Malignant Brenner tumor with extensive clear cell features: A valuable diagnostic lesson learned from molecular tumor profiling and AI-based tumor differential tool.

A 55-year-old woman presented with two weeks of severe constipation, nausea, and vomiting. The imaging revealed a large 17.7 cm complex left ovarian mass without evidence of bladder involvement. Preoperative laboratory tests revealed elevated carcinoembryonic antigen (CEA) at 63.4 and cancer antigen 125 (CA-125) at 84. The patient underwent an exploratory laparotomy. A large, partially cystic and partially solid ovarian mass (20 x 14 x 8 cm) was received for intraoperative consultation. The mass consisted of fleshy and white-tan smooth cut surfaces. A representative frozen section showed tumor cells with cytoplasmic clearing, distinct cellular borders, and moderate to marked cytologic atypia, consistent with a high grade carcinoma with extensive clear cell features ( Fig. 1 A). The intraoperative diagnosis directed surgical staging including total abdominal hysterectomy, bilateral salpingo-oophorectomy, infracolic omentectomy, bilateral pelvic and para -aortic lymphadenectomy, and excision of small bowel mesentery nodules. Fig. 1 Initial pathologic evaluation of the ovarian tumor. (A) The frozen section shows extensive clear cell changes (10x). Inlet: High power view of the marked cytologic atypia in the clear tumor cells. The permanent sections show: (B) Tumor cells with an expansile growth pattern replacing the ovarian stroma (2x). (C) Histologic features characteristic of a clear cell carcinoma (20x). (D) A component of clear cell borderline tumor in the background (2x). Inlet: High power view of the mild cytologic atypia in the clear tumor cells. Initial pathologic evaluation of the ovarian tumor. (A) The frozen section shows extensive clear cell changes (10x). Inlet: High power view of the marked cytologic atypia in the clear tumor cells. The permanent sections show: (B) Tumor cells with an expansile growth pattern replacing the ovarian stroma (2x). (C) Histologic features characteristic of a clear cell carcinoma (20x). (D) A component of clear cell borderline tumor in the background (2x). Inlet: High power view of the mild cytologic atypia in the clear tumor cells. On the permanent sections, the left ovarian tumor showed classic histologic characteristics of ovarian CCC. Large, crowded nests of clear cells showed inner cystic papillary growth and replaced ovarian stroma in an expansile growth pattern ( Fig. 1 B-C). There were also variable sizes of cysts and nests composed of clear cells with mild to moderate cytologic atypia in a fibromatous stroma consistent with a clear cell borderline tumor in the background ( Fig. 1 D). Based on these findings, the tumor was diagnosed as ovarian CCC confined to the ovary (FIGO stage IA). All the pelvic and peri -aortic lymph nodes (total 23) were negative for metastatic carcinoma. In our institution, all malignant ovarian tumors are sent for molecular analysis by commercial next-generation sequencing (NGS). Based on the NGS results, the commercial AI-based tumor differential tool predicted an 88 % phenotypic match of urothelial origin for our ovarian tumor driven by strong expression of GATA3, uroplakin 2, and AMACR, while assigning 0 % phenotypic match to ovarian CCC. To confirm these gene expressions, the commercial laboratory performed immunohistochemical (IHC) urothelial markers of GATA3, uroplakin 2, and AMACR, which showed positive expression in our ovarian tumor ( Fig. 2 ). Pax-8, a Müllerian marker, was negative. Fig. 2 Immunohistochemistry performed at the commercial laboratory collaborator demonstrates that the tumor cells