Evaluation of colposcopic accuracy in diagnosis of cervical neoplasia

To evaluate the accuracy of colposcopy in the diagnosis of cervical neoplasia.

A four year retrospective study from May, 1991 to May 1995 was conducted in 309 patients with abnormal cervical cytology who had undergone colposcopic examination at the Mc Cormick Hospital. The correlation between cervical cytology and colposcopically directed biopsy (CDB), colposcopic impression and CDB, CDB and subsequent conization or hysterectomy specimens were analyzed. The correlation was accepted as accurate when both reports were either the same or within one-step disparity of pathologic diagnosis and no invasive cancer was missed.

In comparison of cervical cytology and CDB, the correlation was accurate in 212 of 241 cases (88%). The CDB was found to be less and more severe than cytology in 16 cases (6.6%) and 13 cases (5.4%) respectively. Colposcopic impression was accurate in 221 of 241 patients (91.7%) when compared to the CDB results. Sixty nine patients had undergone conization. The results of the CDB were comparable with those of the conization specimens in all but 8 patients, giving a total accuracy of 88.4 per cent. Four cases of invasive cancer were missed at the time of colposcopy and biopsy but subsequently diagnosed by conization. Of 67 patients who underwent hysterectomy, the surgical specimens were found to be more advanced than CDB in 4 patients, giving a total accuracy of 94 per cent. Two cases of minimally microinvasive cancer were missed.

This study confirms the diagnostic value of colposcopy in the management of patients with abnormal cervical cytology.