Establishing Prognostic DNA Methylation-Based Chordoma Subgroups in Tissue that are Detectable in Plasma

Objective: Chordomas are bony tumors of skull base and spine that make up 2 to 4% of aggressive primary bone cancers. Despite standard of care treatment with surgery and radiotherapy, around half of patients recur, experience further neurological morbidity, and die within 4 years while the remaining half survive over 10 to 20 years. Unfortunately, these clinically aggressive and clinically benign subsets of chordoma cannot be reliably identified using existing clinical, pathological, or molecular factors/features to guide treatment approaches. Accordingly, this work aims to identify prognostic DNA methylation-based subgroups of chordomas and to establish feasibility for subgroup identification noninvasively, so that patients may be prognosticated at the time of diagnosis to guide treatment decisions. Methods: A multi-institutional 20-year surgical series of 68 patients was identified with available chordoma tissue samples. Extracted, bisulfite-converted DNA from tissue samples underwent whole genome DNA methylation profiling on the Illumina EPIC array. Matched plasma samples underwent sequencing of methylated cell-free circulating tumor DNA where available. Publicly available chordoma methylation and RNAseq datasets were obtained for validation. Results: The 68 chordoma tissue samples underwent consensus clustering which identified two stable patient clusters ([ Fig. 1 ]). Cluster 1 had a statistically significant poorer disease-specific survival than cluster 2 ([ Fig. 2 ]: median 6.0 vs. 17.3 years, p = 0.0062). The prognostic utility of these methylation-based clusters (HR = 14.2, 95% CI: 2.1–94.8, p = 0.0063) was independent of that of extent of resection and adjuvant radiotherapy use in a multivariate Cox's analysis, all three of which were independently prognostic. Cluster 1 was labeled as the “Immune-infiltrated” subtype based on the identification of immune-related pathways with genes hypomethylated at promoters in this cluster along with increased immune cell abundance ([ Fig. 3A ]). Comparatively, cluster 2 was characterized as the “Cellular” subtype based on the identification cell-to-cell interaction plus extracellular matrix pathway hypomethylation and higher tumor cellularity ([ Fig. 3B ]). These characterizations were validated using external DNA methylation data to show similar clusters, in external RNAseq data to show increased gene expression in the hypomethylated pathways for each cluster, and with immunohistochemical staining of immune markers. Differentially methylated regions of the genome in the plasma methylome data were identified that accurately distinguished chordomas from other clinical differential diagnoses by applying fifty chordoma-versus-other binomial generalized linear models in random 20% testing sets ([ Fig. 4A ]: mean AUROC = 0.84, 95% CI: 0.52–1.00). Tissue-based and plasma-based methylation signals were highly correlated and leave-one-out models accurately classified all tumors into their correct cluster using plasma methylome data ([ Fig. 4B ]). Three clinical cases of chordomas accurately diagnosed noninvasively, after alternate nonchordoma diagnoses were made clinically using MR imaging, are reported. Conclusion: This work is the first to establish prognostic DNA methylation-based subtypes of chordoma and to utilize plasma methylomes as noninvasive biomarkers for chordoma diagnosis and prognostication. The ability to identify and subtype chordomas prior to treatment will allow for therapy aggressiveness, including extent of resection, to be tailored to patient prognosis to improve clinical outcomes. Fig. 1 Fig. 2 Fig. 3 Fig. 4 Publication History Article published online: 15 February 2022 © 2022. Thieme. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.