Skip to main content
NCBI home page
AACR Open Access logoLink to AACR Open Access
. 2024 Oct 17;30(24):5657–5665. doi: 10.1158/1078-0432.CCR-24-2105

Genomic Landscape of ctDNA and Real-World Outcomes in Advanced Endometrial Cancer

Pamela Soberanis Pina 1,#, Keelia Clemens 2,#, Adrian Bubie 2, Brooke Grant 1, Ginger Haynes 2, Nicole Zhang 2, Leylah Drusbosky 2,‡, Stephanie Lheureux 1,*,‡
PMCID: PMC11647206  PMID: 39417689

Abstract

Purpose:

ctDNA is a novel technique extensively studied in solid tumors, although not currently well defined in endometrial cancer.

Experimental Design:

A de-identified retrospective analysis of 1,988 patients with advanced/recurrent endometrial cancer was performed. In addition, an analysis of a real-world evidence cohort was completed (n = 1,266). Patients underwent ctDNA testing using Guardant360 during routine clinical care. The objective was to describe and assess molecular landscape using ctDNA.

Results:

Among 1,988 ctDNA samples, at least one somatic alteration was detected in 91.6% (n = 1,821). Most frequently altered genes were TP53 (64%), PIK3CA (29%), PTEN (25%), ARID1A (20%), and KRAS (14%). Overall, 18.5% had amplifications, with the majority identified in CCNE1 (40.9%), PIK3CA (22%), and EGFR (19.3%). From the real-world evidence cohort, those with TP53 mutations had a worse overall survival (OS) versus those without TP53 mutations (P = 0.02) and those with TP53 comutations had an inferior OS in comparison with TP53-mutated only (P = 0.016). Amongst these, patients with a PIK3CA comutation (P = 0.012) and CCNE1 amplification (P = 0.01) had an inferior OS compared with those with only TP53 mutations. Fifty-seven patients with newly diagnosed endometrial cancer had at least two serial ctDNA samples showing evolution in detected variants compared with baseline samples, with TP53 being the most frequent change.

Conclusions:

This study is one of the largest cohorts of ctDNA currently reported in endometrial cancer. The presence of TP53 mutation and other comutations detected by ctDNA have a negative effect on outcomes. This report suggests that ctDNA analysis is feasible and could become a useful biomarker for endometrial cancer.

Translational Relevance.

Liquid biopsy has emerged as a promising and complementary alternative to tissue biopsy to enable identification of genomic alterations in gynecologic cancers; nonetheless, it is not well defined in endometrial cancer. In this study, we present one of the largest retrospective cohorts to date describing the molecular landscape using ctDNA in patients with advanced or recurrent endometrial cancer and evaluating the impact on outcomes. Of the 1,988 ctDNA samples analyzed, at least one somatic alteration was detected in most patients (91.6%). The most frequently altered genes were TP53, PIK3CA, PTEN, and ARID1A. Patients with TP53 comutations had worse survival, especially those with a PIK3CA comutation or CCNE1 amplification within the real-world evidence cohort. This study showed that ctDNA analysis has the potential to accurately detect the genomic landscape and predict patient outcomes in endometrial cancer. Future research is warranted to demonstrate the prognostic value and clinical utility of ctDNA assays in endometrial cancer.

Introduction

Endometrial cancer is one of the most common cancer types among women worldwide (1). The endometrial cancer therapeutic landscape has evolved to personalized medicine by recognizing that endometrial cancer is a heterogeneous disease that harbors diverse histologic subtypes and variable prognoses (2, 3).

Historically, endometrial cancer has been broadly classified according to histologic subtypes (4). A deeper understanding of the endometrial cancer biology has provided insights into its genomic landscape and unique features according to each subtype (5). Endometrioid endometrial cancer is the most frequent type, with somatic mutations involving the PI3K-PTEN-AKT, RAS-RAF-MEK-ERK, and WNT pathways. Serous endometrial cancer is the second most common subtype, with TP53, PIK3CA, FBXW7, PPP2R1A, CCNE1, ARID1A, and PTEN being commonly altered. Less common histologic subtypes such as clear cell, mucinous, and neuroendocrine carcinomas have distinctive molecular landscapes (4, 5).

The genomic heterogeneity among endometrial cancer subtypes has been further elucidated by the integrated analysis of The Cancer Genome Atlas whereby an integrated molecular morphologic classification system was developed for a better comprehension of the endometrial cancer biology. Four molecular subgroups have been defined, including POLE-mutant, mismatch repair–deficient, copy number–low, and copy number–high. This classification is based on surrogate markers in archival paraffin-embedded tissues (6). Alternate algorithms have been proposed to implement in practice The Cancer Genome Atlas classification which is based on IHC (3). These molecular prognostic groups have provided the opportunity to improve risk stratification and potentially guide treatment decisions (6–8). As such, this classification has been included in the latest international guidelines for endometrial cancer management (6, 9).

Liquid biopsy has emerged as a minimally invasive and complementary alternative to tissue biopsy to enable detection of genomic alterations in gynecologic cancers, including endometrial cancer. ctDNA is a novel liquid biopsy technique extensively studied in solid tumors, although not currently well-defined in endometrial cancer (10). In a recent prospective study of cell-free DNA (cfDNA), patients with advanced endometrial cancer had significantly higher levels of cfDNA in plasma in comparison with those with stage I. A positive correlation was also observed between cfDNA concentrations and aggressive histologic subtypes (11). Efforts are underway to define the role of ctDNA for molecular monitoring, detection of progression, and possible impact on treatment approaches (12). Developing novel strategies to actualize personalized treatment and sequencing approaches in the recurrent setting of endometrial cancer by measuring ctDNA is a promising intervention (12, 13). The aim of this study was to describe the molecular landscape using a liquid biopsy assay in patients with advanced or recurrent endometrial cancer and assess the impact on patient outcomes.

