Purpose: Chordomas are rare, malignant neoplasms that arise from remnants of the embryonic
notochord along the axial skeleton, most commonly in the sacrum, skull base, and spine.
Chordomas are characterized by their unique histopathological features, including
physaliphorous cells, and have a distinctive immunohistochemical profile but have
limited information on their somatic mutational landscape. The current treatment paradigm
involves surgery or radiation therapy, as there are no FDA-approved drugs for chordoma.
There remains great potential to discover targeted therapies against specific mutations
in chordomas. In this study, a publicly available genomic database was used to profile
the somatic mutational landscape of chordoma patients and interrogate differences
based on their histological subtype.
Methods: The American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia
Information Exchange (GENIE) database was accessed from cBioPortal (v16.1-public)
on July 22, 2024, to identify all patients with chordoma. The most common gene mutations,
gene correlations, and mutual exclusivities were assessed using a two-sided t-test and nonparametric tests, with Benjamini-Hochberg False Discovery Rate (FDR)
correction.
Results: Of the 1,437 collected bone cancer samples, 133 (9.2%) were chordoma. Of these samples,
14 (10.5%) were conventional chordoma, 3 (2.3%) were dedifferentiated chordoma, and
116 (87.2%) were not otherwise specified. In this cohort, 18 (15.2%) patients were
pediatric and the rest were adults (n = 115; 84.8%). 85 (72.0%) patients were White, 8 (6.8%) were Asian, and 7 (5.9%)
were Black. Sixty-six (55.4%) patients were male and 53 (44.9%) were female. The top
mutations in the cohort were: PBRM1 (n = 10; 7.5%), ARID1A (n = 10; 7.5%), TERT (n = 9; 7.1%), TP53 (n = 8; 6.1%), SETD2 (n = 7; 5.6%), and NOTCH2 (n = 7; 5.6%). The majority (77.8%) of TERT mutations were the 5′ flank type, resulting in promotor changes, and ARID1A (70.0%), TP53 (87.5%), and NOTCH2 (100%) were associated with missense mutations. In this cohort, PBRM1 mutations tended to co-occur with BRCA2 (2/10; p = 0.027) and KMT2D (2/10; p = 0.044), while KMT2D mutations tended to co-occur with BRCA2 (n = 3/6; p < 0.001). There were no additional statistically significant mutual co-occurrence
or mutual exclusivities. ARID1A, TERT, TP53, SETD2, and NOTCH2 were relatively mutually exclusive with only 1 (<0.1%) sample with more than 1 of
these mutations. There were no significant clinical or genomic differences between
pediatric and adult patients.
Conclusion: The study identified a distinct somatic mutational landscape in chordoma patients,
with significant mutations in PBRM1, ARID1A, TERT, TP53, SETD2, and NOTCH2. Co-occurrence patterns, particularly involving PBRM1 with BRCA2 and KMT2D, highlight potential pathways of DNA repair and chromatin modification that may guide
the development of novel therapies. However, there are limited clinical trials evaluating
direct targeted therapies against the most common chordoma mutations. For example,
biologics targeting chromatin remodeling mechanisms may serve as a potential future
strategy for chordoma with PBRM1 mutations. Understanding these mutational profiles may inform the development of
targeted therapies and improve clinical outcomes for chordoma patients.
