Glandular Lesions of the Uterine Cervix: Cytology and Histology Updates

Dear Readers, This issue of the journal spotlights glandular lesions of the uterine cervix and describes updates in their cytology and histology. In health resource–rich countries such as the United States and Canada, the control of cervical squamous cell carcinoma is the principal target of secondary prevention with cytology and/or human papillomavirus (HPV) screening to detect premalignant lesions and, more recently, of primary prevention using mass vaccination to minimize acquisition of the etiological HPV genotypes. Most cervical adenocarcinomas resemble squamous cell carcinoma in that pathogenesis involves a premalignant lesion (adenocarcinoma in situ [AIS]), and etiology involves persistent HPV infection. The prevention of cervical adenocarcinoma is therefore an expected secondary benefit of squamous cell carcinoma control strategies. Primary invasive adenocarcinoma of the uterine cervix is an uncommon cancer that constitutes 20% to 25% of cervical malignancies. The 2010–2014 average annual incidence rate in the United States was 1.8 per 100,000 females based on data from the North American Association of Central Cancer Registries (personal communication with K. Miller and J. Ahmedin, American Cancer Society, Inc, Atlanta, Ga, November 30, 2017). According to data from the 9 oldest Surveillance, Epidemiology, and End Results registries, incidence rates increased from 1975 to 1990, but have since been stable.1 In that same period, the relative incidence rates of squamous and glandular cancers declined from a factor of 8 to 3. While these data support the benefits of cancer control strategies both for squamous and glandular malignancies, stratified analysis of the glandular data shows the incidence of adenocarcinoma in women younger than 50 years is increasing. Reasons for this have not been studied but may relate to increasing estrogen exposure secondary to obesity and which is a cofactor in cervical adenocarcinoma pathogenesis. Some recent developments in the pathology of glandular tumors of the uterine cervix relate to classification and biomarker testing. The current World Health Organization (WHO) classification of cervical “Glandular Tumors and Precursors” and “Other Epithelial Tumors” published in 2014 revised the categories and reduced the number of malignant histotypes included in the 2003 classification.2,3 Some of the changes consisted of the formation of a separate category named endocervical adenocarcinoma, usual type. In the 2003 version, this histotype was named endocervical mucinous adenocarcinoma and classified as a mucinous carcinoma subtype, along with intestinal and signet ring carcinoma. Another change was the addition of the newly described gastric-type carcinoma as a subtype of mucinous carcinoma and a high-grade variant of minimal deviation endocervical adenocarcinoma. Adenocarcinoma admixed with neuroendocrine carcinoma became a separate novel category, and the categories of early invasive adenocarcinoma and glandular dysplasia were deleted. The WHO classifications are predicated on morphology so that they have global applicability and can be used in centers without access to immunohistochemical or molecular technologies. In a study of pathologists' agreements analyzed by κ statistics, WHO 2003 collapsed into 6 categories and WHO 2014 collapsed into 7 categories of individual or bundled histotypes diagnosed on hematoxylin-eosin–stained slides were moderately reproducible between different pathologists and trended toward substantial between the same pathologist.4 Reproducibility between different gynecological pathologists was substantial and contrasted with the moderate reproducibility of anatomical pathologists. These results indicate reproducibility is unaffected by changes in the WHO 2014 classification, but can improve to substantial between different pathologists when they have expertise in gynecological pathology. For morphology-based reproducibility of any future updates of the WHO classification to reach almost perfect κ agreement across the broad breadth of real-life pathology practice, however, more refinement of the WHO categories in terms of their histotypes and their morphological criteria will be needed. A newly proposed classification of invasive adenocarcinomas of the endocervix is based on morphologic features linked to etiology.5 This pathogenic binary system divided adenocarcinomas into those with and without HPV-associated features and which the authors defined as the presence or absence, respectively, of luminal mitoses and apoptosis visible at scanning power. The classification was highly accurate, with a sensitivity of 95% and specificity of 97%, when compared with the criterion standard of HPV chromogenic in situ hybridization testing. Further study of this promising system will be needed as the accuracy of the classification has not been correlated with clinical outcome, nor has it been compared with other HPV testing platforms, or other criterion standards such as the pattern of genomic abnormalities in the tumors. Neither have reproducibility studies or comparisons with the reproducibility of the current WHO classification been published. A pattern-based classification