Multiple myeloma Cancer
Conditions
Interventions
Current methodology as of 15/06/2026:
WISER-M is a multi-centre, multi-cohort retrospective study involving collection and analyses of routinely acquired whole-body MRI scans and clinical diagnostic
Sponsors
Institute of Cancer Research
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: Patients with a confirmed diagnosis of multiple myeloma who underwent a whole-body MRI scan as standard of care during their diagnostic (Training/Internal Validation & External Validation Cohorts) and/or treatment (Training/Internal Validation Cohort only) pathway, within the preceding 8 years, fulfilling all of the following: 1. WBMRI scan acquisition protocol included axial-diffusion-weighted MRI with at least 2 b-values 2. WBMRI scan acquisition protocol included axial Dixon T1-weighted MRI 3. WBMRI scan showed at least two radiographically active focal lesions defined as high signal on diffusion-weighted imaging, low ADC values (500-10000s/mm²) and low relative fact fraction (< 20%)
Exclusion criteria
Exclusion criteria: Inadequate whole-body MRI image quality
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Per-lesion sensitivity of the software to identify focal lesions, compared to a “gold standard” reading provided by a central expert radiologist. Per-lesion sensitivity of the software is defined as the proportion of lesions identified as active focal lesions by the software amongst the total of true focal lesions as defined by the central expert radiologist. This will be reported for each cohort | — |
Secondary
| Measure | Time frame |
|---|---|
| 1. Per-lesion positive predictive value (precision) of the software to identify focal lesions, compared to a “gold standard” (ground truth) reading provided by a central expert radiologist. Per lesion positive predictive value of the software is defined as the proportion of true focal lesions, as determined by the central expert radiologist, amongst all lesions identified as active focal lesions by the software. This will be reported for each cohort 2. Overall disease burden as measured by the total diffuse disease volume (TDV) and apparent diffuse coefficient (ADC) metrics. Overall disease burden will be measured by the total diffuse disease volume (TDV) and apparent diffuse coefficient (ADC) metrics. Quantitative analysis of the diffusion-weighted MRI parameters will be performed for each cohort. Existing software will calculate ADCs using an iterative weighted linear least-square approach. Code for the fitting algorithm will be integrated as a plugin into the OsiriX radiological viewing platform. The region of interest (ROI) will be segmented semi-automatically using highest b-value (b750s/mm2) with reference to T2- weighted images. The ROI will be then transferred to the corresponding ADC map at each timepoint, and first-order histogram measures will be derived from the enclosed ADC voxel values, including mean, median, variance, skewness, kurtosis, and selected percentiles (dwMRI parameters). This will be reported for each cohort 3. Degree of perceived overall helpfulness as measured on a 4-point Likert scale (1=not helpful at all, 4=extremely helpful) (Clinical Usability Sub-Study) | — |
Countries
England, United Kingdom
Contacts
Public ContactNatasha Iles
Outcome results
None listed