Francesco Pacchiano, Filippo Arrigoni, Chiara Doneda, Jacopo Carlo Poli, Giana Izzo, Mario Tortora, Irina Stavarache, Ferdinando Caranci, Cecilia Parazzini, Andrea Righini
Incorporating BB into pediatric lumbar MRI protocols improved interreader agreement and diagnostic confidence while markedly reducing equivocal interpretations. Although diagnostic performance metrics improved, these findings should be interpreted cautiously given the limited CT correlation. BB may represent a practical radiation-free adjunct to standard MRI for evaluating suspected pediatric spondylolysis.
INTRODUCTION: Lumbar spondylolysis is a common cause of low back pain in adolescents, yet early and confident diagnosis often requires CT, exposing patients to radiation. The Black-Bone (BB) MR sequence is a 3D low-flip-angle gradient-echo (GRE) acquisition designed to enhance bone-soft tissue contrast. This study assessed whether adding BB to a standard pediatric lumbar MRI protocol improves diagnostic accuracy, interreader-agreement, and confidence.
METHODS: Forty-four adolescents (mean age, 12 years; 23 girls, 21 boys) with suspected spondylolysis underwent lumbar MRI including BB, with radiographs and/or CT as reference standard. Thirty patients had confirmed spondylolysis and 14 served as controls. Four blinded readers (2 senior neuroradiologists, 2 residents) independently reviewed all cases twice: first with standard MRI only, then with BB included. Readers scored each pars interarticularis level and side (0 = negative, 1 = probable, 2 = definite). Interrater agreement (weighted Cohen's κ), sensitivity, specificity, PPV, NPV, and accuracy were calculated.
RESULTS: BB improved interrater agreement across all pairs (κ 0.81-0.93 vs. 0.59-0.82 pre-BB). Sensitivity increased for all readers (+ 5-12% points), while specificity remained high (0.89-1.00). PPV was consistently excellent (0.93-1.00), and NPV improved with BB (0.83-0.87 vs. 0.75-0.83 pre-BB). Ambiguous ratings dropped from 58 pre-BB to 5 post-BB, resolving > 90% of indeterminate assessments. Accuracy increased by approximately 2-5% points, whereas diagnostic confidence showed a larger improvement.
CONCLUSION: Incorporating BB into pediatric lumbar MRI protocols improved interreader agreement and diagnostic confidence while markedly reducing equivocal interpretations. Although diagnostic performance metrics improved, these findings should be interpreted cautiously given the limited CT correlation. BB may represent a practical radiation-free adjunct to standard MRI for evaluating suspected pediatric spondylolysis.