for HPV-related adenocarcinomas to predict the risk of nodal metastases and recurrence has been proposed.6 This 3-tiered system is based on the absence (pattern A) or presence of destructive stromal invasion and lymph vascular invasion (patterns B and C). Lymph node metastases and recurrence correlate with patterns B and C, but not with pattern A, and this suggests lymphadenectomy is not necessary for the treatment of pattern A tumors. The overall reproducibility of this classification system is moderate, and refinements are being studied in order to improve it.7 If this classification could be accurately and reproducibly diagnosed on biopsy, curettage, or loop electrical excision procedure/cone samples, it would enable planning of definitive patient treatment. This was studied by Djordjevic et al, who reported pattern prediction on biopsy was suboptimal, whereas that of loop electrical excision procedure/cone, although very high, was not perfect.8 Future refinement of the diagnostic criteria may improve the performance of the classification and facilitate acceptance and implementation. Quality health care and clinical research are dependent on pathologists having a high degree of diagnostic accuracy when evaluating samples. The accuracy of glandular lesion diagnosis can improve with the appropriate use of immunohistochemical biomarker testing. A systematic review and meta-analysis of literature published between 1975 and 2015 is the most precise and reliable information on all 56 antibodies that were investigated for diagnostic utility.9,10 Despite the publication of almost 1000 articles, biomarker accuracy in this practice area is understudied mostly because of the limitations of the study designs. Based on the systematic review and meta-analysis, only a few biomarkers emerged as useful in guiding the practicing pathologist in the judicious selection of appropriate diagnostic immunohistochemical panels. The only biomarker with enough studies for a case-control meta-analysis with case defined as positive for malignant glandular tumor and control as negative, and which could be displayed as forest plots, was p16. Sensitivity and specificity of p16 expression were high for AIS and mucinous adenocarcinomas in comparison with normal/benign endocervix and benign endocervical lesions. Sensitivity was also high, but the specificity was unreliable when AIS and mucinous adenocarcinomas were compared with benign endometrioid lesions. This related mainly to variability among studies in the classification of p16 expression as positive or negative. Among biomarkers tested as case-comparator discriminants, for example, AIS from adenocarcinoma histotypes, and the different adenocarcinoma histotypes from each other, there were insufficient data for a meta-analysis. Instead, differences in positive expression of 50% or greater and which were infrequent, were identified as potentially clinically useful. Premalignant case-adenocarcinoma comparator biomarkers were HIK1083, α-SMA, PAX8, VIL1, CEA, p53, p16, and CD10, and only α-SMA had a difference of 100%. The adenocarcinoma case-comparator biomarkers were CEA, p53, claudin 18, HIK1083, p16, calretinin, CD10, PR, chromogranin, MUC6, vimentin, and p63, and none had a difference of 100%. Expression differences of these biomarkers were visually displayed as heat maps of the various case-comparator scenarios and can be used to guide the selection of the most appropriate biomarker panel for a specific differential pathology diagnosis. The issue kicks off with the article by McCluggage entitled “Primary Adenocarcinoma of the Uterine Cervix: Grossing, Staging and Synoptic Reporting.” The positioning reflects their importance in clinical staging and patient prognosis and the critical contributions of pathologists to those assessments. In a previous editorial of this journal on the Lower Anogenital Tract Squamous Terminology workshop in 2013, the controversy and lack of consensus on the measurement of horizontal extent of superficially invasive cervical carcinomas were noted.11 Because nonstandardized and inconsistent measurement of this element could lead to understaging of the cancer, the editorial made some suggestions about future steps or actions that could be pursued in order to resolve the controversy. Thereafter, the International Collaboration on Cancer Reporting undertook a review and updating of cervical cancer grossing and synoptic reporting to address and resolve this and other issues relating to pathological staging. The article in this issue describes the International Collaboration on Cancer Reporting procedure for creating required and recommended data reporting elements for cervical cancer synoptic reporting and customizes them where needed to the challenges of staging glandular malignancies. It clearly and comprehensively addresses each challenge and provides evidence-based solutions or pragmatic ones when no evidence exists. This is a valuable up-to-date document that brings clarity to some of the more controversial areas of cervical cancer, pathological staging. The next 2 articles are a complementary pair as they are comprehensive reviews of the pathology of malignant and benign glandular tumors. These are challenging diagnoses in resection specimens