Abstract Introduction: Primary Central Nervous system lymphomas (PCNSL) account for ~ 4% of newly diagnosed CNS tumors with an annual incidence rate of 0.5 per 100,000 in the United States. There have been varied reports from large databases including SEER stating that incidence of PCNSL is low in Hispanics as compared to other races with better survival. PCNSL is known to have recurrent mutations in JAK-STAT, NF-kB, and B-cell receptor signaling pathways, including hallmark mutations in MYD88 L265P (67%) and CD79B (63%), and CDKN2A deletions (83%). However, difference in genomic landscape related to ethnicity impact on incidence, management, and survival amongst Hispanics vs non-Hispanics is unknown. Methods: This retrospective study evaluated 15 patients with PCNSL treated at University of Texas Health San Antonio from 2008-2023. Somatic variant calling was performed on whole-exome sequencing (WES) data using GATK4 MuTect2 in tumor-only mode. Raw WES reads were aligned to the human reference genome (hg38) using BWA-MEM with default parameters. The resulting BAM files were processed following GATK best practices, including duplicate marking with Mark Duplicates, base quality score recalibration (BQSR), and indexing. Common germline variants were excluded based on a gnomAD allele frequency threshold of 1%. Raw RNA-Seq FASTQ files were aligned to the human reference genome (hg38) using STAR-Fusion (v0.13.0) with GENCODE v44 annotation. In addition to detecting gene fusions, STAR-Fusion outputs a gene-level raw count matrix, similar to standard STAR alignment. Cluster-specific marker genes were obtained from PMID: 36402300, in which the authors defined four molecular subtypes (C1–C4) and reported that C1 and C4 were associated with better overall survival compared to C2 and C3. The EdgeR package was used to normalize read counts and conduct differential expression. Results: We identified 15 patients with PCNSL who had sufficient tumor tissue available for analysis. The mean age was 60.6 years (range: 28–83 years). Of these, 9 were Hispanic and 6 were non-Hispanic. Notably, 1 patient had a history of lung transplant, 2 were HIV-positive, and 4 tested positive for EBV. The most used treatment regimen was RMPV (Rituximab, Methotrexate, vincristine and Procarbazine). The median overall survival (OS) was 26.5 months (range: 2–124 months). We identified a total of 10,373 somatic mutations across 15 PCNSL FFPE samples, with a median of 492 mutations per sample. Patients who had undergone radiation therapy exhibited the highest mutation burden, characterized predominantly by CA substitutions. Among the 21 well-established driver genes analyzed, we observed frequent mutations in MYD88 (53%), CD79B (33%), EP300 (53%), PIM1 (47%), PRDM1 (27%), and others. Notably, the oncogenic MYD88 L265P mutation was present in 53.3% of cases (8 out of 15 samples). In contrast, none of the four EBV-positive cases harbored this mutation. Using the single-sample Gene Set Enrichment Analysis (ssGSEA) and Cluster-specific marker genes from PMID: 36402300, subtype predictions were applied to each sample. Samples from non-Hispanic ethnicity resembled most closely with subtype C3 which is associated with worse survival. The FBXL16 gene had the most significant decrease in expression in the Hispanic group (logFC= -2.5, p=.0001, FDR .35). Conclusion: We identified high rate of somatic mutations in our sample with MYD 88 being the most common mutation, except in EBV-positive PCNSL, which is consistent with previous studies reporting mutual exclusivity between MYD88L265P and EBV-associated PCNSL. Samples from non-Hispanic patients closely resembled to C3 cluster, which may suggest a potential association between non-Hispanic ethnicity and poorer survival in our cohort. However, due to the limited sample size, no conclusions can be drawn. There was no statistically significant difference in survival between Hispanic and non-Hispanic patients in our study due to the small sample size. A study with larger sample size is underway to evaluate the survival differences.
Abhishek Pandya (Mon,) studied this question.