(P 56) CHROME: PREDICTING CHONDROSARCOMA RISK OUTCOME FROM METHYLATION

Objective: Histological grading remains the most important predictor of clinical outcome in patients with central conventional chondrosarcoma and guides surgery. It is, however, subject to high interobserver variability, which can hamper the distinction of low- and high-risk patients and therefore result in over- or under-treatment, respectively. Several attempts have been made to establish biomarkers with prognostic value. The mutation status of IDH1 and IDH2, which is associated with global DNA hypermethylation in chondrosarcoma, is known to have an important role in the development of cartilage tumours but its prognostic significance is unclear. We evaluated whether DNA methylation patterns would be superior prognostic biomarkers to known risk factors like IDH1 and IDH2 mutation status and histological grading.

Methods: We determined the DNA methylation profiles of 69 primary central conventional chondrosarcomas using Illumina’s Human MethylationEPIC Array (850k) platform. Out of 69 cases, 15 patients suffered from at least one of the following events: local recurrence, metastasis, death of any cause. Differential methylation analysis was performed between sample groups based on clusters determined by a graph-based clustering approach. Univariate Cox Proportional Hazard Analysis was applied to identify differentially methylated probes that were individually related to survival. Then, we applied a LASSO Cox regression model to identify a small set of probes with prognostic value from which the risk score CHROME was constructed. This risk score was validated in an independent, publicly available, cohort.

Results: We constructed the risk score CHROME based on the methylation status of 11 sites. Samples were categorized into high- or low-risk. Kaplan-Meier survival estimates of the disease-free survival showed a significant difference between high (n=26) and low (n=43) risk groups (P < 0.0001), with no events present in the low-risk group. Histological grade, IDH mutational status, and sex displayed associations with disease-free survival but their significance was lost once the risk score was incorporated in a multivariate cox proportion analysis. This superior prognostic value of the methylation risk score was confirmed in the independent validation dataset.

Conclusion: We developed CHROME: a methylation-based risk score for central conventional chondrosarcoma that has a superior prognostic value to clinicopathological features including histological grade, IDH mutational status, age at diagnosis and sex. The methylation-based risk score developed in this study can thus confidently stratify patients into low- and high-risk and has great potential to be integrated in the clinical management of chondrosarcoma patients.