Mucinous Tumor Coexisting With Mesonephric-like Proliferation/Tumor in the Ovary: A Novel Association.

Clinicopathologic features are provided in Table 1 and immunohistochemical findings are summarized in Table 2 . This series included 2 cases of MBT/APMT associated with a MLP, as well as 1 case of mucinous cystadenofibroma and 1 case of MBT/APMT each associated with MLA and a background benign MLP. The patients’ age ranged from 55- to 67-year-old (mean, 59; median, 57). All primary tumors presented as pelvic/adnexal masses and all were unilateral. Two cases were consistent with an MBT/APMT associated with a MLP. Case 1 was a 56-year-old woman who presented with a left adnexal mass and subsequently underwent a total hysterectomy and bilateral salpingo-oophorectomy (TH-BSO). The gross examination of the left ovary revealed a 12×9×2.5 cm multiloculated cyst with the cyst wall thickness ranging from 0.1 to 0.4 cm. Microscopic examination revealed that the vast majority of the ovarian tumor was composed of MBT/APMT and a focal area (about 2% of the total tumor volume) consisted of an atypical glandular proliferation ( Fig. 1A ) next to the mucinous glands ( Fig. 1B ). The nonmucinous glandular proliferations were composed of aggregates of small, round tubules lined by cytologically bland, nonstratified cuboidal cells and containing luminal eosinophilic secretions ( Fig. 1B ). The glands and tubules were well-spaced with retained stroma without desmoplastic reaction. Some areas displayed glandular enlargement, crowding and irregular shapes, indicating a hyperplastic process but insufficient to establish a diagnosis of carcinoma ( Fig. 1C ). These glandular proliferations were diffusely positive for Gata3, Pax8, and CK7 ( Figs. 1D - F ). The combined histologic and immunohistochemical features of this component were consistent with a mesonephric lineage. Mucinous epithelium was highlighted by CK7 but was negative for Gata3 and Pax8. In some areas, the mesonephric epithelium was distributed linearly in parallel to the mucinous epithelium with the former component highlighted by Gata3 and Pax8 ( Figs. 1G - I ). On high power, some mesonephric glands exhibited mucinous metaplasia/differentiation ( Figs. 2A - C ). In some glands, Gata3 and Pax8 were positive in both mesonephric epithelium and adjacent metaplastic mucinous epithelium ( Figs. 2D - F ). In other areas, Gata3 expression was lost in the mucinous epithelium and Pax8 expression was decreased ( Figs. 2G - I ). Case 2 was a 67-year-old woman who presented with a pelvic mass suspected to be an advanced gynecologic malignancy. TH-BSO, omentectomy, appendectomy, sigmoid colon segmental resection, and staging biopsies were performed and revealed a 28 cm primary unilateral ovarian MBT/APMT ( Figs. 3A , B ) with pseudomyxoma ovarii and extensive gland rupture. The mucinous tumor was focally positive for Pax8 ( Fig. 3C ). Microscopic foci (< 0.1% of total tumor volume, Fig. 3D ) of the mesonephric glands and mesonephric proliferations (highlighted by Gata3 immunostaining, Fig. 3E ) were found either adjacent to the mucinous epithelium or intimately admixed with mucinous glands ( Fig. 3F ). While the mesonephric component showed diffuse Pax8 staining, its expression was diminished but still retained in most areas of in the mucinous epithelium. Similar to case 1, some mesonephric epithelium displayed a variable degree of mucinous metaplasia/differentiation ( Figs. 3G - I ). Case 3 was a 58-year-old woman with a 12 cm