Biopsy Findings in Patients with Atypical Glandular Cells of Uncertain Significance

ABSTRACT & COMMENTARY

Synopsis: Patients with atypical glandular cells of uncertain significance have a significant risk of malignant lesions, which are nearly all glandular and predominantly arise from the endometrium.

Source: Eddy GL, et al. Am J Obstet Gynecol 1997; 177:1188-1195.

Eddy and associates retrospectively analyzed computerized records over a three-year period. The most significant histologic diagnosis from all biopsy specimens submitted was compared with the subcategory of the first Pap smear obtained showing atypical glandular cells of uncertain significance (AGUS). The purpose of this study was to define the use and limitations of the category of AGUS by studying histologic findings in biopsy specimens obtained from patients.

Biopsy results were available for 531 of 1117 patients with AGUS (48%). Biopsy-proved preinvasive (83%) or invasive (17%) lesions were present in 191 patients (36%). Eighty-nine percent of the preinvasive lesions were squamous, whereas 97% of the invasive lesions were glandular. Glandular lesions were more likely to be invasive, whereas squamous lesions were more likely to be preinvasive. Twenty-eight patients had endometrial carcinoma, which represents 88% of all invasive carcinomas detected. Eddy et al conclude that almost three-fourths of patients with AGUS and with lesions have squamous lesions-not glandular as suggested by the name of the category. Unlike patients with atypical squamous cells of uncertain significance (ASCUS), patients with AGUS have a significant risk of malignant lesions, which are nearly all glandular and predominantly arise from the endometrium.

COMMENT BY DAVID M. GERSHENSON, MD

The category of AGUS arose out of the Bethesda system classification. We now know that the frequency of this diagnosis is approximately one-tenth that of the diagnosis of ASCUS. Although the findings of this study are not particularly new, this series does appear to be the largest reported to date in which biopsy followed the diagnosis of AGUS on Pap smear.

The main message from this report for the clinician is that the diagnosis of AGUS on Pap smear requires further investigation. Interestingly, the vast majority of the lesions found on biopsy are squamous, not glandular. Eddy et al speculate that this is because metaplastic cells associated with squamous intraepithelial lesions frequently retain cytologic features of glandular epithelium. Importantly, this paper also includes a literature survey, which indicates a similar trend. On a practical note, I find no mention of the recommended work-up of a patient with the cytologic diagnosis of AGUS. Until proven otherwise, based on my reading, I would choose to perform colposcopy with directed biopsy of any lesions in combination with endocervical and endometrial biopsies. (Dr. Gershenson is Professor and Deputy Chairman, Department of Gynecology, M.D. Anderson Cancer Center, Houston, TX.)