Abstract

The objective of this study was to define the colposcopic features of the normal anal canal and of anal human papillomavirus (HPV)-associated lesions, including anal intraepithelial neoplasia (AIN), and to correlate the colposcopic impression with the final histopathologic diagnosis. A controled colposcopic screening study of women considered at risk for HPV-associated anal epithelial abnormalities was carried out. All colposcopic assessments included a biopsy with matching histopathologic diagnosis. The study group consisted of 213 women who were considered at risk of anal HPV infection and AIN. A further group of 50 women, who had no previous history of ano-genital HPV infection or AIN and whose recent cervical smear was negative were recruited as controls. Informed consent was obtained from all patients, and the study was approved by the local ethical committee. In the control group of 50 women no AIN was detected. Normal histology was obtained in 45/50 (90%) biopsies where normality had been predicted on colposcopy. Histologic diagnosis in the at-risk group was normal in 143 (67%), subclinical papillomarvirus infection (SPI) in 24 (11%), and AIN of all grades (including three cases of early invasive squamous cancer in a field change of AIN III) in 46 (22%) patients. Nineteen of 24 (79%) cases of SPI were incorrectly predicted as normal on colposcopy, and another one (4%) as AIN I–II. Only four (17%) cases of SPI were correctly predicted at colposcopy. Of the 46 cases of histologically proven AIN, 26 (56%) were AIN I–II, and 20 (44%) were AIN III. Some 50% of AIN I–II were incorrectly predicted as SPI on colposcopy. Of the 20 AIN III lesions, 15 (75%) were correctly predicted by colposcopy. Three (20%) of these lesions contained foci of early invasion, of which in only one case (33.3%) was invasive disease suspected at colposcopy. Some 25% (5/20) of AIN III lesions were incorrectly diagnosed as AIN I–II at colposcopy. As is the experience with colposcopic assessment of the cervix, anal colposcopy predictions correlated well with the final histologic diagnosis, at the normal and high-grade AIN ends of the spectrum. The colposcopic predictive distinction between SPI and low-grade AIN (I–II) was less accurate. It was difficult to distinguish early invasive lesions within a field change of AIN III, from pure AIN III. In these studies there were three cases of early anal squamous carcinoma arising in AIN III lesions, two of which were unsuspected clinically.