takes as input chest CT scans and triages COVID-19 cases and produces multiple descriptive features to train a machine learning based classifier that distinguishes between CO VID-19 and other lung abnormalities (including community acquired pneumonia).

Abstract

The outbreak of COVID-19 has lead to a global effort to decelerate the pandemic spread. For this purpose chest computed-tomography (CT) based screening and diagnosis of COVID-19 suspected patients is utilized, either as a support or replacement to reverse transcription-polymerase chain reaction (RT-PCR) test. In this paper, we propose a fully automated AI based system that takes as input chest CT scans and triages COVID-19 cases. More specifically, we produce multiple descriptive features, including lung and infections statistics, texture, shape and location, to train a machine learning based classifier that distinguishes between COVID-19 and other lung abnormalities (including community acquired pneumonia). We evaluated our system on a dataset of 2191 CT cases and demonstrated a robust solution with 90.8% sensitivity at 85.4% specificity with 94.0% ROC-AUC. In addition, we present an elaborated feature analysis and ablation study to explore the importance of each feature.

Automated triage of COVID-19 from various lung abnormalities using chest CT features

TL;DR

A fully automated AI based system that takes as input chest CT scans and triages COVID-19 cases and produces multiple descriptive features to train a machine learning based classifier that distinguishes between CO VID-19 and other lung abnormalities (including community acquired pneumonia).

Abstract

The outbreak of COVID-19 has lead to a global effort to decelerate the pandemic spread. For this purpose chest computed-tomography (CT) based screening and diagnosis of COVID-19 suspected patients is utilized, either as a support or replacement to reverse transcription-polymerase chain reaction (RT-PCR) test. In this paper, we propose a fully automated AI based system that takes as input chest CT scans and triages COVID-19 cases. More specifically, we produce multiple descriptive features, including lung and infections statistics, texture, shape and location, to train a machine learning based classifier that distinguishes between COVID-19 and other lung abnormalities (including community acquired pneumonia). We evaluated our system on a dataset of 2191 CT cases and demonstrated a robust solution with 90.8% sensitivity at 85.4% specificity with 94.0% ROC-AUC. In addition, we present an elaborated feature analysis and ablation study to explore the importance of each feature.

Paper Structure

This paper contains 12 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Dataset Preparation
  4. Image processing pipeline
  5. Feature analysis pipeline
  6. COVID-19 vs other abnormalities triage
  7. Experiments and Results
  8. Classification optimization process & results
  9. Feature analysis
  10. Feature ablation study
  11. Discussion and Conclusion
  12. Compliance with Ethical Standards

Figures (3)

  • Figure 1: System segmentation output maps (rows 1-4) and final classification decision (row 5). Three cases are shown with varying pathologies (columns). Visual differences include laterality, peripherality, volume and location of the detections.
  • Figure 2: Chest CT scan classification process, depicting the image processing pipeline and output maps followed by the feature analysis pipeline and classification, in which this paper is focused.
  • Figure 3: (a) Features importance; (b) KDE analysis on selected features.