Quantitative assessment of bone marrow attenuation values at MDCT: an objective tool for the detection of bone bruise related to occult sacral insufficiency fractures.

Abstract

To prove the feasibility of using Hounsfield attenuation values at MDCT to detect bone bruises related to sacral insufficiency fractures. Twenty-two patients with acute sacrum trauma and no fracture findings at MDCT were included in our prospective study. Two observers independently reviewed CTs regarding visual signs of bone bruises in 132 defined regions of the sacral alae. Interobserver agreement was tested by κ statistics. Subsequently, HU values were obtained in the same regions, and attenuation differences between the two sides were calculated. Validity and reliability were assessed by intraclass correlation coefficient and Bland-Altman analysis. HU differences were subjected to ROC curve analysis to determine sensitivity, specificity, PPV and NPV. MRI served as standard reference. MRI revealed 19 regions with bone bruises and associated sacral insufficiency fractures. HU measurements demonstrated good validity and reliability (r = 0.989). ROC curve analysis exhibited an ideal cutoff value of 35.7 HU density difference between affected and non-affected regions. Visual evaluation revealed moderate agreement (κ = 0.48); diagnostic accuracy was inferior to objective evaluation. Assessment of differences in bone marrow density by HU measurements is an objective and reliable tool for detection of bone bruises associated with occult sacral insufficiency fractures. • Bone bruising is associated with occult sacral insufficiency fractures. • Assessment of differences in bone marrow CT attenuation appears valid and reliable. • Comparative HU measurements of bone marrow allow detection of bone bruises. • Comparative HU measurements have high specificity and negative predictive values. • Comparative HU measurements may make further diagnostic workup with MRI unnecessary.