P1155: more precise risk stratification for tp53 mutant diffuse large b cell lymphoma

Topic: 19. Aggressive Non-Hodgkin lymphoma - Clinical Background: Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoid malignancy and is characterized by pronounced genetic and clinical heterogeneity. TP53 mutations correlates with inferior survival in many cancers, whereas it as prognostic markers are still controversies in DLBCL. TP53 is the most frequently mutated gene enriched in A53 subtype and other gene subtypes. However, A53 subtype which has fatal prognostic cannot be identified using the next-generation sequencing which used mostly in clinical practice. Aims: In the present study, we aimed to develop the prognostic index (PI) to stratify more accurately for TP53-mut DLBCL patients and reveal the underlying mechanisms of different prognosis. Methods: The available clinical information and corresponding mutation data of B cell lymphoma patients were retrieved and obtained from published articles. Ultimately, 2637 DLBCL patients in six cohorts were enrolled in the final analysis. The Jiangsu Province Hospital (JSPH) study cohort consisted of 109 patients diagnosed with DLBCL. Among the 109 DLBCL patients, all tumor tissue samples were collected to perform next-generation sequencing (NGS), while a total of 104 DLBCL samples were analyzed the gene expression levels using RNA-seq. Results: Among the 2637 DLBCL patients from the integrated cohort, TP53 mutations were found in 370 patients with a frequency of 14.0% lower than previous reports. The distributions of mutation events were mainly located in the DNA binding domain (DBD) (N=333, 83.3%), containing 34 at Arg248, 20 at Gly245 and 20 at Arg273, corresponding to the TP53 hotspots in non-Hodgkin lymphoma described in previous studies (Figure 1A). Compared with TP53-wt patients, significative P value was generated by Breslow test (P=0.0014), whereas Log Rank test uncovered border-line P value for OS (P=0.0840). Such a result would indicate that the survival difference just occurred during early survival, the 10-year OS was even slightly better than the TP53-wt patients (Figure 1B). Accordingly, we sought to construct a model to identify the truly high-risk patients. Based on the univariate Cox regression result, significant variables were further performed multivariate Cox regression analysis. As shown in Figure 1C, three variables retained independent prognostic significance for PFS and OS, which constituted TP53-PI model consisting of age > 60, IPI intermediate- or high-risk groups and MCD subtype by actual regression coefficients as weight. The TP53-PI model could significantly distinguish the prognosis of patients with TP53-mut DLBCL (P<0.0001, Figure 1D). To calculate the weight of each selected factor, a nomogram was generated, which was used to predict the survival rate. Differential expression analysis by using RNA-seq in JSPH cohort offered insights into the underlying biological mechanisms in TP53-mut DLBCL. As expected, the immune-associated biological processes occupied a large proportion in the result of Gene Ontology (GO) functional enrichment analysis by using all the differentially expressed genes between TP53-mut and TP53-wt DLBCL (Figure 1E). Of particular note, the CD8+ T cell, checkpoint, cytolytic activity, Th1 cell and Th2 cell exhausted in the TP53-mut group with a characteristic immune microenvironment. Furthermore, enriched GO terms in the intermediate and high-risk groups of TP53-PI, including B cell mediated immunity and cytokine activity pathways.Summary/Conclusion: In conclusion, the TP53-PI model could further identify the adverse prognosis of patients in TP53-mut DLBCL. The mechanism driving different survival outcomes may be explained by the unique immune microenvironment in TP53-mut DLBCL. Keywords: TP53, DLBCL