MSK-MET (Memorial Sloan Kettering - Metastatic Events and Tropisms) is a pan-cancer cohort of tumor genomic and clinical outcome data from 25,000 patients (12). The dataset identifies associations between tumor genomic alterations and patterns of metastatic dissemination across 50 tumor types; This study retrospectively included 94 Asian-far east/indian subcont patients diagnosed with endometrial cancer in MSK-MET. Except for two patients with missing age data, the remaining patients were divided into two groups based on age: young group (age ≤50 years, 19 patients) and elderly group (age>50 years, 73 patients). Clinical data, including age, histological subtype, metastatic site, mutation count, MSI, and TMB, were obtained from patient records.

The EC-3rd-SYSU (Endometrial cancer - Third Affiliated Hospital of Sun Yat-sen University) cohort, comprising 18 Chinese endometrial cancer patients retrospectively enrolled at the Third Affiliated Hospital of Sun Yat-sen University (2022-2024), underwent next-generation sequencing (NGS). Patients were excluded due to: unavailable matched normal tissue; low sequencing coverage (<100x); low tumor purity (evidenced by absence of somatic alterations, including silent mutations); age <18 years; multiple unique sequenced tumor types. The patients were divided into two groups based on age: young group (age ≤50 years, 10 patients) and elderly group (age>50 years, 8 patients). Clinical data, including age, histological subtype, metastatic site, mutation count, MSI, and TMB, were obtained from patient records.

The sequencing methodology and genomic analysis for the MSK-MET cohort are described in detail in the referenced publication (12). The genomic alterations were analyzed and download using cBioPortal (http://www.cbioportal.org).

For EC-3rd-SYSU cohort matched germline (blood) and tumor DNA samples were processed for targeted NGS using a customized gene panel (90 genes, Supplementary Table S1), which covered exonic regions, known hotspots, and selected intronic/promoter sequences. Sequencing was performed on an Illumina platform with depth-of-coverage thresholds set at ≥500x for tumor DNA and ≥100x for germline DNA to enable somatic variant calling.

Somatic Mutations: Somatic mutations were interpreted according to the Association for Molecular Pathology (AMP), the American Society of Clinical Oncology (ASCO), and the College of American Pathologists (CAP) guidelines. Only Tier I (strong clinical significance) and Tier II (potential clinical significance) variants are reported here. Somatic mutations were detected using MuTect2, and the variant calls were annotated using ANNOVAR. We focused on point mutations and small insertions/deletions (indels) in known cancer-related genes.

Pathway Enrichment Analysis: Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to identify biological pathways significantly altered in young vs. elderly patients.

Tumor Mutational Burden (TMB) and Microsatellite Instability (MSI): TMB was calculated as the number of somatic, coding, nonsynonymous mutations per megabase of the panel’s targeted region. A tumor mutational burden (TMB) of ≥10 mutations per megabase (mut/Mb) was classified as high. MSI status was assessed using the NGS data from the tumor and matched normal samples. The sample was classified as MSI-High(MSI-H) if MSIsensor score >10, MSI-Low (MSI-L)/Microsatellite Stable (MSS) if MSIsensor score ≤10 as previously described (12).

The data was analyzed using SPSS (version 19.0) and cBioPortal. Statistical comparisons between the two groups were performed using the Chi-square test for categorical variables and the Student’s t-test for continuous variables. Differentially expressed genes were considered significant if they had a false discovery rate (FDR) < 0.05. A p-value < 0.05 was considered statistically significant for all other tests. Data visualization was performed with PowerPoint and Sangerbox tools (13).

