Utilizing magnetic resonance imaging to differentiate osteomyelitis and vaso-occlusive crisis in pediatric sickle cell patients: a diagnostic guide

Abstract Background Differentiating between acute osteomyelitis and vaso-occlusive crisis in pediatric sickle cell disease (SCD) patients is challenging due to overlapping clinical and imaging features. Magnetic resonance imaging (MRI) provides a detailed assessment of bone and soft-tissue changes, which may enhance diagnostic accuracy. Purpose To determine the effectiveness of MRI in differentiating acute osteomyelitis (OM) from vaso-occlusive crisis (VOC) in pediatric patients with sickle cell disease (SCD), by analyzing specific MRI imaging features and correlating them with combined MRI features and clinical/laboratory findings. Methods In this prospective study, sixty children with SCD (37 males, 23 females; mean age 7.5 ± 3.2 years) presenting with acute bone pain underwent MRI of the painful region on a 1.5 Tesla (T) scanner between December 2024 and May 2025. Imaging protocols included T 1-weighted, T 2-weighted, fat-suppressed, and contrast-enhanced T 1-weighted sequences in multiple planes. Two blinded musculoskeletal radiologists independently evaluated marrow signal characteristics, cortical integrity, periosteal reaction, bone-marrow and subperiosteal fluid collections, joint effusions, soft-tissue inflammation, abscess formation, and enhancement patterns. Definitive diagnoses were made by combining MRI findings with clinical presentation, laboratory markers, and image-guided aspiration ( n = 12). Statistical analyses included Pearson’s chi-square test and Cohen’s kappa coefficient, with significance set at P value < 0.05. Results Integration of all data classified 18 patients (30%) with OM and 42 patients (70%) with VOC. Cortical defects were observed in 55.6% of OM cases versus 4.8% of VOC ( P < 0.01). Bone-marrow fluid collections occurred in 22.2% of OM and 0% of VOC ( P < 0.01). Subperiosteal fluid collections were present in 66.7% of OM and 33.3% of VOC ( P < 0.05). Soft-tissue abscesses were seen exclusively in OM (44.4%), with P < 0.001. Geographically heterogeneous marrow enhancement was more common in OM (66.7% vs. 14.3% of VOC, P < 0.01). Non-specific findings such as periostitis and joint effusion did not differ significantly between groups. Magnetic resonance imaging demonstrated a sensitivity of 88.9%, specificity of 85.71%, and overall accuracy of 96.7% for OM, with perfect inter-observer agreement ( κ = 0.918). Conclusion Magnetic resonance imaging is a powerful tool for distinguishing OM from VOC in pediatric SCD, with features such as cortical disruption, intramedullary and subperiosteal fluid collections, and adjacent soft tissue abscesses serving as reliable indicators of infection. A definitive diagnosis requires careful integration of these imaging criteria, with clinical and laboratory data, to enable early and accurate diagnosis, guiding appropriate management and improving patient outcomes.