Impact of a Bayesian approach on the diagnosis of Interstitial Lung Diseases

Résumé fourni par la source

Introduction: Current guidelines recommend a multi-disciplinary team (MDT) approach involving experienced pulmonologists, radiologists and pathologists in improving the confidence in diagnosing various interstitial lung diseases (ILD). However, accurately diagnosing fibrotic ILD remains a challenge with significant diagnostic heterogeneity, impacting the management and prognosis of patients. Aims: We evaluate the effects of implementing a Bayesian approach, incorporating clinical judgment, pre-test probability and radiological features, in improving the diagnosis of fibrotic ILD and reducing the number of unclassifiable ILD without the need for a surgical lung biopsy.(1) Methods: This is a prospective study of consecutive patients referred for MDT assessment from April 2015 to December 2018. We compared diagnostic categories and level of confidence between 2 groups of patients before and after the implementation of a Bayesian approach in ILD diagnosis.(1) Results: We report on the first 125 patients. The Bayesian approach was used in 46 (37%) patients. The proportion of cases with a confident diagnosis of idiopathic pulmonary fibrosis (IPF) was 25% and 22%, before and after the use of the approach, respectively (p=0.65). There was a trend towards a decrease in the proportion of cases with unclassifiable ILD, from 17% to 7% (p=0.11). Conclusion: A Bayesian approach in MDT discussion may reduce the number of cases with unclassifiable disease and improve the diagnostic classification of fibrotic ILD without the need for surgical biopsy. Further studies are needed on the utility and impact of this approach on subsequent management and patient-related outcomes. Reference: (1)Ryerson et al. Am J Respir Crit Care Med. 2017;196:1249

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.