LAUSR

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: To report our experience from encountering the COVID-19 pandemic and its repercussion in the imaging of patients during the period ranging from March 2020 to October 2020 in our tertiary military hospital based in a city of 1.2 million people. METHODS: We included all patients that showed up and were examined in our tertiary hospital's specific section of the ER for suspected cases of COVID, excluding those for whom an alternative diagnosis was far more plausible and including those whose molecular testing turned out to be positive despite initial triage. RT-PCR was utilized for the confirmation of COVID cases. The imaging modalities used were chest X-ray, chest CT and CTPA and the images were assessed by two independent researchers. Chest imaging findings were categorized into 9 different radiological patterns including normal appearance. For the statistical analysis we used version 23 of IBM SPSS Statistics ®. We compared the emergence of the forementioned radiological findings in SARS-COV2 positive and negative patients and assessed their correlation with SARS-COV2 infection and possible predictive value. RESULTS: A total of 414 patients were included in the statistical analysis. 31 (7.5%) were confirmed COVID cases, 143 (34.5%) underwent imaging with chest X-ray and 145 (35%) underwent imaging with chest CT or CTPA with 12/145 being confirmed COVID cases. Among the latter, 37 (25.5%) had a normal imaging with 3/37 being confirmed COVID cases and 108 (74.5%, 9/108 COVID) had one or more of the 8 remaining radiological findings, namely 39 (26.9%, 3/39 COVID) had fibrosis, 1 (0.7%, 0/1 COVID) had PE, 31 (21.4%, 4/31 COVID) had consolidation(s), 32 (22.1%, 7/32 COVID) had ground-glass opacity (GGO), 5 (3.4%, 0/5 COVID) had tree-in-bud pattern, 8 (5.5%, 0/8 COVID) had bronchiectasis, 40 (27.6%, 0/40 COVID) had pleural effusion(s) and none had crazy paving. We observed a statistically significant association with SARS-COV2 positivity only in the cases of GGO (p=0.005) and pleural effusion (p=0.037) when assessing the entirety of our cohort. Adjusting for sex brought about the emergence of statistically significant correlation with SARS-COV2 infection only for GGO in male patients (p=0.017), while adjusting for age led to the observation of statistically significant correlation with SARS-COV2 infection in the cases of consolidation (p=0.041) and GGO (p=0.001) in elderly patients exclusively. We also used logistic regression to generate the best models predicting SARS-COV2 infection based on radiological imaging and our population under examination and acquired the best results in terms of sensitivity (58.3%), specificity (81.2%), positive predictive value (21.9%) and negative predictive value (95.6%) when GGO was present. CONCLUSIONS: Ground glass opacity appears to be the most common chest CT finding associated with SARS-COV2 infection. Microbial superinfections are encountered more often in elderly patients. CLINICAL IMPLICATIONS: During the first wave of the SARS-COV2 pandemic (March 2020-October 2020), pretest probability was low in our country and chest imaging could be utilized to contribute to clinical decision in the event of ambiguous or not clear status of SARS-COV2 infection. DISCLOSURES: No relevant relationships by Stergios Gkintikas, source=Admin input No relevant relationships by DIMITRIOS KARAPIPERIS, source=Admin input No relevant relationships by Maria Kipourou, source=Web Response No relevant relationships by KAROZIS KONSTANTINOS, source=Web Response No relevant relationships by Prodromos Koutoukoglou, source=Web Response No relevant relationships by Dimitrios MOLIVAS, source=Web Response No relevant relationships by Dimitrios Sidiropoulos, source=Web Response No relevant relationships by Ioannis Tsanaktsidis, source=Web Response