Abstract
Introduction: Coronavirus disease 2019 (COVID-19) pneumonia is associated with high rate of pulmonary embolism (PE). In patients with contraindications for CT pulmonary angiography or non-diagnostic, perfusion single photon emission computed tomography/CT (Q-SPECT/CT) is an option.
Aim: to develop an artificial intelligence model based on Q-SPECT/CT images of patients during the COVID-19 pandemic which is able to classify lung lesions to optimize PE diagnosis Methods: Single center study with a prospective observational branch with patients who tested positive for COVID-19 and underwent a Q-SPECT/CT for diagnosis of PE. The second branch is retrospective, with patients in pre-pandemic period who underwent Q-SPECT/CT for diagnosis of PE. The entire image pre-processing was conducted with MATLAB. The diagnosis for each patient and the different tissue type segments were validated by two senior nuclear physicians. The Intelligent Radiomic System for identification of patients with PE is based on a local analysis of SPECT-CT volumes that considers both CT and SPECT values for each volume point. A support vector machine model was trained to discriminate among 4 types of tissue: no pneumonia without PE; no pneumonia with PE; pneumonia without PE; pneumonia with PE. Statistical analysis was performed with k-fold (with k=30) scheme
Results: 133 patients, 63 in prospective branch and 70 in retrospective branch. Average sensitivity and positive predictive value over 92%
Conclusions: This represents a first step towards a complete intelligent radiomic system to optimize PE diagnosis by Q-SPECT/CT and encourages the development of a tool in the cloud for clinical use.
Footnotes
Cite this article as: European Respiratory Journal 2021; 58: Suppl. 65, PA359.
This abstract was presented at the 2021 ERS International Congress, in session “Prediction of exacerbations in patients with COPD”.
This is an ERS International Congress abstract. No full-text version is available. Further material to accompany this abstract may be available at www.ers-education.org (ERS member access only).
- Copyright ©the authors 2021