DISCUSSION
Almost all patients in our study were treated uniformly, with only 2/54 (4%) receiving chemotherapy prior to RT and 1/54 (2%) receiving non-myelosuppressive chemotherapy (vincristine) during RT. The majority of the patients (80%, 43/54) also received proton beam RT. Typically, for patients receiving 23.4 Gy CSI, the CSI portion of the RT course is completed by the end of week 3, with CSI extended through week 4 for those receiving ≥36 Gy; by week 5, CSI is normally complete, and the RT field shrinks to cover only the tumor bed. Consequently, late lymphopenia during weeks 4 and 5 may have more clinical significance than lymphopenia that develops earlier in the RT course, as the exposure to CSI is completed at the end of week 3 or later.
As previously mentioned, RT-induced lymphopenia is associated with worse PFS and OS in numerous adult tumors, a phenomenon presumed secondary to RT causing a decrease in TIL.2-6,13-15 Among CNS tumors, higher numbers of both CD3+ T cells generally and specifically CD8+ cytotoxic TIL at the time of tumor diagnosis are associated in the majority of studies with improved prognosis in adult high grade glioma and atypical meningioma.13,14,16-19
Unfortunately, medulloblastoma tumor samples characteristically have low numbers of TIL even prior to the start of RT.7-10,20,21 These findings suggest that medulloblastoma has an immunosuppressive tumor microenvironment. Overall, CD8+ T cells are the largest component of the TIL in medulloblastoma; additionally, there are differences in numbers of CD8+ T cells and TIL in general among medulloblastoma molecular groups, subgroups, and even down to the subtype level.8,9,20,22-24 The few published investigations assessing for a correlation between survival and TIL/CD8+ T cell numbers in pediatric medulloblastoma have had inconsistent results, demonstrating that larger studies are clearly needed.23,24
Our data showed no statistically significant difference in RFS between patients treated with protons vs. photons and in fact showed a trend toward lower RFS in patients that received proton RT. However, there is increasing awareness that proton beam RT may induce less lymphopenia than photons.25-27 When proton vertebral body sparing techniques are utilized for CSI, which in our cohort occurred for skeletally mature patients, data shows that lymphopenia is lessened28,29, as the vertebral bodies contain >25% of the total bone marrow.30 The majority of our patients had their entire vertebral body covered by the CSI dose due to receiving photon beam RT or being skeletally immature during proton beam RT.
To our knowledge, there are no published data regarding the association of molecular group with lymphopenia during RT. It is of value to control for molecular group when assessing for an association between risk of recurrence and lymphopenia over the course of RT which unfortunately we were unable to do in our study as the vast majority were diagnosed prior to the routine use of molecular grouping as standard diagnostic practice.
While our study shows the influence of lymphopenia during week 4 and 5 of RT on risk of medulloblastoma, other groups have shown that early lymphopenia (weeks 1 and 2 of RT) is what impacts the risk of recurrence in medulloblastoma 1 or that there is no effect at all of lymphopenia on RFS in childhood medulloblastoma 31. More data are clearly needed from additional patients with newly-diagnosed childhood medulloblastoma and ideally should be collected in a prospective fashion. Future studies should consider analyzing lymphocyte subsets throughout the RT course to characterize changes in specific lymphocyte populations over time. It would also be worthwhile to correlate the baseline number of TIL from the initial tumor resection with the risk both of lymphopenia throughout the RT course and tumor recurrence.
In summary, lymphopenia during weeks 4 and 5 of RT measured as CTCAE grade 3-4 and during week 5 of RT measured as below median ALC correlates with an increased risk of tumor recurrence in patients with newly-diagnosed childhood medulloblastoma.
CONFLICT OF INTEREST STATEMENT: SLG serves on the advisory board for Chimerix, which has no relationship to this manuscript. HBL, MES, AKA, MC, and ACP have no conflicts of interest to disclose.
ACKNOWLEDGEMENTS: None. No funding was utilized for this work.