Managing Atypical Squamous Cells of Undetermined Significance (ASCUS): Human Papillomavirus Testing, ASCUS Subtyping, or Follow-Up Cytology?

Abstract

Colposcopically directed biopsy finds high-grade squamous intraepithelial lesion (HSIL) in the histologic specimen of 5% to 15% of cervical smears taken from women with a diagnosis of atypical squamous cells of undetermined significance (ASCUS). This study was designed to investigate the ability of human papillomavirus (HPV) testing, ASCUS subtyping, and follow-up cytology to predict the presence of HSIL in the cervical smear. The study material consisted of 918 liquid-based cervical cytologic samples (ThinPrep test, Cytec Corp, Boxborough, MA) collected in a single pathologic laboratory between 1997 and 2000. HPV testing was carried out in 704 of the samples that had sufficient cytologic material remaining for follow-up cytologic examination, including 105 with a normal diagnosis, 426 cases of ASCUS, 126 with low-grade squamous intraepithelial lesion (LSIL), and 47 with HSIL. The cervical samples of 91% of the women with HSIL, 88% of those with LSIL, and 49% of those with ASCUS were positive for all HPV types, but only 27% of those with normal cervical cytology contained all HPV types. When only high-risk HPV types were considered, the percentage of positivity for normal samples dropped to 19% and the percentage for HSIL increased to 96%. The ASCUS category was further sorted into ASCUS, favor reactive (AFR); ASCUS, not otherwise specified (ANOS); ASCUS, favor LSIL (AFLS); and ASCUS, favor HSIL (AFHS). Both high- and low-risk HPV types were found in AFR and ANOS samples slightly more frequently than in normal samples. High- and low-risk HPV types were found in AFLS far more frequently than in normal smears but slightly less than for LSIL. This pattern continued with the progression from LSIL to AFHS and from AFHS to HSIL. Three hundred eighty-six women of the 704 women had a year or more of follow-up data available, including a biopsy or other tests within 1 year of the cervical smear. Eight women with an initial diagnosis of ASCUS progressed to HSIL (3.4%). The presence of HPV was detected in the samples of all of these women, including high-risk HPV in seven of the eight, yielding a sensitivity of 88% and a specificity of 73% for high-risk HPV testing in predicting the development of HSIL in women with ASCUS. The positive predictive value for high-risk HPV testing was 11%. The sensitivity of any type HPV testing was 100%, but the specificity was only 48%. In all, 64% of the women for whom follow-up data were available regressed to normal. Compared with the 3.4% of ASCUS women who were found to have HSIL, 11% of the positive high-risk HPV patients developed HSIL. A designation of AFHS led to HSIL in 25% of the women with an initial diagnosis of ASCUS. Only 1.1% of the ASCUS women designated AFR later developed HSIL.