Concurrent TP53 loss-of-function and epigenetic modifier mutations as a potential predictor of chemoimmunotherapy failure in DLBCL & hgbl Free

Background:

Despite the efficacy of modern chemoimmunotherapy, a significant subset of patients with diffuse large B-cell lymphoma (DLBCL) and high-grade B-cell lymphoma (HGBL) experience primary refractory or relapsed disease. Current risk stratification relies heavily on clinicopathologic scores and ‘double/triple-hit’ (DH/TH) genetic status. However, aggressive outcomes occur in non-DH/TH patients, and while TP53 mutations are established markers of poor prognosis, their predictive value is heterogeneous. Broader genomic signatures distinguishing patients who achieve durable remission from those who fail first-line therapy and require intensive salvage, including chimeric antigen receptor (CAR) T-cell therapy, remain incompletely defined. Identifying high-risk molecular signatures at diagnosis is crucial for guiding novel therapeutic strategies.

Methods:

We conducted a retrospective pilot analysis of 10 patients with DLBCL/HGBL treated with first-line anthracycline-based chemoimmunotherapy at Yale Cancer Center. Pre-treatment diagnostic tumor samples underwent comprehensive genomic profiling (CGP) with the FoundationOne® Heme assay. Patients were stratified based on response to first-line therapy: a durable complete response (CR) group (n=5) and a relapsed/refractory (R/R) group (n=5). Clinicopathologic variables were collected, including age, sex, stage, IPI score, ECOG performance status (PS), LDH, cell of origin (COO), and rearrangement status of MYC, BCL2, and BCL6. Genomic alterations were correlated with clinical trajectories. Statistical comparison of clinicopathologic variables was performed using the Mann-Whitney U test.

Results:

The CR (n=5) and R/R (n=5) groups were similar regarding baseline clinicopathologic variables (CR vs. R/R): Age at diagnosis (median 59.9 vs. 68.8 years, p=0.5), Sex (4M/1F in each group), Stage (median 2.0 vs. 3.0, p=0.9), IPI score (median 2.0 vs. 2.0, p=1.0), ECOG PS (median 1.0 vs. 0, p=0.5), LDH (median 474 vs. 344 U/L, p=0.8), and COO (3 GCB/2 non-GCB in each group). Within the R/R group, two patients (40%) demonstrated treatment resistance requiring CAR-T therapy. One patient required an allogeneic stem cell transplant after three lines of therapy, while the remaining responded to standard second-line salvage. Genomic profiling revealed a distinct molecular signature in the two patients requiring CAR-T, characterized by the co-occurrence of pathogenic, loss-of-function (LOF) TP53 mutations (R282W, R248Q) and concurrent inactivating mutations in key epigenetic modifier genes (KMT2D, CREBBP, EZH2). This dual signature was absent in the other three R/R patients and all durable CR patients (0/8). In contrast, 4/5 patients (80%) in the CR group were TP53 wild-type. The fifth CR patient harbored a TP53 variant (G245D) not classified as a canonical LOF mutation, alongside a KMT2D mutation. While alterations such as CDKN2A/B loss were identified in both cohorts, the dual signature of TP53 LOF and epigenetic pathway disruption appeared highly specific for chemotherapy resistance.

Conclusion:

In this hypothesis-generating pilot cohort, the combination of pathogenic, loss-of-function TP53 mutations and inactivation of critical epigenetic regulatory genes defines a high-risk molecular profile strongly associated with failure of standard chemoimmunotherapy and the necessity for cellular therapy. Our findings highlight that the functional consequence of TP53 mutations (LOF vs. non-LOF), particularly in the context of epigenetic dysregulation, may be a more critical predictor of outcome than TP53 mutation status in isolation. While validation in larger cohorts is essential, upfront identification of this dual pathway inactivation may identify patients unlikely to benefit from standard salvage approaches, warranting early stratification toward CAR-T therapy or clinical trials of novel agents.