Clinical significance of a cytologic diagnosis of atypical glandular cells of undetermined significance

OBJECTIVE: To assess the clinical significance of a cytologic diagnosis of atypical glandular cells of undetermined significance (AGUS) and determine the most appropriate evaluation of these patients. STUDY DESIGN: Between 1993 and 1995, 44,217 Papanicolaou smears were evaluated at Allegheny University Hospitals, Medical College of Pennsylvania Campus. There were 108 (0.24%) cases of AGUS smears during that time. No clinical information was available for 14 patients, and 19 were lost to follow-up. The charts of the remaining 75 cases were retrospectively reviewed. RESULTS: Tissue specimens were available for 62 of the 75 patients. There were 26 (42%) with no significant histopathologic findings, 13 (21%) with polyps, 5 (8%) cases of endometrial hyperplasia, 2 (3%) with endometrial adenocarcinoma, 12 (19%) with cervical intraepithelial neoplasia (CIN), 1 (2%) with adenosquamous carcinoma of the cervix, 2 (3%) with cervical adenocarcinoma in situ and 1 (2%) case of metastatic breast cancer. The total number of patients with significant histopathology other than polyps was 23 (37%). The median age of the patients was 49 years. There were more cases of endometrial hyperplasia and endometrial cancer (19%) in women 49 years or older than in younger women; only one (3%) case of endometrial hyperplasia was detected in the younger age group (P=.057). Patients who underwent more-aggressive evaluation (colposcopy and biopsies plus endometrial sampling, cone biopsy or hysterectomy) had greater numbers of abnormal histopathologic findings (55%) than patients who underwent endometrial sampling only (21%) or those who underwent colposcopy and biopsy only (33%). This difference approaches statistical significance (P=.057). A significant proportion of patients with a history of CIN and a cytologic diagnosis of AGUS were found to have CIN (47%), while 8% of those with no history of CIN were found to have CIN (P = .002). Fifty percent of patients with a history of cancer (all had breast cancer) and AGUS had abnormal histopathology. Patients with a subclassification of AGUS 'favor neoplasia' had a greater proportion of significant histopathology (72%) as compared to AGUS 'unspecified' (26%) and AGUS 'favor reactive' (20%) (P = .003). CONCLUSION: A significant proportion of women with a cytologic diagnosis of AGUS have abnormal histopathology. Heightened awareness should be raised in patients with AGUS and a history of CIN or cancer and in those with the AGUS subclassification 'favor neoplasia'.