Abstract
BACKGROUND: Whole-body MRI is used for staging paediatric Hodgkin lymphoma, commonly using size thresholds, which fail to detect disease in normal-size lymph nodes. OBJECTIVE: To investigate quantitative whole-body MRI metrics for nodal characterisation. MATERIALS AND METHODS: Thirty-seven children with Hodgkin lymphoma underwent 1.5-tesla (T) whole-body MRI using short tau inversion recovery (STIR) half-Fourier-acquisition single-shot turbo-spin-echo and diffusion-weighted imaging (DWI). (18)Flourine-2-fluoro-2-deoxyglucose (FDG) positron emission tomography (PET)/CT was acquired as the reference standard. Two independent readers assessed 11 nodal sites. The readers measured short-axis-diameter, apparent diffusion coefficient, (ADC) and normalised T2-signal intensity of the largest lymph node at each site. We used receiver operating characteristics (ROC)/area-under-the-curve (AUC) analysis for each MRI metric and derived sensitivity and specificity for nodes with short-axis diameter ≥10 mm. Sub-analysis of sensitivity and specificity was performed with application of ADC cut-off values (<0.77, <1.15 and <1.79×10(-3) mm(2) s(-1)) to 5- to 9-mm nodes. RESULTS: ROC/AUC values for reader 1/reader 2 were 0.80/0.80 and 0.81/0.81 for short-axis-diameter measured using DWI and STIR half-Fourier-acquisition single-shot turbo spin echo, respectively; 0.67/0.72 for normalised T2 signal intensity and 0.74/0.67 for ADC. Sensitivity and specificity for a short-axis diameter ≥10 mm were 84.2% and 66.7% for Reader 1 and 82.9% and 68.9% for Reader 2. Applying a short-axis-diameter ≥10-mm threshold followed by ADC cut-offs to normal-size 5- to 9-mm nodes resulted in sensitivity and specificity for Reader 1 of 88.8% and 60%, 92.1% and 56.7%, and 100% and 16.7%; and for Reader 2, 86.1% and 67.2%, 95.3% and 65.6%, and 100% and 19.7%; and ADC thresholds of <0.77, <1.15, and <1.79×10(-3) mm(2) s(-1), respectively. CONCLUSION: Nodal size measurement provides the best single classifier for nodal disease status in paediatric Hodgkin lymphoma. Combined short-axis diameter and ADC thresholds marginally improve sensitivity and drop specificity compared with size classification alone.