Quantitative Investigation of Tibiofemoral Cartilage Early Degeneration After ACL Reconstruction: An Integrated Biomechanical and Longitudinal MRI Analysis

Abstract

Background:

Anterior cruciate ligament reconstruction (ACLR) restores stability but is often followed by early cartilage degeneration. The contribution of altered dynamic contact pressure during gait to this degeneration remains poorly understood.

Purpose:

To investigate the biomechanical mechanisms underlying early cartilage degeneration after ACLR, with a focus on dynamic contact pressure distribution during gait.

Study Design:

Controlled laboratory study.

Methods:

In a 3-year longitudinal magnetic resonance imaging (MRI) study of patients with ACLR (n = 30), cartilage thickness of the tibial plateau was quantitatively assessed using deep-learning-based 3-dimensional segmentation techniques. In parallel, cadaveric knees (n = 8) were tested under intact, ACL-deficient, and ACL-reconstructed conditions using a 6 degree-of-freedom robotic simulator replicating gait cycles. Tibiofemoral contact pressure and pressure center trajectories were recorded using pressure-sensitive film.

Results:

Quantitative MRI analyses revealed significant cartilage thinning in the posterior tibial subregion 3 years after ACLR. On the medial plateau, the central medial tibia, internal medial tibia, and posterior medial tibia subregions exhibited mean reductions of 10% (P = .027), 21% (P = .019), and 13% (P = .041), respectively. On the lateral plateau, significant decreases were observed in central lateral tibia (9%; P = .031), posterior lateral tibia (14%; P = .029), and external lateral tibia (15%; P = .035). In robotic gait simulations, the reconstructed knees exhibited persistent posterior displacement of the whole-gait-cycle contact center of stress on both tibial plateaus (residual shifts: medial, 5.29 mm; lateral, 4.73 mm; both P < .05). Additionally, the contact area was significantly enlarged in early stance (2%-12% gait cycle) and terminal swing (75%-100%), especially in the medial compartment.

Conclusion:

ACLR-induced pressure center displacement coincides with focal posterior cartilage degeneration, forming a spatiotemporal mechanical-pathological chain. This work highlights the potential of dynamic loading biomarkers for early osteoarthritis risk stratification and targeted mechanical intervention.

Clinical Relevance:

Persistent shifts in tibiofemoral pressure centers after ACLR coincide with focal cartilage thinning, suggesting a mechanical pathway to post-ACLR osteoarthritis. Identifying such dynamic loading biomarkers may guide early risk stratification and targeted interventions in sports medicine.

Keywords

anterior cruciate ligament reconstruction biomechanics cartilage degeneration gait simulation magnetic resonance imaging tibial cartilage thickness

Anterior cruciate ligament (ACL) injury is one of the most common sports-related injuries in young adults,¹³ and ACL reconstruction (ACLR) is widely performed to restore knee joint stability.^39,40 However, accumulating evidence indicates that even after successful reconstruction, patients may experience progressive cartilage degeneration within a few years postoperatively.^10,19,26 This suggests that restoring mechanical stability alone does not fully prevent degenerative changes. The mechanisms underlying this cartilage degeneration remain incompletely understood.

Quantitative imaging studies of post-ACLR cartilage degeneration have largely relied on magnetic resonance imaging (MRI) techniques such as T1ρ and T2 mapping,^23,24,41 which indirectly reflect biochemical changes—T1ρ is sensitive to early proteoglycan loss,⁵ and T2 mapping to collagen network integrity.¹² Although these methods demonstrate sensitivity to early cartilage degeneration, their clinical interpretation remains subjective, and they fail to adequately represent objective morphological degeneration. For instance, Howe et al¹⁵ reported significantly elevated T1ρ values in articular cartilage after ACL deficiency (ACLD), yet macroscopic histological assessments revealed no significant structural changes.¹⁵ In contrast, direct 3-dimensional (3D) cartilage modeling and thickness measurements based on MRI provide more accurate, region-specific evaluations of cartilage wear and allow intuitive spatial interpretation of joint degeneration.^4,50 Despite its advantages, this methodology has been rarely applied in long-term post-ACLR degenerative studies.

Walking, as the most frequent and fundamental form of human locomotion, subjects the knee joint to repetitive loading that may significantly contribute to postoperative cartilage degeneration.^36,43 Consequently, investigating kinematic and kinetic alterations during walking in ACLD knees is essential for elucidating the biomechanical mechanisms underlying post-ACLR cartilage degeneration. While many in vitro studies have characterized intra-articular mechanics after ACL injury, highlighting its effects on tibial plateau load distribution, most employ static loading protocols (eg, contact pressure measurements at fixed flexion angles of 30° or 60°),^17,28,52 which fail to replicate the dynamic, continuous mechanical environment of actual gait.

Robotic testing systems often use simplified motion inputs—such as flexion-extension, axial compression, or anterior-posterior shear—while neglecting the internal-external rotation and varus-valgus motions crucial for walking.^6,27,52 Given the knee's inherently asymmetric structure, its contact mechanics critically depend on coordinated 6 degree-of-freedom motion patterns. Simplified inputs, therefore, cannot accurately reproduce true gait kinematics, compromising the simulation of medial-lateral tibial plateau loading. Current biomechanical research inadequately establishes the mechanical pathways linking ACLR to cartilage degeneration, and a causal relationship between abnormal loading and structural cartilage changes is noticeably lacking.

