Abstract
Background: Traditionally, bone marrow aspirate/biopsy (BM-Bx) has been used for diagnosis and quantification of the extent of disease as well as response to therapy in MM patients. This information is used to make decision for treatment initiation. Since MM is not a contiguous disease, marrow involvement can be patchy and BM-Bx may be misleading in assessing the true extent of the disease. BM-MRI is a non-invasive technique that can evaluate a large amount of marrow for tumor infiltration. We prospectively investigated the sensitivity of BM-MRI and compared its results with those obtained by BM-Bx as well as with the clinical stage of disease.
Methods: All patients with the diagnosis of MM who had BM-MRI at our center were evaluable. Patients must have received a BM-Bx within 4 weeks of the BM-MRI. In these patients sagittal T1 and fast spin echo inversion recovery sequences of the cervical, thoracic and lumbosacral spine and coronal T1 and fast spin echo inversion recovery sequences of the sacrum and pelvic bones were reviewed on the MRI. Durie-Salmon (DS) staging criteria were used for correlation. To study the statistical relationship between pairs of ordinal variables the test corresponding to the Spearman correlation was used. To study the statistical relationship between nominal and ordinal variables, the Wilcoxon or Kruskal-Wallis test was used. A 0.05 nominal significance level was used in all testing. Following staging system was defined for evaluation of the involvement of the marrow by BM-MRI: A (0%), B (< 10%), C (10%–25%), D (26%–50%), E (> 50%). Results of this were then compared with the extent of involvement reported on histological evaluation of the BM-Bx.
Results: A total of 50 patients (23 females and 27 males) were identified. Median age was 61.5 years (range 35–82 years) with 23 (46%) having stage IIIA disease. As per the staging system defined above, 6%, 10%, 10%, 22% and 52% of the patients had categories A, B, C, D and E involvement as per BM-MRI, respectively. Similarly, involvement observed on the BM-Bx was 8%, 24%, 22%, 16% and 30%, respectively. Categories of marrow involvement on BM-Bx and BM-MRI were concordant in 23 (46%) and discordant in 27 (54%) patients. Of the patients that showed discordance, 89% had a more extensive BM involvement detected by the BM-MRI and 11% had a higher reading on BM-Bx. The estimated Spearman correlation coefficient between MRI involvement and MM stage was 0.4849 (95% CI; 0.2494, 0.7205), showing a significant association between BM-MRI involvement and MM stage (p =0.0002). The estimated Spearman correlation coefficient between BM-Bx involvement and MM stage was 0.1775 (95% CI; −0.1406, 0.4956), showing no significant association between BM-Bx involvement and MM stage (p = 0.2764).
Conclusions: We demonstrate for the first time that BM-MRI is a more sensitive technique to assess the true disease burden in MM and is significantly better than BM-Bx. We also observe that the extent of marrow infiltration noted on the BM-MRI correlates significantly with other prognostic characteristics like DS stage. Based on this observation we recommend that BM-MRI should be considered as part of the pre-treatment evaluation of patients with multiple myeloma.
Effective blood signal suppression using double inversion-recovery and slice reordering for multislice fast spin-echo MRI and its application in simultaneous proton density and T2 weighted imaging
