Atypical glandular cells of undetermined significance: the experience at long island jewish medical center.
Objective: Our aim was to determine the rate, accuracy and histological correlates of cytological smears re-ported as atypical glandular cells of undetermined significance (AGUS).
Methods: Cervical cytological smears reported as AGUS from January 1991 to December 1995 were reviewed and correlated with subsequent biopsies and Papanicolaou (Pap) smears.
Results: The AGUS rate was 0.53% (n = 114 of 21,468). Sixty-nine of the 114 cases (61%) had corresponding tissue specimens (biopsies and curettage), and 24 (21%) cases had subsequent Pap smears. No follow-up was available in 17 cases (15%). Four AGUS diagnoses represented recurrent high-grade squamous intraepithelial lesions. These lesions did not have biopsy confirmation but were treated on the basis of the Pap smear and colposcopic findings. Twenty-seven (23.6%) cases represented premalignant or malignant lesions, with a mean age of 52 years. Those patients with endometrial pathology had a mean age of 63 years, and those with squamous neoplastic lesions had a mean age of 39 years. Findings included 12 endometrial lesions (5 hyperplasias and 7 carcinomas); 2 adenocarcinomas in situ; 11 squamous lesions (10 squamous intraepithelial lesions and 1 squamous carcinoma); and 2 lesions with concomitant glandular and squamous abnormalities. Eleven of the 12 cases re-ported as AGUS-favor endometrial cells and 7 of the 12 cases reported as AGUS-favor neoplastic (AGFN) showed significant pathology. One screening error and 10 interpretive errors occurred (6 overcalled, 4 undercalled); however, no adverse out-comes were reported.
Conclusions: AGUS-favor endometrial cells usually occur in older patients and almost always is associated with premalignant or malignant lesions. Endometrial curettage should be performed on these patients. AGFN often indicates a significant lesion, usually occurs in younger patients, and colposcopically directed biopsy or endocervical curettage should be performed. AGUS-favor reactive and AGUS unspecified are less likely to represent significant lesions, and a conservative clinical approach (interval Pap smears) may be appropriate.