left ovarian tumor who underwent TH-BSO and omentectomy. Histologically, the tumor was consistent with a MLA with an adjacent benign MLP ( Fig. 4A ). The carcinoma displayed a mixture of tubular/glandular, papillary, and solid architectural patterns. Intraluminal eosinophilic material was present in some glands. The glandular epithelium was composed of cuboidal or columnar cells with vesicular nuclei, irregular nuclear membranes, and scant eosinophilic cytoplasm. The tumor cells were diffusely positive for Gata3 ( Fig. 4B ) and Pax8, focally positive for TTF-1 and showed a luminal staining pattern for CD10 (focal), a wild-type expression pattern for p53, and a heterogeneous staining of p16. ER, PR, and WT-1 were negative. In some areas, the MLA was located next to the mucinous cystadenofibroma ( Fig. 4C , 5% of total tumor volume) and some of the malignant mesonephric epithelium showed mucinous metaplasia/differentiation ( Figs. 4D , E ). The mucinous cystadenofibromatous component was focally positive for Pax8. In some sections, scattered benign MLPs were adjacent to or intimately admixed with mucinous glands ( Figs. 4F , G ). Some of the MLPs focally displayed mucinous metaplasia/differentiation ( Figs. 4H , I ). Endometriosis was found in the contralateral ovary. The last case (case 4) was a 55-year-old woman with a 13 cm left adnexal mass who underwent TH and left salpingo-oophorectomy, omentectomy, lymph node dissection, and pelvic staging biopsies. Histologically, the tumor was composed of a MLA and adjacent MBT/APMT ( Fig. 5A ). The former component accounted for ~20% to 30% of the total tumor volume. Benign MLPs were found in multiple peripheral areas of the MLA ( Fig. 5B ) and of the MBT/APMT ( Fig. 5C ). Similar to case 3, some sections demonstrated scattered benign MLPs intermixed with enlarged cysts and mucinous glands ( Fig. 5D ). Some of the MLPs displayed mucinous metaplasia/differentiation ( Figs. 5E , F ). Immunohistochemically, the MLA component ( Fig. 6A ) was diffusely immunoreactive for Gata3 ( Fig. 6B ) and Pax8, displayed focal positivity for TTF-1 and a luminal staining pattern for CD10, whereas the MBT/APMT component was negative for these markers. A p16 immunohistochemistry showed complete loss of staining in the MLA but retained patchy expression in the MBT/APMT ( Fig. 6C ). MLPs ( Fig. 6D ) were also positive for Gata3 ( Fig. 6E ); however, some glands were completely negative for p16 while others showed patchy expression. ( Fig. 6F ). The p16-negative pattern was likely due to a CDKN2A/P16 gene deletion and the patchy-positive pattern had a wild-type genotype. NGS for this case was performed to compare the genetic alterations of the MBT/APMT and MLA components. A KRAS c.35G > T (p.Gly12Val) somatic mutation was detected in both components ( Fig. 7A ), indicating a clonal origin. This mutation was further confirmed by Sanger sequencing ( Fig. 7B ). Interestingly, the KRAS mutation, although present with a low mutational percentage, was also detected in the benign MLPs. A CTNNB1 c.98C > T (p.Ser33Phe) somatic mutation, FGFR2 amplification, and CDKN2A/p16 deletion were also detected by NGS; these genetic alterations were only present in the MLA but not in the MBT/APMT component. As a result of CDKN2A/p16 deletion, a p16 immunohistochemistry showed complete loss of staining in the MLA ( Fig. 6C ). Based on the morphology and the observed molecular changes, a model recapitulating the histopathogenesis of case 4 was proposed ( Fig. 7C ).

