Molecular Classification of the PORTEC-3 Trial for High-Risk Endometrial Cancer: Impact on Prognosis and Benefit From Adjuvant Therapy

PURPOSE The randomized Adjuvant Chemoradiotherapy Versus Radiotherapy Alone in Women With High-Risk Endometrial Cancer (PORTEC-3) trial investigated the benefit of combined adjuvant chemotherapy and radiotherapy (CTRT) versus radiotherapy alone (RT) for women with high-risk endometrial cancer (EC). Because The Cancer Genome Atlas defined an EC molecular classification with strong prognostic value, we investigated prognosis and impact of chemotherapy for each molecular subgroup using tissue samples from PORTEC-3 trial participants. METHODS Paraffin-embedded tissues of 423 consenting patients were collected. Immunohistochemistry for p53 and mismatch repair (MMR) proteins, and DNA sequencing for POLE exonuclease domain were done to classify tumors as p53 abnormal (p53abn), POLE-ultramutated (POLEmut), MMR-deficient (MMRd), or no specific molecular profile (NSMP). The primary end point was recurrence-free survival (RFS). Kaplan-Meier method, log-rank test, and Cox model were used for analysis. RESULTS Molecular analysis was successful in 410 high-risk EC (97%), identifying the 4 subgroups: p53abn EC (n = 93; 23%), POLEmut (n = 51; 12%), MMRd (n = 137; 33%), and NSMP (n = 129; 32%). Five-year RFS was 48% for patients with p53abn EC, 98% for POLEmut EC, 72% for MMRd EC, and 74% for NSMP EC (P < .001). The 5-year RFS with CTRT versus RT for p53abn EC was 59% versus 36% (P = .019); 100% versus 97% for patients with POLEmut EC (P = .637); 68% versus 76% (P = .428) for MMRd EC; and 80% versus 68% (P = .243) for NSMP EC. CONCLUSION Molecular classification has strong prognostic value in high-risk EC, with significantly improved RFS with adjuvant CTRT for p53abn tumors, regardless of histologic type. Patients with POLEmut EC had an excellent RFS in both trial arms. EC molecular classification should be incorporated in the risk stratification of these patients as well as in future trials to target specific subgroups of patients.


INTRODUCTION
The endometrial cancer (EC) molecular classification introduced by The Cancer Genome Atlas 1 has initiated a transition toward molecular-based classification with clear prognostic value and thus a potential impact on the clinical care of patients with EC. The significant prognostic differences among the 4 molecular subgroups have been replicated using surrogate markers in formalin-fixed, paraffin-embedded (FFPE) tissues, identifying analogous subgroups: p53-abnormal (p53abn), POLE-ultramutated (POLEmut), mismatch repair-deficient (MMRd), and no specific molecular profile (NSMP) EC. The integration of the molecular classification with clinicopathological features has resulted in improved prognostic accuracy in intermediate-risk EC as well as unselected cohorts, [2][3][4][5] highlighting the potential of the molecular classification to refine and further individualize patients' risk stratification.
Although patients with EC generally have a good prognosis, 15%-20% have high-risk disease with increased incidence of distant metastases and cancerrelated death. Characteristics defining high-risk EC are high-grade disease, advanced stage, and/or nonendometrioid histology. Adjuvant pelvic external beam radiotherapy (EBRT) is a standard of care for patients with high-risk EC. 6 The randomized Adjuvant Chemoradiotherapy Versus Radiotherapy Alone in Women With High-Risk Endometrial Cancer (PORTEC-3) trial investigated the benefit of combined adjuvant chemotherapy and EBRT (CTRT) versus EBRT alone (RT) in patients with high-risk EC. 7 PORTEC-3 showed a significant benefit in both overall survival (OS) and failure-free survival (FFS) with CTRT, although the absolute benefit was limited (5% for 5-year OS and 7% for 5-year FFS), and significantly more adverse events occurred with CTRT. The greatest benefit was observed in serous cancers and stage III disease. 8 However, because there is substantial interobserver variability in assessment of pathologic factors that define high-risk, especially in high-grade EC, 9,10 it remains a challenge to identify patients who will benefit from chemotherapy. In this context, the molecular classification might help determine appropriate adjuvant treatment.
