According to the recommendations of the Bethesda System, the diagnosis of
atypical squamous cells of undetermined significance (
ASCUS) should be further qualified, when possible, as to whether a reactive or a
squamous intraepithelial lesion (SIL) is favored. To determine the utility of this recently proposed terminology, we undertook this study to correlate the diagnosis of
ASCUS (with or without qualifiers) with results obtained from examination of biopsy specimens. All patients were identified for the study who had a coloposcopically obtained cervical biopsy specimen or endocervical
curettage specimen recorded in the surgical pathology files at Beth Israel Hospital, Boston, Massachusetts, from April 1994 through September 1994 and had either prior or concurrent Pap smear(s) reported as
ASCUS,
ASCUS-favor reactive,
ASCUS-favor SIL, or SIL-low grade (SIL-LG). Patients with a cytologic diagnosis of SIL-LG served as a reference group. A total of 435 patients with 485 Pap smears were included. The prevalence rates of biopsy-proven SIL in patients with a cytologic diagnosis of
ASCUS-favor reactive,
ASCUS,
ASCUS-favor SIL, and SIL-LG were 10, 28, 36, and 55%, respectively. The difference between cases diagnosed as
ASCUS (with or without qualifiers) and SIL-LG, with respect to the presence of SIL at examination of the biopsy specimen, was statistically significant (P < 0.001 for all correlations). Cases diagnosed as
ASCUS-favor reactive had a significantly lower rate of biopsy-proven SIL compared with those diagnosed as
ASCUS and
ASCUS-favor SIL (P < 0.01 for both correlations). A significant proportion of biopsy specimens with Pap smear diagnosis of
ASCUS-favor SIL had SIL-high grade (15%). In contrast, an underlying SIL-high grade is much less likely in patients with
ASCUS (unqualified) (3%) or when a reactive process is favored (3%). It seems justified to manage patients with
ASCUS-favor SIL in a manner similar to those with SIL-LG. A
conservative management seems appropriate for patients with
ASCUS and
ASCUS-favor reactive.