Final Analysis of JCOG0203 for Advanced-Stage Indolent B-Cell Lymphoma 15 Years after the End of Enrollment: Pooled Analysis of Arms a and B for Follicular Lymphoma

Optimal immunochemotherapy as first-line treatment for patients with follicular lymphoma (FL) remains undetermined. FL frequently relapses after frontline treatment; thus, fatal secondary malignancies (SMs) and treatment-related death, including death due to infection, which will be accumulated by subsequent treatments, should be avoided. To evaluate the cumulative incidence of SMs as well as histologic transformation (HT), a follow-up of more than 15 years is necessary. However, such data are available only in the pre-rituximab era by retrospective and single-institute studies. In addition, the real incidence of HT was probably underestimated in the studies that estimated only biopsy-proven cases. In JCOG0203, previously untreated patients with stage III/IV indolent B-cell lymphoma were randomly assigned to receive six cycles of R-CHOP (Arm A) or R-CHOP-14 (every 2 weeks with G-CSF, Arm B) without rituximab maintenance. All histopathological specimens were centrally reviewed by three hematopathologists. To evaluate long term follow-up data, analyses for overall survival (OS), progression-free survival (PFS), HT, SMs, and progression of disease within 24 months after treatment initiation (POD24) were performed. OS and PFS were calculated using Kaplan-Meier method. Hazard ratios (HRs) were estimated by proportional hazards model. Cumulative incidence function of HT and SM were estimated. Exploration of risk factors for frequency of HT and POD24 were conducted by logistic regression model. Between Sep 1, 2002 and Feb 28, 2007, 300 patients, including 248 with grade (G)1-3A FL, were enrolled from 44 centers. Nineteen (6.3%) patients were lost to follow-up. The 10- and 15-year OS of patients with FL in Arm A vs. Arm B was 83.8% vs. 85.3% and 76.4% vs. 75.9%, respectively. No protocol treatment-related death or fatal late infection was reported. Multivariable analysis for the 299 patients eligible for OS revealed that significant adverse prognostic factors were age ≥61 (HR=2.54, p<0.0001), male sex (HR=2.06, p=0.002), and ECOG PS ≥1 (HR=1.71, p=0.044). As both arms revealed similar outcomes, pooled analyses of Arms A and B were performed for FL. The incidence of HT, including biopsy-proven and clinically diagnosed HT, was 9.8% and 14.4% at 10 and 15 years after enrollment, respectively, and increased thereafter. Seventy-seven of 245 (31.4%) patients, excluding two who died and one lost to follow-up patients without POD24, experienced POD24. HT occurred in 19 of 76 (25.0%) patients. Multivariable analysis for POD24 identified high Follicular Lymphoma International Prognostic Index (FLIPI) score (Odds ratio (OR)=4.40) and G3A histology (OR=2.63) as risk factors for HT and intermediate (OR=2.53) and high (OR=2.24) FLIPI score, B symptoms (OR=2.09), and G3A (OR=1.83) as risk factors for POD24. Additional analysis suggested that, restricted to within a 10-year period, high FLIPI (OR=3.24, p=0.046) and G3A (OR=3.30, p=0.054) were more related to HT than in the whole study period in univariable analysis. The cumulative incidence of SMs increased after the 10-year follow-up. The incidence of solid tumors was 12.8% {10.6% (fatal [f], n=3/17 events) in Arm A vs. 15.0% (f, n=10/23 events) in Arm B} at 15 years, whereas that of hematologic malignancies was 3.7% {4.1% (f, n=2/5) in Arm A vs. 3.3% (f, n= 1/4) in Arm B}. These results are the longest follow-up data of more than 200 patients with FL homogeneously treated in a multicenter prospective trial. R-CHOP was safe and efficacious, reducing HT, without increasing fatal hematologic malignancies in the long-term and may, therefore, be a reasonable option as frontline therapy for advanced-stage FL.