A total of 35,000 somatic mutations were identified across all samples. Samples from all 94 Asian patients were profiled by targeted sequencing, and all exhibited at least one genetic alteration. The most frequently mutated 5 genes were PTEN (55.3%; total number of mutations, n=83; number of samples with one or more mutations, n=52), PIK3CA (54.3%; n=78; n=51), ARID1A (46.8%; n=72; n=44), TP53 (45.7%; n=50; n=43), and PIK3R1 (29.8%; n=35; n=28). These were followed by genes ranked 6-15 — including CTNNB1 (24.5%), ZFHX3 (22.3%), KRAS (22.3%) and others — with mutation frequencies ranging from 24.5% to 16.0%.” (Figure 1A).

After excluding two patients with missing age information, all remaining patients were divided into two groups based on age: young group (age ≤50 years, 19 patients) and old group (age>50 years, 73 patients). The clinical characteristics of Asian endometrial cancer patients in the MSK-MET cohort were compared between younger (n = 19) and older (n = 73) groups. A significant age difference was observed, with means of 39.93 ± 5.95 years in the younger group and 62.67 ± 7.51years in the older group (p = 2.25E-11). The distribution of histological subtypes differed significantly between groups (p = 4.04E-03), with all younger patients presenting with endometrioid carcinoma, while the older group included cases of serous carcinoma (n = 17) and carcinosarcoma (n = 12). No significant difference in metastasis rate (p = 0.052) and mortality (p = 0.140) were observed (Table 1).

Among young patients, the most frequently mutated 5 genes were PTEN (84.21%; number of samples with one or more mutations, n=16), CTNNB1 (68.42%, n=13), PIK3CA (63.16%, n=12), ARID1A (52.63%, n=10) and FAT1 (31.58%, n=6) (Figure 1B; Supplementary Table S2). In elderly patients, mutations in genes such as TP53 (54.79%, number of samples with one or more mutations, n=40), PIK3CA (52.05%, n=38), PTEN (50.68%, n=37), ARID1A (45.21%, n=33) and PIK3R1 (28.77%, n=21) (Figure 1C) were more prevalent.

Statistical analysis revealed that the frequencies of CTNNB1 and PTEN mutations in the young group were significantly higher compared with that in the old group (68.42% vs 12.33%, p = 3.03e-6 and 84.21% vs 50.68% p = 2.709e-4, respectively), while TP53 alterations were potentially associated with old age (54.79% vs 15.79%, p = 1.156e-4) (Figure 2).

To investigate the biological implications of age-related genomic differences in EC, we analyzed the 15 most frequently mutated genes (Figures 1B, C), followed by KEGG pathway enrichment analysis for each group.

Enrichment analysis revealed significant terms across multiple categories in young group: for cellular processes, signaling pathways regulating pluripotency of stem cells and focal adhesion; for environmental information processing, phosphatidylinositol signaling system; for human diseases, endometrial cancer, central carbon metabolism in cancer, bacterial invasion of epithelial cells, EGFR tyrosine kinase inhibitor resistance, Human T-cell leukemia virus 1 infection, PD-L1 expression and PD-1 checkpoint pathway in cancer and Insulin resistance; and for organismal systems, neurotrophin signaling pathway, thyroid hormone signaling pathway, inflammatory mediator regulation of TRP channels and Leukocyte transendothelial migration (Figure 3A).

In old group, cellular processes were enriched for signaling pathways regulating pluripotency of stem cells and cellular senescence. Environmental information processing were enriched for PI3K-Akt signaling pathway and sphingolipid signaling pathway. Human diseases featured central carbon metabolism in cancer, EGFR tyrosine kinase inhibitor resistance, endometrial cancer, PD-L1 expression and PD-1 checkpoint pathway in cancer, endocrine resistance, human papillomavirus infection, human T-cell leukemia virus 1 infection and MicroRNAs in cancer. These processes were anchored in organismal systems including thyroid hormone signaling pathway, B cell receptor signaling pathway and longevity regulating pathway (Figure 3B).