Cases were identified in the files of the authors’ institutions (3 cases from the consultation files of Johns Hopkins Hospital, MD, and 1 from Sky Ridge Medical Center, CO). The patients’ age, clinical presentations, procedures, and specimen gross descriptions including tumor site and size were retrieved and reviewed. Histologic sections of these cases were rereviewed by 2 pathologists (N.N. and D.X.) to confirm the diagnosis. The study was approved by the Institutional Review Board at the Johns Hopkins Hospital. Immunohistochemical staining was performed in the Johns Hopkins Immunopathology Laboratory on formalin-fixed, paraffin-embedded (FFPE) tissue sections, using Ventana Benchmark automation and the Ultra View detection kit (Ventana Medical Systems, Tucson, AZ) as previously described. 11 , 12 Some stains were performed at the time of original diagnosis and some were done for this study. Markers used included Pax8 (Polyclonal, BD Pharmingen, San Jose, CA; 1:100 dilution), Gata3 (L50-823, Biocare, Concord, CA; 1:100 dilution) and SATB2 (EP281, Cell-marque, Hot Springs, AZ; prediluted). Prediluted antibodies from Ventana, included: ER (SP-1), P16 (INK4a), p53 (BP53-11), TTF-1 (8G7G3/1), and WT-1(6F-H2). Prediluted antibodies from Leica, Bannock Burn, IL, included: PR (1E2) and CD10 (MS/56C6). Prediluted antibodies from DAKO, Carpinteria, CA, included: CK7 (OV-TL), CK20 (Ks20.8), and CDX2 (EPR2764Y). FFPE tumor and corresponding normal tissues were identified by H&E staining and subsequently were macro-dissected (with tumor elements accounting for about 60% or more of the section area), and genomic DNA was extracted using a QIAamp DNA FFPE Tissue Kit with an adopted protocol (Qiagen, Valencia, CA). Briefly, slides bearing paraffin-embedded tissue were baked at 68°C for 20 to 30 seconds; the tissue was deparaffinized 3 times with xylene, and residual xylene was removed by washing through serial dilutions of ethanol. The rest of the procedure followed the manufacturer’s instruction. The extracted DNA was amplified by the Oncomine Comprehensive Panel (OCP) v2 and subjected to NGS using the Ion Torrent S5 system (Life Technologies) according to the vender recommendations. 13 The Oncomine Comprehensive assay was developed and its performance characteristics were determined by the Clinical Genomics Laboratory, Department of Pathology and Laboratory Medicine at Weill Cornell Medicine/New York-Presbyterian Hospital 14 and approved by the New York-State Department of Health (NYS-DOH). The targeted gene panel interrogates 143 unique cancer genes including the hotspots of 73 genes, 49 copy number alteration genes, entire coding regions of 26 genes, and 22 fusion driver genes. The data obtained were analyzed with the Ion Reporter Software 5.6 including Coverage Analysis and Torrent Variant Annotator v2.3 plug-ins. The mutation nomenclature is based on the recommendations from the Human Genome Variation Society ( http://www.hgvs.org/mutnomen ). KRAS c.35G > T (p.Gly12Val) mutation was further assessed by Sanger sequencing. Briefly, 50 ng of DNA was amplified by PCR with Taq DNA Polymerase and Standard Taq Buffer (New England BioLabs, MA). The following primers were used for amplification: 2F1: 5′-GTGTATTAACCTTATGTGTGACA-3′, 2R1: 5′-TGGTCAGAGAAACCTTTATCTG-3′, and 2R2: 5′-TGGTCCTGCACCAGTAATATGC-3′. The reaction for amplification of first-round PCR primers (2F1/2R1) was carried out in the following conditions: an initial melting step of 2 minutes at 95°C, followed by 30 cycles of 30 seconds at 94°C, 30 seconds at 51°C and 45 seconds at 72°C, and a final elongation of 7 minutes at 72°C. The PCR products amplified by 2F1/2R1 primers were diluted 10 times and used as templates of a second PCR amplification (nested PCR) with another pair of primers 2F1/2R2 that was carried out in similar reaction conditions except for an annealing temperature of 56°C. DNA sequencing of the purified DNA products was performed using the ABI 3730 high-throughput DNA sequencer. The mutations and variations were analyzed using Unipro UGENE software.