Materials and Methods

Retrospective data analysis

We performed a de-identified retrospective analysis of patients with advanced newly diagnosed or recurrent endometrial cancer who underwent ctDNA testing using Guardant360 (Guardant Health) during routine clinical care between April 2019 and April 2023. Guardant Health is a Clinical Laboratory Improvement Amendments-certified, College of American Pathologists-accredited, New York State Department of Health-approved laboratory. Sample shipment, plasma isolation, and ctDNA extraction procedures were performed in accordance with previously described and validated methods (14). Ordering physicians provide patient age, sex, cancer type, and confirmation of advanced disease (stage IIIB or higher). Only patients with endometrial cancer diagnosis were included in the analysis. The timing of testing in relation to the patient’s diagnosis of endometrial cancer was indicated by the ordering provider as “newly diagnosed” or “progressing on therapy” (recurrent). For patients with more than one Guardant360 test, only the first test was included.

Guardant360 utilizes hybrid capture technology and next-generation sequencing to detect single-nucleotide variants (SNVs), indels, fusions, and copy-number variations (CNVs) in up to 83 genes (14). The reportable range of SNVs, indels, fusions, and CNVs is ≥0.06%, ≥0.04%, ≥0.06%, and ≥2.12 copies, respectively. Microsatelite instability–high (MSI-H) status is determined utilizing next-generation sequencing reads across 90 loci that are integrated into a MSI-H bioinformatics caller, which has been already described (15). A blood tumor mutational burden (bTMB) score is generated from a previously detailed computational algorithm examining the total number of synonymous and nonsynonymous SNVs and indels across a 1 Mb footprint (16).

Real-world study population, data source, and extraction

A retrospective analysis of real-world outcomes was performed using the GuardantINFORM database, including anonymized data from patients with advanced cancer who underwent genomic profiling using Guardant360 in the United States as part of clinical care between 2014 and 2023 and had at least one treatment claim after their first (index) Guardant360 test. The GuardantINFORM database includes genomic information paired with structured commercial payer claims data from inpatient and outpatient facilities in both academic and community settings. It does not include clinical features not coded as claims, such as tumor biomarkers assessed using other tests or response to anticancer therapy. Information on deaths is obtained from third parties and aggregated with the administrative claims data.

Real-world outcomes

To assess outcomes, real-world time to treatment discontinuation (rwTTD) and real-world time to next treatment (rwTTNT) were used as surrogates for progression free survival (PFS), expressed in months. rwTTD was defined as the time from the first day of therapy to the estimated last day of therapy. Patients were censored if the last claim date was before the end of therapy. rwTTNT was defined as the time from start of one line of therapy until initiation of a new line of therapy. Death was considered an event if the patient died before the end of therapy or within 90 days of the end of therapy or if the patient was lost to follow-up and died before the end of therapy. Patients were censored due to a lack of follow-up if the last claim activity date was prior to the start of the next line of therapy. Real-world overall survival (rwOS) was defined as the time from the ctDNA report date until death. Patients were censored at the last claim activity date.

Statistical analyses

Baseline characteristics included age, gender, and lines of therapy prior to the index Guardant360 test. Descriptive statistics were reported for the overall population as well as the genomic subgroups. Proportions were compared using the χ2 test, whereas continuous variables were compared using a two-sided t test, with significance defined as P < 0.05. FDR correction for multigene and variant prevalence comparisons across groups by the χ2 test were made using the Benjamini–Hochberg model (notated as “BH”).

For the comparison of mutational changes between patient serial tests, patient-normalized limit of detection (LoD) values were calculated based on the 0.1% mutation allele frequency (MAF) of the serial test with the lowest max mutation allele frequency value (17). Patient-specific LoD was used to filter variants compared across serial tests to account for detection stochasticity due to assay limitations and fluctuations in input ctDNA quantity.

Cox regression models were computed with packages in R [(v4.2.1; RRID: SCR_001905; survminer (RRID: SCR_021094); survival (RRID: SCR_021137); and ggplot2 (RRID: SCR_014601)]. Models adjusted for therapy lines (LOT) since the terms advanced disease and recurrent disease were used to assess differences in outcomes between patients with and without TP53 mutation and with or without CCNE1 amplification. HRs were reported for each comparison with 95% confidence intervals (95% CI).

Ethical statement

Guardant Health has obtained Institutional Review Board approval to generate de-identified datasets from Advarra Institutional Review Board. GuardantINFORM is a fully de-identified database that complies with sections 164.514 (a)-(b)1ii of U.S. Health Insurance Portability and Accountability Act, which defines the determination and documentation of statistically de-identified data. Given the nature of the de-identification of this dataset, informed consent was not obtained from patients analyzed by this study. The study was conducted in accordance with the Declaration of Helsinki.

Data availability

The data generated in this study are not publicly available to protect patient privacy. They are available upon reasonable request from the corresponding author.

