Synchronous endometrial/cervical and ovarian/fallopian tube carcinoma: a genome-wide mutation analysis.

LL and MY made equal contributions to this work. LL and YS conceived the study. LL and MY carried out experiments. LL was involved in analyzing and interpreting the data. LL wrote the paper. SS sponsored the study. SS and YS are co‐corresponding authors.

Our study evaluated a cohort comprising 12 SEOC and 5 SCOC patients treated at the Cancer Hospital, Chinese Academy of Medical Sciences from 2012 to 2021. The clinicopathological details are methodically tabulated in Table  1 and supplementary material, Table  S4 for SEOC, and Table  2 and supplementary material, Table  S4 for SCOC cases. Clinicopathological presentation of 12 SEOC cases CCC, clear cell carcinoma; DC, dedifferentiated carcinoma; EC, endometrial carcinoma; EEC, endometrioid endometrial carcinoma; EOC, endometrioid ovarian carcinoma; FIGO, International Federation of Gynecology and Obstetrics; HGSC, high‐grade serous carcinoma; OFC, ovarian/fallopian tube carcinoma; OS, overall survival time; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma; USC, uterine serous carcinoma. Clinicopathological presentation of SCOC cases AC, adenocarcinoma; CC, cervical cancer; FIGO, International Federation of Gynecology and Obstetrics; HASCC, HPV‐associated SCC; HIAC, HPV‐independent adenocarcinoma; OFC, ovarian/fallopian tube carcinoma; OS, overall survival time; SCC, squamous cell carcinoma; SCOC, synchronous cervical and ovarian/fallopian tube carcinoma. FIGO 2019. FIGO 2014. The SEOC cohort, with a median age of 50 years (range: 37–67 years). Regarding EC, two patients were diagnosed at FIGO stage (2023) IA2, one at FIGO stage IAm‐ POLE mutant, one at FIGO stage IB, three at FIGO stage IA3, and five at FIGO stage IIIA. OFC staging (FIGO 2014) showed seven patients at stage I, two at stage II, and three at stage III. Lymph node dissection was performed in eight patients, with one showing lymph node involvement (SEOC‐7). Notably, two patients (SEOC‐6 and SEOC‐10) underwent platinum‐based chemotherapy due to the aggressive nature of their carcinomas (HGSC and CCC). Only one patient (SEOC‐1) exhibited endometriosis in the ovary adjacent to the ovarian tumor. Additionally, two patients (SEOC‐1 and SEOC‐5) had atypical hyperplasia/endometrial intraepithelial neoplasia (AH/EIN) in close proximity to the endometrial tumor. The median tumor size for OFC was 3.6 cm, ranging from 1.3 to 13 cm. As illustrated in Figure  1 and classified according to Scully's criteria, six cases were categorized as metastatic. SEOC‐6 demonstrated both uterine serous carcinoma (USC) and ovarian HGSC, while SEOC‐14 displayed USC and fallopian tube HGSC. These two cases exhibited strongly positive WT1 expression in both EC and OFC. SEOC‐10 and SEOC‐11 both presented with endometrial and ovarian CCC, with SEOC‐11 additionally harboring a minor endometrioid component, while SEOC‐8 manifested as EEC and EOC. SEOC‐7 involved EEC and ovarian CCC. The remaining six cases were classified as double primary cases: SEOC‐1, SEOC‐4, and SEOC‐5 featured EEC and EOC, with SEOC‐4 and SEOC‐5 further featured by mucinous differentiation and papillary structures in the ovarian tumors, respectively. SEOC‐3 displayed EEC and ovarian DC, SEOC‐9 showed endometrial DC and ovarian CCC, and SEOC‐12 presented with endometrial EC and fallopian tube HGSC. Overview of the study design and analysis workflow. This figure illustrates the step‐by‐step process employed in the study to analyze the SEOC and SCOC samples, from sample collection to data analysis. The workflow includes stages from patient selection through sample collection to classification based on Scully's criteria and WES. AdC, HPV‐independent adenocarcinoma; CCC, clear cell carcinoma; DC, dedifferentiated carcinoma; EEC, endometrioid endometrial carcinoma; EOC, endometrioid ovarian carcinoma; HGSC, high‐grade serous carcinoma; SCC, HPV‐associated squamous cell carcinoma; SCOC, synchronous cervical and ovarian/fallopian tube carcinoma; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma; USC, uterine serous carcinoma; WES, whole‐exome sequencing. Among the five women included in the study with SCOC, the median age was 52 years (range: 49–60 years). SCOC‐1 was diagnosed with HPV‐independent adenocarcinoma in the cervix and OFC. Similarly, SCOC‐2 was identified as HPV‐independent adenocarcinoma with neuroendocrine differentiation in both the cervix and OFC. SCOC‐1 and SCOC‐2 exhibited patchy p16 immunohistochemistry (IHC) expression. SCOC‐3 to SCOC‐5 presented as early‐stage HPV‐associated SCC as the p16 IHC exhibited diffuse, block‐like staining, with histological similarities observed between cervical and ovarian tumors. SCOC‐3 and