OBJECTIVETo investigate the value of the liquid-based cytology (LBC) of brushing specimens obtained via fiberoptic bronchoscopy for clinical diagnosis of lung cancer.

METHODSWe retrospectively analyzed the LBC cases in our hospital from January 2011 to May 2012, and evaluate its role in the diagnosis of lung cancer.

RESULTSThe clinical data of a total of 4 380 cases were reviewed and 3 763 of them had histopathological or clinical follow-up results (including 3 306 lung cancer cases and 457 benign lesion cases). The sensitivity, specificity, and accuracy of LBC diagnosis for lung cancer were 72.4% (2 392/3 306), 99.3% (454/457) and 75.6% (2 846/3 763), respectively. Of the 1 992 lung cancer cases diagnosed by brushing LBC, 528 cases (26.5%) were failed to take forceps biopsy and 113 cases (5.7%) showed negative forceps biopsy results. The accurate rate of subtyping of LBC for non-small cell carcinoma and small cell carcinoma was 99.0% (1 487/1 502) (P < 0.001). Take the resection histopathology as gold standard, the accurate rates of subtyping squamous cell carcinoma, adenocarcinoma and small cell carcinoma by LBC were 95.6% (351/367), 95.6% (351/367) and 100% (367/367), respectively, (P < 0.001). The accurate rates of subtyping of squamous cell carcinoma, adenocarcinoma and small cell carcinoma by forceps biopsy were 97.0% (293/302), 97.4% (294/302) and 99.7% (301/302), respectively, (Kappa = 0.895, P < 0.001). There was no significant difference in subtyping respectively between forceps biopsy and brushing LBC (P > 0.05).

CONCLUSIONSFiberoptic bronchoscopic brushing liquid-based cytology can significantly improve the detection rate of lung cancer, and have a high specificity and accurate rate of subtyping. It is an effective tool for the diagnosis and subtyping of lung cancer.