Results

Patient characteristics

From April 2019 through April 2023, a total of 1,988 patients with advanced endometrial cancer underwent ctDNA testing and were included. The clinical characteristics of the entire cohort are summarized in Supplementary Table S1A. The median age was 68 (23–95) years. Nearly half of patients were reported as newly diagnosed with advanced endometrial cancer (45.2%; n = 898) and 43.3% (N = 860) had recurrent/progressive disease; staging was not available for 11.5% (n = 230).

Genomic landscape

Among the 1,988 ctDNA samples, at least one somatic alteration was detected in 91.6% of patients (n = 1,821); the average number of alterations detected per sample was 6.7, and the median maximum variant allele frequency was 2.6%.

Excluding variant of uncertain significance (VUS) and synonymous alterations, the most frequently altered genes were TP53 (64%), PIK3CA (29%), PTEN (25%), ARID1A (20%), and KRAS (14%; Figs 1A and 2A). Alterations in TP53 were enriched in patients with recurrent endometrial cancer (P < 0.0001) compared with newly diagnosed patients with endometrial cancer. There were no other significant differences between the two groups (Fig. 2B).

Figure 1.

Figure 1.

Open in a new tab

Endometrial cancer ctDNA genomic landscape. A, Most common genomic alterations identified in the overall cohort of patients with endometrial cancer. B, Most common genomic alterations identified in patients with endometrial cancer without MSI-H. Variants of uncertain significance were excluded.

Figure 2.

Figure 2.

Open in a new tab

Frequency of genomic alterations. A, In patients with advanced and recurrent endometrial cancer. The most commonly altered genes identified were TP53, PIK3CA, PTEN, ARID1A, and KRAS. Variants of uncertain significance were excluded. B, In newly diagnosed and recurrent endometrial cancer patient samples, alterations in TP53 were more frequently identified in recurrent samples (P < 0.001). C, Incidental pathogenic germline alterations. Sixty-two incidental germline alterations were identified in 61 patient samples, most commonly in BRCA1, BRCA2, and ATM. D, Frequency of gene amplifications in endometrial cancer patient samples. Amplifications were identified in 18.5% (337/1,821), with the majority identified in CCNE1, PIK3CA, EGFR, and ESR1.

When restricting to samples analyzed on the 83-gene panel (n = 672), alterations in homologous recombination repair (HRR) genes were identified in 24.3% (163/672), most frequently in ATM (49%, 80/163), CHEK2 (31%, 51/163), BRCA2 (17%, 28/163), and BRCA1 (14%, 23/163). Alterations in mismatch repair genes occurred in 6.3% (42/672), most commonly in MSH6 (71%, 30/42).

Overall, 11.9% of patients had MSI-H tumors (n = 237) detected on Guardant360. From the 553 patients with bTMB scores available, 12.3% of samples were bTMB-high, classified as a bTMB score above the 80th percentile (48.75 mut/Mb), and 87.7% were bTMB-low (18–20). All but three samples (95.6%) of bTMB-high samples also had MSI-H detected. MSI-H status and bTMB status were similar between patients with newly diagnosed and recurrent disease (Supplementary Fig. S1). As the present panel did not include POLE testing, a bTMB score of more than 100 mut/Mb was explored as a potential surrogate for this mutation (typically used to define ultramutation; ref. 21), and no patient was identified with a bTMB >100 mut/Mb.

In the cohort without MSI-H (88.1%), the most frequently altered genes were TP53 (66%), PIK3CA (25%), PTEN (17%), and ARID1A and KRAS (both 11%). Driver alterations in MMR genes (MLH1, MSH2, MSH6, and PMS2) in this cohort were rare (1%; Fig. 1B). Therefore, the genomic landscape was similar between the overall cohort and when restricting to samples without MSI-H.

A total of 62 incidental pathogenic germline alterations were detected in 61 patients (3%), most commonly in BRCA1 (34%), BRCA2 (21%), ATM (18%), and MSH6 (10%), with lower frequencies in FANCA, PALB2, MLH1, MSH2, and CHEK2 (Fig. 2C).

Amplifications were identified in 18.5% (337/1,821), with the majority identified in CCNE1 (40.9%), PIK3CA (22%), EGFR (19.3%), and ESR1 (16.2%; Fig. 2D). There were 64 patient samples (4%) with driver alterations in ERBB2. Of these, 54.7% (35) had an amplification, 42.2% (27) had a missense mutation, 3.1% (2) had both a missense mutation and amplification, and 1.6% (1) had an inframe mutation.

There were no differences in amplification frequencies between newly diagnosed or recurrent patients. Structural rearrangements and fusions were uncommon, observed in only five patients (0.3%).

TP53 comutations and coamplifications

Given the high frequency of TP53 alterations, we investigated the rates of TP53 comutations and coamplifications. Comutations in FBXW7 (P < 0.0001), PIK3CA (P < 0.001), CHEK2 (P = 0.0014), CDKN2A (P = 0.0028), CDK12 (P = 0.0035), RB1 (P = 0.0049), and MPL (P = 0.0123) were frequently identified in the TP53-mutated cohort. Mutations in CTNNB1 (P < 0.0001), PTEN (P < 0.0001), ARID1A (P < 0.0001), KRAS (P < 0.001), ESR1 (P = 0.0284), CCND1 (P = 0.0327), and EGFR (P = 0.0432) were more frequently detected in the cohort in which TP53 mutations were not detected (Fig. 3A). When assessing amplifications, CCNE1 amplifications were significantly enriched in patients with TP53 mutations (P < 0.0001); EGFR (P < 0.0001) and RAF1 (P = 0.0393) amplifications were significantly enriched in the cohort in which TP53 mutations were not detected (Fig. 3B).

