Can current guidelines improve the diagnosis of a usual interstitial pneumonia among general radiologists?

Abstract

Current ATS/ERS/JRS/ALAT guidelines on Idiopathic Pulmonary Fibrosis propose criteria to diagnose Usual Interstitial Pneumonia (UIP) on a chest High Resolution Computed Tomography (HRCT); and state that a definite UIP pattern on HRCT obviates the need for surgical biopsy. However, these criteria have not been prospectively validated. Objective: To assess the impact of applying current guidelines in the agreement of UIP diagnosis between general radiologists. Methods: Two general radiologists at a tertiary care academic center in Canada reviewed 128 HRCTs of patients with Interstitial Lung Diseases before and after applying current guidelines to determine the diagnosis of UIP. Results: Before applying the guidelines, the agreement between radiologists of definite UIP was 75% (Kappa 0.475); possible UIP 66% (Kappa 0.126) and inconsistent UIP pattern 78% (Kappa 0.516). After applying the guidelines, the agreement between radiologists of definite UIP was 78% (Kappa 0.525); possible UIP 72% (Kappa 0.378) and inconsistent UIP 88% (Kappa 0.709). The agreement for the presence of reticulation was 95% (Kappa 0.640); honeycombing 85% (Kappa 0.697); subpleural predominance of abnormalities 82% (Kappa 0.385); basilar predominance 73% (Kappa 0.463). Conclusions: Applying the guidelines improved the agreement in the diagnosis of possible UIP and inconsistent UIP, but not of definite UIP. Disagreement in the diagnosis of a definite UIP pattern between radiologists could results in an important variation in the number of patients requiring surgical biopsy to confirm the diagnosis (up to 1 in 4 patients).