DNA methylation-based biomarkers and prediction models for the survival of patients with colorectal cancer: systematic review and external validation study

Abstract Background DNA methylation biomarkers (e.g., methylation level at CpG sites) have the potential to improve prognostic accuracy for patients with colorectal cancer (CRC). We identified existing DNA methylation-based prognostic biomarkers and prediction models for CRC prognosis and validated them in a large external cohort. Methods Epigenome-wide studies reporting CpGs associated with survival among CRC patients were retrieved by systematic literature search until October 2022. Validation data were drawn from the 2310 CRC patients of the DACHS study recruited from 22 hospitals in the Rhine-Neckar region in the southwest of Germany. Cox regression analyses with adjustment for clinical variables were performed to validate individual CpGs and prognostic models. Model discriminiation and calibration accuracy were assessed by area under the receiver operating characteristic curves (AUC) and calibration plots. Results We identified 200 unique CpGs and 10 CpG-based prognostic models derived from 15 studies. In the external validation analysis, 1252 of 2310 patients died during follow-up (median 10.4 years). Thirty-nine CpGs (20%) and five prognostic models (50%) were independently associated with overall survival after adjustment for clinical variables. The discrimination ability for the five models was unsatisfactory, with AUCs ranging from 0.53 to 0.60. The calibration accuracy of the five models using recalibrated baseline survival was also poor. All five models can provide only small added prognostic value to traditional clinical variables. All prognostic models were rated as high risk of bias. Conclusions Only a fraction of published CpGs associated with survival in CRC patients could be externally validated. So far derived published CpG-based prognostic models for CRC do not seem to be useful for clinical practice.