Figure 3.

Figure 3.

Open in a new tab

Coalterations with TP53 mutations. A, Coalterations with other types of mutations. Variants of uncertain significance, fusions, and CNVs were excluded. Comutations in FBXW7, PIK3CA, and CHEK2 were frequently identified in the TP53-mutated cohort; meanwhile, comutations in CTNNB1, PTEN, ARID1A, and KRAS were more frequently in those without TP53 mutation. B, Coamplifications with TP53 mutations. CCNE1 amplifications were significantly enriched with TP53 mutations, and EGFR and RAF1 were significantly enriched in those without TP53 mutations.

Variant and patient response in the real-world evidence cohort

We utilized real-world data to investigate the correlation to patient outcomes. A total of 1,266 patients with endometrial cancer with reported variant and matched treatment outcomes were identified and included in this analysis. The median age of patients was 67 (23–85) years. Approximately a third of patients had newly diagnosed advanced endometrial cancer (37.8%, n = 479), 44.1% had recurrent disease (n = 558), and 18.1% had no diagnostic staging available (n = 229; Supplementary Table S1B).

TP53 was the most common gene mutated across the cohort, detected in 776 patients (61.3%, 776/1,266). Those with TP53 mutations detected by ctDNA had a worse rwOS in comparison with those without TP53 mutations [n = 479; 71.2 vs. 108 months; HR, 0.75 (95% CI, 0.59–0.96); P = 0.02; Fig. 4A]. There was no significant difference in rwTTNT [14.4 vs. 15.8 months; HR, 0.98 (95% CI, 0.79–1.22); P = 0.9] or rwTTD [4.1 vs. 4.1 months; HR, 0.98(95% CI, 0.87–1.1); P = 0.8] for patients with TP53 mutations versus those without TP53 mutations.

Figure 4.

Figure 4.

Open in a new tab

Kaplan–Meier curves representing OS. A,TP53-mutated vs. TP53 nonmutated, (B) TP53 comutated vs. TP53-only patients, (C) TP53-PIK3CA vs. TP53-only patients, and (D) CCNE1-amplified vs. CCNE1 nonamplified. Amp, amplified; Comut, comutated; mut, mutated; WT, wild-type.

Patients with TP53 comutations were analyzed, as defined by oncogenic and variants of possible clinical significance, including CNVs, in one or more genes besides TP53. Almost three quarters (74%, 574/776) of TP53-mutated patients had at least one comutation and were more often detected in patients with recurrent disease (55.9%, 321/574) than at the time of diagnosis (32%, 184/574; Supplementary Table S1B). Those with TP53 comutations had an inferior rwOS in comparison with only TP53 mutation [66.4 vs. 78.1 months; HR, 0.66 (95% CI, 0.47–0.92); P = 0.016; Fig. 4B]. Within comutations, the most frequently mutated genes were PIK3CA (38.7%, 222/574), PTEN (32.9%, 189/574), ARID1A (30.7%, 176/574), KRAS (12.2%, 70/574), and ATM (11.7%, 67/574; Supplementary Table S2). Amongst these, the most common mutations were PIK3CA H1047R (n = 33), ARID1A D1850 fs (n = 21), PIK3CA E542K (n = 20), PIK3CA E545K (n = 20), and KRAS G12D/V (n = 19, 17).

Of the 222 TP53 patients comutated with PIK3CA, nearly one quarter (23.8%, 53/222) also had an additional HRR-related gene comutation, whereas the majority (76.1%, 169/222) were not HRR-mutated. As observed with TP53 comutations generally, those patients with a PIK3CA comutation had inferior rwOS compared with those with only TP53 mutation [65.9 vs. 78.1 months; HR, 1.68 (95% CI, 1.12–2.51); P = 0.012; Fig. 4C; nine patients were excluded due to missing endpoint data], as well as expedited rwTTD [4.1 vs. 4.1 months; HR, 1.24 (95% CI, 1.0–1.51); P = 0.04] but not rwTTNT [not reached vs. 15.7 months; HR, 1.25 (95% CI, 0.85–1.83); P = 0.3].

CCNE1 amplifications were reported in 7.81% (99/1,266) of patients. Thirty nine CCNE1 amplifications (39.4%, 39/99) were detected in patients with newly diagnosed endometrial cancer, with 55 detected in patients with recurrent disease (55.6%, 55/99; Supplementary Table S1B). Survival analyses included 98 patients as one patient did not have data available for the survival endpoint. Of the 98 CCNE1-amplified patients included in the analysis, 93 (93.9%) were comutated with TP53. The presence of CCNE1 amplification was associated with significantly shorter rwOS [49.5 vs. 89.8 months; HR, 0.50 (95% CI, 0.36–0.71); P < 0.001; Fig. 4D]. The median rwTTNT was shorter for CCNE1-amplified patients compared with CCNE1 nonamplified patients [8.4 vs. 15.7 months; HR, 0.52 (95% CI, 0.38–0.720); P < 0.001]; however, rwTTD was observed to not significantly differ between patients [3.9 vs. 4.1 months; HR, 0.86 (95% CI, 0.7–1.06); P = 0.2].

