J L Benedet1, J P Matisic, M A Bertrand. 1. Divisions of Gynaecologic Oncology and Pathology, BC Cancer Agency and The University of British Columbia, Vancouver, BC, Canada V5Z 4E6. lbenedet@bccancer.bc.ca
Abstract
OBJECTIVE: To determine the diagnostic correlation between referral cytology, initial biopsies and colposcopic impression in patients assessed in a provincial cytology screening program. METHODS: A retrospective review of the computerized cytology screening database for British Columbia (BC), to identify all patients having their first colposcopy between 1986 and 2000 in 24 participating clinics constituted the study population. 84244 patient records were identified for analysis. Colposcopies were performed mainly by 37 general gynecologists as part of a province-wide colposcopy program. Correlation of cytology, colposcopic impression and directed biopsies was performed. RESULTS: The colposcopic impression correlated with the referral cytology within one degree in over 90% of cases. Colposcopists felt cytology underestimated disease in 1.5% and overestimated disease in 8.3%. Cytology-histology correlation within one degree occurred in 82%. Cytology underestimated the result of the biopsies in 2.3% and appeared to overestimate disease in 16.1% of patients. Patients with HSIL cytology had corresponding lesions in 77%, with a further 4.9% having LSIL disease. The predictive accuracy of colposcopy increased with advancing severity of disease expected. As the degree of cytological abnormality worsened, the predictive accuracy of colposcopic diagnosis increased. CONCLUSIONS: Both cytology and colposcopy have high sensitivity but low to moderate specificity. Colposcopy is most accurate in identifying high-grade diseases. Colposcopic impression correlates closely with the cytology diagnosis and combining the two produces optimum results.
OBJECTIVE: To determine the diagnostic correlation between referral cytology, initial biopsies and colposcopic impression in patients assessed in a provincial cytology screening program. METHODS: A retrospective review of the computerized cytology screening database for British Columbia (BC), to identify all patients having their first colposcopy between 1986 and 2000 in 24 participating clinics constituted the study population. 84244 patient records were identified for analysis. Colposcopies were performed mainly by 37 general gynecologists as part of a province-wide colposcopy program. Correlation of cytology, colposcopic impression and directed biopsies was performed. RESULTS: The colposcopic impression correlated with the referral cytology within one degree in over 90% of cases. Colposcopists felt cytology underestimated disease in 1.5% and overestimated disease in 8.3%. Cytology-histology correlation within one degree occurred in 82%. Cytology underestimated the result of the biopsies in 2.3% and appeared to overestimate disease in 16.1% of patients. Patients with HSIL cytology had corresponding lesions in 77%, with a further 4.9% having LSIL disease. The predictive accuracy of colposcopy increased with advancing severity of disease expected. As the degree of cytological abnormality worsened, the predictive accuracy of colposcopic diagnosis increased. CONCLUSIONS: Both cytology and colposcopy have high sensitivity but low to moderate specificity. Colposcopy is most accurate in identifying high-grade diseases. Colposcopic impression correlates closely with the cytology diagnosis and combining the two produces optimum results.
Authors: Judith R Mourant; Thérese J Bocklage; Tamara M Powers; Heather M Greene; Maxine H Dorin; Alan G Waxman; Meggan M Zsemlye; Harriet O Smith Journal: J Low Genit Tract Dis Date: 2009-10 Impact factor: 1.925