To improve the efficacy of fibreoptic bronchoscopy in the diagnosis of peripheral lung cancer, we evaluated the effectiveness of various techniques for obtaining samples for cytological examination. Between January 1984 and December 2000, flexible fibreoptic bronchoscopy under fluoroscopic guidance was performed in 1372 patients with lung cancer having no visible endoscopic findings. Histological examination of specimens obtained by forceps biopsy and cytological examinations on imprints of biopsy specimens, brushing, selective bronchial lavage, curettage, transbronchial needle aspiration, rinse fluids of the forceps, brush, curette, and aspiration needle, and all fluids aspirated during the bronchoscopic examinations were evaluated for diagnostic power. Using these techniques, the overall diagnostic rate with bronchoscopy was 93.4%. The sensitivity of the histological examination was 76.9%; additional imprint cytology increased the sensitivity to 84.8% (Po0.0001), while additional cytology on the rinse fluid of the forceps increased the sensitivity to 83.7% (Po0.0001). The addition of both imprint cytology and cytology on the rinse fluid of the forceps increased the diagnostic rate to 86.2% (Po0.0001). Our results indicate that cytological examinations of the imprints of biopsy samples and the rinse fluids of the forceps and the brush improve the efficacy of fibreoptic bronchoscopy in the diagnosis of peripheral lung cancer. British Journal of Cancer (2003) Fibreoptic bronchoscopy is now routinely used for obtaining specimens from which to diagnose various respiratory diseases. Bronchoscopic techniques for the diagnosis of lung cancer consist of histological examination of specimens obtained by transbronchial biopsy and cytological examination procedures such as bronchial washing, brushing, and needle aspiration. Some combinations of these techniques have been reported to increase the diagnostic sensitivity for lung cancer compared with that of transbronchial biopsy alone. Harvesting the cells shed from the tumour during bronchoscopic needle aspiration or brushing, by rinsing the needle or brush, has already been reported to be useful (Steinmann and Creul, 1978;Smith et al, 1980). The cytological examination of the brushing samples or bronchial washing fluid in addition to the histological examination of specimens obtained by forceps biopsy has been shown to increase sensitivity in the diagnosis of peripheral lung tumour, by an estimated 12 -30% (Chaudhary et al, 1978;Mak et al, 1990;Govert et al, 1996). Furthermore, cytology on the rinse fluids has been previously reported to elevate the diagnostic sensitivity from 65.8 to 70.7% (Zavala, 1975;Rosell et al, 1998). On the basis of these results, we planned to evaluate the effects of a histological examination combined with cytological techniques such as imprints of biopsy specimens and cytology on the rinse fluids of the forceps and brush; the effects of such combinations have not been previously reported. The primary objective in this study was to evaluate the...