SCOC‐5 were diagnosed at FIGO stage IA for CC and FIGO stage I for OFC. SCOC‐4 was diagnosed at FIGO stage IA for CC and FIGO stage II for OFC. Additionally, two patients with adenocarcinoma were at FIGO stage IB for CC and FIGO stage II for OFC. Lymph node dissection was performed in all patients, with two showing lymph node involvement (SCOC‐1 and SCOC‐3). Additionally, one patient (SCOC‐1) had received platinum‐based chemotherapy prior to surgery due to the poorly differentiated nature of the cancer. The median size of ovarian/fallopian tube tumors in SCOC was 1.5 cm, ranging from 0.5 to 20 cm. Our study encompassed WES of 51 samples. Detailed exome coverage, presented in supplementary material, Table  S1 , averaged 225×, ranging from 48× to 563× across lesions. Somatic mutations in the samples were defined according to the following criteria: (1) the mutation was identified in four or more distinct reads; the mutation depth was greater than 15 reads, (2) the mutation frequency was greater than 0.05, (3) the mutation was uncommon in dbSNPs (frequency < 0.1%), and (4) the mutation was not present in any of the reads in the matched control sample or in the other control samples analyzed in this study. This cohort revealed 11,085 somatic mutations encompassing single nucleotide variations (SNVs), deletions, and insertions, alongside 509 driver gene mutations in SEOC cases. In contrast, SCOC cases exhibited 1,981 mutations and 122 driver gene mutations, detailed in supplementary material, Table  S2 . Notably, all cases, with the exception of one SEOC‐8 endometrial lesion, contained canonical cancer driver gene mutations (supplementary material, Table  S3 ). Analysis across 12 SEOC cases revealed seven instances (SEOC‐3, 4, 6, 9–11, 14) with significant mutation overlaps, indicating clonal relatedness and a metastatic origin, confirmed by a CI exceeding 0.995. These cases exhibited a mutation overlap ranging from 0.99% to 58.53% (Figure  2 ). Among them, in three (43%) of seven cases, there was a substantial percentage (>20%) of shared somatic mutations, while the remaining cases had fewer mutations in common. Among the four cases that shared less than 20% of somatic mutations, SEOC‐3 and SEOC‐9 had different histology; shared percentage is 0.09% and 15.9%, respectively. In one case (SEOC‐9), shared mutations occurred in less than 1% of total mutations. Notably, the percentage of shared somatic mutations was higher in mismatch repair‐deficient (dMMR) cases compared to proficient (pMMR) cases, although without statistical significance ( p  = 0.08, Student's t ‐test). The remaining SEOC cases (SEOC‐1, 5, 7, 8, 12) showed no shared mutations, suggesting they are DPT. Photomicrographs of H&E staining in all metastatic pairs of EC and OFC are shown in Figure  3 and all DPT pairs of EC and OFC are shown supplementary material, Figure  S1 . Comparison of shared somatic mutations, gene copy number variations, and mutational profiles in paired EC and OFC of metastatic SEOC cases. Venn diagrams (left), gene copy number alterations (middle), and mutational signatures (right) illustrating the overlap in somatic mutations, CNVs, and mutational signatures between paired EC/CC and OFC. (Left) The Venn diagram includes both non‐synonymous and synonymous mutations, with percentages indicating the proportion of mutations shared between EC/CC and OFC. (Middle) CNV patterns of paired EC and OFC generally exhibit similarities. However, unique CNV events, such as the gains and amplifications seen in SEOC‐3 and SEOC‐10, indicate variability in certain cases. These events are highlighted in red for gains and amplifications, and in green for losses on the chromosome plots. Log2 ratios are plotted on the y ‐axis, and genomic positions are plotted on the x‐axis. (Right) Mutational signatures of all non‐synonymous SNVs in EC/CCs and OFCs of a given case are displayed based on the 96 substitution classification defined by substitution classes (C>A, C>G, C>T, T>A, T>C, and T>G bins) and the 5' and 3' sequence context. All mutational signatures are normalized to the trinucleotide frequency of the human genome. CNV, copy number variation; EC, endometrial carcinoma; OFC, ovarian/fallopian tube carcinoma; SNV, somatic single nucleotide variants. Histologic examination of all metastatic SEOC cases with hematoxylin and eosin (H&E) staining, original magnification ×100. Photomicrographs depicting metastatic SEOC cases, including EC and OFC. Note: SEOC‐4 has the mucinous differentiation in the EOC, and SEOC‐11 has a small clone with endometrioid carcinoma histology, as the figure shows. CCC, clear cell carcinoma; DC, dedifferentiated carcinoma; EC, endometrial carcinoma; EEC, endometrioid endometrial carcinoma; EOC, endometrioid ovarian