The NCCN-IPI Helps to Identify Very High-Risk Patients with DLBCL and Can be Improved By Other Independent Clinical Factors Including Bulky Disease and beta2-Microglobulin

The International Prognostic Index (IPI) has been the basis for predicting the outcome of diffuse large B cell lymphoma (DLBCL). With the improvement in prognosis with immunochemotherapy the predictive power of the IPI has been questioned.

Recently, the enhanced NCCN-IPI has been proposed to provide a better discrimination of risk groups. Additionally, other patient- and disease-specific factors not incorporated in the prognostic algorithms and easily assessed at diagnosis have been shown to impact on the clinical course. The prognostic value of the NCCN-IPI together with other clinical characteristics remains to be widely validated.

We aimed to assess the power of NCCN-IPI in risk group discrimination and compare it with IPI, as well as to investigate if the Body Mass Index (BMI), gender, diameter of the largest mass and levels of β₂-microglobulin have independent prognostic impact when controlling for NCCN-IPI risk.

We conducted a retrospective analysis of consecutive R-CHOP treated de novo DLBCL cases diagnosed between January 2002 and December 2013 in a single center.

Kaplan-Meier method was used to estimate the overall (OS) and progression-free survival (PFS) distributions for risk groups. Logrank test with p-value adjustment by Hochberg method was used for pairwise comparisons between risk groups within each index. Independent prognostic impact of BMI, gender, bulky disease and β2-microglobulin was assessed by multivariate analysis using Cox model.

410 patients were included in the analysis, with a median age of 64 (16-89) years; 51% were male. The distribution according to age, stage, extranodal (EN) sites, performance status, β₂-microglobulin, BMI and bulky disease (Table 1) were similar to published data.

With a median follow-up of 59 months, 5-yr OS and PFS were 69% and 65%, respectively. The distribution of patients according to IPI and NCCN-IPI risk groups is detailed in Table 1. Both scores adequately stratified patients according to risk of progression and death. In our cohort OS and PFS of low and low-intermediate risk groups using the NCCN-IPI were not significantly different. We documented a migration of patients from low (62%) and high (48%) to intermediate risk groups using the NCCN-IPI compared to the IPI. This score allowed a better discrimination between high-intermediate and high-risk disease (pairwise comparison of OS and PFS within NCCN and IPI risk scores of p=0.0087 vs p=0.044 and p=0.032 vs p=0.096, respectively). Only 12 patients classified as intermediate-risk IPI migrated to high NCCN risk group. Of these, 5 relapsed and 6 died (4 of lymphoma).

Multivariate analysis showed that bulky disease and β2-microglobulin are independent prognostic factors for OS and PFS when controlling for NCCN-IPI.

Using a large series of unselected R-CHOP treated patients, we pursued the validation of the NCCN-IPI as a better prognostic tool for DLBCL. In contrast to the original series, the NCCN-IPI showed limited discrimination capacity for OS and PFS between low- and low-intermediate risk groups. However, the new prognostic index helps to identify particularly poor prognosis patients that might benefit from alternative therapies. We also identified bulky disease and β2-microglobulin as independent prognostic factors in addition to NCCN-IPI and that could be incorporated in future prognostic scores. Additional studies are needed to evaluate the ability of biomarkers to improve the clinical risk stratification systems.