Artificial intelligence in radiodiagnosis assessment of circulatory disorders in cases of community-acquired pneumonia before and in the COVID-19 pandemic

Abstract

Introduction. The COVID-19 pandemic not only caused a surge of viral pneumonia patients, but also opened up new opportunities in the field of radiology. One of the conditional «pluses» was the quantitative damage assessment of the lung parenchyma and microcirculation in COVID-19 using an artificial intelligence program. Objective. To evaluate the AI capabilities to detect the severity of anatomical and microcirculatory post-inflammatory disorders by CT and SPECT of pneumonic patients data. Materials and Methods. We analyzed prospective and retrospective data obtained by radiological studies of 187 patients with community-acquired pneumonia from 2006 to 2022 in the clinics of the St. Petersburg State Medical University named after acad. I.P. Pavlov. The follow-up period varied from 3 months to 8 years. The mean age of the patients was 34.3±9.2 years (w/m – 107/80). All patients underwent CT scan, a comprehensive functional examination of external respiration (KFIVD), SPECT. Results. The community-acquired pneumonia before the COVID-19 pandemic was generally characterized by signs of exudative bronchiolitis/bronchopneumonia, infectious viral alveolitis and pleuropneumonia. The first two forms in contrast to pleuropneumonia were associated with microcirculation disturbances. Signs of lung damage in COVID-19 had staging pattern based on morphological changes: edema, reticulation (increased severity of edema, cellular infiltration, intraalveolar fibrin), organization=consolidation (cellular infiltration, intraalveolar fibrin, fibroblast proliferation). Residual anatomical changes were accompanied by clinical symptoms (shortness of breath of varying severity, dry cough, weakness, intoxication). Conclusions. Using the artificial intelligence for post-processor CT and SPECT image processing seems to be relevant to assess the postinflammatory anatomical and microcirculatory disorders severity. Experience accumulation in combined AI and radiological study of patients with community-acquired pneumonia is capable to quantify residual anatomical and microcirculatory changes and useful for treatment tactics.