Mesonephric-like Adenocarcinoma (MLA) Diagnostic Criteria and Controversies: Perspectives and Guidance From Pathologists in the MLA Consortium

MLA was first described in 2016 in a study coauthored by one of us (W.G.M.); in this study, 12 cases were reported (7 uterine corpus, 5 ovary). 1 The term MLA was recommended since, although the morphologic features and immunophenotype strongly suggested a mesonephric neoplasm, there were other features which suggested a Müllerian derivation. These included the absence of mesonephric remnants associated with the neoplasm and the observation that in all of the uterine cases, the tumor extensively involved the endometrium and appeared to arise there with subsequent invasion into the myometrium; this would not be expected with a true mesonephric adenocarcinoma (MA) arising from mesonephric remnants in the outer aspects of the uterine corpus. As such, the term MLA was recommended until the histogenesis was firmly established. The cases in this series had originally mostly been reported as endometrioid carcinomas, but several of the ovarian neoplasms were diagnosed as malignant struma ovarii (papillary-type thyroid carcinoma), given the nuclear features and the presence of diffuse TTF1 immunoreactivity (discussed below). The discovery of MLA highlights the important point that “new” pathologic entities can still be identified through careful observational studies. Although this represented the first publication of MLA, in hindsight, the original observations of this neoplasm were several years earlier, when one of us (W.G.M.) reported several cases where the patients had a history of endometrioid carcinoma of the uterine corpus and were subsequently discovered to have a pulmonary tumor. 2 The pulmonary tumors were TTF1-positive and estrogen receptor (ER)-negative, understandably resulting in designation as an independent lung primary. However, the clinical scenario was thought to be more in keeping with metastasis from the tumor within the uterine corpus and this was supported by the subsequent demonstration of PAX8 positivity within the lung neoplasms. The uterine corpus neoplasms were then stained with TTF1 and were found to be positive. It was speculated whether the propensity for pulmonary metastasis of these TTF1-positive endometrial carcinomas might in some way be due to a “homing” mechanism whereby TTF1-positive tumors had a tendency to metastasize to the lungs. 2 In the ensuing years, additional studies brought MLA into focus as a distinctive gynecologic primary with aggressive behavior, frequent expression of TTF1 and GATA3, absent to minimal hormone receptor expression, and morphologic features mimicking a variety of other tumor types. 3 – 5 MLA was then included in the 2020 World Health Organization (WHO) Classification of Endometrial and Ovarian Carcinomas.

Clinically, the identification of MLA as a histologic subtype of endometrial cancer has allowed for further prognostication of these high-risk tumors from the general no specific molecular profile (NSMP) molecular cohort. Uterine MLA accounts for only 0.7% to 3% of endometrial cancer cases in unselected patient populations; however, retrospective series show that patients are more likely to present with stage III/IV disease and more likely to have metastatic disease to regional lymph nodes. 2 , 5 , 6 , 8 , 10 , 12 , 23 Importantly, in a retrospective cohort study comparing uterine MLA to endometrioid and serous cases, 71% of uterine MLA cases had a documented recurrence compared with 7.8% of endometrioid and 29% of serous cases, with a propensity for distant metastatic disease to the lungs. 2 When considering adjuvant therapy options and known prognostic factors for uterine cancer, MLA presents with higher stage disease, is twice as likely to exhibit lymphovascular space invasion, and should be considered high-risk disease for clinical management. The consideration of uterine MLA as a high-risk disease highlights the critical clinical implications that hinge on the proper pathologic classification of these tumors, where adjuvant chemotherapy and/or radiation are the standard of care. In review of cases sent for expert consultation, ∼80% of MLA cases were originally diagnosed as a histology other than MLA, with the majority being classified as FIGO grade 1-2 endometrioid carcinoma, which may be considered low-risk or intermediate-risk disease based on other pathologic factors, where adjuvant chemotherapy is not standard of care. 2 In addition, post-treatment surveillance of uterine cancer does not routinely include imaging. However, given the increased risk for pulmonary metastases, specifically in MLA, the proper classification of these tumors allows for consideration of periodic chest imaging in surveillance of these patients. Extrauterine MLA has been estimated to account for 0.6% to 0.9% of epithelial ovarian cancer. Experience is limited by the fact that the total number of reported clinical cases in the literature is <100, and clinical management is currently based on the standard of care for epithelial ovarian cancer with platinum-based chemotherapy. However, the proper classification of MLA again is key in the decision for adjuvant therapy, as MLA is diagnosed with stage I/II disease in 64% of patients, where adjuvant therapy may be de-escalated for other histologic subtypes such as endometrioid tumors of the ovary. 