Discussion:
Previous studies have demonstrated a clear correlation between TMB and DM in pediatric brain tumors.9 This multi-institutional analysis assessed whether higher TMB tumors correlate with outcomes in children with a broad range of brain tumors. We summarized genomic characteristics of pediatric CNS tumors and outcome data in 119 pediatric patients across 4 academic medical institutions between 2008-2021. We report a clear link between higher TMB and worse outcome. Higher TMB was a significant and independent predictor of reduced OS. Death events and tumor progression were more common in high TMB tumors as compared to low TMB tumors; high TMB tumors also had shorter TTD and TTP. These represent novel findings in understanding the genomics as they relate to outcomes in pediatric CNS tumors.
TMB and outcomes have previously been studied in many adult, and some pediatric, tumors. TMB has not always correlated to OS; there are wide ranges of solid tumor types and data has been inconsistent.5 TMB has been associated with shorter OS in adult patients with gliomas, BRCA1 or BRCA2 mutated high-grade serous ovarian cancer, and lung cancer. Worse outcomes in high TMB tumors are likely attributed to mutations that inhibit cell cycle checkpoints and DNA repair mechanisms, leading to high proliferative activity.6,9,12,13 These mutations contribute to more severe and rapid metastatic disease in high TMB tumors.6 Our data further demonstrates poor prognostication of TMB.6,14,15 Poorer outcomes may also be indicative of inadequate treatment options for this high-risk pediatric CNS tumor patient subset.
Similar to data presented by Patel et al., high TMB tumors in this analysis were more likely to have a TP53 inactivation and low TMB tumors were more likely to have a BRAFmutation.9 Patel et al., postulated this was becauseTP53 inactivation causes an inability to repair DNA mutations andBRAF has a minimal effect on DNA repair.9TP53 inactivation is likely associated with high TMB tumors given its facilitation of a multi-step malignant progression.16 The current data further demonstrates that high TMB tumors often do not have driver mutations that can be therapeutically targeted; identifying a need for alternative therapies.
Although most high TMB tumors in our dataset were treated with systemic therapy, only a small fraction of patients were treated with targeted or immunotherapy. Targeted and immunotherapy may not have had a significant association with survival because such therapies were most commonly used second-line, post-tumor recurrence or progression; this is true for all three patients who received immunotherapy in our study. Literature supports that first-line immunotherapy may have better outcomes than when immunotherapy is used as second-line; the hypothesis being that upfront chemotherapy may destroy the host immune system, decreasing the efficacy of subsequent immunotherapy.17 A lack of statistical significance may also be due to low sample size. The use of ICIs may be indicated for such patients, given our findings that high TMB tumors are more likely to be associated with loss of tumor suppressor genes and less likely to be associated with oncogene mutations.
Furthermore, TMB is a biomarker for ICI enrollment; TMB has been found to predict ICI response in 20-60% of patients.3,7,18-21 High TMB (>3 mut/Mb) has also been demonstrated to predict improved progression-free survival in patients on ICI.11 Although TMB has been associated with worse survival, tumors with high TMB tumors are often extremely responsive to ICI therapy. Additionally, high TMB tumors associated with poor OS were found to have improved OS with ICI when compared to low TMB tumors. These responses have been demonstrated in non-small cell lung cancer, malignant melanoma, cervical cancer, breast cancer, gastric, esophageal and head and neck cancers.18,19,22-24 In pediatrics, however, responses have been inconsistent, possibly because of the relatively low TMB of pediatric cancers.25,26
There are over 2,000 clinical trials investigating ICI in patients of all ages and with all tumor types (accessed June 2022). Of the 141 trials investigating ICI in pediatric patients, five of which have or are currently investigating ICI in pediatric CNS tumor patients.27-31 Only one trial, specifically recruits pediatric patients with high TMB CNS and solid tumors to analyze the effect of ICI.28 Pembrolizumab has been approved for high TMB solid tumors, including pediatric CNS tumors with PD-L1 expression.32 In a post-hoc analysis of the HERBY trial, hypermutated pediatric HGGs treated with bevacizumab (anti-VEGF) had significantly more CD8+ tumor-infiltrating lymphocytes and improved OS as compared to non-hypermutated types.33 This analysis supports that bevacizumab may be a treatment option for pediatric CNS tumor patients with high TMB and poor outcomes in need of more effective treatment regimens. It also emphasizes the importance of understanding tumor microenvironment as a means of further stratifying patients, which may be a future direction for studies.34
Our multi-institutional analysis shows a clear correlation between higher TMB and poor outcomes for children with brain tumors.Although variables were standardized, discrepancies may exist in data collection amongst institutions. Additionally, DNA methylation profiling has increasingly been used for pediatric CNS tumor classification.2,35,36 Only two patients in our dataset had DNA methylation profiling and only 75% of patients had TMB data. In addition, the brain tumor tissue obtained at these institutions was not routinely sent to FoundationOne for TMB testing for all children so it is possible our sample is not representative of all children with brain tumors. We recommend an increase in TMB and DNA methylation testing, as both may be applicable toward stratifying treatments and also used to improve research and treatment developments in the future.2 Lastly, although our analyses were pre-determined prior to data collection, results should be interpreted in the context of the retrospective nature of this study.
We identify pediatric CNS high TMB tumor patients as a unique high-risk patient subset that has poor outcomes: higher progression rates and poorer overall survival. Our data may help in identifying such patients who may require alternative treatments, such as ICI.
Conflict of Interest Statement: All authors have no conflicts of interest to disclose.
Acknowledgements : This study was supported in part by Albany Medical Center Children’s Hospital Grant, Albany, NY, 2021.