Longitudinal collection and evolution in endometrial cancer

A total of 118 patients (9.3%, 118/1,266) had ≥2 serial ctDNA assays. Fifty-seven patients (48.3%, 57/118) were newly diagnosed with advanced disease at the time of their first serial test, presenting the opportunity to investigate variant evolution in these patients. For these 57 patients, the median time between serial tests was 229 days (mean = 301). A median of three somatic oncogenic mutations were detected per patient in the first serial ctDNA test (mean = 5.12), with a total of 292 mutations detected across the patients.

To interrogate whether mutational changes between patient serial testing are associated with recent treatment regimens, patients were striated as molecular response–like (receiving a serial test within 3 months of the start of a LOT) or tumor evolution–driven (serial test more than 3 months from LOT start). Twelve patients (21.1% 12/57) had a serial test defined within the molecular response–like window, with a median time between serial tests of 84 days (mean = 139). A total of 26 new mutations were detected at serial timepoint 2 in these patients (7 clonal and 19 subclonal), and 45 mutations detected at the first serial test were not detected at serial test two (16 clonal and 29 subclonal). TP53 mutations were the most commonly gained (n = 10) although exclusively subclonal in these patients. TP53 mutations were also the most commonly lost mutations (three clonal, four subclonal), followed closely by PIK3CA (five clonal; Supplementary Tables S3–S6).

Forty-five patients (78.9% 45/57) had serial testing reflecting mutational differences likely tumor evolution–driven, with a median time between serial tests of 268 days (mean = 347). From these patients, a total of 86 mutations were gained between serial timepoints 1 and 2 (29 clonal and 57 subclonal), and 162 mutations were lost between serial tests (74 clonal and 88 subclonal). TP53 subclonal mutations were the most commonly gained mutation (11 subclonal, 3 clonal), whereas ARID1A mutations were the most commonly lost mutation (10 clonal, 6 subclonal), along with TP53 clonal mutations (nine clonal, seven subclonal; Supplementary Table S3).

Discussion

This is one of the largest cohorts of ctDNA in patients with advanced and recurrent endometrial cancer demonstrating feasibility of detecting alterations in ctDNA. The mutation profiles elucidated in this patient population align with the known landscape of somatic tumor tissue alterations in previously published studies (5, 13). In this report, we described the most frequently altered genes in a cohort of 1,821 ctDNA samples, including TP53, PIK3CA, PTEN, ARID1A, and KRAS. In addition, RWE demonstrates that the presence of TP53 mutation ± other comutations (i.e., PIK3CA and CCNE1 amplification) in ctDNA can have a deleterious effect on patient outcomes, consistent with data from tumor tissue (22–24). ctDNA profiling can also reflect underlying endometrial cancer tumor heterogeneity (13).

Challenges in the clinical application of ctDNA have involved, in part, limitations in detecting a sufficient number of mutant DNA molecules from blood samples in endometrial cancer (12). Detection of ctDNA in endometrial cancer varies significantly between cohorts (18%–65%), and this is also impacted by the wide array of available sequencing techniques (10, 25). In this cohort, the detection rate of genomic alterations in ctDNA was high, with at least one somatic alteration detected in 91.6% of samples, similar to previous reports (13, 26). The high sensitivity in this cohort provides rationale to assess feasibility of screening patients for targeted therapies and/or evaluating therapy response. Indeed, assessment of molecular alterations within the tumor is often required for patient selection in the personalized medicine era but not always available. Thus, ctDNA analysis has been proposed as an alternative minimally invasive method in endometrial cancer that can potentially decrease the timeline for screening and assess tumor evolution (27). It has become routine clinical practice in other solid tumors such as non–small cell lung cancer, for which a modified TNM staging system has been proposed to incorporate ctDNA analysis (28).

Among the alterations detected by ctDNA in this cohort, TP53 was the most commonly mutated gene. The proportion of patients with TP53 mutations was slightly higher compared with other reports, likely due to patients’ stage and bias in patient recruitment (11, 29). Within the group of TP53-mutant endometrial cancer, co-ocurrence with somatic mutations in FBXW7 and PIK3CA were significantly more frequently found, whilst alterations in PTEN and ARID1A were commonly seen in the TP53 mutation not detected group. This is consistent with previous data from tissue biopsies; however, frequency may vary according to endometrial cancer histologic subtypes (22, 30). Additionally, rwOS was significantly worse in patients with a TP53 mutation (P = 0.02), consistent with prior studies showing that TP53 somatic mutations are associated with poor prognosis in endometrial cancer (31). The presence of TP53 mutations in endometrial cancer has been included in current molecular classification, and its presence requires escalation of treatment to improve outcomes (32, 33). The TP53 comutated group had an inferior rwOS compared with the group with only TP53 mutation (P = 0.016). Specifically, TP53 comutated with PIK3CA (P = 0.01) seems to have worse rwOS compared with other TP53 comutated genes. The H1047R mutation (exon 20) in PIK3CA was one of the most common mutations reported in the RWE cohort, and this has been previously correlated with high tumor grade and reduced relapse-free survival (23). Mutations in PIK3CA exon 9 have been associated with reduced survival as well, and these were also among the most common mutations (PIK3CA E542K and E545K; ref. 34). In addition, CCNE1-amplified tumors had a shorter rwOS (P < 0.001) and TP53 comutation with CCNE1 amplification (P = 0.012) also had worse OS in comparison with other TP53 comutated genes. CCNE1 amplification has been linked to poor prognosis and platinum resistance and is frequently seen in high-grade endometrial cancer. A high frequency of TP53 mutations in CCNE1-amplified tumors has been found in endometrial cancer (22). A relevant clinical question is whether liquid biopsy can identify development of chemotherapy resistance through subclonal/de novo CCNE1 alterations and guide treatment strategy. CCNE1 and PI3K pathways are two of the most frequently deregulated molecular pathways in endometrial cancer (24). Targeting the PI3K pathway is still under investigation, and new agents are being currently developed to target CCNE1 amplification. In addition, ERBB2 alterations represent a promising target for novel agents, and numerous correlative studies have demonstrated that the ERBB2 gene is amplified in 17% to 33% of high-grade endometrial cancer, including uterine serous carcinoma, carcinosarcoma, and a subset of high-grade endometrioid endometrial cancer (35). Of those, a large proportion of tumors exhibit ERBB2 protein overexpression (35–37). Within our cohort, 4% of ctDNA samples had driver alterations in ERBB2, with the most frequent alterations being amplifications and missense mutations. This rate might seem relatively low in comparison with the existing literature, yet several factors may impact the result (35, 38), including the unselected endometrial cancer population and potential limitation on CNV analysis due to variable DNA shed from the tumor (38, 39) and the potential challenge of calling CNVs via liquid biopsy (40).

