From retrospective analysis to real-world impact: Mismatch repair deficiency detection in gliomas by tissue and liquid ctDNA NGS in glioma management.

Abstract

2070 Background: Tissue biopsy remains the gold standard for Microsatellite Instability and MMR (mismatch repair) gene alteration assessment. The regulatory approval of immune checkpoint inhibitors is for mismatch repair-deficient cancers, regardless of tumor type. Surgical resection or biopsy is challenging when the glioma is located deep in the brain or brainstem. Circulating tumor DNA (ctDNA) next-generation sequencing (NGS) offers a non-invasive alternative, garnering attention in extracranial cancers, however not so in gliomas due to the lower concentrations of tumor-derived biomarkers. While Cerebrospinal fluid (CSF) provides superior sensitivity and specificity for gliomas, it is not a standard test in gliomas with its own technical issues. Methods: A retrospective analysis was conducted using databases to evaluate the prevalence of pathogenic inactivating alterations in MMR genes in glioma tissue samples. Data were queried from the MSK, Clin Cancer Res 2019 database for targeted sequencing on MSK-IMPACT and FMI panels, comprising 1004 tissue samples (837 with matched normal) from 923 glioma patients through the cBioPortal platform. Frequencies of MMR gene alterations were assessed. Results: A total of 850 patients (out of 923) were retrospectively analyzed for MMR gene alterations, with 40.3% (343) being female and 59.6% (507) male. Primary samples constituted 79.8% (679), and recurrent samples 20.2% (172). OncoKb level alterations were categorized as Level 1 (0.5%), Level 2B (16.6%), Level 3B (13.6%), Level 4 (32.8%), and none (36.4%). MMR gene alterations were found in 35 samples (4%), with MSH2 and MSH6 each detected in 2%, MLH1 in 1%, and PMS1, PMS2, and MSH3 in less than 1% of samples. In a specific case, ctDNA NGS was performed on a 9-year-old boy diagnosed with diffuse intrinsic pontine glioma as tissue biopsy was not feasible. Survivals are 9 to 11 months despite multimodality treatment. ctDNA NGS identified a truncating MSH6 alteration at 100% Variant Allele Frequency, suggesting biallelic inactivation of MSH6. Additionally, an IDH R132C activating mutation, a TP53 splice site SNV and high tumor mutational burden (bTMB) at 132.33 Mut/Mb were detected. Post radiation resulted in no change in tumor size. Injection pembrolizumab 3 weekly was initiated. Follow up MRI revealed further reduction in size and tumor has remained stable with ongoing therapy. Conclusions: Identifying MMR alterations potentially broadens the therapeutic options for glioma. The compelling case of the 9-year-old boy highlights the clinical utility of ctDNA NGS in identifying actionable MMR gene alterations, leading to successful immunotherapy with pembrolizumab and continued stable disease beyond 13 months. While few studies exist on utility of ctDNA in gliomas, it is time for bigger studies in both primary and recurrent gliomas where biopsy is not feasible.