carcinoma; HGSC, high‐grade serous carcinoma; OFC, ovarian/fallopian tube carcinoma; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma; USC, uterine serous carcinoma. In all SCOC cases, our comparative analysis revealed a broad range (0.5–59.0%) of shared mutations between paired CC and OFC, indicating a strong clonal relationship and metastatic pattern (Figure  4 ). This finding was consistent across all SCOC cases, with three showcasing significant mutation sharing (>20%). Histological details for these cases are provided in supplementary material, Figure  S2 . CNV patterns and mutational profiles, compared across all lesions, highlighted similarities in seven metastatic SEOC cases and five metastatic SCOC cases (Figures  2 and 4 ). Supplementary material, Figure  S3 presents CNV patterns and mutational profiles in five DPT SEOC cases. Comparison of shared somatic mutations, gene copy number variations, and mutational profiles in paired CC and OFC. Venn diagrams (left), gene copy number alterations (middle), and mutational signatures (right) illustrating the overlap in somatic mutations, CNVs, and mutational signatures between paired EC/CC and OFC. (Left) The Venn diagram includes both non‐synonymous and synonymous mutations, with percentages indicating the proportion of mutations shared between EC/CC and OFC. (Middle) CNVs of paired CC and OFC exhibit similar patterns, with gains and amplifications highlighted in red and losses in green on the chromosome plots. Log2 ratios are plotted on the y‐axis, and genomic positions are plotted on the x ‐axis. (Right) Mutational signatures of all non‐synonymous SNVs in EC/CCs and OFCs of a given case are displayed based on the 96 substitution classification defined by substitution classes (C>A, C>G, C>T, T>A, T>C, and T>G bins) and the 5′ and 3′ sequence context. All mutational signatures are normalized to the trinucleotide frequency of the human genome. CC, cervical cancer; CNV, copy number variation; OFC, ovarian/fallopian tube carcinoma; SNV, somatic single nucleotide variants. Analysis revealed a significant correlation between pathogenic POLE mutations and metastatic disease in SEOC cases, particularly noted in SEOC‐9 and SEOC‐11, which harbored mutations deemed pathogenic (Tables  1 and 3 , Figure  5 ). Specifically, SEOC‐9's mutation ( POLE P286R) is recognized as a hotspot mutation, and SEOC‐11's mutation ( POLE L698Cfs*94) has a significant POLE score [ 34 ] of 4. The absence of such mutations in DPT patients suggests an increased metastasis risk associated with these genetic alterations. Molecular characteristics of SEOC cases dMMR, deficient mismatch repair; EC, endometrial carcinoma; NSMP, no specific molecular profile; OFC, ovarian/fallopian tube carcinoma; pMMR, proficient mismatch repair; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma. WHO molecular classification. Landscape of somatic driver gene mutations in paired EC/CC and OFC. Oncoplot illustrating the distribution of somatic driver gene mutations identified in the current study, including recurrent mutations in well‐known endometrial and cervical cancer‐associated genes such as TP53 , POLE , PTEN , ARID1A , ARID1B, PPP2R1A, KRAS, CTNNB1, MTOR , and PIK3CA . Cases are stratified into SEOC (left) and SCOC (right) groups. Each row signifies a specific gene, while each column represents an individual sample. The upper histogram delineates the tumor mutation burden (TMB) per sample, whereas the right histogram delineates the frequency and count of mutations across each gene. The plot encompasses a total of 34 positive samples; the samples are located below the oncoplot. Additionally, pertinent clinical data including metastatic or independent status, vital status, primary diagnosis, lesion type, mismatch repair (MMR) status, FIGO stage, lymph node metastasis (LNM) status, lymphovascular invasion (LVSI) status, and prior treatment status are provided in the lower panel. For the mutations in the same gene of the same case, most are the same, except for the PIK3CA mutations in SEOC‐9, which are marked with asterisks inside the box. AdC, HPV‐independent adenocarcinoma; CC, cervical cancer; CCC, clear cell carcinoma; DC, dedifferentiated carcinoma; EC, endometrial carcinoma; EEC, endometrioid endometrial carcinoma; EOC, endometrioid ovarian carcinoma; HGSC, high‐grade serous carcinoma; LNM, lymph node metastasis; LVSI, lymphovascular invasion; MMR, mismatch repair; OFC, ovarian/fallopian tube carcinoma; SCC, HPV‐associated squamous cell carcinoma; SCOC: synchronous cervical and ovarian/fallopian tube carcinoma; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma; USC, uterine serous carcinoma. The comparison revealed no significant differences in TMB between EC and OFC within metastatic