Cervical cancer screening programs have significantly decreased the incidence of cervical squamous cell carcinoma; however, the percentage and total number of cervical adenocarcinoma cases has been increasing. Several hypotheses have been proposed for this increased incidence of cervical adenocarcinoma, including the low sensitivity of the Papanicolaou (Pap) test in diagnosing glandular lesions and the difficulty in sampling cervical glandular lesions due to their location higher in endocervical canal.
The term “atypical glandular cells (AGC)” is an interpretative category used for specimens with cellular features that exceed those of reactive changes but fall short of the criteria for glandular neoplasia (in situ or invasive adenocarcinoma). This term, introduced in the 2001 Bethesda system for reporting cervical cytology, has replaced the term “atypical glandular cells of undetermined significance” from the 1988 Bethesda system. Approximately 0.3% to 0.5% of Pap tests have a reported result of AGC,[4, 5] which can be further subcategorized into AGC, not otherwise specified; AGC, endocervical cells; and AGC, endometrial cells. Retrospective studies with both the conventional Pap test and liquid-based cytology have demonstrated that 10% to 39% of patients with AGC demonstrated malignant or premalignant diseases in histologic follow-up biopsies, including both squamous lesions (squamous dysplasia and squamous cell carcinoma) and glandular lesions (adenocarcinoma in situ [AIS], cervical adenocarcinoma, endometrial adenocarcinoma, or metastatic carcinoma).[4, 6] Due to the substantial risk of premalignant or malignant diseases associated with AGC, current American Society for Colposcopy and Cervical Pathology guidelines recommend colposcopy with endocervical sampling for women with all subcategories of AGC except atypical endometrial cells. Endometrial sampling is recommended in conjunction with colposcopy and endocervical sampling in women aged ≥ 35 years with all subcategories of AGC.
To the best of our knowledge, the Pap test has been the most successful screening test to detect cervical squamous lesions; however, its sensitivity and specificity for the detection of endocervical/endometrial glandular lesions remain low, despite previous reports that a significant percentage of patients with AGC demonstrated significant lesions on histological follow-up.[4, 6, 7] Reasons for the insufficient sensitivity and specificity include the higher location in the endocervical canal, multiple foci, and smaller size of glandular lesions; irregular shedding of abnormal glandular cells; and the low accuracy of both cytologic and histologic interpretations for glandular lesions. Some studies have shown that a diagnosis of AGC was not well reproduced when interpretations between different laboratories and pathologists were compared. In addition, to our knowledge there is no consensus among pathologists/cytotechnologists pertaining to the cell of origin of atypical glandular cells.[8-12] To improve the sensitivity and specificity of AGC Pap tests, cell block preparation has been suggested as a complementary tool for evaluating gynecologic cytology specimens, similar to the evaluation of nongynecologic cytology specimens.[13, 14]
The study by Xing et al in this issue investigated the usefulness of the Cellient Automated Cell Block System (Hologic Inc, Bedford, Mass) in the differential diagnosis of AGC and demonstrated that the rate of negative/low-grade squamous intraepithelial lesions was significantly lower (from 85% to 55%) and that of endocervical or endometrial adenocarcinomas was significantly higher (from 8% to 36%) in cases with automated cell block preparation compared with cases without cell block preparation. Their data also suggested that cell block preparation is able to aid pathologists in further classifying AGC into other categories. For example, 30% of AGC samples originally interpreted by cytotechnologists were finally diagnosed as either negative or neoplasms after pathologists reviewed cell block sections. Second, cell block preparation was able to increase the positive predictive value of AGC for detecting endocervical or endometrial adenocarcinoma because 36% of cases demonstrated endocervical or endometrial adenocarcinoma in histologic follow-up results.
However, this study has limitations similar to those of other retrospective studies. As mentioned in the article by Xing et al, cell block was more likely to be ordered in cases in which the pathologist favored either a negative or malignant process; therefore, this would create sample selective bias in the cell block group, resulting in an increased positive predictive value.
In addition to cell block preparation, other ancillary studies, including high-risk human papillomavirus (hrHPV) testing and immunohistochemical dual-immunostaining (p16/Ki-67), have been suggested to have value in the differential diagnosis of AGC. For example, a study by Zeferino et al found that neoplasia was significantly associated with positive hrHPV testing in women with AGC, not otherwise specified. Castle et al found that hrHPV-positive women were at a 10.4% risk of cervical cancer and a 0% risk of endometrial cancer, but women who were hrHPV negative were at a 10.5% risk of endometrial cancer among those women aged ≥ 50 years. Our recent study demonstrated that an AGC Pap test with a positive hrHPV result was strongly associated with grade 2 or 3 cervical intraepithelial neoplasia and AIS in women aged < 50 years and that AGC and hrHPV-negative results among older women may be associated with endometrial adenocarcinomas. Similar findings were also reported by Chen and Yang from the Cleveland Clinic. Overall, the negative predictive value of HPV testing for detecting clinically significant cervical lesions (grade 2 or 3 cervical intraepithelial neoplasia positive/AIS positive) was reported to approach 100%.[6, 8, 17]
Traditional cell block preparation techniques are mostly manual, difficult to standardize, and not feasible for hypocellular samples. The Cellient cell block system as used in the study by Xing et al is an automated system that allows for the efficient and rapid processing of micro-sized cytology samples without manual handling, and with the collection of cytologic samples taking place at one defined plane in paraffin for histologic sectioning.[18-20] In this study and others by the same group, the authors found that this technique minimized or eliminated the potential for cross-contamination, reduced turnaround times, and produced histology sections with excellent quality that was comparable to that of routine paraffin sections.
Cell block preparation provides histologic sections with which to evaluate morphology and materials on which to perform immunostains, which are very helpful under certain circumstances, such as p16/Ki-67 dual-immunostaining for high-grade squamous dysplasia, AIS, and cervical carcinoma; estrogen receptor/progesterone receptor for endometrial carcinoma; and p53/WT1 for high-grade ovarian serous carcinoma, which is the most common metastatic carcinoma detected in Pap tests.
Although cell block preparation can be valuable in the differential diagnosis of AGC cytology, especially for hypercellular cytologic specimens, it is debatable whether it is practical for all AGC cytologic specimens when cost and effectiveness are considered. More studies are warranted to validate the data and standardize the criteria for cell block adequacy. At this time, cell block preparation is best ordered by pathologists on a case-by-case basis, instead of as a routine practice.