are positive for GATA3, Uroplakin, AMACR, and negative for PAX8. Immunohistochemistry performed at the commercial laboratory collaborator demonstrates that the tumor cells are positive for GATA3, Uroplakin, AMACR, and negative for PAX8. The case was re-evaluated by the primary pathology and clinical team. Additional tissue from the ovarian tumor was submitted by pathology. One of the 22 sections showed a small benign Brenner tumor adjacent to the tumor area with extensive clear cell changes ( Fig. 3 A). IHC staining showed diffuse positivity of GATA3 and p63 (urothelial markers) in both the tumor cells with clear cell changes and benign Brenner tumor nests ( Fig. 3 B-C). The tumor also showed wild-type expression of p53, intact nuclear expression of MMR, and negative expression of Napsin A, WT-1, ER, PR, and Pax-8. Intraluminal mucinous material, often observed in benign Brenner tumors, was focally present ( Fig. 3 D). Based on the histologic and IHC evidence of urothelial differentiation and the presence of a component of benign Brenner tumor, the final diagnosis was amended to MBT with extensive clear cell changes. The clinical team also reviewed the radiographic and surgical findings and confirmed there was no tumor involvement in the bladder. Fig. 3 Pathologic re-evaluation of the ovarian tumor. (A) Nests of benign Brenner tumor are found adjacent to the tumor with clear cell changes (10x). Inlet: High power view of benign Brenner tumor. (B) GATA3 is diffusely positive in both the benign Brenner nests and the tumor with clear cell changes (10x). (C) p63 is diffusely positive in both the benign Brenner nests and the tumor with clear cell changes (10x). (D) Intraluminal mucin is present in the tumor with clear cell changes (20x). Inlet: High power view of intraluminal mucin. Pathologic re-evaluation of the ovarian tumor. (A) Nests of benign Brenner tumor are found adjacent to the tumor with clear cell changes (10x). Inlet: High power view of benign Brenner tumor. (B) GATA3 is diffusely positive in both the benign Brenner nests and the tumor with clear cell changes (10x). (C) p63 is diffusely positive in both the benign Brenner nests and the tumor with clear cell changes (10x). (D) Intraluminal mucin is present in the tumor with clear cell changes (20x). Inlet: High power view of intraluminal mucin. The NGS data was further reviewed with the molecular liaison team from the commercial laboratory, which showed CDKN2AB loss and a molecular triple phenotype (MDM2 amplification/TP53wt/TERTwt) that have been described for MBT but not for ovarian CCC. Wild-type TERT promoter status also disfavored metastatic urothelial carcinoma. The patient initiated the first cycle of chemotherapy with carboplatin plus paclitaxel after the initial diagnosis of ovarian CCC. After the diagnosis amendment and shared decision making, she opted to complete six cycles of chemotherapy as she had tolerated the treatment well. Her CA-125 levels had been stable at 10 to 14 mL/unit throughout treatment without clinical signs of recurrence.

John A Steinharter: Writing – review & editing, Writing – original draft, Visualization, Data curation. Corinne Jansen: Writing – original draft, Data curation. Hassan Ghani: Writing – review & editing, Software, Formal analysis. Apsra Nasir: Writing – review & editing, Visualization, Formal analysis. Marzia Capelletti: Visualization, Formal analysis. Ashley Stuckey: Visualization, Resources. M.Ruhul Quddus: Writing – review & editing, Formal analysis, Conceptualization. Yun-An Tseng: Writing – review & editing, Writing – original draft, Visualization, Supervision, Project administration, Formal analysis, Data curation, Conceptualization.