or DPT contexts. However, bilateral involvement of the ovary or fallopian tube was significantly associated with metastatic status ( p  = 0.04, chi‐squared test), while other clinical factors, such as LVSI, tumor size <5 cm, and multiple nodular ovarian or fallopian tube lesions showed no correlation with metastatic behavior. Endometriosis and AH/EIN were only observed in DPT cases, as shown in supplementary material, Table  S4 . As illustrated in Figure  6 , our findings suggest that in SEOC‐3—which involved EEC and ovarian DC—the OFC likely originated from a minor subclone of the EC. Similarly, in SCOC‐4, the CC appeared to derive from minor subclones of the OFC. Cancer cell fraction (CCF) distribution of non‐synonymous somatic mutations in metastatic SEOC and SCOC cases. Each block represents a patient sample, with individual boxes denoting mutations that are color‐coded according to their CCF values (refer to the legend on the right and below). The CCF values have been adjusted for tumor purity and ploidy. Patient identifiers (SEOC/SCOC) and histologic subtypes (EC/CC and OFC) are clearly labeled for reference. A detailed view of subclonal mutations that became clonal is displayed on the right. The directionality of progression from EC/CC to OFC in SEOC‐3 and SCOC‐4, and from OFC to EC in SEOC‐6, 10, 11, and 14, is indicated based on the clonality of mutations. SEOC‐4, 9, SCOC‐1‐3, and SCOC‐5 exhibited complex clonal architecture, making it difficult to determine the direction of metastasis. CC, cervical cancer; CCC, clear cell carcinoma; EC, endometrial carcinoma; EEC, endometrioid endometrial carcinoma; EOC, endometrioid ovarian carcinoma; HGSC, high‐grade serous carcinoma; OFC, ovarian/fallopian tube carcinoma; SCOC, synchronous cervical and ovarian/fallopian tube carcinoma; SEOC, synchronous endometrial and ovarian/fallopian tube carcinoma; USC, uterine serous carcinoma. Conversely, in cases SEOC‐6, 10, 11, and 14, the progression pattern was reversed, with EC emerging from minor subclones of the OFC. This was substantiated by immunohistochemical findings, particularly noted in SEOC‐6 and SEOC‐14, where diffuse WT1 positivity indicated likely dissemination of OFC to the endometrium. SEOC‐11 exhibited mixed histology of CCC and EOC in the ovary. The metastatic lesion in the endometrium was identified as CCC, indicating that only the CCC component had spread to the endometrium. This suggests a selective metastatic potential of the CCC subclone in this case. The results underscore the bidirectional potential of cancer metastasis between primary and metastatic sites. MMR testing highlighted that four samples displayed dMMR status, impacting both the EC and OFC in three of the seven metastatic cases analyzed (Figure  5 , Tables  1 and 3 ). This distribution of dMMR status across both metastatic and DPT cases suggested that MMR proficiency alone may not conclusively determine the origin of ovarian/fallopian tube malignancies. Investigating the impact of metastatic versus DPT categorizations on patient survival, we noted a median follow‐up duration of 42 months. Of these, patients SEOC‐6, SEOC‐10, and SEOC‐7 succumbed to the tumor, while SEOC‐3 died due to other reasons. The remaining SEOC patients were alive and free from relapse. In the SCOC group, SCOC‐1 died from the tumor, while SCOC‐3 was alive but experienced relapse with retroperitoneal lymph node metastasis. The other SCOC patients were alive and free from relapse. Nevertheless, the majority of patients in both cohorts remained alive and relapse‐free, suggesting nuanced survival dynamics contingent on the specific nature and treatment of their conditions. To assess survival outcomes in SEOC patients, we conducted follow‐up for all 54 patients in our cohort. Eighteen patients were in FIGO stage (2023) IA3, and 36 were in FIGO stage IIIA. Seven patients were lost to follow‐up, leaving 47 cases for final analysis with a median follow‐up duration of 70 months (supplementary material, Table  S4 ). The SEOC FIGO 2023 classification effectively stratifies patients in our cohort into distinct prognostic groups (IA3 versus IIIA) ( p  < 0.05 in for both PFS and OS, as shown in supplementary material, Figure  S4 ). The 5‐year survival rate for the IA3 group was 100%. Despite being stage III, the IIIA subgroup demonstrated a relatively favorable 5‐year survival rate of approximately 75%. TP53 status was associated with poorer survival ( p  < 0.05). However, within the IIIA subgroup, neither p53 status, histologic grade (low versus high), LVSI, nor deep myometrial infiltration provided additional prognostic stratification. Notably, TP53 ‐mutated SEOC was exclusively observed in advanced‐stage disease, with no IA3 cases exhibiting TP53 mutations.

