Whole-body MRI for staging and interim response monitoring in paediatric and adolescent Hodgkin’s lymphoma: a comparison with multi-modality reference standard including 18F-FDG-PET-CT

Arash Latifoltojar, Shonit Punwani, Andre Lopes, Paul D. Humphries, Maria Klusmann, Leon Jonathan Menezes, Stephen Daw, Ananth Shankar, Deena Neriman, Heather Fitzke, Laura Clifton-Hadley, Paul Smith, Stuart A. Taylor

To prospectively investigate concordance between whole-body MRI (WB-MRI) and a composite reference standard for initial staging and interim response evaluation in paediatric and adolescent Hodgkin's lymphoma. Fifty patients (32 male, age range 6-19 years) underwent WB-MRI and standard investigations, including 18F-FDG-PET-CT at diagnosis and following 2-3 chemotherapy cycles. Two radiologists in consensus interpreted WB-MRI using prespecified definitions of disease positivity. A third radiologist reviewed a subset of staging WB-MRIs (n = 38) separately to test for interobserver agreement. A multidisciplinary team derived a primary reference standard using all available imaging/clinical investigations. Subsequently, a second multidisciplinary panel rereviewed all imaging with long-term follow-up data to derive an enhanced reference standard. Interobserver agreement for WB-MRI reads was tested using kappa statistics. Concordance for correct classification of all disease sites, true positive rate (TPR), false positive rate (FPR) and kappa for staging/response agreement were calculated for WB-MRI. There was discordance for full stage in 74% (95% CI 61.9-83.9%) and 44% (32.0-56.6%) of patients against the primary and enhanced reference standards, respectively. Against the enhanced reference standard, the WB-MRI TPR, FPR and kappa were 91%, 1% and 0.93 (0.90-0.96) for nodal disease and 79%, < 1% and 0.86 (0.77-0.95) for extra-nodal disease. WB-MRI response classification was correct in 25/38 evaluable patients (66%), underestimating response in 26% (kappa 0.30, 95% CI 0.04-0.57). There was a good agreement for nodal (kappa 0.78, 95% CI 0.73-0.84) and extra-nodal staging (kappa 0.60, 95% CI 0.41-0.78) between WB-MRI reads CONCLUSIONS: WB-MRI has reasonable accuracy for nodal and extra-nodal staging but is discordant with standard imaging in a substantial minority of patients, and tends to underestimate disease response. • This prospective single-centre study showed discordance for full patient staging of 44% between WB-MRI and a multi-modality reference standard in paediatric and adolescent Hodgkin's lymphoma. • WB-MRI underestimates interim disease response in paediatric and adolescent Hodgkin's lymphoma. • WB-MRI shows promise in paediatric and adolescent Hodgkin's lymphoma but currently cannot replace conventional staging pathways including 18 F-FDG-PET-CT.