BACKGROUND Parenchymatous lung lesions often present as peripheral non-endobronchial lesions, which are not visible through conventional flexible fiberoptic bronchoscopy. Tissue diagnosis from these lesions is usually obtained by transbronchial lung biopsy (TBLB). The lesion location is estimated by chest radiograph (CXR) or computerized tomography (CT) of the chest. The diagnostic yield of TBLB is limited and variable, ranging from 16-80%. Fluoroscopy is used simultaneously with FOB during TBLB for better localization of parenchymatous lesions. Nevertheless, fluoroscopy-guided transbronchial biopsy (Flu-TBLB) is not widely used at present. The advantages and safety of Flu-TBLB have not yet been verified. OBJECTIVE To compare the diagnostic yields and complications of TBLB with and without fluoroscopy guidance for non-endobronchial lung lesion. METHODS Descriptive study with subgroup analysis. METHODS Medical and bronchoscopic data records of patients who underwent TBLB at Siriraj Hospital from January 2001 to June 2005 were reviewed. The patients were divided into two groups according to the use of fluoroscopy during TBLB. Patient demographic data, underlying diseases, CXR findings, diagnoses, complications and yields of TBLB of the two groups were compared. METHODS Student t-test and chi-square test. RESULTS Six hundred and fifty patients were included in the present study. Three hundred and thirty-one patients were in Flu-TBLB group, 319 patients were in non fluoroscopy-guided transbronchial biopsy (NFlu-TBLB) group. The overall diagnostic yield of Flu-TBLB group was statistically significantly higher than NFlu-TBLB group (43.8% vs. 32.9%; p = 0.003). When comparing the diagnostic yields of the 2 groups by CXR findings, the yields of Flu-TBLB group were statistically significantly higher than NFlu-TBLB group for lung masses (41.4% vs. 29.5%; p = 0.036) and focal infiltrative lesions (46.2% vs. 29.4%; p = 0.008), respectively. The yield of Flu-TBLB group was slightly higher than NFlu-TBLB group for diffuse infiltrative lesion (45.1% vs. 40%; p = 0.289). No significant difference in the rate of pneumothorax discovery between the two groups (1.2% in Flu-TBLB group and 0.6% in NFlu-TBLB group) was observed. CONCLUSIONS Flu-TBLB significantly increases the diagnostic yields of non-endobronchial lung masses and focal infiltrates compared to NFlu-TBLB. There is no clinical significant difference in the rate of pneumothorax discovery between the two groups. Flu-TBLB is also more cost-saving.