Despite a limited number, our case series characterizes novel associations between MLP, MLA, and mucinous tumor in the ovary. In particular, the presence of MLP in this scenario has not been described previously. The recently defined MLAs are rare tumors with morphology and immunoprofile similar to that of MCs of the uterine cervix but arise in the endometrium and ovary without associated mesonephric remnants. 1 , 15 - 19 Since its first description, 1 it has been debated whether these tumors represent true MCs that originate from occult mesonephric remnants in the uterine corpus and ovary or whether they are carcinomas of Müllerian origin with mesonephric-lineage differentiation. The concept of the latter, as well as the term “MLA,” have been well accepted in the literature since these tumors are occasionally associated with Müllerian-type lesions in the ovary including endometriosis, 2 , 3 , 7 serous cystadenoma, 5 endometrioid borderline tumor, 2 and SBT/LGSC. 4 - 6 , 8 Particularly, the presence of identical KRAS, NRAS , or PIK3CA mutations in both SBT/LGSC and MLA provide convincing evidence that these different components are clonally related and suggest that at least some MLAs may arise from Müllerian precursors. 4 - 6 , 8 In our series, however, a benign MLP was identified in all 4 cases. Morphologically, these proliferations were composed of small, round tubules lined by cytologically bland cuboidal cells with eosinophilic secretions and were surrounded by stroma without desmoplastic reaction. The cells were diffusely positive for Pax8 (all 4 cases) and focally (1 case) or diffusely (3 cases) positive for Gata3. While case 2 only showed pure benign proliferations of mesonephric-like tubules, there were mesonephric-like hyperplasia in case 1 and MLAs in cases 3 and 4 associated with these benign proliferations. In theory, these benign MLPs still could have arisen from Müllerian-type tissue that underwent transdifferentiation; however, evidence of an associated Müllerian-type lesion is not identified in these cases. As such, our observations raise the likelihood that these proliferations are derived from true mesonephric remnants. Although commonly distributed deep in the lateral cervical wall stroma, these embryological remnants also exist in the rete ovarii of the ovarian hilum, paratubal tissue, and the epoophoron and paroophoron in the broad ligament. The rete may communicate with mesonephric tubules within the mesovarium. 20 Female adnexal tumor of probable Wolffian origin, thought to arise in the vestiges of the mesonephric duct, commonly occurs in a paraovarian location but rarely in the ovary. 21 , 22 Theoretically, MLA, similar to female adnexal tumor of probable Wolffian origin, could originate from mesonephric remnants around the ovary, and subsequently be incorporated into the ovary. Unexpectedly, we found mucinous tumors (1 mucinous cystadenofibroma and 3 MBT/APMTs) coexisting with mesonephric lesions (1 benign proliferation, 1 mesonephric hyperplasia, and 2 MLAs) in all 4 cases. Consistent with our findings, a recent large series of MCs and MLAs of the female genital tract reported 2 cases of mixed ovarian MLA and MBT/APMT. 8 In one case (OV75), the cystic portion of the mass was composed of mucinous cystadenoma and MBT/APMT lined by gastrointestinal-type mucinous epithelium whereas the solid areas contained a mixture of mucinous cystadenofibroma intimately admixed with scattered foci of MLA with variable architecture. In another case (OV21), the right ovary consisted of MLA intimately admixed with a MBT/cystadenofibroma. Similar to when combined LGSC and MLA in the ovary was first described, 4 the question of what is the etiologic relationship in between these 2 components naturally arises in this study. This biphasic lesion may represent a collision of 2 unrelated mucinous tumor and mesonephric lesion. However, similar to the coexistence of LGSC and MLA in the ovary, these morphologically distinct mucinous and mesonephric populations may be clonally related and our results support this interpretation. In case 4, a NGS-based molecular analysis was performed to compare the genetic alterations between the mucinous and the mesonephric components. In this case, NGS analysis revealed a common KRAS p.G12V driver mutation in both tumor components. This mutation was also detected in benign MLPs, suggesting that the KRAS G12V mutation was an early event in this carcinogenic process. This shared genetic alteration indicates that these 2 morphologically divergent tumors are clonally related rather than representing collision tumors. Additional genetic alterations, including CTNNB1 p.Ser33Phe somatic mutation, FGFR2 amplification, and CDKN2A/p16 deletion, were only present in the MLA but not in the MBT/APMT, indicating an accumulation of lineage-specific acquired genetic changes that drive the development of MLA. Complete loss of p16 staining, correlating with CDKN2A/p16 gene deletion, was present in some benign MLPs but not in others. We suspect that MLPs with loss of p16 expression continued to evolve into MLA, whereas those with intact p16 developed into MBT/APMT, forming 2 morphologically divergent components. Similar findings have been described in the published cases discussed above. 