Our study's exploration of the clonal relationships in SEOCs and SCOCs through WES has illuminated significant insights into the genetic continuum of these cancers. Our analysis unveiled clonal relatedness in seven pairs of SEOCs and all five SCOCs. Remarkably, the presence of different histological types may also be associated with clonal relatedness. Furthermore, we investigated the metastatic pathways observed in the metastatic cases through clone analysis based on the CCF in non‐synonymous mutations. Our observations highlight the critical role of WES in resolving the ambiguity between metastatic and DPT classifications. The ability of WES to detect shared mutations at levels as minimal as 0.5% underscores its superiority over traditional panel sequencing methods, which may overlook subtle yet significant genetic overlaps. The presence of shared mutations across histologically distinct tumors suggests a nuanced spectrum of tumor origin, challenging conventional diagnostic criteria based solely on morphology and IHC. We further delve into morphological variances and instances of dedifferentiation, particularly in metastatic cases such as SEOC‐3 and SEOC‐9, where the shared mutation percentages were notably low. Such variances may imply that these cases underwent early‐stage dissemination or that the metastatic lesions originated from a small clone within the primary site, contributing to significant divergence in histology. Throughout the metastatic process, a myriad of genetic and environmental factors could contribute to these observed differences in histology. In our larger cohort survival analysis, the FIGO 2023 classification demonstrated excellent prognostic stratification. Notably, the IIIA subgroup showed a favorable 5‐year survival rate of approximately 75%, suggesting that SEOC outcomes may be better than previously recognized. Similar to previous findings [ 1 ], TP53 status predicted OS but did not contribute to risk stratification in the IIIA subgroup. Other conventional histopathological markers – including tumor grade, LVSI, and myometrial invasion – failed to provide additional risk stratification within this IIIA subgroup, underscoring the necessity for novel biomarkers to further classify IIIA patients. A striking observation was the complete absence of TP53 mutations in IA3 tumors compared to their high frequency in advanced‐stage disease, supporting their potential role in tumor progression. Patients with multiple transitional cell carcinomas of bladder and ureter, including low grade and high grade, have been found to be clonally related [ 35 , 36 , 37 , 38 ], but are treated as independent tumors; and intramural metastasis is frequently observed in the gastrointestinal tract [ 39 ], but stage is not upgraded. Thus, as this may considered a form of benign ‘metastasis’, the term ‘trans‐tubal spread’ emerges as a more accurate descriptor of the dissemination observed as the prognosis is favorable, challenging the traditional notion of metastasis as inherently aggressive. Dedifferentiation may be acquired during the spread process, for example the SEOC‐3 featured EEC and ovarian DC, SEOC‐9 displayed endometrial DC and ovarian CCC. We do not rule out the alternative possibility that the discordance may be a result of intralesional heterogeneity, in which dedifferentiation lesion develops from a minor subpopulation within the primary tumor, which shows dedifferentiation histology. Spread within organs such as the esophagus, ureter, and bladder typically does not affect staging. However, there is a notable distinction between spread within the urothelium and the gynecological‐ and tract. Concerning tumor progression from EC to OFC, one might speculate that ovarian tumors originating from the uterus could have a greater propensity to spread to the peritoneum compared to primary endometrial tumors, as part of the ovary is exposed to the peritoneum. Furthermore, our study illuminates the bidirectional potential of tumor spread between the endometrium/cervix and the ovary/fallopian tubes. The strong WT1 expression in EC and OFC noted in specific cases underscores the genetic underpinnings of this spread, supporting the hypothesis of ovarian/fallopian tube carcinomas as the origin of metastasis in certain instances [ 40 ]. Our study delves into the debated territory of tumor dissemination mechanisms, examining traditional models such as hematogenous and lymphatic spread alongside the less commonly acknowledged but biologically plausible model of trans‐tubal implantation [ 41 ]. The absence of a correlation between LVSI and spread status, coupled with the less common occurrence of lymph node and distant metastases in our cohort, strongly supports trans‐tubal implantation as a key mechanism for tumor cell dissemination within the gynecological tract; a schematic diagram is shown in supplementary material, Figure  S5A . This inference is further reinforced by the observation of tumor spread along the fallopian tube and endometrium in cases of ovarian CCC, depicted in supplementary material, Figure  S5B . This spread pattern has been likened to the ‘seed and soil’ phenomenon in disease