Endobronchial ultrasonography with guide sheath versus computed tomography guided transthoracic needle biopsy for peripheral pulmonary lesions: a propensity score matched analysis.

BACKGROUND: Computed tomography guided transthoracic needle aspiration (CT-TTNA) and endobronchial ultrasonography with guide sheath (EBUS-GS) transbronchial lung biopsy are important methods for the diagnosis of peripheral pulmonary lesions (PPLs). Without enough evidence, it is controversial which is a better choice for diagnosing PPLs. In this study, we hypothesized that the complication rate of EBUS-GS would be lower than CT-TTNA and the diagnostic accuracy of EBUS-GS were not inferior to CT-TTNA.

METHODS: This study was a retrospective analysis of a prospective registry with propensity matching. Patients with PPLs were divided into EBUS-GS group and CT-TTNA group according to patients' intent to treatment. Pathological results and procedure related complications of EBUS-GS and CT-TTNA were recorded. Propensity score matching(PSM) was used to eliminate the intergroup bias. Diagnostic yields and complications of two groups were compared. Subgroup analysis was performed to conclude the indications of different procedures.

RESULTS: A total of 187 patients (CT-TTNA: 130; EBUS-GS: 57) were enrolled. After propensity score matching, 54 paired patients were included. Diagnostic yield was 81% (44/54) for EBUS-GS and 87% (47/54) for CT-TTNA (P=0.43), respectively. Diagnostic sensitivity in malignancy was 93% (42/45) for CT-TTNA and 79% (37/47) for EUBS-GS (P=0.04). Subgroup analysis revealed that the sensitivity of CT-TTNA was significantly higher in diagnosing of lesions close to the chest wall (100% vs . 80%, P=0.04), and bronchus sign on CT was a predictive factor for accurate diagnosis by EBUS-GS. The overall complication rate was 13% (7/54) for CT-TTNA group, which was not significantly higher than that of EBUS-GS group (2%, P=0.06). Subgroup analysis showed that patients combined with pulmonary comorbidities and lesions apart from chest wall were risk factors for complications of CT-TTNA.

CONCLUSIONS: Both of the two procedures are effective for the diagnosis of PPLs, but CT-TTNA has a higher diagnostic sensitivity for malignancy. EBUS-GS has fewer complications, and it is more suitable for patients combined with pulmonary comorbidities and lesions with bronchus signs.