Diagnostic likelihood thresholds that define a working diagnosis of idiopathic pulmonary fibrosis

Abstract

Rationale: The level of diagnostic likelihood at which physicians prescribe antifibrotic therapy without requesting surgical lung biopsy (SLB) in patients suspected of idiopathic pulmonary fibrosis (IPF) is unknown. Objectives: To determine how often physicians advocate SLB in patient subgroups defined by IPF likelihood and risk associated with SLB, and to identify the level of diagnostic likelihood at which physicians prescribe antifibrotic therapy with requesting SLB. Methods: An international cohort of respiratory physicians evaluated 60 cases of interstitial lung disease, giving: 1) differential diagnoses with diagnostic likelihood; 2) a decision on the need for SLB; and 3) initial management. Diagnoses were stratified according to diagnostic likelihood bands described by Ryerson and colleagues. Measurements and Main Results: A total of 404 physicians evaluated the 60 cases (24,240 physician-patient evaluations). IPF was part of the differential diagnosis in 9,958/24,240 (41.1%) of all physician-patient evaluations. SLB was requested in 8.1%, 29.6%, and 48.4% of definite, provisional high-confidence and provisional low-confidence diagnoses of IPF, respectively. In 63.0% of provisional high-confidence IPF diagnoses, antifibrotic therapy was prescribed without requesting SLB. No significant mortality difference was observed between cases given a definite diagnosis of IPF (90-100% diagnostic likelihood) and cases given a provisional high-confidence IPF diagnosis (hazard ratio, 0.97; P = 0.65; 95% confidence interval, 0.90-1.04). Conclusions: Most respiratory physicians prescribe antifibrotic therapy without requesting an SLB if a provisional high-confidence diagnosis or "working diagnosis" of IPF can be made (likelihood>70%). SLB is recommended in only a minority of patients with suspected, but not definite, IPF.