Clinical validation of a DNA methylation biomarker to predict overall survival of relapsed ovarian cancer patients.

BackgroundAbout 70% of ovarian cancer (OC) patients relapse after initial chemotherapy, making it crucial to predict survival before second-line treatment. Our previous work discovered a blood-based DNA methylation prognostic signature (PLAT-M8) that uses 8 CpG sites related to chemoresistance. We aim to validate this biomarker and its correlation with clinicopathological features and treatment profiles in additional cohorts. MethodsExtracted DNA from whole blood was provided from the BriTROC 1 (n=47) and OV04 cohorts (n=57) upon the first relapse. Additional samples from Hammersmith Hospital (n=100) were collected during first-line chemotherapy (cycles 3-4 and 6). Bisulphite pyrosequencing was used to quantify DNA methylation at the previously identified 8 CpG sites. The methylation data obtained were combined with previous data from ScoTROC 1D and 1V (n=141) and OCTIPS (n=46). Cox regression was used to assess overall survival (OS) after relapse concerning clinicopathological characteristics. The DNA methylation Class (Class 1 vs 2) was determined by consensus clustering. FindingsBlood DNA methylation at relapse predicts better clinical outcomes. Methylation Class shows no association with outcome during first-line chemotherapy treatment. Methylation Class 1 is associated with shorter survival, as indicated by a meta-analysis of five cohorts (OS: HR 2.54, 1.67-3.85). Class 2 patients on carboplatin monotherapy have the best prognosis, while Class 1 patients on the same treatment have the poorest prognosis (OS: aHR 9.69, 2.38-39.47). Class 1 is linked to older patients (>75 years) with advanced-stage, platinum-resistant cases, correlating with residual disease, and shorter progression-free survival. In contrast, Class 2 of PLAT-M8 is linked to platinum-sensitive patients, and higher complete response rates by RECIST criteria, but shows no correlation with CA-125. These findings emphasise the potential of PLAT-M8 in guiding second-line chemotherapy decisions. InterpretationPLAT-M8 methylation biomarker is associated with survival in OC patients with relapse and hypothetically may predict platinum treatment response at second-line chemotherapy. FundingThis work was supported by funding from Ovarian Cancer Action ("Risk and Prevention" programme grant), Cancer Research UK programme grant (A13086) with support from the Cancer Research UK Imperial Centre, the National Institute for Health Research Imperial Biomedical Research Centre and the Ovarian Cancer Action Research Centre. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSThere is a strong association between platinum-based chemotherapy and DNA methylation changes in blood DNA during ovarian cancer relapse. Previous findings identified eight specific CpG methylation changes (known as PLAT-M8) in blood at relapse following platinum-based chemotherapy that were associated with overall survival in patients enrolled in the ScoTROC 1 trial and the OCTIPS cohort. Using an ovarian cancer cell line model, the study also showed that functional DNA mismatch repair increased the frequency of platinum-induced methylation, providing insights into the observed epigenetic changes. Added values of this studyOur current study validates in five large relapsed ovarian cancer cohorts that: (1) PLAT-M8 is associated with various clinicopathological characteristics, such as age, stage, platinum sensitivity, RECIST response, and progression time; (2) PLAT-M8, particularly from blood samples taken at the time of the first relapse before second-line chemotherapy, can serve not only as prognostic indicators for overall survival but also time to death after relapse in ovarian cancer patients; (3) PLAT-M8 does not have prognostic value when blood samples are taken during first-line chemotherapy before relapse, after initial diagnosis; and (4) PLAT-M8 may stratify overall survival and time to death after relapse based on the second-line treatment received by patients. These findings pave the way for our ongoing research, showcasing the potential of this non-invasive approach in predicting second-line treatment response, guiding decisions, and enhancing outcomes for relapsed ovarian cancer patients. Implications of all the available evidenceThe lack of biomarkers guiding treatment decisions during second-line therapy highlights the need for more reliable biomarkers. As a prognostic biomarker, PLAT-M8 is considered simple yet impactful, as it only requires one blood sample taken before second-line treatment at the time of relapse. The advantages of this research include developing personalised treatment approaches, minimizing side effects and wasted time from ineffective medications, reducing the likelihood of subsequent relapse episodes, and improving clinical outcomes for patients. Ultimately, the use of biomarkers has the potential to reduce hospital stays and healthcare costs by optimizing treatment effectiveness and efficiency, while also enhancing the quality of life for patients.