Current ultrasound techniques face several challenges to measure strains when translated from large tendon to in-situ knee collateral ligament applications, despite the potential to reduce knee arthroplasty failures attributed to ligament imbalance. Therefore, we developed, optimized and validated an ultrasound speckle tracking method to assess the in-situ strains of the medial and lateral collateral ligaments. Nine cadaveric legs with total knee implants were submitted to varus/valgus loading and divided into two groups: “optimization” and “validation”. Reference strains were measured using digital image correlation technique, while ultrasound data were processed with a custom-built speckle tracking approach. Using specimens from the “optimization” group, several tracking parameters were tuned towards an optimized tracking performance. The parameters were ranked according to three comparative measures between the ultrasound-based and reference strains: R², mean absolute error and strains differences at 40 N. Specimens from the “validation” group, processed with the optimal parameters, showed good correlations, along with small mean absolute differences, with correlation values above 0.99 and 0.89 and differences below 0.57% and 0.27% for the lateral and medial collateral ligaments, respectively. This study showed that ultrasound speckle tracking could assess knee collateral ligaments strains in situ and has the potential to be translated to clinics for knee arthroplasty-related procedures.

Inroduction This cadaveric study aimed to assess the accuracy of the Next-AR navigation platform in terms of implant alignment, kinematics, and collateral ligament elongations in comparison to gold standard measurement tools. Materials and methods Computed tomography of 7 fresh-frozen legs were acquired for patient-specific guide design, identification of bony landmarks and collateral ligaments to acquire real-time feedback for bone resection, alignment and collateral ligament elongations using Next-AR system. The specimens underwent medially-stabilized TKA (Medacta) based on the mechanical alignment technique. The difference between target and achieved implant alignment parameters was assessed. The relative change in tibiofemoral kinematics, medial collateral ligament- (MCL), and lateral collateral ligament (LCL) elongations acquired with the Next-AR system were compared to reference values acquired with bone-pin mounted motion trackers (Vicon, UK) while performing quasistatic squatting in an ex vivo knee simulator. Results The mean differences between achieved and target hip-knee-ankle angle, femoral varus, femoral flexion, femoral rotation, tibial varus, and tibial slope measured -1.5°, 0.5°, 0.07°, -0.14°, 0.79° and -0.36°, respectively, with no statistically significant differences found (p ≥ 0.12). Knee flexion (p=0.32) and valgus orientation (p=0.63), obtained from Next-AR and Vicon systems were comparable. Collateral ligament elongations derived from Next-AR exhibited similar values for the anterior bundle of MCL (p=0.18) and LCL in the anterior and middle bundles (p>0.3) obtained from the Vicon system. Conclusions Our findings showed that the Next-AR system achieved postoperative implant alignment close to the planned targets, with small systematic deviation (mean bias ≤ 1.5°), limited total error (RMSE ≤ 2.3°), and variability of paired differences ≤ 1.87°, while additionally providing quantitative real-time assessment of tibiofemoral kinematics and collateral ligament elongations.