Non-invasive Detection and Assessment of Treatment Response in Multiple Myeloma using Efficient Whole-body MRI
Introduction
Multiple myeloma (MM) is the second most common hematological malignancy and is characterized by significant morbidity that lowers the patient’s quality of life. Almost all MM patients develop bone lesions which lead to unremitting pain, hypercalcemia, and increased incidence of fractures. Whole-body [18F] fluorodeoxyglucose positron emission tomography (FDG-PET) is often used for the assessment of lesions in MM, but the International Myeloma Working Group (IMWG) currently recommends whole-body MRI (WBMRI) as the preferred imaging modality for pretreatment assessment of MM. Current MRI sequences such as Short Tau Inversion Recovery (STIR) and Diffusion Weighted Imaging with Background Suppression (DWIBS) demonstrate high lesion contrast, but they suffer from geometric distortions limiting lesion localization and prolonged acquisition times. To address these shortcomings, we have developed a novel dual-echo T2-weighted acquisition for enhanced conspicuity of tumors (DETECT). We have an ongoing research study to demonstrate the improved lesion detection capability of WBMRI with DETECT compared to DWIBS and STIR.
Methods
Patients: We recruited adult patients with diagnosed MM and confirmed bone lesions. Patients were imaged using WBMRI and FDG-PET/CT. WBMRI was performed on a 3T Ingenia MR scanner (Philips Healthcare) at five anatomical stations – Head & Neck, Chest, Abdomen, Pelvis, and Upper Extremities – using the following sequences: STIR, DWIBS, DETECT, along with pre- and post-contrast 3D T1-mDixon. Approximate scan times for WBMRI were, DETECT = 7:00 min; STIR = 17:43 min; DWIBS = 18:40 min; T1-mDixon = 1:05 min each for pre-contrast and post-contrast. FDG-PET/CT imaging was performed on a Discovery MI (GE Healthcare) within 1 week of WBMRI. Image Analysis: A radiologist (A.C.) with 20 years of experience identified lesions on all MR images. Another radiologist (O.O.) with 31 years of experience identified lesions on FDG-PET/CT. Quantitative fat fraction (FF) maps from DETECT and apparent diffusion coefficient (ADC) maps from DWIBS were calculated. SUVmax values were calculated using subject dose information. Correlations were performed using simple linear regression between the following pairs: FF and ADC; FF and SUVmax; ADC and SUVmax in lesions identified at baseline.
Results
To date, 17 MM patients have been enrolled and 8 MM patients have been evaluated for this study. WBMRI-DETECT has better lesion conspicuity than STIR and DWIBS and significantly less geometric distortions than DWIBS, while also requiring shorter scan times. Approximately 40 lesions were identified across the eight subjects. Examples of lesions found in WBMRI or FDG-PET have demonstrated that WBMRI identified more lesions than FDG-PET in situations where both modalities have been clinically evaluated. Biomarker maps of ADC, FF, and SUVmax show examples of identified lesion. Linear regression of ADC and SUVmax in 21 lesions showed a slight negative correlation, while the correlation between FF and SUVmax in 23 lesions showed a stronger negative correlation.
Conclusion
Our results demonstrated that WBMRI-DETECT provides higher SNR, improved lesion conspicuity, and minimal geometric distortions in shorter scan times compared to STIR and DWIBS. More lesions were uniquely identified in WBMRI when directly compared to FDG-PET in the same subjects. Quantitative FF from WBMRI showed a weak correlation with ADC, suggesting both biomarkers provide complementary information. Both ADC and FF showed negative correlation with SUV values, demonstrating potential use of quantitative MRI biomarkers for the assessment of therapy response. Our efficient WBMRI-DETECT along with contrast-enhanced T1-weighted mDixon can be reliably performed in less than 30 minutes of table time and shows promise as a robust and efficient protocol for comprehensive treatment planning and assessment of therapy response in MM.