Analysis of Tumor Mutation Burden (TMB) revealed no statistically significant difference between the young and old patient cohorts (p = 0.324, Figure 4A). Descriptive statistics further illustrated this similarity. The mean TMB was nearly identical at 30.00 mutations/Mb (SD = ± 49.65) in the young group compared to 31.86 mutations/Mb (SD = ± 90.27) in the old group. Notably, the median TMB was precisely the same at 6.05 mutations/Mb for both cohorts. However, the elderly group exhibited a substantially wider range (1.73 - 549.95 mutations/Mb) and a larger standard deviation, suggesting greater heterogeneity in TMB values compared to the younger group (range: 2.59 - 218.77 mutations/Mb).

We next compared the distribution of microsatellite instability (MSI) status between the two age groups. In the young group, 3 patients (15.8%) were classified as MSI-H, compared to 16 (84.2%) as MSI-L/MSS. Similarly, the old group comprised 10 MSI-H patients (13.7%) and 63 MSI-L/MSS patients (84.9%), and 1 missing data (1.37%). Consistent with this distribution, statistical analysis confirmed that the difference in MSI scores between younger and elderly patients was not significant (p = 0.641, Figure 4B).

Kaplan-Meier survival analysis revealed a pronounced age-dependent disparity in long-term clinical outcomes among endometrial carcinoma (EC) patients. Younger patients demonstrated significantly superior overall survival compared to elderly counterparts, with5-year survival rates of 76.92% versus 36.38%, respectively HR: 0.353(95% CI:0.130-0.985, p = 0.136) (Figure 5). The substantial survival difference observed between the young and elderly EC patients may reflect a true clinical disparity, though the lack of statistical significance, likely due to the limited sample size, warrants further investigation with larger cohorts.

Given this context, we sought to determine if these observations could be generalized to a specific patient population. Consequently, we conducted an identical analysis within our internal cohort, the EC-3rd-SYSU dataset (n=18), which comprises genomic data from Chinese endometrial carcinoma patients.

The most frequently mutated 5 genes were PTEN (88.89%; total number of mutations, n=23; number of samples with one or more mutations, n=16), PIK3CA (72.22%; n=15; n=13), ARID1A (66.67%; n=23; n=12), KRAS (33.33%; n=6; n=6), and FGFR2 (22.22%; n=5; n=4). These were followed by genes ranked 6-15 — including TP53 (16.67%), MSH6 (16.67%), CTNNB1 (16.67%), RNF43 (16.67%) and others — with mutation frequencies ranging from 16.67% to 11.11%.” (Figure 6A).

A comparison of the genomic profiles from both cohorts confirms PTEN, PIK3CA and ARID1A as the most frequently mutated genes in endometrial carcinoma, consistent across populations. However, the EC-3rd-SYSU cohort showed notably higher PTEN mutation rates (88.89%) and greater prominence of KRAS and FGFR2 alterations, while TP53 and PIK3R1 mutations were more common in the MSK-MET cohort. These differences suggest potential population-specific molecular features that could inform targeted therapy strategies.

The patients were divided into two groups based on age: young group (age ≤50 years, 10 patients) and old group (age>50 years, 8 patients). In the EC-3rd-SYSU cohort, younger (n = 10) and older (n = 8) endometrial cancer patients differed significantly in mean age (43.80 ± 6.37 vs. 56.75 ± 2.12 years, p = 5.04E-05). However, no significant differences were detected in histological subtype (p = 0.358), pathological grading (p = 0.656), metastasis (p = 0.637), FIGO stage (p = 0.638), molecular subtype (p = 0.456), or MSI type (p = 0.410) (Table 2).

Among young patients, the most frequently mutated 5 genes were PTEN (80.00%, n=8), PIK3CA (70.00%, n=8), ARID1A (50.00%, n=5), FGFR2 (30.00%, n=3) and CTNNB1 (30.00%, n=3), (Figure 6B). In elderly patients, mutations in genes such as PTEN (100.00%, number of samples with one or more mutations, n=8), ARID1A (87.50%, n=7), PIK3CA (75.00%, n=6), KRAS (50.00%, n=4), and MSH6 (37.5%, n=3) (Figure 6C; Supplementary Table S3) were more prevalent.

