Results of a Prospective Study Comparing Whole-Body Diffusion-Weighted Magnetic Resonance Imaging with Skeletal X-Ray and Magnetic Resonance of the Spine for Assessing Bone Disease in Multiple Myeloma.

Abstract 2913 Introduction: Standard assessment of bone disease in multiple myeloma (MM) is based on skeletal X-ray (XR) and magnetic resonance (MR) of the spine (MRS). Diffusion-weighted MR (DW-MR) is a novel functional MR that detects changes of water diffusion through cells in tissues. To assess the value of DW-MR to detect bone lesions in MM, we designed a prospective study comparing whole-body DW-MR with XR and MRS. The study included symptomatic patients (pts) at diagnosis or at relapse before the start of the treatment; they performed XR, MRS, conventional whole-body MR (WB-MR), and whole-body DW-MR at enrolment (time point 1, T1), after treatment (T2), and after 6 months of follow-up (T3). Clinical and hematologic, including bone marrow (BM), disease evaluations were done at the same time points. The study was approved by the Institutional Review Board in 2008 (protocol 44/08). Methods: The primary objective was to assess whether DW-MR could detect more focal lesions (FL) than XR and MRS. Secondary objectives were to correlate the changes of FL detected by DW-RM with response, to assess the prognostic value of DW-RM, and to compare DW-MR with WB-MR. MRS, WB-MR and DW-MR were done in a single 45-minute session on a standard 1.5 Tesla MR scanner. DW-MR consisted of multiple stacked axial Echo Planar Imaging sequences at 4 b-values, evaluated by PET-like Maximum Intensity Projection and Multi-Planar reconstructions at the highest b-value (1000). Each exam was independently read by 3 radiologists experienced in MM. 53 bone segments per exam were evaluated in whole-body imaging (XR, WB-MR and DW-MR); 25 segments were evaluated in spine imaging (MRS and DW-MR). All the patterns (focal, diffuse, mixed, and salt-and-pepper) of bone lesions were recorded. Matching FL detected by >=2 radiologists were counted for the present analysis. Statistics were carried out with the Wilcoxon signed rank test for methods comparisons and the Kruskal-Wallis test to assess intra-patient changes through the time points. Survival and relapse were analyzed by Kaplan-Meier and Cumulative Incidence method with log-rank and Gray's tests. All tests were 2-sided. Results: Between 2008 and 2010, 36 symptomatic pts were enrolled: 43% were at diagnosis, 57% at relapse; 71% of pts had ISS stage 1 MM. The most frequent isotype was IgG (57%), median BM infiltration was 30%. FISH on selected CD138+ plasma cells detected t(4;14) and del(17) in 9 and 6% of pts. At T1, the DW-MR detected more FL than standard XR (306 vs 117 FL, p<0.01), WB-MR (306 vs 225 FL, p=0.02), and MRS (165 vs 116 FL, trend, p=0.08). At T2, a similar number of FL was detected by DW-MR and XR (97 vs 104 FL, p=0.99) and MRS (20 vs 20 FL, p=1.00); DW-MR detected more FL than WB-MR (97 vs 60 FL, p=0.01). At T3, the DW-MR detected more FL than WB-MR (88 vs 45 FL, p<0.01) and MRS (24 vs 11 FL, p=0.05), and similar FL compared to XR (88 vs 62 FL, p=0.27). Considering all the time points, the DW-MR detected more FL than XR (p=0.01), WB-MR (p<0.01) and MRS (p=0.02). Between T1 and T2, all pts were treated with IMIDS or bortezomib–based regimens, 33% underwent a stem cell transplant. Overall response rate (ORR) was 73%. DW-MR detected significant changes of FL according to disease response at T2 (from 79 to 15 FL in >=VGPR, from 69 to 27 in PR, and from 34 to 55 FL in SD or PD, p=0.04 [whole body]; p=0.02 [spine]). Also MRS consistently detected response (p=0.04), whereas WB-MR showed only a weak correlation (p=0.13); XR did not detect response (p=0.55). Between T2 and T3, pts had minor changes of disease status (72% ORR), and, accordingly, all the radiological exams did not show significant changes in FL. One-, 2- and 3-year progression-free survival (PFS) was 80, 62 and 37% (median, 30 months), OS was 88, 79 and 76% (median not reached), and relapse incidence was 15, 32, and 54% (median, 21 months). Since the median number of FL detected by DW-MR at T1 was 4 (range, 0–49 FL), we compared PFS, relapse, and OS by the presence of <=4 FL or >4 FL before treatment. Patients with <=4 FL at DW-MR had better PFS (72 vs 50% at 2 years, p=0.02) and less relapse incidence (17 vs 50%, p<0.01) than those with >4 FL, whereas OS was not different (84 vs 75%, p=0.76). Conclusions: DW-MR is superior to XR, MRS, and WB-MR in detecting FL in MM. The number of FL detected by DW-MR before treatment predicts PFS and relapse incidence. DW-MR is a functional imaging that effectively detects the bone disease changes according to treatment response and can be used to monitor disease response. Disclosures: No relevant conflicts of interest to declare.