5 Surveillance of extrauterine MLA follows the current standard of care for epithelial ovarian cancer. While the identification of MLA has allowed for improved prognostication, many questions remain in the clinical management of these tumors. In consideration of uterine MLA, the impact of chemotherapy and/or radiation on the risk of recurrence is largely unknown, as the largest cohort studies to date do not include adjuvant treatment data. In addition, the impact of radiation in the treatment of high-risk, advanced-stage uterine cancer is debated among gynecologic oncology and radiation oncology experts, with practice variation across centers. The MLA Consortium is actively accruing a tumor registry in an effort to further evaluate response to therapies in a larger patient cohort. However, the prospective evaluation of chemotherapy and radiation is needed in this rare patient population. Similarly, the impact of platinum-based chemotherapy on the risk of recurrence in uterine MLA is largely unknown. The molecular underpinnings of MLA do allow for consideration of exciting novel therapeutics, including small molecule KRAS or NRAS inhibitors, other targeted therapies capitalizing on the Ras/Raf/MEK pathway, vaccine therapies, and cellular therapies, with several early phase studies open to accrual for patients. The first is a single-arm phase II study of avutometinib (VS-6766) and defactinib in advanced or recurrent mesonephric gynecologic cancer. This study capitalizes on the success of the study regimen in low-grade serous ovarian cancer (LGSOC), with recent FDA approval for the regimen in patients with KRAS-mutated LGSOC ( NCT05787561 ). In addition, patients may be eligible for ongoing early-phase basket trials of PAN-KRAS or PAN-RAS inhibitors ( NCT04111458 , NCT05379985 ). Lastly, cellular therapy is an interesting treatment strategy with an actively accruing phase I/II trial of TROP2-CAR/IL-15-transduced CB-NK cells delivered intraperitoneally in tumors with at least 1+ TOP2 expression ( NCT05922930 ) in addition to T-cell therapy targeting mutant KRAS proteins specifically ( NCT06043713 ). All these drugs remain investigational, and there are no current Food & Drug Administration approvals for their use in this setting.

MLA is an uncommon, underrecognized, and potentially overdiagnosed malignancy in the endometrium and extrauterine sites. Its proper identification is critical due to its aggressive behavior and the future potential for unique treatment avenues. The MLA Consortium has been convened to better understand and approach these challenging tumors, and herein provides practical diagnostic guidance for pathologists who are likely to encounter this entity. In particular, we emphasize situations in which expert consultation is desirable, including cases with challenging morphology (such as mixed tumors, including cases with dedifferentiated and carcinosarcomtaous components), unexpected immunohistochemical patterns (such as cases with TTF1 and GATA3 negativity or more than limited hormone receptor staining) and nonclassical molecular profiles (such as cases with TP53 abnormalities or MMR deficiency). Pathologists interpreting gynecologic samples as well as specimens from potential metastatic sites—including lung—should be familiar with MLA and the criteria for its diagnosis.

The protean morphologic appearance of MLA results in overlap with a host of other tumors at uterine and extrauterine sites. Prior to the description of MLA as a specific tumor type, these neoplasms were likely diagnosed as a wide variety of other histotypes. More recently, increased awareness of MLA also means that there is a risk of overdiagnosis, particularly when more common entities express MLA-associated markers such as TTF1/GATA3. It is therefore critical for pathologists to have a robust understanding of the differential diagnosis of MLA in both uterine and extrauterine settings. Occasionally, there is a diagnostic issue as to whether a tumor represents an endometrial MLA or a cervical MA. This may be a consideration in a small biopsy specimen, especially if clinically and radiologically it is not clear whether a tumor is arising from the cervix or the endometrium or where the biopsy has been taken from. This may also be an issue in a resection specimen if tumor involves both the uterine corpus and the cervix. Usually in such cases, the distribution of the tumor will help determine the site of origin, but sometimes this is not the case and a constellation of pathologic features, including immunohistochemistry and occasionally molecular testing, may be necessary to distinguish between the 2 tumor types. Similarities and differences between MLA and MA are detailed in Table 4 and discussed below. Mesonephric Adenocarcinoma (MCA) Versus MLA MLA within the uterine corpus arises from the endometrium, and within the ovary or, less commonly, at extrauterine and extraovarian sites, mostly from endometriosis; these neoplasms are not associated with mesonephric remnants. They are not “centered” in the myometrium with limited or no endometrial involvement, as would be expected with a true MA