ASCUS is the most common epithelial abnormality diagnosed in cytology laboratories in the US. Recently, the clinical importance of this diagnosis has been seriously questioned, with some investigators advocating elimination of this diagnostic category. This might be inappropriate if the ASCUS designation does define a population that is at significant risk for the development of dysplasia. Cytology and surgical pathology reports for all patients diagnosed as ASCUS in our laboratory during 1990 were reviewed. Patients with previous dysplasia or carcinoma were excluded from analysis. The pathology reports for the subsequent 9.1 yr were obtained and follow-up data collected. In 1990, 15,860 cervical cytology cases were examined in our laboratory. A diagnosis of ASCUS was made in 1,117 cases (7.0%). After excluding 345 patients with previous dysplasia or human papillomavirus-related diagnoses and 129 patients with no follow-up specimens, 643 study patients remained. Among these, the mean number of subsequent cervical smears was 4.3 (range, 1-18). Subsequent histologic material was available for 134 (20.8%) patients and the mean number of surgical specimens was 1.5 (range, 1-10). Squamous intraepithelial lesion (SIL) or dysplasia was subsequently diagnosed in 197 patients (30.6%). High-grade squamous intraepithelial lesion (HSIL) or at least moderate dysplasia was diagnosed in 64 patients (10.0%). In 21 cases (3.3%) the high-grade dysplasia developed more than 2 yr after the first ASCUS diagnosis. Follow-up disclosed no cases of invasive carcinoma. Among ASCUS patients followed for up to 9 yr, 20% develop only low-grade SIL or mild dysplasia and 10% develop HSIL or moderate or severe dysplasia. ASCUS should be retained as a diagnostic category since it identifies a significant percentage of patients who are at an increased risk for the development of cervical dysplasia.