The associations of radiological features of high-resolution computed tomography with the outcomes of transbronchial cryobiopsy in interstitial lung diseases: A cohort study

BackgroundTransbronchial cryobiopsy (TBCB) is a critical procedure in the diagnosis of interstitial lung diseases (ILD). The associations between high-resolution computed tomography (HRCT) features and outcomes of TBCB were unknown.MethodsThis study was conducted as a single-center prospective cohort study between September 2018 and January 2020 (NCT04047667). HRCT was obtained before performing TBCB. The clinical and radiological characteristics, complications, pathological and multidisciplinary discussion (MDD) diagnoses were recorded. The relationships between HRCT features and outcomes of TBCB were analyzed.ResultsTBCB was performed on 216 ILD patients. The radiological features usually interstitial pneumonia (UIP) or probable UIP, indeterminate for UIP, ground-glass opacities (GGO) and cysts were found in 55 (25.5%), 38 (17.6%), 84 (38.9%) and 9 (4.2%) patients, respectively. And 118 (54.6%) patients had high HRCT score (involved lung proportion ≥50%) in the biopsied lobe. Multivariate analysis suggested radiological probable UIP pattern may be an independent risk factor for moderate bleeding (OR = 4.304; 95% CI: 1.383–13.393; P = 0.012), while GGO may be a protective factor from moderate bleeding (OR = 0.173, 95% CI: 0.043–0.687; P = 0.013). The pathological diagnostic yield in patients presenting cysts on HRCT was significantly lower than others (44.4 vs. 87.9%; P = 0.009). While performing TBCB in the lobe with high HRCT score increased pathological diagnostic yield (91.5 vs. 79.6%; P = 0.022). Neither pneumothorax nor MDD diagnostic yields were found to be associated with HRCT features.ConclusionsHRCT features were associated with moderate bleeding and pathological diagnosis. Pre-TBCB assessments of HRCT pattern and scores were helpful for bronchoscopists to make a better patient selection and procedure planning.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.