8 In one case (OV75), molecular analysis revealed that both the MLA and MBT/APMT shared clonal CD79A p.T140N, NOTCH2 p.V1633I, and POLD1 p.I927L missense mutations as well as similar copy number alterations. Both components harbored a subclonal PIK3CA p.P539R hotspot as well as a KRAS p. G12V hotspot mutations. On the other hand, a SETD8 p.A21V mutation was detected only in the MLA component and a PIK3R1 p.R386G mutation was detected only in the MBT/APMT component. In another case (OV21), both the pure MBT/APMT and the mixed MLA and mucinous cystadenofibroma shared clonal KRAS p.G12D hotspot mutations, SPOP p.M117T missense mutations, and gains of chromosomes 10, 12, and 1q. Subclonal CTNNB1 p.D32V/p.T41A and AKT1 p.E17K hotspot mutations were found only in the latter component and a subclonal NOTCH3 p.F1327S missense mutation was found only in the former component. The presence of shared somatic alterations between the mixed MLA and the MBT/APMT provides further evidence to support the notion of clonal relatedness of both components, while the presence of private mutations in each component indicates subsequent differentiation into lineage-specific lesions illustrated by distinct morphology. MLPs in all 4 cases displayed variable degree of mucinous metaplasia/differentiation and were intimately associated with the mucinous glands and epithelial lining of the mucinous tumor component. We suspect that the mesonephric epithelium undergoes lineage-specific differentiation to mucinous epithelium through metaplasia, which then can give rise to a mucinous tumor. We have proposed a histopathogenic model for our case 4 ( Fig. 7C ) and this model can be expanded to all similar cases including 2 recently reported ones. 8 In theory, a single progenitor cell or pluripotent stem cell, either derived from mesonephric remnants or from transdifferentiation of Müllerian tissue, acquires KRAS or other driver mutations, proliferates, and subsequently differentiates into 2 distinct morphologic lineages. Some cells may differentiate into a mucinous tumor through metaplasia, whereas others may accumulate additional genetic alterations and give rise to MLA. In our view, it is unlikely that MLA originated primarily from MBT/APMT for as the mucinous epithelium is thought be the more terminally differentiated tissue type. Likewise, it is also unlikely that the mucinous tumor was derived primarily from the MLA, which would have needed to lose additional genetic alterations. It has been reported that the most common genetic alteration in both ovarian and endometrial MLAs is KRAS somatic mutation 8 , 10 , 16 - 18 , 23 which was also detected in case 4 in this series. Interestingly, a KRAS -activating mutation is also thought to be the most common single molecular genetic alteration in ovarian mucinous tumors. 24 - 27 KRAS mutations have been found in up to 56% of mucinous cystadenomas. 24 Of note, KRAS mutations have been identified in benign, borderline, and malignant components of the same mucinous neoplasm, suggesting that KRAS mutations are an early event in progression from benign tumors to MBT/APMT to carcinoma. 24 , 27 We suspect that the presence of common KRAS mutations in both MLAs and ovarian mucinous tumors may indicate some intrinsic link between these 2 entities. We propose that this link can be explained in part by mesonephric proliferations being the origin of some ovarian mucinous tumors, and we believe that our case series provides some evidence in favor of this hypothesis. In summary, we report 4 cases of mucinous tumors (1 mucinous cystadenofibroma and 3 MBT/APMTs) with co-existent mesonephric lesions in the ovary. All cases contained benign MLPs which showed variable degree of gastrointestinal metaplasia/differentiation. These MLPs are postulated to give rise to lineage-specific mucinous and mesonephric lesions and our results support their clonal relationship. We propose that MLPs are a possible new origin of some ovarian mucinous tumors, however, this is largely speculative. Whether these benign MLPs arise through transdifferentiation of Müllerian tissue or represent true mesonephric remnants, however, remains unknown.