P002 | COMPARISON OF BONE MARROW BIOPSY, FDG-PET AND BCL2::IGH PCR, TO EVALUATE THE BONE MARROW INVOLVEMENT IN FOLLICULAR LYMPHOMA: PRELIMINARY ANALYSIS FROM THE FONDAZIONE ITALIANA LINFOMI FOLL12 PHASE III TRIAL

Background: Bone marrow (BM) infiltration at staging represents a prognostic factor in follicular lymphoma (FL). Beyond conventional bone marrow biopsy (BMB), it can be assessed through BCL2::IGH detection by PCR in BM and PET bone staging (PETb)+. We report preliminary data on a comparison of BMB, PETb, and BCL2::IGH detection by PCR in assessing BM involvement among FL patients enrolled in the FIL-FOLL12 trial. Methods: Untreated stage II–IV FL patients with intermediate-high FLIPI2 risk requiring therapy were included. BMBs were assessed by morphology. BCL2::IGH rearrangements were evaluated using Euro-MRD consortium assays targeting major (MBR) and minor breakpoints by nested-PCR, and quantified by RQ-PCR on diagnostic BM samples. PET/CT scans were reviewed by two expert nuclear medicine physicians. PETb positivity was defined as FDG focal uptake in bone segment higher than liver activity. Results: This analysis included 175 patients with available baseline PETb. Median age was 58 years (range 27–82), 52% male, median follow-up 68 months (95% CI: 64–71). High-risk FLIPI2 was noted in 35.4%. BMB involvement was present in 58.5% (median infiltration 30%, range 0.5–95%), PETb positivity in 31.5%, and BCL2::IGH marker detectable in 62.2%: 91 MBR (53%), 8 mcr (5%), 10 (5%) others.Comparison of BMB and PETb identified: 29% BMB+/PETb+ ("infiltrated metabolic"), 29% BMB+/PETb– ("infiltrated non-metabolic"), 25.1% BMB–/PETb– but BCL2::IGH+, and 12% triple-negative cases. PETb identified BM involvement in 3% of cases missed by BMB. Figure 1 shows univariate analysis of PFS for each BM positivity by different methods, in pooled arms of FOLL12 trial. Compared to BMB which represents the gold standard, PETb had the highest specificity (93%), sensitivity was 50% (PPV 90.9%, NPV 55.5%) while BCL2::IGH through PCR had highest sensitivity (60%) with low specificity (32%) (PPV 56%, NPV 36%). Combining the two methods sensitivity was 80.4% with a PPV of 63.6%. Conclusions: Our preliminary results show good concordance between BMB, PETb, and BCL2::IGH detection in assessing BM involvement in FL. Ongoing analyses are refining PETb criteria using quantitative radiomic features, including metabolic bone assessments. While no single technique identifies all involved cases, combination may improve BM assessment accuracy. Extended results on quantitative PET features and impact on outcome will be presented.