Ovarian mesonephric-like adenocarcinoma arising in serous borderline tumor: a case report with complex morphological and molecular analysis

Mesonephric-like adenocarcinomas (M-LAC) are rare, recently described tumors occurring in the uterine corpus and ovary, which share the same morphological and immunohistochemical features with more common mesonephric adenocarcinoma (MAC), most commonly arising in the uterine cervix [1‐5]. Despite the similarities between these tumors, the histogenesis of M-LAC is still disputable, as this type of tumor occurs outside the anatomical areas in which mesonephric remnants/hyperplasia may normally occur. Recently, two cases of ovarian M-LAC arising in a serous borderline tumor (S-BTO) and a low grade serous carcinoma (LGSC) have been described. In both of these cases, the Müllerian tumors shared the same mutations as M-LAC, which proved their clonal origin [6, 7]. Based on this finding, the authors suggest that at least some of the M-LACs may be of Müllerian origin. In our report, we present another case of M-LAC arising in association with a S-BTO, with molecular evidence of the clonal origin of both tumors sharing the same KRAS and PIK3CA mutations.

Sixty-one-year-old female was referred in June 2019 to our Gynecologic Oncology Centre for interval debulking surgery (IDS) after previous administration of 5 cycles of neoadjuvant chemotherapy (NAC) for a primarily non-resectable left ovarian tumor FIGO (the International Federation of Gynecology and Obstetrics) classification [8] stage IV (presence of liver metastases). After NAC (paclitaxel plus carboplatin) the computed tomography findings revealed a partial regression tumor response according to RECIST (The Response Evaluation Criteria in Solid Tumors) criteria [9]. The size of the primary ovarian tumor mass and peritoneal carcinomatosis was reduced. However, the intraparenchymatous liver metastases remained unchanged in number and size. In July 2019, the patient underwent optimal IDS with no visible residual disease at the end of surgery. The surgery consisted of a hysterectomy with bilateral salpingo-oophorectomy, resection of liver metastases and the diaphragm, total omentectomy, appendectomy and a resection of an umbilical metastasis. After the surgery, the patient received 3 additional cycles of adjuvant chemotherapy (PTX/CBDCA), with the addition of bevacizumab from the second post-operative cycle for a period of 12 months. Recently, she underwent her 11th cycle of bevacizumab and is currently showing signs of complete clinical remission.

MACs of the female genital tract are rare tumors, usually arising in the uterine cervix and vagina [2, 3, 11]. Histologically, these tumors are characterized mostly by a mixture of heterogeneous architectural patterns. The most common is tubular or glandular with the formation of small glands, with a common presence of eosinophilic intraluminal colloid-like material. However, other patterns may occur, including solid (with possible spindled morphology), slit-like, papillary, glomeruloid, sex cord-like, retiform, and glandular with the formation of medium-sized or large glands [3, 12‐14]. The tumor cells may be flattened, cuboidal or columnar, with usually scant pale eosinophilic or, rarely, clear cytoplasm. The nuclei are usually vesicular, commonly resembling those of papillary thyroid carcinoma, and may show grooving, irregular nuclear membranes, and in some cases nuclear pseudoinclusions. Hobnail cells can be present as well. Immunohistochemically, these tumors may be positive for CD10, GATA3, TTF1, HNF1B, PAX2 and PAX8 [12, 13, 15]. The p16 expression is patchy in most cases, and p53 expression is of the wild-type. WT1, estrogen and progesterone receptors are negative, although a focal expression of the estrogen receptor has been described in rare cases. On a molecular level, these tumors are characterized by the KRAS mutation which occurs in most cases (∼80%) [14, 16]. The ARID1A mutation is also common (∼50% of cases), as well as the gain of chromosome 1q detected by copy number analysis (∼80% of cases). Other mutated genes such as NRAS (1/16 cases), ARID1B (3/16 cases), BCOR (2/16 cases), BCORL1 (2/16 cases), SMARCA4 (2/16 cases), TP53 (1/16 cases), and PTCH2 (4/12 cases) have also been described [14, 16, 17]. A single case report of mutation in MAC and CTNN1B (occurring together with the KRAS mutation) was reported [11].

