Predictive Power of FDG-PET Parameters at Diagnosis and after Induction in Patients with Mantle Cell Lymphoma, Interim Results from the LyMa-PET Project, Conducted on Behalf of the Lysa Group

INTRODUCTION. FGD-PET has emerged as an important predictor of clinical outcome in lymphomas. However, its utility in everyday clinical practice in mantle cell lymphoma (MCL) remains uncertain because there is a lack of large prospective trials including FDG-PET results. To address this question, we conducted the LyMa-MRD project as an ancillary study in a prospective phase III trial in MCL (NCI NCT00921414; LyMa Trial). From Sept 2008 to Aug 2012, 299 previously untreated MCL patients (<66yrs) were enrolled in the LyMa trial (a phase III international prospective trial, NCT00921414). Briefly, all patients received 4 courses of R-DHAP followed by ASCT using an R-BEAM conditioning regimen (n=257). After ASCT, patients were randomized between observation (obs) (n=120) versus Rituximab maintenance (RM) (n=119). The first planned interim-analysis, with a median follow-up of 40.6 months was presented at ASH 2014 and indicated superior progression-free survival (PFS) in the RM versus Obs arms (Le Gouill et al. ASH 2014). Sequential FDG-PET monitoring was optional and a predefined secondary objective and was performed throughout. Treatment strategy was not modified by FDG-PET results. Indeed, the population of the LyMa trial is the ideal population to investigate the predictive power of FDG-PET parameters at diagnosis and after induction in MCL. Herein, we present the first results performed from the database of first planed interim-analysis.

METHOD. FDG-PET of 94 MCL patients have been independently and centrally reviewed by 2 lymphoma expert nuclear physicians. Quantitative metrics including SUV_max, SUV_mean, SUV_peak, total lesion glycolysis (TLG) were extracted from the area with the highest uptake, at diagnosis and before ASCT (iPET). Visual analysis by the Deauville scale was also performed at iPET. The best cut-off values were determined for each metric using X-tile® analysis. Prognostic value was assessed using univariate analysis by Kaplan-Meier estimates of PFS.

RESULTS. The studied population did not differ from the whole population of the LyMa trial (baseline characteristics, demographic data, staging, balance between observation vs maintenance and outcome). At diagnosis, univariate analysis showed a prognostic value on PFS of 4 metrics: SUV_max (p<0.001), SUV_mean (p< 0.001), SUV_peak (p<0.001), TLG (p=0.03). The best cut off for these 4 indices, were 11.4, 7.7, 8.7, and 65 respectively. Indeed, median PFS was not reach (NR) for patients with low SUV_max (n=63) as compared to only 26.3 months for patients with high SUV_max (p<0.0001) (n=31). Results were similar for SUV_peak, SUV_mean and TLG. The prognostic value of SUV_max was reinforced when combined with MIPI. Indeed, patients can be separated in 3 prognostic groups including a group of patients with a very good outcome (low SUV_max plus MIPI inter/low). Results of iTEP showed that ΔSUV before ASCT (defined as a decrease of SUVmax values from diagnosis to end of induction > 29.65%) was predictive of PFS: median NR vs 42.3m (p=0.0051). In contrast FGD-PET before ASCT according to Deauville scale was only predictive for score 5 vs 1/2/3/4.

Conclusion. The LyMa-PET project is the largest study addressing the question of FDG-PET in a homogeneously treated population of MCL. It shows for the first time that SUVmax (</>11.4) at diagnosis and ΔSUV (</>29.65%) provide predictive parameters for outcome in MCL. Update results and patients' outcome according to randomization arms will be updated for the time of the meeting.