Using tissue samples donated by PORTEC-3 clinical trial participants, we investigated the prognostic relevance of the molecular classification and the relationship between the molecular subgroups and benefit from adjuvant CTRT in patients with high-risk EC.

Patient Selection and Study Design
FFPE tissue was collected from 423 consenting patients from 5 of the 6 clinical trial groups participating in the PORTEC-3 clinical trial. 7 The design and results of the PORTEC-3 trial have been reported previously 7 11 Abdominal or laparoscopic hysterectomy with bilateral salpingooophorectomy was performed on all patients. Patients were randomly allocated 1:1 to EBRT alone or CTRT (2 cycles of cisplatin followed by 4 cycles of carboplatin and paclitaxel). The study was approved by the Dutch Cancer Society and ethics committees of participating groups.
DNA isolation was performed as described previously. 12 POLE mutational status was assessed by next-generation sequencing using the AmpliSeq Cancer Hotspot Panel, version 5 (Thermo Fisher Scientific, Waltham, MA), a panel including POLE exonuclease domain and TP53, as described in the Data Supplement. If sequencing with the panel failed, KASPar competitive allele-specific polymerase chain reaction (LGC Genomics, Berlin, Germany) assays were used to screen for POLE hotspot variants at codons 286, 297, 411, 456, and 459 (primer sequences are available upon request). POLE exonuclease domain mutations (EDMs) were considered pathogenic (in this context, mutations causative of ultramutation) following previously defined criteria. 13 Sequencing and immunohistochemistry results were evaluated blinded for patient outcome.
If $ 1 of the molecular features (p53, MMR, and/or POLE status) could not be determined (eg, not enough tumoral material or testing failed), and thus the molecular subgroup could not be determined, the case was classified as EC, not otherwise specified 14 and excluded from the study (n 5 13). Tumors with . 1 classifying feature (multiple-classifier EC) were allocated in 1 of the 4 molecular subgroups, as described previously. 13,15 Briefly, ECs with a pathogenic POLE EDM with p53 abnormal expression and/or MMR protein loss were classified as POLEmut EC, whereas ECs with loss of any MMR protein or MSI-high with a p53 mutant staining pattern were classified as MMRd EC (Data Supplement). 13,15 Statistical Analysis The primary end point was RFS, defined as time from randomization to date of first relapse or death, whichever occurred first. The secondary end point was OS, defined as the time from randomization to date of death of any cause.

Key Objective
To determine, using tissue samples from the PORTEC-3 clinical trial, the prognostic value of the endometrial cancer (EC) molecular classification in high-risk EC and the possible benefit from chemotherapy within each molecular subgroup.

Knowledge Generated
The molecular classification has a strong prognostic value in high-risk EC. Additionally, the molecular classification may guide adjuvant treatment decisions for patients with high-risk EC and supports adjuvant chemotherapy for p53abn ECs. Relevance This study shows that incorporation of the molecular classification into risk stratification systems is essential and future clinical trials should address specific molecular subgroups of EC. Adjuvant treatment decisions based on the molecular classification are supported; specifically, patients with p53abn EC should be considered for adjuvant chemoradiotherapy, whereas for those with POLEmut cancers, de-escalation of adjuvant treatment should be considered.
Statistical analysis was performed using SPSS statistics, version 25 (IBM, Armonk, NY) and R, version 3.6.1). Clinicopathological characteristics of patients included in this study and the complete PORTEC-3 cohort, as well as between molecular subgroups, were compared with a x 2 test for categorical variables or 1-way ANOVA for continuous variables. All analyses were based on intention to treat. Analysis was also done by treatment received, without relevant differences (data not shown).