dissemination. The phenomenon reported here may also have implications for precursor lesions. Several studies indicate that serous endometrial intraepithelial carcinoma (SEIC) [ 42 , 43 ], serous tubal intraepithelial carcinoma (STIC), and other early‐stage gynecological carcinomas can follow the same route of spread [ 44 , 45 , 46 ]. This highlights the importance of considering these precursor lesions in future research and clinical practice. Aligning with a previous study [ 47 ], the observed pathogenic POLE mutation rate of 29% (two out of seven) in metastatic cases, compared to the absence of such mutations in DPT cases, strengthens the hypothesis that pathogenic POLE mutations are more likely to be associated with metastatic status. This highlights the importance of evaluating the pathogenicity of POLE mutations, which is crucial for understanding their biological impact and potential for contributing to metastasis. But since the POLE mutant ECs are considered mostly low stage (mostly stage I), our hypothesis needs to be confirmed with a larger study and downstream verification in the future. While one study suggested that the majority of SEOCs do not arise from hereditary germline mutations but rather as sporadic cancers [ 48 ], other research has delved into the association between SEOCs and Lynch syndrome [ 7 , 10 , 31 ] and suggested that patients with Lynch syndrome may have the possibility of DPT or trans‐tubal spread. This study provides valuable insights into the clonal relationships between SEOCs and SCOCs, although its conclusions are moderated by several limitations. The small cohort size may not adequately represent the broader patient population, and the retrospective nature of the analysis introduces potential biases and limits control over confounding variables. Our case selection, which includes heterogeneous cases with divergent histology and excludes those with similar histology, could indeed introduce biases. These factors underscore the necessity for larger, prospective studies with comprehensive genomic profiling to confirm our findings and further elucidate the complex biology of these cancers. Nonetheless, this research makes significant contributions by analyzing metastatic pathways, particularly emphasizing trans‐tubal implantation, and incorporating cases with diverse histologies. In conclusion, our analysis advocates for a nuanced understanding of SEOC and SCOC spread. By integrating molecular diagnostics into clinical decision‐making, we can pave the way for more personalized and effective treatment strategies, ultimately improving patient outcomes in gynecological oncology.

The concurrent presentation of endometrial and ovarian/fallopian tube carcinoma, collectively referred to as synchronous endometrial and ovarian/fallopian tube carcinoma (SEOC), has emerged as a focal point in gynecologic oncology, representing 3.1–10% [ 1 , 2 , 3 ] of all endometrial and ovarian cancers. In most situations, both the endometrial carcinoma (EC) and the ovarian/fallopian tube carcinoma (OFC) are endometrioid carcinomas. However, occasionally, they may exhibit similar but different differentiation, and rarely, the histologic type of the tumor differs between the two organs. The differential diagnosis between metastatic spread and dual primary tumors (DPTs) has profound implications for staging and treatment. Specifically, metastatic involvement of the ovary by EC necessitates comprehensive treatment due to its advanced staging (III), contrasting sharply with the management of DPTs, which might not require additional intervention owing to their generally low‐grade nature. The Ulbright and Roth criteria, alongside the Scully criteria, are utilized to differentiate synchronous tumors from metastatic counterparts [ 4 , 5 , 6 ]. Despite their application, distinguishing between these tumor types presents considerable challenges, primarily due to the risks of either over‐treatment or under‐treatment. This underscores the critical need for a more nuanced comprehension of their genomic intricacies. In response, recent studies have delved into the genetic underpinnings of these cases, exploring clonal relationships through sequencing. SEOC cases have been bifurcated based on histotype: those with the same histotype and those with different histotypes. To date, extensive genomic analyses have been conducted on roughly 70 cases of endometrioid endometrial carcinoma (EEC) and endometrioid ovarian carcinoma (EOC), employing both targeted sequencing (using 39 and 409 gene panels) and whole‐exome sequencing (WES). These methodologies aimed to scrutinize somatic mutations, copy number variations (CNVs), and mutational signatures to ascertain clonality. Impressively, more than 95% of these cases revealed a clonal relationship, underscoring a significant genetic linkage [ 7 , 8 , 9 , 10 , 11 , 12 ]. Conversely, a minority exhibited histotype diversity, particularly between EEC and ovarian clear cell carcinoma (CCC), with only two cases among them confirming clonal congruence. Furthermore, even in cases where tumors were clonally related or identified as metastatic, patients with simultaneous low‐grade