A Quantitative Lung Computed Tomography Image Feature for Multi-Center Severity Assessment of COVID-19

Ghosh, B.; Kumar, N.; Sadhu, A. K.; Ghosh, N.; Mitra, P.; Chatterjee, J.

The COVID-19 pandemic has affected millions and congested healthcare systems globally. Hence an objective severity assessment is crucial in making therapeutic decisions judiciously. Computed Tomography (CT)-scans can provide demarcating features to identify the severity of pneumonia, commonly associated with COVID-19, in the affected lungs. Here, a semi-quantitative severity assessing lung CT image feature is demonstrated for COVID-19 patients. An open-source multi-center Italian database was used, among which 61 cases were incorporated in the study (age 27-86, 71% males) from 27 CT imaging centers. Lesions in the form of opacifications, crazy-paving patterns, and consolidations were heuristically marked and the severity determining feature L_norm was quantified and established to be statistically distinct for the three classes i.e., mild, moderate, and severe (p-value < 0.0001). The thresholds of L_norm for a 3-class classification were determined based on the optimum sensitivity/specificity combination from ROC analyses. The feature L_norm classified the cases in the three severity categories with 89.34% accuracy when L_norm was calculated based on experts' heuristic annotations. Substantial to almost-perfect intra-rater and inter-rater agreements were achieved involving expert and non-expert based evaluations ({kappa}-score 0.79-0.97). We trained machine learning based classification models and showed L_norm alone has a superior diagnostic accuracy over standard image intensity and texture features. Classification accuracy was further increased when L_norm was used to delineate the severe cases from non-severe ones (98.9%) with high sensitivity (97.7%), and specificity (97.49%). Therefore, key highlights of this severity assessment feature are accuracy, lower dependency on expert availability, and wide utility across multiple imaging centers.