Abstract
PURPOSE: Ultrashort echo time (UTE) MRI enables direct imaging of cortical bone and quantification of its water compartments via bicomponent T(2)* modeling. However, conventional approaches require multiple separate dual-echo scans due to limitations in gradient power. This approach is prone to inter-scan inconsistencies such as motion and signal drift, which degrade fitting accuracy. This study proposes an interleaved dual-echo acquisition sequence that acquires multiple echo time (TE) images in a single scan to improve bicomponent T(2)* quantification in cortical bone. METHODS: The proposed UTE sequence utilizes interleaved dual-echo acquisitions with flexible TE spacings. This sequence was tested on five healthy subjects' tibial midshafts and compared to conventional separate dual-echo scans with and without image registration. Bicomponent T(2)* modeling was performed, and fitting accuracy was evaluated using normalized-root-mean-squared error (NRMSE). Three subjects were scanned three times to evaluate the scan repeatability. RESULTS: The interleaved method significantly reduced NRMSE (3.2% ± 2.3% vs. 6.2% ± 3.1%, p = 0.0231) and yielded lower and more stable T(2)* (T(2s)*; 0.50 ± 0.10 ms vs. 0.76 ± 0.13 ms, p < 0.0001) and fraction (F(s); 78.2 ± 5.1 vs. 84.2% ± 7.1%, p = 0.0006) of short T(2) components compared to separate scans without registration. Image registration had a minimal improvement on mapping results for separate scans. Parameter maps from the interleaved scans confirmed more homogeneous distributions of T(2s)* and F(s) with lower fitting errors. The much lower coefficients of variance of the interleaved scans demonstrated improved repeatability compared with separate scans. CONCLUSION: The proposed interleaved UTE dual-echo sequence improves the robustness of bicomponent T(2)* mapping of the cortical bone by reducing inter-scan inconsistencies.