Raghu G, et al. - Researchers present prospective results for the clinical validity and usefulness of a machine learning algorithm in less-invasive transbronchial lung biopsy samples for the detection of a molecular usual interstitial pneumonia signature for idiopathic pulmonary fibrosis (IPF) diagnosis. This study included prospectively selected 237 patients from the Bronchial Sample Collection for a Novel Genomic Test (BRAVE) study. With specificity and sensitivity of 88% and 70%, respectively, usual interstitial pneumonia was identified by the classifier in transbronchial lung biopsy samples from 49 patients. Classifier-aided clinical diagnoses showed 86% agreement with clinical diagnoses made using histopathology data. IPF can be diagnosed objectively with the help of the molecular test, especially in patients without a clear radiological diagnosis.