arising from mesonephric remnants in the uterine corpus. True MAs arise from the cervix or rarely from the vagina, often in association with benign mesonephric remnants, although these are not always seen. Although there have been reports of primary uterine corpus MAs, it is likely that most or all of these actually represent MLAs. 30 In support of this is the observation that, although it is often stated that mesonephric remnants may occur in the outer aspects of the myometrium, this is extremely rare. The morphologic features are similar between MLA and MA. Both neoplasms typically exhibit a variety of architectural patterns, often admixed within a single tumor. However, MLAs are characterized by clear angulated vesicular nuclei, sometimes with overlapping and grooves. These nuclear features are not a particular feature of MA, and the nuclei in these neoplasms are often more hyperchromatic. Admixture with another tumor type may be a clue to an MLA diagnosis since MAs almost always occur as pure neoplasms. 16 , 17 The immunophenotypes of MLA and MA are similar, but not identical. Both tumor types are typically diffusely positive with PAX8 and CK7, exhibit wild-type p53 expression, non-diffuse p16 expression, are negative for WT1, and are MMR-intact. They are both often positive with variable distribution for TTF1, GATA3, and CD10. However, MLAs are more likely than MAs to be positive with TTF1. 5 , 53 ER and PR staining may be useful in distinguishing between these neoplasms as MA are almost always entirely negative, while MLAs may show limited stainingwith both markers in some cases. From a molecular standpoint, both MLA and MA commonly exhibit KRAS and, to a lesser extent, NRAS mutations. 16 However, as discussed, MLAs more commonly exhibit additional mutations which are characteristic of Müllerian carcinomas, such as PIK3CA and PTEN , whereas MAs usually lack such alterations. MLA and endometrioid carcinoma, particularly FIGO grade 1 to 2 tumors, may closely resemble one another, and before the description of MLA most were likely diagnosed as low-grade endometrioid carcinoma 2 , 5 , 9 , 10 , 12 , 22 , 23 (Table 5 ). Both neoplasms may be composed of glandular structures of varying degrees of complexity lined by moderately atypical cells. While both may have other architectural patterns, including cribriform and papillary growth, well-formed glomeruloid structures are much more typical of MLA, although they are not seen in all cases. Intraluminal eosinophilic secretions may also be present in both MLA and endometrioid carcinoma, although MLA-associated secretions typically have a denser and brighter appearance than the secretions typical of endometrioid carcinomas. Some MLAs have dense hyaline stromal material that may prompt consideration for the corded and hyalinized variant of endometrioid carcinoma. 54 Finally, mucinous and squamous metaplasia are common in endometrioid carcinomas, but are not a feature of MLA. MLA Versus Endometrioid Carcinoma and Serous Carcinomas ER and PR are among the most valuable ancillary markers in this setting, as negative/very low staining is uncommon in low-grade endometrioid carcinoma and should prompt an expanded differential that includes MLA. Indeed, ensuring that MLA and other rare tumors are not missed is among the arguments for reflex ER/PR testing. GATA3, TTF1, and CD10 are most often negative in endometrioid carcinomas, although exceptions occur as highlighted in the Immunohistochemistry section. 12 , 27 PTEN has potential value in this context, with loss of expression providing some support for an endometroid carcinoma diagnosis though notablysome endometrioid carcinomas lack PTEN pathogenic variants and some MLAs will bear these mutations. Although MLA lacks the marked nuclear atypia that is characteristic of uterine serous carcinoma, it typically exhibits moderate nuclear atypia which could provoke concern for an serous carcinoma diagnosis in this location. More often, MLA mimics serous neoplasia in the tube/ovary, where it can convincingly masquerade as a low-grade or high-grade serous carcinoma due to its cytologic features and predilection for papillary growth (Fig. 8 , Table 5 ). The paucity of psammomatous calcifications should be a clue to consider MLA in this context, as should brisk mitoses without associated severe nuclear atypia. MLA mimicking serous carcinoma. This omental mass in a patient with an ovarian tumor and carcinomatosis was originally interpreted as high-grade serous carcinoma (A). Histology showed papillary structures with areas of background hyalinization (B) with areas of glomeruloid-like architecture (C). Although mitotic activity was robust, cytologic atypia was less prominent than is typical of high-grade serous carcinoma (D). Tumor board review led to additional work-up and revealed wild-type p53 staining (E), which then prompted consideration for low-grade serous carcinoma; however, the lack of psammoma bodies and completely negative ER (F) confounded this. TTF1 and GATA3 were subsequently performed to assess for mesonephric-like differentiation and showed strong diffuse (G) and moderate patchy (F) staining, respectively, supporting an MLA diagnosis. Immunostaining can also be very valuable in this differential. Mutation-type p53 staining would generally support an endometrial serous or tubo-ovarian high-grade serous carcinoma, while MLAs are classically p53 wild-type (with rare exceptions discussed in the “Molecular” section). TTF1 and GATA3 expression are uncommon in serous carcinomas, and tubo-ovarian high-grade serous carcinoma and low-grade serous carcinomas will typically demonstrate diffuse WT1, which MLAs will lack. Moreover, complete negativity for ER/PR is fairly uncommon among both high and low-grade serous carcinomas, although it can occur in occasional high-grade serous carcinomas. MLAs sometimes demonstrate nuclear hobnailing and/or a dense hyalinized stroma reminiscent of some clear cell carcinomas. 