Differences in RFS and OS between molecular subgroups were tested with log-rank test and Cox regression analysis, which included molecular subgroups, age (as a continuous variable), histology (EEC grade 1-2 v EEC grade 3 or mixed v nonendometrioid serous carcinoma, clear-cell carcinoma, or other), stage (stage I-II v III), and adjuvant treatment received (CTRT v RT). To identify differences in treatment effect between molecular subgroups, Cox regression analysis was performed with treatment received, molecular subgroup, and their interaction. Similarly, Cox regression with treatment received, stage, and their interaction was performed within each molecular subgroup to analyze treatment effect per stage. Median follow-up time was calculated using the reverse Kaplan-Meier method. Reported P values were based on 2-sided tests, with P , .05 considered statistically significant.

RESULTS
A total of 423 EC samples from PORTEC-3 participants were collected and available for molecular analysis. Molecular testing was successful for 410 tumors (97%; Fig 1). Patient and tumor characteristics of these 410 ECs were comparable to the trial population not included in this study (Data Supplement). Median follow-up was 6.1 years (range, 0.52-11.03 years).
The 410 ECs were classified in 1 of the 4 molecular subgroups (Data Supplement):  Fig 2). All 30 patients with a multiple-classifier EC were alive without recurrence of their cancer at the time of data analysis (median follow-up, 5.3 years).
The prognostic value of the molecular EC classification was evaluated in univariable and multivariable analysis (  Table 2).
We analyzed the differences in adjuvant treatment effect (CTRT v RT) among the molecular subgroups (Fig 3).  Fig 3). Test for interaction between molecular subgroups and treatment arm did not reach significance (RFS: P 5 .072; OS: P 5 .113). In exploratory subanalyses, the benefit from CTRT remained significant for RFS in early-stage p53abn EC but lost significance in stage III disease; however, numbers were small and testing for interaction between stage and adjuvant treatment in p53abn EC was not significant (Data Supplement stage, grade, histologic type, LVSI, and age. Challenges with the current system include the lack of interobserver agreement in the evaluation of pathologic features, especially on histologic type in high-grade EC, where discrepancies reach 36% of cases examined. 9 In contrast, the molecular EC classification is a highly reproducible system with strong prognostic value. The assessment of MMR proteins and p53 immunohistochemistry are highly concordant with MSI and TP53 mutational status, respectively, 16,17 and the interobserver agreement is . 95%. 17  shown strong prognostic value of the molecular classification even in high-risk and advanced-stage EC and regardless of histologic subtype, indicating that this classification should be incorporated in standard clinical diagnostics, treatment decisions, and future studies. This study showed that patients with p53abn EC have a highly significant benefit from CTRT. These results are consistent with the PORTEC-3 clinical trial analysis, in which patients with serous cancers had a greater absolute benefit from CTRT. 8 Indeed, 71% of serous cancers in the current study were classified as p53abn EC. However, the serous cancers that were molecularly classified as POLEmut (n 5 6) and MMRd EC (n 5 7) had no recurrences and patients were still alive at time of analysis. Only 2 of the 6 patients with a serous NSMP EC had a recurrence and ultimately died of their cancer. Additionally, 47 of 93 p53abn ECs (51%) had nonserous histology (23% EEC grade 3; 13% clear-cell cancers; 7% mixed; 4% grade 1 EEC; and 4% other histologies), and there was no difference in clinical outcome between this group and serous p53abn ECs. These results show that the molecular EC classification identifies a broader group of patients with a higher specificity who might benefit from CTRT compared with traditional histotyping.
