Over-expression of P53-related protein detected by immunohistochemistry (IHC) and absence of p21waf1— expression, the main downstream target following P53 activation, are useful surrogate markers in identifying diffuse large B-cell lymphoma (DLBCL) patients with mutated P53-gene. Following recent studies on gene expression, tissue micro-array technology has been shown to represent a reliable technique in discriminating DLBCL with germinal center B-cell like (GC), or post- germinal center (non-GC) origin.

We analyzed 80 consecutive patients with de-novo DLBCL, homogeneously treated with cure-intent (CHOP-like regimens +/− Rituximab) in a single Institution, for expression of p53/p21waf1—, with regard to the expression of GC (CD10+, bcl-6+/mum-1−/CD138−), and post-germinal center (CD10−, mum-1+, CD138+) IHC markers. Cases with 50% or more p53-positive neoplastic cells were defined as p53+++. All patients were provided with complete clinical information. Median age of our patients was 50 years, AAS was III or IV in half of patients, IPI was high or int/high in 36%. Median follow-up for survivors was 77 months. A p53+++/p21waf1— phenotype (corresponding to loss of function of the P53-gene) was detected in 19 of 80 patients (23%), while a GC phenotype characterized 42 patients (52%). Patients with P53+++/p21waf1— phenotype were more frequently resistant to induction therapies (p<0.0001) and had a significantly lower 5-year Progression Free Survival (PFS, 27% vs 48%, p=0.01). Patients with GC phenotype were associated with a better 5-year PFS than non-GC patients (57% vs 26%, p=0.02). When analysis was restricted to the 42 patients with GC phenotype, p53+++/p21waf1— phenotype could discriminate 28% of patients with PFS of 26% vs 68% of patients with GC pattern and other p53/p21waf1— phenotypes (p=0.005). In contrast, the outcome of patients with a non-GC phenotype was not modified by p53/ p21waf1— expression. At multivariate analysis, p53+++/p21waf1— phenotype, non-GC phenotype, and high IPI resulted the strongest independent factors in worsening PFS of our patients. Combining such adverse prognostic factors we could identify 60% of patients with an extremely poor prognosis (5-year PFS of 28%).

Our results strongly point to the usefulness of a combined molecular and clinical approach for prognostic considerations in DLBCL, and should be validated in larger series.

Author notes

Corresponding author

Sign in via your Institution