Comparison and review of abrasive bronchial brushing versus non-abrasive aspiration, lavage and washing – Higher sensitivity but with risk of over-diagnosis for bronchial brushing

Abstract

Bronchial exfoliative cytology is classified as non-abrasive (washing, aspiration and bronchoalveolar lavage) and abrasive (brushing). Brush abrasion dislodges epithelial cells but can induce bleeding and cytomorphologic artifacts. In this study, the largest cohort to date of bronchial cytology specimens were referenced against bronchial biopsy as the reference standard. Findings in the study will be useful for selecting biopsy modality and reducing necessary procedural risks. All consecutive bronchial cytology and bronchial biopsy from 1995 to 2022 were retrieved. The diagnoses were reviewed and categorized into five-tiered diagnostic categories to compare diagnostic agreement and concordance. Review of 14,148 specimens yielded 3963 non-abrasive, 2378 abrasive cytology specimens matched to biopsy, with 4355 matches between non-abrasive and abrasive cytology specimens. Agreement between non-abrasive and abrasive cytology was moderate (κ = 0.580), and similar when referenced against biopsy (κ = 0.456 (non-abrasive), κ = 0.498 (abrasive)). Abrasive bronchial cytology showed a higher percentage of malignant diagnosis (20.95 % vs. 12.63 %, p < 0.001) and over-diagnosis rate (36.40 % vs. 29.79 %, p < 0.001), but higher sensitivity (0.747 vs. 0.572, p = 0.002). For subgroup analysis of transbronchial biopsies, matched abrasive cytology showed higher discordant rates (p < 0.05) and lower accuracy (0.907 vs. 0.873, p = 0.020). With the added bleeding risk associated with brushing, abrasive techniques may only be preferable in cases with clinical or bronchoscopic suspicion of malignancy, in particular endobronchial mucosal lesions. For routine bronchoscopy, non-abrasive bronchial cytology appears to be adequate.