Objectives: Bone contusions consist of posttraumatic marrow change resulting from a combination of hemorrhage, edema, and microtrabecular injury and are depicted on MR images as geographic, nonlinear areas of abnormally increased T2 marrow signal intensity. Detection of bone contusions is important, not only in that they are frequently a sign of more serious associated injury, but in that isolated contusions may account for clinical symptoms and thereby obviate further workup. We evaluated two promising rapid imaging sequences, T2-weighted fast spin-echo with fat saturation and fast spin-echo short inversion time (T1) inversion recovery (STIR) and compared their efficacy to that of conventional STIR for the diagnosis of bone contusions in knees.
Subjects and methods: Fifty-one consecutive military, retired military, and military-dependent patients were prospectively entered. Imaging consisted of three experimental sequences: coronal T2-weighted fast spin-echo with spectroscopic fat saturation, fast spin-echo STIR, and STIR, as well as conventional T1, proton density, and T2. Each examination was interpreted by three radiologists, for a total of 153 interpretations. Each interpretation consisted of the conventional sequences plus one of the three experimental sequences, blinded to the other two, specifically addressing menisci, ligaments, bones, and soft tissues. A positive test result was defined as a diagnosis of bone contusion on any one of the three independent interpretations of a given examination. Our gold standard was the consensus opinion of all three radiologists in a retrospective, unblinded evaluation of all imaging sequences. Sensitivity and specificity of fast spin-echo with fat saturation and fast spin-echo STIR were then calculated and compared with STIR.
Results: Sensitivity and specificity for bone contusion diagnosis in the three experimental sequences were as follows: fast spin-echo with fat saturation, 91% sensitivity and 86% specificity; fast spin-echo STIR, 96% sensitivity and 96% specificity; and conventional STIR, 83% sensitivity and 93% specificity. Our results showed no statistically significant difference between either of the fast spin-echo sequences and conventional STIR (p > .05).
Conclusion: Because of their comparable accuracy and short imaging time characteristics, coronal fast spin-echo with fat saturation and fast spin-echo STIR sequences are superior to conventional STIR sequences for the diagnosis of bone contusions in knees.