The reliability of a cytological prediction of cervical adenocarcinoma in situ.

A spectrum of changes precedes clinical invasive adenocarcinoma of the cervix. Cytological and histological criteria for the diagnosis of adenocarcinoma in situ (ACIS) are becoming more clearly defined. A 5-year prospective study was undertaken to test the accuracy of a cytological prediction of ACIS. From a total of 290,000 cervical smears 54 predictions of ACIS were made: 33 (61%) alone and 21 (39%) with associated squamous carcinoma in situ (SCIS). The rate of reporting ACIS was compared to the rates for intraepithelial and invasive squamous lesions and for frank invasive adenocarcinoma. The findings suggest that ACIS is being underdiagnosed. Forty-seven patients were adequately investigated; 46 had intraepithelial or invasive malignancy. There were 10 cases of ACIS, 10 ACIS with SCIS, 9 microinvasive adenocarcinoma, 5 invasive adenocarcinoma, 2 microinvasive adenosquamous carcinomas, 1 invasive adenosquamous carcinoma, 8 SCIS and 1 endometrial carcinoma. There was one true false positive report. Thus cervical glandular neoplasia was confirmed in 37 patients (79%), 13 of these having adenosquamous tumours. Because 98% of patients had in situ or invasive malignancy and because 36% of cases were invasive (though mostly microinvasive) prompt investigation, by cone biopsy, must follow a cytological report of ACIS.