Multivariate study of predictive factors for clearly defined lung lesions without visible endobronchial lesions in transbronchial biopsy

Abstract

Although the diagnostic validity and predictive factors for the diagnostic yield of transbronchial biopsy (TBB) of clearly defined pulmonary lesions with no visible endobronchial lesion have been analyzed in numerous studies, very few have used multivariate analysis techniques to evaluate the validity of TBB as a diagnostic tool or to analyze the independent influence of clearly dependent variables, such as the bronchus sign and lesion size. We retrospectively analyzed all cases in which this type of lesion underwent TBB under fluoroscopic control between 2006 and 2008. The analyzed variables included lesion size, localization, the presence of the bronchus sign, and the final result obtained. We performed a descriptive analysis of the TBB results and a multivariate analysis of the predictive factors for the results using logistic regression techniques. A total of 273 patients (206 males, 75.5%) were included in the study. The average lesion diameter was 34 (± 16) mm, with 24% 2 cm or smaller. Twenty-eight percent of the lesions were localized in the lower lobes and 32% in the peripheral third of the lung. The bronchus sign was present in 28% of the patients. Seventy-eight percent of the patients had primary or metastatic malignant lung lesions, the rest were benign lesions of diverse etiology. TBB was diagnostic in 45.4% of cases. In the multivariate analysis, the only independent predictors of outcome were malignant etiology (OR = 4.8; 95% CI = 2.210.4), diameter >20 mm (OR = 3.6; 95% CI = 1.8-7.3), and the presence of the bronchus sign (OR = 2.4; 95% CI = 1.3-4.3). TBB of lesions clearly delimited without an endobronchial lesion can lead to diagnosis in almost half of the patients. The nature of the lesion, diameter >20 mm, and the presence of the bronchus sign are independent predictors of outcome.