THE RELIABILITY OF PRECONIZATION DIAGNOSTIC EVALUATION IN PATIENTS WITH CERVICAL INTRAEPITHELIAL NEOPLASIA AND MICROINVASIVE CARCINOMA

The accuracy of preconization cytology and histology was evaluated in 536 patients undergoing combination laser conization. Exact agreement between cytology and cone diagnosis was observed in 41.8% of the patients. The lowest agreement, 13.6% was demonstrated in cytologic cervical intraepithelial neoplasia (GIN) I, the highest in cytologic CIN III, 64.3%. Concerning microinvasive disease, the positive predictive value of cytology was only 27.3%; the negative predictive value 94.6%. When cytology showed CIN II or worse, the cone biopsy showed CIN or invasive disease in 92.8%. Exact agreement between preconization histology and the cone biopsy was found in 59.5%. Highest agreement was observed in CIN III lesions, 67.1%, and the lowest agreement in CIN II lesions, 42.7%. When preconization showed CIN II, a higher grade of lesion was found in the cone biopsy in 29.1%. Cone biopsy revealed invasive disease in 38 cases. In 24 cases, invasive disease was not demonstrated prior to conization, corresponding totally to 4.7% of patients not suspected to have invasive disease prior to conization. Regarding invasive disease, the sensitivity of preconization histology was 36.8%, the positive predictive value 58.3%, and the negative predictive value 95.3%. Kappa statistics demonstrated rather low agreement between cone diagnoses and preconization diagnoses. These results confirm the potential risk of over-looking invasive disease by conventional preconization evaluation and demonstrate the need for excisional methods in the management of cervical intraepithelial neoplasia to provide a sufficient specimen for diagnostic purposes. Combined with the therapeutic results, combination laser conization was a reliable diagnostic and therapeutic method in the management of patients with CIN and microinvasive cancer of the cervix. (C) 1995 Academic Press, Inc.