The ability of ctDNA to detect tumor evolution through serial sample collections is an active area of interest in different solid tumors that could support the development of personalized patient management (25). In a study of longitudinal ctDNA samples of 25 patients treated with ICIs, including 12 patients with endometrial cancer, the ctDNA detection rate while on-treatment was 68%, and seven patients with increasing ctDNA in serial collections had progressive disease, with an average time of 2.5 months prior to imaging (41). Longitudinal measurements of ctDNA for treatment response proved to be effective in 16 patients with serous endometrial cancer and carcinosarcoma. The genomic profile of one patient identified in the primary tumor invariably changed after treatment, detected by ctDNA sampling (42). Interestingly, we performed an exploratory analysis of 57 patients with newly diagnosed endometrial cancer with longitudinal testing. TP53 mutations were most frequently lost or gained compared with the patient baseline sample at both clonal and subclonal levels. This was true among patients with serial testing subsequent to a new line of therapy and testing at a distal time from regimen start, suggesting consistent tumoral evolution in advanced disease and potential treatment pressure. However, given the lack of standard time windows for sample collection (ctDNA was collected as per physician request), no definitive conclusions can be drawn, requiring further evaluation in a prospective study with standardized sample collection.

Limitations

Despite the large number of patients, the main limitation of this dataset is the lack of extensive clinical information about ethnicity, histologic, and molecular subtypes and the lack of comparison with tumor molecular characteristics. We also recognize a bias in patient selection being advanced or recurrent stage, potentially explaining the higher proportion of TP53 mutations in this population. This study represents a POLE mutation-agnostic cohort, and this should be taken into account as well. Germline gene prevalence may be impacted by the genes tested by each patient (e.g., panel differences). This test is validated for somatic findings and, as such, germline findings should be interpreted with caution. This would also apply to somatic HRR gene rates. RWE has unique limitations given it uses claims-based data. Although these data have been shown to be reliable, they are surrogate and should be interpreted as such. Although descriptive in nature, the present study demonstrated the feasibility of genomic detection by using a clinically available liquid biopsy assay in endometrial cancer. Prospective studies to demonstrate the prognostic value and clinical utility of ctDNA assays in endometrial cancer are warranted.

In conclusion, this is the largest cohort of ctDNA collected in advanced endometrial cancer describing the landscape of somatic alterations detected by liquid biopsy. This study highlights the interest to incorporate ctDNA assessment in clinical trials for patient selection and biomarker validation.

Supplementary Material

Supplementary Data1

Supplementary Data

ccr-24-2105_supplementary_data1_suppds1.pdf (252.6KB, pdf)

Acknowledgments

We all agree to be responsible for each aspect of the job to ensure that questions related to the accuracy or completeness of any part of the job are properly investigated and resolved.

Footnotes

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

Authors’ Disclosures

K. Clemens reports being an employee and shareholder at Guardant Health. A. Bubie reports being an employee of Guardant Health, a liquid biopsy company focused on cancer detection. The technologies of Guardant Health were used to generate the data that were presented as part of this study. N. Zhang reports being a full-time employee and shareholder of Guardant Health. L. Drusbosky reports other support from Guardant Health during the conduct of the study. S. Lheureux reports fees from AstraZeneca, GSK, Repare Therapeutics, Merck, Eisai, Zai Lab, Gilead, Seagen, Abbvie and Schrodinger outside the submitted work. No disclosures were reported by the other authors.

Authors’ Contributions

P. Soberanis Pina: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. K. Clemens: Conceptualization, resources, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. A. Bubie: Resources, data curation, software, formal analysis, validation, visualization, methodology, writing–original draft, project administration, writing–review and editing. B. Grant: Data curation, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. G. Haynes: Conceptualization, data curation, software, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. N. Zhang: Data curation, software, formal analysis, methodology, writing–review and editing. L. Drusbosky: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. S. Lheureux: Conceptualization, resources, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing.

