Diagnostic performance of an artificial intelligence-driven cardiac-structured reporting system for myocardial perfusion SPECT imaging

• Type of Material: Article
• Authors: Ernesto Garcia, Emory University; Larry J Klein, University of North Carolina; Valeria Moncayo, Emory University; Charles Cooke, Emory University; Christian Del'Aune, Syntermed Inc; Russell Folks, Emory University; Liudmila Verdes Moreiras, Emory University; Fabio Esteves, Emory University
• Language: English
• Date: 2020-10-01
• Publisher: SPRINGER
• Publication Version: Author Accepted Manuscript (After Peer Review)
• Copyright Statement: © 2020 American Society of Nuclear Cardiology. Published by ELSEVIER INC. All rights reserved. Published by Mosby, Inc. All rights reserved.
• Final Published Version (URL): http://dx.doi.org/10.1007/s12350-018-1432-3
• Title of Journal or Parent Work: JOURNAL OF NUCLEAR CARDIOLOGY
• Volume: 27
• Issue: 5
• Start Page: 1652
• End Page: 1664
• Grant/Funding Information: This work was supported by the NHLBI grant number R42HL106818.
• Keywords: CORONARY-ARTERY-DISEASE; Radiology, Nuclear Medicine & Medical Imaging; Cardiovascular System & Cardiology; NEURAL-NETWORK; Science & Technology; IMPROVED ACCURACY; SCINTIGRAMS; quantitative analysis; structured reporting; STATEMENT; EXPERT-SYSTEM; Expert systems; artificial intelligence; myocardial perfusion SPECT; Life Sciences & Biomedicine; TEMPORAL TRENDS; HEART; Cardiac & Cardiovascular Systems; CARDIOLOGY

Abstract

Objectives: To describe and validate an artificial intelligence (AI)-driven structured reporting system by direct comparison of automatically generated reports to results from actual clinical reports generated by nuclear cardiology experts. Background: Quantitative parameters extracted from myocardial perfusion imaging (MPI) studies are used by our AI reporting system to generate automatically a guideline-compliant structured report (sR). Method: A new nonparametric approach generates distribution functions of rest and stress, perfusion, and thickening, for each of 17 left ventricle segments that are then transformed to certainty factors (CFs) that a segment is hypoperfused, ischemic. These CFs are then input to our set of heuristic rules used to reach diagnostic findings and impressions propagated into a sR referred as an AI-driven structured report (AIsR). The diagnostic accuracy of the AIsR for detecting coronary artery disease (CAD) and ischemia was tested in 1,000 patients who had undergone rest/stress SPECT MPI. Results: At the high-specificity (SP) level, in a subset of 100 patients, there were no statistical differences in the agreements between the AIsr, and nine experts’ impressions of CAD (P = .33) or ischemia (P = .37). This high-SP level also yielded the highest accuracy across global and regional results in the 1,000 patients. These accuracies were statistically significantly better than the other two levels [sensitivity (SN)/SP tradeoff, high SN] across all comparisons. Conclusions: This AI reporting system automatically generates a structured natural language report with a diagnostic performance comparable to those of experts.