endometrioid carcinoma affecting both the endometrium and ovary reported favorable outcomes [ 13 ]. The 2023 International Federation of Gynecology and Obstetrics (FIGO) staging guidelines [ 14 ] recommend that patients with low‐grade endometrioid carcinoma involving the endometrium and ovaries (stage IA3) can be managed without adjuvant therapy, provided they fulfill specific criteria: (1) no more than superficial myometrial invasion is present (<50%); (2) absence of extensive/substantial lymphovascular space invasion (LVSI); (3) absence of additional metastases; and (4) the ovarian tumor is unilateral, limited to the ovary, without capsule invasion / rupture (equivalent to pT1a) [ 15 ]. This guidance reflects a nuanced approach to treatment, tailoring therapy to the genetic and histopathological characteristics of the tumors. Despite numerous studies in the field, a mere six cases have exhibited histological uniqueness. In clinical settings, we often encounter cases with different histological types, not classified as EEC and ovarian CCC. We have also encountered high‐grade tumors, including high‐grade serous carcinoma (HGSC), CCC, and dedifferentiated carcinoma (DC), manifesting as SEOC. This highlights a significant gap in our understanding, pointing to the urgent need for comprehensive investigations to ascertain whether these instances indeed constitute synchronous DPTs. Synchronous cervical and ovarian/fallopian tube carcinoma (SCOC) occurs less frequently than SEOC. However, documented instances of ovarian metastasis from cervical carcinoma (CC) have been reported, affecting 1.5–3.61% of the cases studied [ 16 , 17 , 18 ]. The likelihood of ovarian involvement varies with the histological subtype of the cervical cancer; approximately 5% of adenocarcinomas and less than 1% of squamous cell carcinomas (SCCs) lead to ovarian metastases [ 16 ]. The spread of CC to the ovaries is frequently linked to tumor extension into the uterine corpus [ 19 , 20 ] and mucosal involvement of the fallopian tube [ 19 ]. Predominantly, research indicates that ovarian preservation is feasible in patients diagnosed with stage IB–IIA SCC, whereas oophorectomy is recommended for those with adenocarcinoma [ 16 ]. Despite these guidelines, there has been a lack of studies investigating the clonal relationships between cervical cancer and concurrent ovarian/fallopian tube carcinoma through sequencing. Addressing this gap, our study engaged in a detailed genomic analysis of five patients with stage IB–IIB cervical carcinoma. By sequencing both cervical and ovarian lesions, we aimed to elucidate their clonal linkage and pathways of metastasis. This study seeks to bridge this gap by employing WES to analyze the clonal relationships within SEOC and SCOC cases. By elucidating the genomic signatures that underpin these synchronous cancers, we aim to refine diagnostic criteria, improve staging accuracy, and inform targeted therapeutic strategies, ultimately enhancing patient outcomes in gynecologic oncology.

This retrospective study was approved by the Institutional Review Boards of the Cancer Hospital, Chinese Academy of Medical Sciences (Beijing, PR China) with approval number 24/105‐4385. Formalin‐fixed paraffin‐embedded (FFPE) tissue blocks were chosen from patients who underwent hysterectomy and bilateral salpingooophorectomy between 2012 and 2021. Inclusion criteria were strictly followed to ensure a focused and relevant sample population: (1) pathologically diagnosed SEOC or SCOC; (2) sufficient tissue quality and quantity for WES; and (3) absence of uterine serosa involvement and direct tumor spread between the ovaries and the corpus, to omit cases of extensive invasion. We established a cohort comprising 54 SEOC patients and 7 SCOC patients (all Lynch syndrome‐negative). From these, we selected 12 SEOC and 5 SCOC cases through a stratified approach: (1) We excluded 42 low‐risk endometrioid carcinomas (redundant cases) that lacked papillary or mucinous features in both sites; (2) intentional enrichment for histologic diversity (including high‐grade, discordant, and rare subtypes); (3) proportional representation of FIGO 2023 stages [3 IA3 and 9 IIIA cases, closely approximating (though not perfectly matching) the 18:36 ratio in the full cohort]; and (4) prioritization of cases with complete clinical data. This selection strategy was validated by comparative analysis of the full cohort's clinicopathologic features. Notably, in cases SEOC‐12, SEOC‐14, and SCOC 3–5, the malignancies originated within the fallopian tubes, whereas the remaining cases were primarily ovarian. Final diagnoses were validated by two experienced pathologists (Lihong Li and Yan Song). Histopathological diagnosis, grade, and stage were reviewed and confirmed. Guided by hematoxylin and eosin (H&E) stained sections, we extracted a 1‐mm‐in‐diameter core from each tissue block. For cases SEOC‐8, SEOC‐14, and SCOC‐1 through SCOC‐5, which presented a substantial stromal component, we isolated tumor tissues using laser‐capture microdissection. Subsequently, these samples underwent WES. To detect somatic sequence mutations accurately, matched normal smooth muscle