Rare tumors with similar morphological and immunohistochemical features have also been described in the uterine corpus and ovary [1, 14]. However, the mesonephric origin of these tumors has not yet been proven as these tumors arise in anatomical areas not associated with mesonephric remnants. In a recent study of 12 cases (7 endometrial and 5 ovarian) none of the endometrial cases were associated with mesonephric remnants in the uterine cervix, and all arose in endometrium. Three of the five ovarian cases were associated with endometriosis [18]. Due to these histogenetic uncertainties, the authors of the study prefer the use of term “mesonephric-like adenocarcinoma”. Molecular aberrations occurring in M-LAC are very similar to MAC, with common KRAS mutation occurring in ∼90% of cases [1, 4, 16, 17]. However, contrary to MAC, in ∼35% of M-LAC the mutation of PIK3CA has been reported, which has not yet been described in MAC. This was also observed in our case, where we detected both KRAS and PIK3CA mutations. Supporting the assumption of the Müllerian origin of M-LAC, there are two recent reports of ovarian cases associated with Müllerian tumors (in both cases with S-BTO and LGSC), with molecular analysis confirming the same clonal origin of these tumors [6, 7]. Based on these results, at least some M-LACs of the ovary are Müllerian in origin and the term “mesonephric-like adenocarcinoma” therefore seems to be appropriate. In the first case, the authors described a tumor consisting of LGSC and M-LAC areas [6]. Both tumor components had metastasized. The LGSC arose in association with a serous borderline tumor. All components (S-BTO, LGSC and M-LAC) were proven to be clonal in origin and shared the same mutation in the NRAS gene p.(Q61R). Moreover, additional aberrations were detected in the LGSC metastasis (KDM5A, STAG2) and the M-LAC (primary – BCOR, AMER1, metastasis – BCOR). The second case describes an M-LAC arising in an S-BTO in a patient with bilateral S-BTO and LGSC, with invasive implants [7]. All components (serous BTO, LGSC and M-LAC) were proven to be clonal in origin and shared the same KRAS mutation p.(G12D). In our case, the serous BTO and M-LAC were also confirmed to be clonal in origin and shared the same KRAS p.(G12C) and PIK3CA p.(E545K) mutations. In addition, in the M-LAC the mutation of the MYCN gene was found.

The differential diagnosis of M-LAC includes other epithelial tumors, especially endometrioid carcinoma (EC) [1]. The distinction between M-LAC and EC may be difficult and should be based on morphological and immunohistochemical features. M-LAC is characterized by common heterogeneity of architectural patterns. These patterns are not specific, but if tubular areas with intraluminal eosinophilic secretions are present, the suspicion of M-LAC is high. Cytological features can be helpful as well, because nuclei with vesicular chromatin and nuclear grooves typical for M-LAC are not a characteristic feature of EC. Immunohistochemically, M-LAC is characterized by GATA-3 and TTF1 expression, which is rare in EC. On the contrary, EC is characterized by common expression of estrogen and progesterone receptors, which are usually absent in M-LAC [1].

In conclusion, we report a case of ovarian M-LAC arising in a S-BTO. The clonal origin of both components suggests that S-BTO may be a precursor of M-LAC. This is in concordance with two previously reported ovarian M-LAC cases associated with Müllerian tumors [6, 7]. Our findings, and the results of the two previous reports, suggest that at least some ovarian M-LACs are of Müllerian origin and the term “mesonephric-like adenocarcinoma” instead of “mesonephric adenocarcinoma” seems to be appropriate. However, M-LACs of the uterine corpus and ovary share the morphology, immunophenotype, molecular aberrations (with the exception of PIK3CA mutation) and, based on the limited experience, also the aggressive behavior with the more common MAC of the uterine cervix (and cases classified as MAC of the uterine corpus). Despite the different origin of these tumors (Müllerian vs. mesonephric), the lineage of “mesonephric-like adenocarcinoma” differentiation seems to be identical to MAC, as evident by the sharing of all mesonephric features including molecular aberrations, which can simply reflect the multipotential differentiation capacity of Müllerian structures. Based on this, the classification of these tumor as “mesonephric-like” seems to be rather arbitrary and reflective more of the histogenetic origin and primary location of the tumor, than of the morphology, molecular aberrations, and biological potential.