Objective:

This study investigates if aSHM is a driving event in recurrent/refractory CNS lymphoma by comparing genomic alterations in archival tumor from diagnosis with circulating tumor DNA (ctDNA) in the cerebrospinal fluid (CFS) at recurrence. 

Background:

Primary CNS lymphoma is a lymphoma exclusively in the CNS. Over half of patients with CNS lymphoma have disease recurrence following initial response to therapy. Aberrant somatic hypermutation (aSHM) is a mechanism of genomic instability in systemic DLBCL and PCNSL. 

Design/Methods:

10 paired PCNSL patient samples (archival tumor biopsy and CSF at recurrence) were collected. Samples analyzed using MSK-HemePACT panel. For concordance analysis of pairs, mutations designated as “present” when independently detected. Secondary mutation analysis performed to detect the compilation of SNVs detectable within samples (CSF and tumor). Mutations considered “present” if the Variant Allele Depth was ≥ 2 and Variant Allele Frequency ≥ 1%. We used previously published criteria (Khodabakhshi et al., Oncotarget 2012) to determine whether a mutation was the result of aSHM. 

Results:

Interval between tumor biopsy and CSF collection ranged from 1-146 months.  Median number of mutations in tumor tissue was 23 vs. 16 in the CSF.  46.8% of SNVs shared between tumor tissue and CSF ctDNA. Most private mutations were found in tumor tissue>CSF samples (35.5% vs. 13.5%). Private mutations in the tumor sample were more frequently a result of aSHM (7.3%) than those in the CSF (0.6%). Shared mutations did not show the hallmarks of aSHM. Driver mutations in the B-cell receptor pathway were generally shared. 

Conclusions:

aSHM was observed in the archival tumor samples that were not shared with the CSF samples. Moreover, aSHM seemed to be a minor contributor to mutations identified in CSF samples. Possibly, aSHM does not play a role in PCNSL after initial diagnosis and therapy, accounting for sparsity of mutations caused by aSHM in recurrent/refractory CSF samples.