Diagnostic concordance of thoracic images for Covid-19 disease

Abstract

Introduction: Reverse Transcription Polymerase Chain Reaction (RT-PCR) is the gold standard for the diagnosis of coronavirus disease-19 (SARS-CoV-2). In the context of a pandemic with limited accessibility to this test, diagnostic images such as X-rays and Computed Tomography provide findings that support diagnostic suspicion, avoiding delays in medical care. Objective: Determine the sensitivity, specificity, positive and negative predictive value of diagnostic images and their agreement with the RT-PCR result. Methodology: Analytical cross-sectional study, 138 patients were evaluated (sampling error of 0.05 and 95 % CI), double-blind results of the imaging report were compared with RT-PCR. The sensitivity, specificity, positive predictive value and negative predictive value were calculated for chest x-rays and tomography for the diagnosis of SARS-CoV-2 infection. The measure of agreement between variable categories was carried out through the Kappa index of Cohen and its strength of association through the Bayes factor (FB). Results: The computed tomography / X-rays presented: Sensitivity 92.9 % / 86 %, Specificity 64 % / 52.9 %, Positive predictive value 92.1 % / 92.9 % and Negative predictive value 66.7 % / 34.6 % respectively, (95 % confidence interval). Conclusion: The tomography showed moderate diagnostic agreement, its usefulness is greater in cases of moderate-high clinical suspicion, diagnostic discrepancy or confirmation of complications. X-rays showed low diagnostic agreement. This method will be useful in cases of high clinical suspicion, but will require verification with reverse transcription polymerase chain reaction for a definitive diagnosis.