Introduction and objectives:  Cervical cancer is the only gynaecological cancer that is staged clinically. The clinical stage of cervical cancer relies largely on the pelvic examination. The aim of this study is to analyse the discrepancy between clinical stage and pathological results, and to explore the accuracy of pelvic examination.
Methods:  We collected retrospective data from 818 patients with cervical carcinoma staged IB–IIB, who were treated with primary surgery from January 1999 to June 2007. Clinical stages of those patients were determined by pelvic examination without anaesthesia. After surgery, all the patients were assigned to pT category according to the pathological findings. Comparisons were made between these two stages.
Results:  The total concordance between clinical stage and pT category for stage IB–IIB was 53.1%, with an overestimation of 37.3% and an underestimation of 9.7%. The concordance in stage IB1, stage IB2, stage IIA and stage IIB were 85.4%, 77.4%, 35.3% and 20.5%, respectively. The most significant discrepancy was caused by the failure to detect the parametrial invasion accurately in stage IIB. The accuracy of pelvic examination to determine vaginal and parametrial disease was 70.2% and 74.0%, respectively.
Conclusions:  There are significant discrepancies between clinical stage and pathological results. Pelvic examination has its limitations in staging determination. Thus for operable cervical cancer, clinical stage alone is not reliable for selecting postoperative therapies and surgical staging system may be considered.