when the entire tumor is available for examination and even more challenging in a sample obtained by curettage or biopsy and that may or may not be representative. The article by Pirog on “Primary Adenocarcinoma of the Uterine Cervix: Classification and Grading” describes and illustrates malignant histotypes as classified by WHO 2014. The WHO grading of adenocarcinoma, usual type, and details of the proposed pattern-based classification of HPV-positive adenocarcinomas are also included. The HPV-positive status of cervical adenocarcinoma ranges from 62% to 100% and varies by histotype. Human papillomavirus–positive adenocarcinomas include endocervical usual type, endometrioid, intestinal, and villoglandular. Human papillomavirus–negative adenocarcinomas are serous, mesonephric, and gastric type and probably also include clear cell. This comprehensive article describes and illustrates the key pathological features of each histotype and the differential diagnoses and lists the most appropriate biomarkers to assist in classification. It also describes the criteria used to diagnose invasion arising from AIS. The Harrison and Nucci article on “Glandular Pseudoneoplasms of the Uterine Cervix” highlights the most diagnostically challenging of the benign glandular lesions. Similar to the Pirog article, they describe and illustrate the key pathology features and differential diagnoses and list the most appropriate immunohistochemical biomarkers to assist in ruling out a malignant lesion and to classify the phenotype. The benign proliferations can be classified as mucinous, endometrioid, mesonephric, gastric, or prostatic. The section on the interpretation of microglandular hyperplasia in small samples provides valuable information on the morphological criteria and possible value of vimentin staining to distinguish it from microglandular hyperplasia-like endometrioid carcinoma of the endometrium. This is a very challenging diagnosis particularly in biopsy or curetted samples, and the authors highlight the value of detecting recurrent mutations in KRAS and NRAS by next-generation sequencing to support a diagnosis of endometrial malignancy. The authors also describe how that test can also aid in distinguishing mesonephric hyperplasia from mesonephric carcinoma and from mesonephric-like endometrial carcinoma. Both articles reference gastric-type adenocarcinomas and their precursor lesions, and this spectrum of pathology is described in even greater detail and illustration in the next article, which is the case review of Park entitled “Gastric-Type Adenocarcinoma of the Uterine Cervix and Precursor Lesions.” This newly described and rare HPV-negative tumor is considered a high-grade variant of minimal deviation endocervical adenocarcinoma. It can be challenging to diagnose on biopsy or curettage as the p16- and HPV-negative status is not consistent with most cervical primaries, and the high-grade morphology combined with p53 overexpression may lead to its misclassification as a high-grade serous carcinoma of noncervical origin. The gastric-type benign lesions include pyloric gland metaplasia, type A tunnel clusters, and lobular endocervical glandular hyperplasia, and the probable premalignant lesions are atypical lobular endocervical glandular hyperplasia and gastric-type AIS. While the majority of adenocarcinomas of the uterine cervix are primary tumors, metastatic spread can occur from müllerian sites such as the endometrium, fallopian tube, and ovary and nonmüllerian sites such as the colon and breast. Because treatment and patient outcome can be very different for a primary versus a metastatic adenocarcinoma, pathologists diagnosing adenocarcinoma based on a cervical biopsy or curettage must not only histotype the adenocarcinoma, but also establish the primary site of origin of the malignancy. The case review by Djordjevic et al entitled “Metastatic Adenocarcinoma to the Uterine Cervix Versus Primary Endocervical Adenocarcinoma” describes and illustrates this diagnostic challenge and lists the useful immunohistochemical biomarkers. The role of HPV testing is also addressed, and the critical importance of clinicopathological correlation is emphasized. The final article by Li and Zhao is a case review–based description and illustration of “The Atypical Glandular Cell Pap Test: Recent Updates.” With the relative and absolute increase in the incidence of cervical adenocarcinoma, the recognition and classification of atypical glandular cells (AGCs) in a Papanicolaou test are becoming more important. This comprehensive article addresses the cytological criteria of The Bethesda System AGC categories and describes the histopathologic correlates. In the past, most AGC tests on tissue follow-up were a squamous abnormality. In recent times, this has changed, and now most are a glandular abnormality. This may relate to more accuracy in the cytological diagnosis as well as changes in the prevalence of malignant glandular lesions of the uterine cervix. In closing, I am extremely grateful to all the authors for making this issue a reality. I hope you find the topics as interesting and educational as I did and that they will be of benefit to you in your everyday practice.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.