MR imaging data were acquired from patients participating in the
“Bridge Enhanced ACL Repair (BEAR) I” (NCT02292004; IRB-P00012985; mean age 24.1 ± 4.9 years,
40% male) and “BEAR II” (NCT02664545; IRB-P00021470; mean age 19.4 ± 5.1 years,
42% male) clinical trials evaluating a new ACL repair procedure.28 (link)–30 (link) MR scans of the contralateral intact ACL
were obtained at one timepoint for the BEAR I (24 months post-surgery) and at up
to three timepoints for the BEAR II (6, 12, and 24 months post-surgery)
patients. After accounting for patients lost to follow up and removing patients
with prior or subsequent contralateral ACL tears, a total of 246 sagittal CISS
scans (FA=35°; TR=12.78ms; TE=6.39ms; FOV=140mm; 384x384 acquisition
matrix with voxel size 0.365mm x 0.365mm x 1.5mm) of intact ACLs were included
in the combined dataset. All MR images were acquired on a 3T TRIO (Siemens;
Erlangen, Germany) with a 15-channel transmit/receive knee coil (Siemens). Data
were randomly divided into training (~70%; n=171), validation
(~20%; n=46), and test (~10%; n=29) sets. This split was
stratified by subject to ensure that the model would not learn and make
predictions on MR images from the same subject.