MRI versus bone scintigraphy. Evaluation for diagnosis and grading of stress injuries

This study was set up to determine the value of magnetic resonance imaging (MRI) and bone scintigraphy (BS) for the diagnosis of stress injuries in athletes, and furthermore to assess reliability and prediction of healing time.

Imaging data was analyzed retrospectively from 28 athletes who had received MRI and BS examinations for suspected stress injuries. MRI- and BS-data were rated by three specialists each in a blinded read, using a 5-point score (i.e. 0-4: inconspicuous to high-grade stress fracture). An interdisciplinary expert truth-panel set the reference standard. Standard statistical parameters, Fleiss' kappa (κ), and group comparisons were calculated.

The sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) for detection of stress injuries were 71.4%, 85.7%, 78.6%, 83.3% and, 75.0%, for MRI and 92.9%, 73.8%, 83.3%, 78.0% and, 91.2% for BS, respectively. Interobserver reliability for the diagnosis of a stress injury was κ = 0.9 for BS and κ = 0.85 for MRI. Mean healing times of mild (grades 1 and 2) and severe (grades 3 and 4) stress injuries were 88 days (d) versus 142d for BS and 57d versus 116d for MRI. No significant difference in healing time could be shown.

MRI and BS reliably detect stress injuries. MRI is to be recommended as the primary imaging modality due to its potential for assessment of differential diagnoses and the lack of radiation exposure, the value of BS lies in the exclusion of stress fractures after inconclusive MRI examinations.