MBTs share similar clinical and radiographic features with other ovarian epithelial tumors. They have been reported in patients between 23 and 94 years of age (a mean age of 49.5 – 66 years) with abdominal pain and distension ( Wang et al., 2024 ). They often appear on imaging as a solid lesion with mild to moderate enhancement and amorphous calcification ( Wang et al., 2024 ). There is no established serological tumor marker for MBTs, but CA-125 is the most widely used. CA125 is elevated in 30–75 % of patients with MBT, and is limited by low sensitivity (50–62 %) and moderate specificity (94–98.5 %) ( Wang et al., 2024 ). Despite lacking specific clinical and radiographic features, MBT has well-defined histologic criteria. The differential diagnosis usually includes high-grade serous carcinoma with urothelial differentiation or metastatic urothelial carcinoma ( Wang et al., 2024 ). Extensive clear cell features have not been reported in MBT and their presence can mislead the pathologists to diagnose ovarian CCC. When ovarian neoplasms show clear cell changes, the common differential diagnoses usually include secretory changes seen in endometrioid adenocarcinomas, microcystic pattern and cytoplasmic vacuoles observed in yolk sac tumors, and papillary architectural growth patterns associated with serous borderline tumors or carcinomas ( Malpica, 2016 , Sangoi et al., 2008 ). However, these tumor entities usually have other histologic clues that would direct the pathologists to the necessary pathologic work-up to reach the correct diagnosis. The MBT in this case posed a diagnostic challenge on both intraoperative consultation and final diagnosis because there were no conventional features of a MBT. Not only the component with overtly malignant features exhibit histologic morphology resembling an ovarian CCC, but the lower grade component also resembled a clear cell borderline tumor. With this case report, now the differential diagnosis for clear cell ovarian neoplasms will be further expanded to include MBT with extensive clear cell features. The important diagnostic lessons learned from this case include extensive tumor sampling to look for histologic evidence of urothelial differentiation, which includes components of benign or borderline Brenner tumors and intraluminal mucinous material often observed in benign Brenner tumors. IHC urothelial markers (e.g. GATA3, uroplakin, high molecular weight cytokeratin, etc) should also be performed up-front to help differentiate MBT with extensive clear cell features from CCC. Positive expression of urothelial markers is observed in MBT, but not in ovarian CCC. Diagnosing MBT is typically achieved by histologic and immunohistochemical assessment alone. In this case report, an AI-based tumor differential tool contributed the first clue of urothelial differentiation in the tumor, triggering histologic re-evaluation. The molecular data showed that our tumor harbored CDKN2AB loss and a molecular triple phenotype of MDM2 amplification/TP53 wt/TERT wt, which has been documented for other MBTs in the literature ( Pfarr et al., 2017 , Lin et al., 2021 ). Ovarian CCCs are known to be associated with endometriosis and carry recurrent somatic mutations in PIK3CA and ARID1A in 46–57 % and 33–50 % of ovarian CCC, respectively ( Bolton et al., 2022 ). In this case, the tumor did not have these molecular alterations or a histologic association with endometriosis. This case demonstrated the powerful impact of molecular tumor profiling and AI-based diagnostics tools on tumor diagnosis, especially when the histologic morphology is unconventional. AI is currently limited by the differential it is trained on and the diversity of samples within each diagnostic label. Therefore, new, rare, or obscure diagnoses can easily have errors in AI output. The ovarian tumor in this case was interpreted as urothelial carcinoma because MBT was not part of the differential implemented in the AI-based diagnostic tool used to analyze the case. However, AI can provide rapid analysis of whole exome and whole transcriptome data. Integration of AI into a workflow allows for a second opinion to flag potential misdiagnoses ( Rakha et al., 2021 ). Another important molecular finding in this ovarian tumor is the lack of TERT promoter mutations. It has been reported that all Brenner tumors, regardless of grading, lack TERT promoter mutations, which are the most frequently identified molecular alteration in invasive urothelial carcinomas ( Pfarr et al., 2017 , Khani et al., 2016 ). This suggests that Brenner tumors and urothelial carcinomas likely have distinct molecular pathogeneses. The TERT promoter wild-type feature and the presence of a benign Brenner tumor component helped confirm an ovarian origin tumor in this case. Excluding a metastatic urothelial carcinoma of the bladder/urinary tract origin was important as