Novel criteria for evaluating early treatment response with whole body magnetic resonance imaging and prognostic implications in patients with newly diagnosed multiple myeloma

Abstract Introduction: Whole-body magnetic resonance imaging (WBMRI) is a highly sensitive technique to detect focal lesions in patients (pts) with newly diagnosed multiple myeloma (NDMM). However, its role and the criteria for evaluating treatment response prospectively are not well defined. Methods: Consecutive NDMM pts who underwent WBMRI at diagnosis and two months after treatment initiation, were prospectively enrolled. Those with normal imaging pattern at baseline were assessed for any changes, while those with diffuse/variegated pattern were categorized as responders or non-responders. For those with focal lesions (FL) an Interim MRI-based Response Grade (iMR-RG) was implemented. Definite response (DR) was defined as: a) complete dissapearance of the FL or b) increase of apparent diffusion coefficient (ADC) to ≥1400 or ≥40% increase compared to baseline and c) disappearance of paramedullary disease (PMD), if present at baseline. Partial response (PR) included: a) an increase of ADC to 1200-1400 or >25% but <40% increase compared to baseline and b) a decrease in size of PMD. Non-response (NR) was characterized by no significant change of ADC compared to baseline and no change in size of PMD. In the case of multiple FLs, if the majority showed evidence of DR and a minority of PR, then the response was labeled as “definite response dominant” (DRD) while the vice-versa as “partial response dominant” (PRD). Moreover, if all or most FLs showed evidence of NR while a minority showed evidence of PR or DR, then this was labeled as “non-response dominant” (NRD). Definite progression was characterized by: a) new FL(s) or b) increase in size of the FL(s) or c) increase in size of the PMD or d) new PMD or EMD. All pts provided written informed consent and the study was approved by the IRB. Results: 93 pts with NDMM were included, with a median age of 69 years (range 31-89), whereas 39.8% were females. At baseline the distribution of imaging patterns was as follows: 61 (65.6%) focal, 20 (21.5%) normal, 11 (11.8%) diffuse and 1 (1.1%) variegated. PMD was detected in 17 (18.3%) pts and fractures in 23 (24.7%). 74 pts (79.6%) received monoclonal antibody-based upfront treatment and 19 (20.4%) a proteasome inhibitor and/or an immunomodulatory drug. The median follow-up since the first WBMRI was 19.4 months. At the second WBMRI, evidence of imaging response was present in 9/12 (75.0%) patients with diffuse/variegated pattern. Serum responses were: 2 (16.7%) (stringent) complete response [(s)CR], 4 (33.3%) very good partial response (VGPR), 5 (41.7%) partial response (PR) and 1 (8.3%) stable disease (SD). We also observed an association between imaging and hematological response rates (p=0.005). All patients with normal pattern remained as such. Among the 61 patients with FLs, imaging responses to treatment two months post initiation were as follows: 7 (11.5%) definite response, 35 (57.4%) definite response dominant, 16 (26.2%) partial response dominant, 2 (3.3%) non-response dominant and 1 definite progression (1.6%). 13/17 (76.5%) pts with FLs and PMD responded. Best hematological responses to first-line treatment included: 3 (s)CR (4.2%), 38 (62.3%) VGPR, 18 (29.5%) PR and 2 (3.3%) minor response. Importantly, the disease progression event was identified before documented hematological progression or clinical deterioration. The depth of imaging response was significantly associated with the depth of hematological response (p=0.027). More specifically, definite response dominant WBMRI responses were associated with VGPR serum responses (OR=4.48, 95% CI: 1.34-16.23, p=0.008). No association between the type of first line of therapy and WBMRI response was observed (p=0.725). Overall, 9 pts progressed and 7 died during the follow up; among pts with FLs, 6 progressed and 6 died. Achieving at least a partial response dominant at second WBMRI was associated with significantly superior PFS outcomes (HR=0.19, 95% CI: 0.04-0.89, p=0.035), while a trend was also observed for TTP (HR=0.16, 95% CI: 0.02-1.41, p=0.098). Similarly, pts achieving at least a definite response dominant at second WBMRI seemed to have superior PFS (HR=0.33, 95% CI: 0.11-1.02, p=0.055) and significantly superior TTP (HR=0.15, 95% CI: 0.03-0.84, p=0.031). OS data are immature. Conclusion: The proposed novel criteria for evaluating early treatment response with WBMRI in NDMM pts seem to have both predictive and prognostic value. Updated results will be presented at the meeting.