Atypical Endometriosis: Comprehensive Characterization of Clinicopathologic, Immunohistochemical, and Molecular Features

Abstract

Atypical endometriosis (A-EMS), defined by cytologic atypia and/or crowded glands resembling endometrial intraepithelial neoplasia, remains poorly understood. We aimed to refine the morphologic, immunohistochemical, and molecular features of A-EMS in an institutional series. Cases were identified through a structured search and reviewed by 2 pathologists. Immunohistochemistry and comprehensive sequencing using a panel 447-gene coverage were performed in suitable cases. A-EMS with synchronous and/or subsequent EMS-related neoplasia were compared with those without. Of 4598 EMS cases over an 11-yr period, 36 A-EMS were identified. The mean age at presentation was 46 (range 26-68) yr. Locations included the ovary (24, 66%), tubo-ovary (6, 17%), fallopian tube (3, 8%), and peritoneum (3, 8%). The mean size was 6.5 (range 0.5-40) mm. Cytologic atypia was mild in 4 (11%), moderate in 21 (58%), and severe in 11 (31%). Most lesions were partially or completely flat (28, 78%); of these, 66% showed hobnail nuclei. Crowded/cribriform and micropapillary/papillary patterns were seen in 11 (31%) and 16 (44%) A-EMS, respectively. Immunohistochemistry, performed in 33 A-EMS, showed wildtype p53 (100%) retained PMS2/MSH6 (100%), and positive estrogen receptor (97%, mean 65% cells), progesterone receptor (76%, mean 30% cells), and Napsin A (39%). Ki67 labelling was <1% to 10% (median 5%). Nine (25%) patients presented with concurrent or subsequent ipsilateral endometrioid, seromucinous, or clear cell neoplasia (4 borderline tumors and 4 carcinomas). The only A-EMS feature statistically more frequent in this subset was crowded/glands (6/9 vs. 2/27 A-EMS without, P = 0.001 Fisher exact test). Sequencing showed pathogenic variants in 5 of 6 cases analyzed, involving ATM, BRCA2, KRAS, AKT, CTNNB1, PTEN, and ARID1A among other genes. In 2 cases, synchronous neoplasia showed an accumulation of additional variants. A-EMS is characterized by cytologic atypia and crowded architecture but low proliferation index, positive estrogen receptor, and normal p53 and MMR, which can be helpful in the distinction from malignancy. The prevalence of synchronous/subsequent tubo-ovarian neoplasia in our series was 25%, significantly higher than the reported 1% in conventional EMS. Moreover, A-EMS harbors genomic alterations seen in EMS-related tumors and shares pathogenic variants with synchronous ipsilateral neoplasia. Therefore, it is important to report A-EMS as currently defined and describe its architectural features, especially gland crowding as this appears to increase the risk of EMS-related epithelial neoplasia. Napsin-A is often positive in A-EMS and should be interpreted with caution.

Atypical Endometriosis: A Comprehensive Systematic Review of Pathological Patterns and Diagnostic Challenges

“…Finally, 134 articles were considered eligible for the systematic review; of these, 118 were also excluded for the absence of reported outcomes related to AE or the design of the study not in line with the inclusion criteria. Overall, 16 studies were finally selected for this review, 14 retrospective studies [ 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ], and 2 prospective studies [ 27 , 28 ]. The selection process is summarized in the PRISMA flowchart in Figure 1 .…”

“…A total of 6 articles evaluated the prevalence and the histopathology of atypical endometriosis ( Table 1 ). The general prevalence of AE in patients with endometriosis was found to range from <1% [ 13 ] to 5.8% [ 17 ], while in the case of EAOC, the prevalence was significantly higher, ranging from 22.8 to 34.6% [ 15 , 27 ]. For what concerns the histologic definition of AE, some authors discerned between cellular atypia and architectural atypia.…”

Whole-exome sequencing reveals candidate high-risk susceptibility genes for endometriosis

“…Also, sequencing of deep endometriotic lesions has revealed known somatic cancer driver mutations in 26% of the samples, including mutations in ARID1A and PIK3CA [ 55 ]. In addition, rare, atypical cytological features harboring endometriotic lesions have been shown to increase the risk for synchronous or subsequent borderline or carcinoma tumors, with prevalence ranging from 12,2 to 25% in the cohorts [ 56 , 57 ]. Despite cancer-associated mutations being relatively common in non-cancerous endometrial tissue, and malignant transformation is rare, accumulating evidence indicate endometriotic lesions as precursors for malignancy.…”

Genetic Links between Endometriosis and Endometriosis-Associated Ovarian Cancer—A Narrative Review (Endometriosis-Associated Cancer)

“…Due to these histologic findings and subsequent molecular data suggesting a clonal relationship between atypical endometriosis and EAOC (discussed in more detail below), atypical endometriosis has been suggested as the putative precursor of EAOC, but nevertheless is not identified in all cases of EAOC [ 48 ]. Wepy et al recently reviewed and analyzed over 4500 cases of endometriosis and identified 36 cases of atypical endometriosis [ 49 ]. Of the 36 cases, 9 had associated endometrioid, sero-mucinous, and clear-cell tumors (borderline and invasive).…”

“…Sequencing showed pathogenic variants in five of six cases analyzed, involving mutations in ATM , BRCA2 , KRAS , AKT , CTNNB1 , PTEN , and ARID1A . TP53 was wild-type in all cases, and mismatch repair was intact [ 49 ]. Interestingly, proteomic studies have suggested that ovarian endometrioid carcinoma arises from the secretory cells of endometriosis, while clear-cell carcinomas arise from ciliated cells [ 50 ].…”