Purpose: Evaluate diagnostic accuracy of portable chest radiograph in mechanically ventilated patients taking autopsy findings as the gold standard and the interobserver agreement among intensivists and radiologists.
Materials and methods: Retrospective study of 422 patients over 22 years who died in the ICU, underwent an autopsy, and had at least one portable chest radiograph 72 h prior to death. Two intensivists and two radiologists independently read each chest radiograph. Sensitivity, specificity, positive and negative likelihood ratios were evaluated. Overall performance metrics accuracy between intensivists and radiologists were compared using a generalized estimating equation. Cohen's kappa coefficient was used to evaluate the interobserver agreement with the following values: <0.20:poor, 0.21-0.40:fair, 0.41-0.60:moderate, 0.61-0.80:good, 0.81-1.00:excellent.
Results: Overall sensitivity and specificity for pneumonia was 24% and 91% respectively. Overall sensitivity and specificity for ARDS was 68% and 74% respectively. Sensitivity for pneumonia was higher among radiologists (p < 0,05). Specificity for ADRS was higher among radiologists (p < 0,05). Good interobserver agreement among radiologists and poor correlation between intensivists was found.
Conclusions: Chest radiographs has a moderate specificity for ARDS and a high specificity for pneumonia, with limited sensitivity in both entities. Interobserver agreement of portable chest radiograph in the mechanically ventilated patients is higher between radiologists than intensivists.
Keywords: Autopsy; Intensive care units; Lung; Observer variation.; Radiography.
Copyright © 2020 Elsevier Inc. All rights reserved.