Abstract
BACKGROUND: Misdiagnosis and underdiagnosis of pulmonary hypertension caused by fibrosing mediastinitis (PH-FM) are considerably prevalent due to unspecific symptoms and as well as the lack of awareness of this fatal disease. OBJECTIVES: The aim of this study was to evaluate the diagnostic accuracy of the chest X-ray (CXR) for screening the patients with PH-FM from those with pulmonary hypertension (PH). DESIGN: This was a retrospective observational cohort study. METHODS: The patients with suspected PH were recruited between October 2014 and October 2020. All the clinical data and CXR findings were collected. The sensitivity, specificity, and likelihood ratio of the CXR features were calculated. Logistic regression was used to identify the factors associated with the CXR characteristics and FM and to generate a prediction model. Finally, the diagnostic efficiency of the prediction model was evaluated using nomogram and internal validation. RESULTS: The patients with PH-FM (n = 36) and PH caused by the diseases other than FM (PH-non-FM, n = 62) were enrolled. The CXR features, including atelectasis, pleural effusion, consolidation, nodules, calcification, interlobular septal thickening, and interstitial reticulation, were more prevalent in patients with PH-FM than in those with PH-non-FM (all p < 0.05). Atelectasis had a specificity of 97%, a sensitivity of 50%, and a greater accuracy for diagnosing of PH-FM [area under the curve (AUC) = 0.720; 95% CI: 0.634-0.806] than the other factors did. The combination of tuberculosis, natural logarithmic NT-proBNP (lnBNP), atelectasis, pleural effusion, and prominent right heart border constituted a prediction model to distinguish the PH-FM from the PH-non-FM, with a sensitivity of 91.7% and a specificity of 83.9%. The model demonstrated good prediction performance by showing an AUC of 0.922 (95% CI: 0.861-0.983) in the internal validation. CONCLUSION: In this study, atelectasis was the most specific and accurate CXR characteristic for identifying PH-FM in the PH patients. The combination of atelectasis, pleural effusion, prominent right heart border, tuberculosis, and lnBNP constituted a prediction model that distinguished the PH-FM patients from the PH-non-FM ones with good performance.