clear cell changes are well-documented features in urothelial carcinomas ( MacLeod et al., 2021 ). The clinical impact of an accurate determination of tumor origin was also profound as the management would have been drastically different. From a clinical perspective, accurate diagnosis is essential for prognostic counseling. Like other ovarian epithelial tumors, the prognosis for MBT is stage-dependent. Two studies performed on SEER database showed that it has a worse prognosis for localized and regional diseases (except for carcinosarcoma) and a similar to better prognosis for distant diseases when compared to other high-grade ovarian carcinomas ( Lan and Yang, 2019 , Peres et al., 2018 ). Currently MBT is managed similarly to other ovarian carcinomas with complete surgical staging and tumor debulking. Limited studies showed that lymph node dissection for MBT may have a lower yield compared to other high grade ovarian carcinomas and is not associated with improved disease outcome ( Zhang et al., 2019 , Nasioudis et al., 2016 ). Small studies have also suggested that conservative management with surveillance may be considered for patients with stage IA or IB disease ( Han et al., 2015 , Gezginç et al., 2011 ). Adjuvant platinum-based chemotherapy can be given to patients with at least stage IC disease or recurrence. While there is controversy in treating stage IA MBT with chemotherapy, our patient elected to continue her adjuvant chemotherapy based on the NCCN guidelines, her good tolerance to chemotherapy, and the uncertain biologic behavior of this rare tumor. While changing the diagnosis did not impact the clinical management significantly, an accurate diagnosis still holds importance as correct tumor classification is the foundation for future studies on prognosis, treatment response, and management guidelines.

In conclusion, this case underscores the diagnostic challenges of ovarian tumors with clear cell features and expands the histologic spectrum of MBT to include clear cell morphology. Multidisciplinary collaboration among pathologists, molecular scientists, and surgical oncologists remains the gold standard for resolving challenging cases and optimizing patient care. AI-based tumor differential tools are valuable adjuncts to tumor diagnostics. When carefully investigated with other pathological and clinical findings, AI-second opinion can act as a strong safety net for cancer diagnostics, especially when the histologic morphology is not conventional. Consent When a patient receives treatment at this academic center, Women & Infants Hospital of Rhode Island, a general consent form is signed, which allows the use of the residual material for academic purposes, provided that unique Identifiers are not revealed. No new sample was collected from the patient. The residual paraffin tissue used in these additional studies was used for clinical purposes without exhausting the tissue in the paraffin block for the patient's potential future needs.

Brenner tumors are rare epithelial ovarian neoplasms, accounting for 1–3 % of all ovarian tumors. They are classified as benign, borderline, or malignant, with malignant Brenner tumors (MBTs) representing fewer than 5 % of all Brenner tumors ( Organisation mondiale de la santé, 2020 ). According to the 2020 WHO classification, MBTs are defined as high-grade carcinomas exhibiting urothelial differentiation with an associated benign or borderline Brenner tumor component ( Organisation mondiale de la santé, 2020 ). In contrast, ovarian clear cell carcinoma (CCC), another aggressive high-grade carcinoma that constitutes 10–11 % of ovarian carcinomas, has different histologic features characterized by clear cell morphology and a frequent association with endometriosis ( Organisation mondiale de la santé, 2020 ). In this case report, we present a diagnostically challenging case initially diagnosed as ovarian CCC but later reclassified as an MBT with extensive clear cell features following molecular profiling with next-generation sequencing (NGS) and an artificial intelligence (AI)-based tumor differential tool. This is the first reported case of a MBT mimicking ovarian CCC, which provides an important diagnostic lesson, highlights the value of molecular tumor profiling and AI-based diagnostic tools, and draws attention to the limited clinical data for this rare ovarian epithelial tumor.

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: The Dr. John Steinharter, Dr. Corinne Jansen, Dr. Ashley Stuckey, Dr. M. Ruhul Quddus, and Dr. Yun-An Tseng declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Hassan Ghani is an employee of Caris Life Sciences, a commercial lab that offers the AI tumor of origin assay (“MI GPSai”), and is the primary developer of said assay. Apra Nasir and Marzia Capelletti are employees at Caris Life Sciences (members of the molecular liaison team).