References

  • 1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7–33. [DOI] [PubMed] [Google Scholar]
  • 2. Tronconi F, Nero C, Giudice E, Salutari V, Musacchio L, Ricci C, et al. Advanced and recurrent endometrial cancer: state of the art and future perspectives. Crit Rev Oncol Hematol 2022;180:103851. [DOI] [PubMed] [Google Scholar]
  • 3. Jamieson A, Barroilhet LM, McAlpine JN. Molecular classification in endometrial cancer: opportunities for precision oncology in a changing landscape. Cancer 2022;128:2853–7. [DOI] [PubMed] [Google Scholar]
  • 4. Yen T-T, Wang T-L, Fader AN, Shih I-M, Gaillard S. Molecular classification and emerging targeted therapy in endometrial cancer. Int J Gynecol Pathol 2020;39:26–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Hong B, Le Gallo M, Bell DW. The mutational landscape of endometrial cancer. Curr Opin Genet Dev 2015;30:25–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Concin N, Matias-Guiu X, Vergote I, Cibula D, Mirza MR, Marnitz S, et al. ESGO/ESTRO/ESP guidelines for the management of patients with endometrial carcinoma. Int J Gynecol Cancer 2021;31:12–39. [DOI] [PubMed] [Google Scholar]
  • 7. Raffone A, Travaglino A, Mascolo M, Carbone L, Guida M, Insabato L, et al. TCGA molecular groups of endometrial cancer: pooled data about prognosis. Gynecol Oncol 2019;155:374–83. [DOI] [PubMed] [Google Scholar]
  • 8. Arciuolo D, Travaglino A, Raffone A, Raimondo D, Santoro A, Russo D, et al. TCGA molecular prognostic groups of endometrial carcinoma: current knowledge and future perspectives. Int J Mol Sci 2022;23:11684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. NCCN.org . Referenced with permission from the NCCN clinical practice guidelines in oncology (NCCN Guidelines®) for guideline uterine neoplasms V.2.2024. Plymouth Meeting (Pennsylvania): National Comprehensive Cancer Network, Inc.; 2024. [Google Scholar]
  • 10. Soberanis Pina P, Lheureux S. Novel molecular targets in endometrial cancer: mechanisms and perspectives for therapy. Biologics 2024;18:79–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Ashley CW, Selenica P, Patel J, Wu M, Nincevic J, Lakhman Y, et al. High-sensitivity mutation analysis of cell-free DNA for disease monitoring in endometrial cancer. Clin Cancer Res 2023;29:410–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Grant BM, Pugh TJ, Oza AM. Molecular monitoring in endometrial cancer-ready for prime time? Clin Cancer Res 2023;29:305–8. [DOI] [PubMed] [Google Scholar]
  • 13. Bolivar AM, Luthra R, Mehrotra M, Chen W, Barkoh BA, Hu P, et al. Targeted next-generation sequencing of endometrial cancer and matched circulating tumor DNA: identification of plasma-based, tumor-associated mutations in early stage patients. Mod Pathol 2019;32:405–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Odegaard JI, Vincent JJ, Mortimer S, Vowles JV, Ulrich BC, Banks KC, et al. Validation of a plasma-based comprehensive cancer genotyping assay utilizing orthogonal tissue- and plasma-based methodologies. Clin Cancer Res 2018;24:3539–49. [DOI] [PubMed] [Google Scholar]
  • 15. Willis J, Lefterova MI, Artyomenko A, Kasi PM, Nakamura Y, Mody K, et al. Validation of microsatellite instability detection using a comprehensive plasma-based genotyping panel. Clin Cancer Res 2019;25:7035–45. [DOI] [PubMed] [Google Scholar]
  • 16. Rizvi NA, Cho BC, Reinmuth N, Lee KH, Luft A, Ahn M-J, et al. Durvalumab with or without tremelimumab vs standard chemotherapy in first-line treatment of metastatic non-small cell lung cancer: the MYSTIC phase 3 randomized clinical trial. JAMA Oncol 2020;6:661–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Lanman RB, Mortimer SA, Zill OA, Sebisanovic D, Lopez R, Blau S, et al. Analytical and clinical validation of a digital sequencing panel for quantitative, highly accurate evaluation of cell-free circulating tumor DNA. PLoS One 2015;10:e0140712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Drusbosky L, Bilen MA, Azzi G, Barata PC, Boland PM, Bryce AH, et al. Blood-based tumor mutational burden from circulating tumor DNA (ctDNA) across advanced solid malignancies using a commercially available liquid biopsy assay. J Clin Oncol 2021;39(Suppl 15):3040. [Google Scholar]
  • 19. Chen EX, Jonker DJ, Loree JM, Kennecke HF, Berry SR, Couture F, et al. Effect of combined immune checkpoint inhibition vs best supportive care alone in patients with advanced colorectal cancer: the Canadian Cancer Trials Group CO.26 study. JAMA Oncol 2020;6:831–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Samstein RM, Lee C-H, Shoushtari AN, Hellmann MD, Shen R, Janjigian YY, et al. Tumor mutational load predicts survival after immunotherapy across multiple cancer types. Nat Genet 2019;51:202–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. León-Castillo A, Britton H, McConechy MK, McAlpine JN, Nout R, Kommoss S, et al. Interpretation of somatic POLE mutations in endometrial carcinoma. J Pathol 2020;250:323–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Momeni-Boroujeni A, Dahoud W, Vanderbilt CM, Chiang S, Murali R, Rios-Doria EV, et al. Clinicopathologic and genomic analysis of TP53-mutated endometrial carcinomas. Clin Cancer Res 2021;27:2613–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Garcia-Dios DA, Lambrechts D, Coenegrachts L, Vandenput I, Capoen A, Webb PM, et al. High-throughput interrogation of PIK3CA, PTEN, KRAS, FBXW7 and TP53 mutations in primary endometrial carcinoma. Gynecol Oncol 2013;128:327–34. [DOI] [PubMed] [Google Scholar]
