The Influence of Contrast Media on Calcium-based Imaging of the Spine in Dual-layer CT

Abstract

Objectives To evaluate the impact of contrast media application on a novel calcium-only imaging technique (VCa) using dual-layer spectral detector CT (DLCT) and to expand the applicability of VCa within CT-based opportunistic osteoporosis screening. Methods True non-contrast (TNC) and venous phase images (VP) of n = 97 patients were acquired with a DLCT scanner. CT attenuation of the first lumbar vertebra (L1) was measured in TNC-VCa, venous phase (VP-VCa), and in virtual non-contrast images (VNC). The difference in L1 CT attenuation between TNC-VCa and VP-VCa was determined. A statistical model for the prediction of TNC-VCa L1 CT attenuation values was established and the capacity of VP-VCa L1 CT attenuation to detect osteoporosis was examined. Results CT attenuation was significantly higher in VP-VCa than in TNC-VCa (p < 0.001). Regression analyses revealed a strong linear association between TNC-VCa and VP-VCa CT attenuation values (R²=0.84). The final statistical model for the prediction of TNC-VCa CT attenuation (TNC-VCa[HU]=-6.81 + 0.87*VP-VCa[HU]-0.55*body weight[kg]) yielded good agreement between observed and predicted values. A L1 CT attenuation threshold of 293 HU in VP-VCa showed a sensitivity of 90% and a specificity of 96% for detecting osteoporosis. Conclusion The application of contrast media leads to an overestimation of L1 CT attenuation in VCa. However, CT attenuation values from VP-VCa can be used to reliably predict TNC-VCa CT attenuation. VCa can be used in the context of contrast-enhanced CT acquisitions and might help to improve the precision of CT-based opportunistic osteoporosis screening in clinical routine.