Objectives:Post-reconstruction imaging of the anterior cruciate ligament (ACL) is needed to non-invasively assess in vivo graft maturity before release to play. The purpose of this study was to evaluate autograft remodeling up to 12 months after ACL reconstruction using 3D ultra-short time to echo (UTE) T2* MRI. We hypothesized that the T2* values of ACL autografts would progressively change over time, initially resembling the values for the native contralateral graft source and gradually approaching that of the intact contralateral ACL.Methods:After IRB approval, 12 patients (ages 14-45 years) who underwent primary ACL reconstruction (ACLR) with semitendinosus (SemiT) or bone-patellar tendon-bone (BTB) autograft were enrolled. Patients with a history of prior injury or surgery to either knee were excluded. Subjects returned for UTE MRIs at 1, 3, 6, 9, and 12 months after ACLR. Imaging at 1-month included the contralateral knee. All scans were performed on a Philips Ingenia 3T system. MRI pulse sequences included high-resolution 3D T2 (slice thickness: 0.6mm, TR: 18.7ms; TE: 11.5ms) and a four-echo T2 UTE (slice thickness: 1mm, TR: 20ms; TE: 0.3, 3.3, 6.3, and 9.3ms). Using high-resolution 3D T2 sequences at 1 month, all slices containing the intra-articular ACL were segmented semi-automatically to generate volumetric regions of interest (ROIs) (Materialise, Inc.; Leuven, Belgium). ROIs were divided into proximal/distal and core/peripheral sub-ROIs using standardized methods. Each ROI was co-registered voxel-by-voxel to T2* maps also obtained at 1 month. These T2* maps were then co-registered with those of subsequent time points. The segmentation process was repeated for the 1-, 3-, and 6-month time points by a second reviewer for inter-observer reliability (κ). Statistical differences among the ROIs and sub-ROIs of ACL autografts were assessed with repeated-measures ANOVA and two-tailed non-parametric t-tests. P<0.05 represented statistical significance.Results:Twelve subjects were enrolled in this prospective study, and after 2 subjects withdrew, 10 subjects were included in the analysis (n=7 SemiT; n=3 BTB). Nine of 10 subjects attended all postoperative imaging sessions at the time of this writing. Inter-observer reliability for T2* values was found to be excellent (κ=0.832; 95% CI [0.70-0.91]; p<0.001). Average T2* relaxation times increased from 3.2ms (standard deviation [SD], 1.0) at 1 month, to 4.8ms (SD, 1.8) at 6 months, and 4.1ms (SD, 1.6) at 12 months (Figures 1 and 2), with statistically significant differences between the intact ACL and 1-month time point (p=0.01), and between the 1- and 6-month time points (p=0.004). The ratio of T2*[inj]/T2*[intact] increased from 59.2% at 1 month to 79.8% at 12 months for SemiT autografts, and from 86.9% at 1 month to 97.2% at 12 months for BTB autografts (p=0.001). Significant differences were found between the following pairs of sub-ROIs when combining all time points: proximal/distal, proximal-peripheral/proximal-core, distal-peripheral/distal-core (p<0.02 each).Conclusions:ACL autografts exhibited progressively increasing T2* values and T2*[inj]/T2*[intact] ratios up to 12 months after ACLR, and these changes appear to be region-dependent. UTE T2* MR imaging can provide unique insights into the condition of remodeling ACL grafts, and may improve our ability to non-invasively assess graft maturity before allowing patients to resume high-intensity activities.