  • 24. Cocco E, Lopez S, Black J, Bellone S, Bonazzoli E, Predolini F, et al. Dual CCNE1/PIK3CA targeting is synergistic in CCNE1-amplified/PIK3CA-mutated uterine serous carcinomas in vitro and in vivo. Br J Cancer 2016;115:303–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Moss EL, Gorsia DN, Collins A, Sandhu P, Foreman N, Gore A, et al. Utility of circulating tumor DNA for detection and monitoring of endometrial cancer recurrence and progression. Cancers (Basel) 2020;12:2231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Danziger N, Lin D, Tukachinsky H, Oxnard G, Elvin J. Landscape of genomic alterations and biomarkers in endometrial cancer (EC) identified by comprehensive genomic profiling (CGP) of peripheral blood biopsies (181). Gynecol Oncol 2022;166:S104. [Google Scholar]
  • 27. Shen Y, Shi R, Zhao R, Wang H. Clinical application of liquid biopsy in endometrial carcinoma. Med Oncol 2023;40:92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Herbreteau G, Vallée A, Charpentier S, Normanno N, Hofman P, Denis MG. Circulating free tumor DNA in non-small cell lung cancer (NSCLC): clinical application and future perspectives. J Thorac Dis 2019;11(Suppl 1):S113–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Kandoth C, Schultz N, Cherniack AD, Akbani R, Liu Y, Shen H, et al. Cancer Genome Atlas Research Network . Integrated genomic characterization of endometrial carcinoma. Nature 2013;497:67–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Schultheis AM, Martelotto LG, De Filippo MR, Piscuglio S, Ng CKY, Hussein YR, et al. TP53 mutational spectrum in endometrioid and serous endometrial cancers. Int J Gynecol Pathol 2016;35:289–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Muller PA, Vousden KH. p53 mutations in cancer. Nat Cell Biol 2013;15:2–8. [DOI] [PubMed] [Google Scholar]
  • 32. Morice P, Leary A, Creutzberg C, Abu-Rustum N, Darai E. Endometrial cancer. Lancet 2016;387:1094–108. [DOI] [PubMed] [Google Scholar]
  • 33. Mitamura T, Dong P, Ihira K, Kudo M, Watari H. Molecular-targeted therapies and precision medicine for endometrial cancer. Jpn J Clin Oncol 2019;49:108–20. [DOI] [PubMed] [Google Scholar]
  • 34. Mjos S, Werner HMJ, Birkeland E, Holst F, Berg A, Halle MK, et al. PIK3CA exon9 mutations associate with reduced survival, and are highly concordant between matching primary tumors and metastases in endometrial cancer. Sci Rep 2017;7:10240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Diver EJ, Foster R, Rueda BR, Growdon WB. The therapeutic challenge of targeting HER2 in endometrial cancer. Oncologist 2015;20:1058–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Groeneweg JW, Hernandez SF, Byron VF, DiGloria CM, Lopez H, Scialabba V, et al. Dual HER2 targeting impedes growth of HER2 gene-amplified uterine serous carcinoma xenografts. Clin Cancer Res 2014;20:6517–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Zhou J, Sukov WR, Carter JM, Schoolmeester JK. HER2 gene amplification and protein overexpression in uterine clear cell carcinoma and its implications in targeted immunotherapy. J Clin Oncol 2017;35(Suppl 7):29. [Google Scholar]
  • 38. Konecny GE, Santos L, Winterhoff B, Hatmal M, Keeney GL, Mariani A, et al. HER2 gene amplification and EGFR expression in a large cohort of surgically staged patients with nonendometrioid (type II) endometrial cancer. Br J Cancer 2009;100:89–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Husain H, Pavlick DC, Fendler BJ, Madison RW, Decker B, Gjoerup O, et al. Tumor fraction correlates with detection of actionable variants across > 23,000 circulating tumor DNA samples. JCO Precis Oncol 2022;6:e2200261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Tsui DCC, Drusbosky LM, Wienke S, Gao D, Bubie A, Barbacioru C, et al. Oncogene overlap analysis of circulating cell-free tumor DNA to explore the appropriate criteria for defining MET copy number-driven lung cancer. Clin Lung Cancer 2022;23:630–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Castaneda K, Saldivar JS, Chae YK, Sindhu H, Diaz-Arrastia C, Grzankowski K, et al. Immunotherapy response monitoring using personalized circulating tumor DNA analysis in patients with relapsed gynecologic malignancies. Int J Gynecol Cancer 2022;32(Suppl 2):A410. [Google Scholar]
  • 42. Bellone S, McNamara B, Mutlu L, Demirkiran C, Hartwich TMP, Harold J, et al. Monitoring treatment response, early recurrence, and survival in uterine serous carcinoma and carcinosarcoma patients using personalized circulating tumor DNA biomarkers. Int J Mol Sci 2023;24:8873. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data1

Supplementary Data

ccr-24-2105_supplementary_data1_suppds1.pdf (252.6KB, pdf)

Data Availability Statement

The data generated in this study are not publicly available to protect patient privacy. They are available upon reasonable request from the corresponding author.

Articles from Clinical Cancer Research are provided here courtesy of American Association for Cancer Research

RESOURCES