To further improve the survival of patients with p53abn EC, research may be directed at the addition of targeted adjuvant treatment to CTRT. Recent studies have shown homologous recombination deficiency in p53abn EC, 20,21 suggesting that these patients could benefit from the addition of poly(ADP-ribose) polymerase inhibitors to CTRT. Alternatively, for patients with p53abn EC and HER2/neu amplification (reported in 20%-25% of serous cancers 22 ), trastuzumab (and possibly pertuzumab) in combination with chemotherapy is promising, as indicated by results of a recent phase II clinical trial. 23 This analysis confirms the excellent survival of patients with POLEmut EC even in those with advanced-stage and nonendometrioid histologies, with no differences between adjuvant treatment received. This favorable clinical outcome is thought to be the result of the patients' enhanced T-cell response due to their high mutational burden. 24 Previous studies have described small groups of patients with POLEmut EC who received no adjuvant treatment and had no recurrences. 25,26 Additionally, no increased sensitivity of POLE-mutant, mouse-derived embryonic stem cells to radiotherapy was found, nor to a selection of chemotherapeutics, including cisplatin. 25 Together with our results, these data support the hypothesis that the excellent prognosis of patients with POLEmut EC is independent of adjuvant treatment. By implementing POLE testing in routine diagnostics, overtreatment of a substantial group of patients would be avoided, with clear impact on the patients' quality of life. 27 The lack of benefit observed from the addition of chemotherapy to RT in patients with MMRd EC suggests a favorable outcome with EBRT alone in stage I-II disease, as also found in the GOG-249 trial. 28 Furthermore, a recent study reported benefit of early-stage MMRd EC from RT compared with no adjuvant treatment. 29 Adjuvant treatment regimens other than chemotherapy should be explored in higher-stage MMRd ECs. Recently, immune checkpoint inhibition has been shown effective against  MMRd solid cancers, which has led to US Food and Drug Administration approval of pembrolizumab for MSI cancers. Phase II clinical trials have shown high response rates to immune checkpoint inhibitors in advanced stage EC, 30 which might be increased by the combination of these therapeutic agents with RT. 31 The analysis of NSMP ECs showed only a trend toward benefit from adjuvant CTRT. Still, the HRs of CTRT versus RT for NSMP cancers were similar to those observed in the PORTEC-3 trial. 8 The lack of significant effect of CTRT may be due to the relatively small numbers and molecular heterogeneity that characterizes NSMP EC. 1 Additional studies directed at refinement of NSMP EC, such as characterization of CTNNB1 exon-3 mutations, the potential fifth molecular subgroup, 3,32 may elucidate which patients with NSMP EC might benefit from intensified adjuvant treatment.
PORTEC-3 analysis showed a greater benefit of added CT for patients with stage III EC and those with serous cancers. 8 The current study provides insight into the biologic origins of this observation, identifying differences between the molecular subgroups. The high benefit of added CT observed across patients with p53abn EC in all stages (RFS: HR, 0.52; P 5 .022) corresponds partly to the benefit seen in serous cancers. The small numbers available per stage suggest that larger cohorts may identify significant differences in stage III p53abn EC (a nonsignificant benefit from CTRT was seen in this group of patients; RFS: HR, 0.58, P 5 .172), as well as in stage III NSMP cancers (RFS: HR, 0.68, P 5 .246). We observed no significant benefit from  CTRT for patients with MMRd EC, even after analysis by stage, indicating that this molecular subgroup may not be contributing to the stage III results observed in PORTEC-3. 8 Finally, although data were limited, patients with stage III POLEmut EC appeared to have a good clinical outcome in both treatment arms: only 1 of 12 patients randomly assigned to the RT arm had a recurrence and died of EC. Prospective observational studies will help define adequate de-escalation of adjuvant treatment of POLEmut EC, especially for stage III cancers.
Our study has limitations. Although it was a predefined translational research analysis within the context of the PORTEC-3 clinical trial, it was not originally powered for analysis by molecular subgroup. Additionally, we were able to obtain molecular results of 62% of the trial population. Nonetheless, the characteristics of the included cases were comparable to those of the excluded trial population.
In conclusion, our study shows the strong prognostic information the EC molecular classification carries, as well as its great potential to guide adjuvant treatment. It is essential to implement the molecular EC classification in clinical diagnostics and decision-making. Patients with p53abn EC may be considered for adjuvant treatment including chemotherapy, whereas adjuvant treatment de-escalation should be considered for those with POLEmut EC; additional studies are needed especially for MMRd and NSMP EC. Future clinical trials should include molecular subgroups in their design and study specific targeted adjuvant treatments.
tissue material, as well as the technicians T. Rutten, N.T. ter Haar, M. Osse, and E.J. Dreef at Leiden University Medical Center for excellent technical support.