OBJECTIVEThe aim of this study was to establish a standard protocol for detection of EGFR mutations in cytologic specimens.

METHODS287 cytologic samples were collected from the patients who were suspected of having lung cancer at six hospitals in Beijing. A detection protocol for EGFR mutations was designed. Two comparative experiments were carried out for the coincidence in EGFR mutation rates between direct sequencing (Seq) and amplification refractory mutation system (ARMS) methods, and between 40 matched cytologic samples with formaldehyde-fixed paraffin embedded (FFPE) cytologic blocks and cytospin slides.

RESULTSTumor cells were found in 236 out of 287 cases (82.2%, 236/287) . Among them, there were 31 cases (13.1%, 31/236) of low tumor cell content samples and 205 cases (86.9%, 205/236) of high tumor cell content samples. 180 cases in the high tumor cell content samples (87.8%, 180/205) were diagnosed to be consistent with NSCLC. 25 out of 194 cases were ruled out or indefinite to be diagnosed as NSCLC by immunohistochemistry. By direct sequencing, the mutation rate of EGFR was 27.8% (50/180) in NSCLC samples and 28.2% (50/177) in adenocarcinoma samples (high tumor content samples) . By ARMS, the mutation rate of EGFR was 45.6% (82/180) in NSCLC samples and 46.3% (82/177) in adenocarcinoma samples (high tumor content samples). The EGFR mutation rate in low tumor content samples was 38.7% (12/31) , there was no significant difference in EGFR mutation rates between the groups of low tumor cell content samples and high tumor cell content samples (P = 0.12). The concordance rate of EGFR mutation rates was 100% between scraping tumor cells from slides samples and from FFEP blocks in the 40 matched samples. Forty-eight out of 180 definitive NSCLC patients received Gefitinib therapy. The FPS was 12 months in the gefitinib-treated ARMS⁺ group and 2 months in the ARMS⁻ group (P < 0.001), and the OS was 19 months in the gefitinib-treated ARMS⁺ group and 7 months in the ARMS⁻ group (P = 0.003), but no significant differences were found in the efficacy (PFS and OS) of Gefitinib between Seq⁺ and Seq⁻ groups (P = 0.227, P = 0.510, respectively), and Seq⁺/ARMS⁺ and Seq⁻/ARMS⁺ groups (P = 0.354, P = 0.334, respectively).

CONCLUSIONSThe detection protocol for EGFR mutations in cytological specimens introduced in this study is tested to be reliable and feasible. Pathological evaluation and immunohistochemistry are important in the detection procedure of EGFR mutations in cytologic specimens. High sensitivity methods should be selected for detection of EGFR mutations in cytologic samples.