2 , 5 This overlap is immunohistochemically complicated by the fact that these entities both lack significant ER/PR expression. Moreover, a recent study by Mirkovic et al. 55 showed that over one-third of MLAs exhibit some Napsin A positivity, underscoring the importance of a broad immunopanel (including TTF1 and GATA3) when addressing this differential diagnosis. Occasional examples of MLA show prominent spindled growth, which can mimic carcinosarcoma. 2 , 5 In such cases, it is key to assess whether the spindled morphology qualifies as true mesenchymal differentiation consistent with a sarcomatous element. Immunostaining may be of value here, with significant cytokeratin expression within the spindled component supporting a carcinomatous contribution. This differential is further confounded by the recent description of MLAs bearing bona fide heterologous (typically chondroid but sometimes rhabdomyosarcomatous) differentiation consistent with a mesonephric-like carcinosarcoma diagnosis; 16 such cases are discussed in further detail in the “Morphology” section. In addition to the carcinoma histotypes discussed above, extrauterine MLAs may provoke an expanded differential that includes rare and easy-to-miss entities arising in and around the ovary (Fig. 9 ). For example, MLAs with tubular growth may be morphologically suggestive of Sertoli or Sertoli-Leydig cell tumors. While recognition of a Leydig cell component is helpful for supporting sex cord differentiation, other ovarian tumors, including MLA, may sometimes induce adjacent steroid cell hyperplasia. Fortunately, this differential can typically be resolved with SF1 or inhibin staining, which should be positive in sex cord tumors and negative in MLA. Uncommon extrauterine mimickers of MLA. A variety of rare neoplasms may masquerade as MLA in the ovary and adnexa. Well to moderately differentiated Sertoli Leydig tumors show tubular architecture (A), which is reminiscent of the tubules formed by some examples of MLA (B). Sieve-like growth and compressed glands may be seen with Wolffian tumors (C) as well as MLA (D). STK11 adnexal tumor (E) can also show similar architecture and cytomorphology to MLA (F). Well-differentiated peritoneal mesotheliomas with papillary architecture (G) can mimic MLAs showing similar growth (H). The relatively “low-grade” appearance and tubular to sieve-like growth of MLA can also mimic a Wolffian tumor [female adnexal tumor of Wolffian origin (FATWO)], particularly when the tumor is centered around the adnexa. FATWOs can also express luminal CD10, 56 prompting further confusion with MLA. However, most FATWOs are fairly mitotically inactive, whereas MLA typically exhibits brisk mitotic activity. Therefore, the identification of prominent mitoses should prompt a pause before rendering a FATWO diagnosis, and conversely, low mitotic activity should prompt reconsideration of MLA. PAX8 expression can also be useful for supporting MLA in this setting, as FATWOs are almost never positive with this marker. The recently described STK11 adnexal tumor may also result in confusion with MLA. 57 , 58 Immunohistochemistry can be very helpful in addressing this diagnosis, as STK11 adnexal tumors will typically display some degree of sex cord marker expression and are usually hormone receptor and WT1-positive. As with Wolffian tumors, PAX8 is typically negative in STK11 adnexal tumors, and can be another helpful discriminatory biomarker. While occasional GATA3 staining is reported in STK11 adnexal tumors, TTF1 is consistently negative in this entity. 57 Molecular testing will also be of value in this distinction. Finally, the clinical presentation of peritoneal malignant mesothelioma may sometimes suggest an ovarian primary, and these tumors can have considerable histologic similarities with MLA. In particular, the papillary and tubular growth that characterizes some peritoneal malignant mesotheliomas can closely resemble MLA. Although its utilization is not recommended by the MLA Consortium, one could imagine how calretinin immunostaining might cause further confusion at this diagnostic interface since it may be positive in both tumors, although it is typically more widespread in malignant mesothelioma. WT1, conversely, can be of value in this setting as mesothelioma should be WT1-positive, whereas MLA is consistently negative. BAP1 staining is lost in some peritoneal malignant mesotheliomas but would be expected to be retained in MLA. As discussed, some MLAs are focally positive with hormone receptors, while malignant mesotheliomas are negative.