from seven cases and negative lymph nodes from 10 cases were also collected and analyzed through WES. DNA was extracted from FFPE specimens using the QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, CA, USA) as per the manufacturer's protocol. DNA quantification followed, utilizing Qubit 3.0 (Life Technologies, Carlsbad, CA, USA). Extracted genomic DNA from both tumor and matched normal tissues underwent library construction and capture with the SureSelectXT Target Enrichment System and Agilent Sure Select Human All Exome V6 Kit (Agilent Technologies, Santa Clara, CA, USA) by Annoroad Gene Technology (Beijing, PR China). Despite rigorous attempts, library preparation for SEOC‐2 and SEOC‐13 was unsuccessful. Pair‐end sequencing was achieved through the Illumina NovaSeq 6000 S4 (Illumina, San Diego, CA, USA), generating 150 base pair reads from each fragment end. Sequencing data, in FASTQ format, was aligned to the human reference genome (hg19) using the Burrows‐Wheeler aligner [ 21 ] (BWA, v0.7.15). The Genome Analysis Toolkit [ 22 ] (GATK, v3.6), Picard ( http://picard.sourceforge.net ; v2.7.1), and Samtools [ 23 ] (v1.3.1) were used for the basic processing and management of marking duplicates, including local realignments and score recalibration. Somatic mutations were detected by comparing tumors and matched normal samples' sequencing data using MuTect2 [ 24 ] for single‐nucleotide variants and InDels. Both non‐synonymous and synonymous mutations were reported. Putative driver mutations were determined based on previously listed cancer driver genes [ 25 ]. All candidate somatic mutations were validated by visual inspection using the integrated genome viewer [ 26 ]. Moreover, ANNOVAR [ 27 ] was used for the functional annotation of each variant in the coding regions. PolyPhen‐2 [ 28 ] was utilized to predict the impact of an amino acid substitution on protein function and structure. All the non‐synonymous mutations were applied to assess tumor mutation burden using the R package ‘maftools’ [ 29 ]. CNV kit was applied to detect CNVs [ 30 ]. To assess the clonal relationship between SEOC and SCOC tumors of a given case, we computed the ‘clonality index’ to quantitate the probability of two tumors sharing mutations not expected to have co‐occurred by chance. The clonality index (CI) is the likelihood of two tumors sharing mutations not expected to have co‐occurred by chance [ 7 , 8 , 9 , 31 ]. Both non‐synonymous and synonymous mutations were included in the clonality analysis. The CI was defined as: CI = 1 − ∏ k = 1 n f k , n > 0 0 , n = 0 The detailed calculation process has been extensively explained in prior literature [ 7 , 8 , 9 , 31 ]. In brief, f k is the proportion of EC for SEOC and CC for SCOC from TCGA harboring a given mutation and n is the number of mutations shared by two tumors. The mutation data from TCGA were sourced from the Pan‐Cancer Atlas. The CI was calculated separately for the 17 sets of tumors subjected to WES using all shared mutations. Tumors were deemed clonally related when the CI exceeded 0.995. Our analysis utilized the cancer cell fraction (CCF) to infer the directionality of tumor progression between primary and metastatic sites. This method hinges on the observation that subclonal mutations at primary sites often become clonal in metastatic lesions, thereby exhibiting a higher CCF in the latter [ 10 ]. Such changes provide insights into the clonal evolution and the pathways of metastasis in both SEOC and SCOC cases. The CCFs of somatic mutations were computed using PyClone [ 32 ] (v0.13.1). This methodology enabled a nuanced understanding of the complex dynamics of tumor spread. Immunostaining on tissue sections was conducted on a Leica bond‐max immunostainer. The staining process employed the following primary antibodies: mouse anti‐human MLH1 (ZM‐0154, ZSGB‐BIO), rabbit anti‐human MSH2 (ZA‐0622, ZSGB‐BIO), rabbit anti‐human MSH6 (ZA‐0541, ZSGB‐BIO), rabbit anti‐human PMS2 (ZA‐0542, ZSGB‐BIO). Data analysis was performed using R software (version 4.0.5). Group comparisons were executed through the independent t ‐test, paired t ‐test, and Mann–Whitney U test. The analysis of clinicopathological variables was conducted employing the chi‐square test. Statistical significance was determined at a p value <0.05. To estimate survival time, overall survival (OS) was employed, representing the mean duration of survival after the diagnosis or treatment of the disease. R package ‘sigminer’ were applied to estimate the mutational signature [ 33 ].

Figure S1. Histologic examination of DPT cases Figure S2. Histologic examination of all SCOC cases Figure S3. Comparison of gene copy number variations and mutational profiles in paired EC and OFC of DPT SEOC cases Figure S4. Prognostic stratification of SEOC by FIGO (2023) staging and clinicopathological features Figure S5. Proposed mechanism of trans‐tubal tumor spread Table S1. Sequence analysis summary of whole‐exome sequencing Table S2. Somatic mutation list Table S3. Driver mutation list Table S4. Clinicopathological characteristics and survival data of 54 SEOC cases, and additional clinicopathological data of 12 SEOC cases