Abstract Rationale Bronchoscopic evaluation of peripheral pulmonary lesions (PPLs) often relies on cytopathology (transbronchial needle aspirations, or TBNA) for diagnosis, but the added value of histopathology from forceps biopsy or cryobiopsy is unclear. We aimed to determine whether adding histopathology improves the diagnostic yield beyond cytopathology alone in a secondary analysis of the RELIANT randomized controlled trial (RCT), using a causal inference framework to account for procedural variability in the decision to obtain histopathology. Methods RELIANT was a single-center cluster-RCT comparing robotic versus electromagnetic navigational bronchoscopy in patients with PPLs. Per protocol all patients underwent TBNA first, and the addition of histopathology using forceps or cryobiopsy (1.1 mm cryoprobe, Erbe, Tubingen, Germany) was at the bronchoscopists’ discretion. Because the decision to obtain histopathology was driven by certain variables (like diagnosis of malignancy on rapid on-site evaluation, or ROSE), a causal directed acyclic graph was constructed to inform confounder selection. We then conducted two inverse probability of treatment weighting (IPTW) analyses to simulate comparisons of: (1) any histopathology (forceps or cryobiopsy) after TBNA vs only TBNA and (2) cryobiopsy after TBNA vs only TBNA. Covariates for propensity modeling included ROSE result, lesion size, lobar location, peripheral location, CT bronchus sign, and radial EBUS ultrasound view. Stabilized IPTW was applied to create weighted cohorts. Diagnostic yield was strictly defined as obtaining lesional tissue (malignancy or specific benign diagnosis). Weighted logistic regression models were used to estimate adjusted odds ratios (OR) for diagnostic yield, with robust standard errors. Results 315 of 411 patients had a specific diagnostic result (diagnostic yield 76.6%), including 233 malignancies and 82 specific benign lesions. Median lesion size was 19mm. Cytopathology alone (TBNA) contributed to the diagnosis in 254 of 315 cases (80.6%). However, 59 cases (18.7%) were diagnosed solely by histopathology (forceps and/or cryobiopsy) when TBNA was nondiagnostic. In IPTW analysis, diagnostic yield was increased by performing any histopathology in addition to TBNA (OR = 3.7, 95% CI 2.1-6.4, p 0.01), or by performing cryobiopsy in addition to TBNA (OR = 4.27, 95% CI = 2.3-8.08, p 0.01). All included covariates were well balanced after weighting (all standardized mean differences 0.1). Conclusions Histopathologic tissue sampling during navigational bronchoscopy provides significant added diagnostic yield (10% improvement) beyond cytopathology alone, supporting routine addition of histopathology to maximize yield in peripheral bronchoscopy. This abstract is funded by: Vanderbilt Learning Healthcare System (LHS) Initiative
Goodfellow et al. (Fri,) studied this question.