Statistical analysis revealed that the frequencies of CTNNB1, FGFE2 and ERBB3 mutations in the young group were higher compared with that in the old group (30.00% vs 0%, 30.00% vs 12.50% and 20.00% vs 0%, respectively), while ARID1A, KRAS, MSH6, RNF43 alterations were associated with old age (87.50% vs 50.00%, 50.00% vs 20.00%, 37.5% vs 0% and 37.5% vs 0%), (Figure 7). However, the small sample size resulted in low statistical power, limiting our ability to detect significant differences.

Comparing the age-related mutation spectra across two cohorts reveals consistent biological patterns. CTNNB1 mutations consistently show age-specific distributions, while PTEN maintains high frequency in younger patients across both populations. Notably, TP53 enrichment in older MSK-MET patients finds parallel in our cohort with ARID1A, KRAS and MSH6 showing similar age-associated trends. These conserved patterns suggest fundamental age-related molecular mechanisms in endometrial carcinogenesis that transcend ethnic boundaries.

To investigate the biological implications of age-related genomic differences in EC, we analyzed the 15 most frequently mutated genes (Figures 6B, C), followed by KEGG pathway enrichment analysis for each group.

Enrichment analysis revealed significant terms across multiple categories in young group: for cellular processes, cellular senescence and signaling pathways regulating pluripotency of stem cells; for environmental information processing, PI3K-Akt signaling pathway, sphingolipid signaling pathway and mTOR signaling pathway; for human diseases, endometrial cancer, central carbon metabolism in cancer, microRNAs in cancer, EGFR tyrosine kinase inhibitor resistance, human papillomavirus infection, proteoglycans in cancer, pathways in cancer and human cytomegalovirus infection; and for organismal systems, longevity regulating pathway and thyroid hormone signaling pathway (Figure 8A).

In old group, cellular processes were enriched for cellular senescence. Environmental information processing was enriched for PI3K-Akt signaling pathway and sphingolipid signaling pathway; Human diseases featured central carbon metabolism in cancer, EGFR tyrosine kinase inhibitor resistance, endometrial cancer, endocrine resistance, pathways in cancer, platinum drug resistance, microRNAs in cancer, PD-L1 expression and PD-1 checkpoint pathway in cancer and human T-cell leukemia virus 1 infection. These processes were anchored in organismal systems including thyroid hormone signaling pathway, prolactin signaling pathway and longevity regulating pathway (Figure 8B).

Pathway enrich analysis of the two cohorts revealed conserved age-related pathway distinctions. Younger patients consistently exhibited enrichment in pathways governing developmental plasticity (e.g., stem cell pluripotency), whereas older patients showed preferential involvement of therapy resistance pathways (e.g., endocrine resistance, platinum drug resistance). A key inter-cohort difference was the specific activation of PI3K-Akt/mTOR signaling in younger patients of the EC-3rd-SYSU cohort, suggesting potential population-specific oncogenic mechanisms in this age group.

Analysis of Tumor Mutation Burden (TMB) revealed no statistically significant difference between the young and old patient cohorts (p = 0.741, Figure 9A). Descriptive statistics further illustrated this similarity. The mean TMB was nearly identical at 17.02 ± 21.55 mutations/Mb in the young group compared to 20.05 ± 15.18 mutations/Mb in the old group.

We next compared the distribution of microsatellite instability (MSI) status between the two age groups. In the young group, 2 patients (20.0%) were classified as MSI-H, compared to 8 (80.0%) as MSI-L/MSS. Similarly, the old group comprised 3 MSI-H patients (37.5%) and 5 MSI-L/MSS patients (62.5%). Consistent with this distribution, statistical analysis confirmed that the difference in MSI scores between younger and elderly patients was not significant (p = 0.608, Figure 9B).