To bridge existing knowledge gaps in the mechanisms underlying post-ACLR cartilage degeneration, we developed an integrated framework that combines clinical MRI-based 3D modeling with in vitro biomechanical testing to systematically investigate both structural and mechanistic aspects of cartilage degeneration. The study has 2 primary objectives: first, to quantify temporal cartilage degeneration patterns using MRI scans from patients with ACLR, collected preoperatively and at 3-year postoperative follow-up, with the nnUNet deep learning framework¹⁸ enabling automated 3D segmentation and thickness mapping of medial and lateral tibial cartilage; second, to validate these findings in vitro by collecting gait data from healthy controls, those with ACLD, and 3-year post-ACLR patients to drive cadaveric knee specimens under ACL-intact (ACLI), ACLD, and ACL-reconstructed conditions, using integrated pressure sensors to measure tibial plateau contact pressures, contact areas, and contact center of stress trajectories.

Methods

Participants

From December 2020 to January 2022, a total of 30 patients with acute unilateral ACL rupture were enrolled at Guangdong Provincial People's Hospital (16 men and 14 women; 15 right-sided, 15 left-sided injuries; mean/median age, 31.3/31.5 years; age range, 21-45 years; mean/median weight, 62.6/64.8 kg; weight range, 52.1-83.4 kg). All patients underwent ACLR within 1 month of injury. Baseline knee MRI scans were performed within 30 days after injury, and follow-up MRIs were obtained 36 months after ACLR.

The inclusion criteria required skeletal maturity confirmed by MRI and evidence of closed growth plates. The exclusion criteria were as follows: (1) a history knee injury or disease; (2) inflammatory joint conditions, such as rheumatoid arthritis; (3) evidence of preexisting osteoarthritis on baseline³⁸ MRI (Outerbridge grade ≥2); (4) multiligamentous injuries requiring additional surgical intervention; and (5) any concomitant intra-articular procedures at the index surgery, including meniscal repair, meniscal debridement or partial meniscectomy, and cartilage-related procedures.

The study was approved by the Ethics Committee of Guangdong Provincial People's Hospital (No. 2019-226H-1) and the Department of Anatomy, School of Basic Medical Sciences, Southern Medical University. Written informed consent was obtained from all participants.

Surgical Procedure and Postoperative Rehabilitation

All patients underwent standardized primary anatomic single-bundle ACLR^1,37 performed by senior orthopaedic surgeons (Y.Z.). Autologous semitendinosus and gracilis tendons were harvested in all cases and prepared as a quadrupled graft.

With the patient in the supine position and the knee flexed to approximately 90°, diagnostic arthroscopy was first performed to confirm intra-articular pathology. The tibial tunnel was positioned at the center of the native ACL tibial footprint, corresponding to the middle third of the line connecting the medial tibial spine and the anterior horn of the lateral meniscus. It was created using a tibial aiming guide. The femoral tunnel was drilled through the anteromedial portal with the knee in high flexion and placed within the anatomic footprint on the lateral femoral condyle, while preserving the posterior cortical wall.

The graft was passed through the tibial tunnel into the femoral tunnel. Femoral fixation was achieved using a suspensory fixation device, and tibial fixation was performed with an interference screw at 20° of knee flexion under manual graft tensioning. To isolate the biomechanical and structural effects of ACLR, no concomitant procedures—including meniscal repair, meniscal debridement, or cartilage-related surgeries—were performed in any enrolled patient.

Postoperatively, all patients followed a standardized rehabilitation protocol.⁴⁷ Immediate weightbearing as tolerated was encouraged, along with early range-of-motion exercises aimed at restoring full knee extension. Progressive strengthening and neuromuscular training were implemented under supervision during subsequent rehabilitation phases to facilitate functional recovery and reduce the risk of reinjury.

MRI Examinations

MRI examinations of all ACLD knees were performed using a 3T scanner (MAGNETOM Verio, ATim System; Siemens). Scans were acquired preoperatively and at 36 months after ACLR. An 18-channel dedicated knee coil was used. Scan parameters for the 3D sagittal proton density Cube (3D Sag PD Cube) were as follows: repetition time, 1502 ms; echo time, 31.466 ms; flip angle, 25°; slice thickness, 0.8 mm; slice spacing, 0.4 mm; and field of view, 160 × 160 mm, resulting in an in-plane resolution of 0.3125 × 0.3125 mm², matrix of 512 × 512, and bandwidth of 250 Hz/pixel. The raw magnitude images were used for cartilage segmentation and morphological measurement.

Automatic Cartilage Segmentation

The nnU-Net framework was used to train a model for automatic cartilage segmentation from MRI images. Details of the nnU-Net architecture are thoroughly described by Isensee et al.¹⁸ The specific procedures for model training and cartilage segmentation were as follows (Figure 1A):

Figure 1.

Methodological approach of this study. (A) Workflow for quantifying morphological wear severity of tibia cartilage. (B) Workflow for monitoring intra-articular mechanical changes during gait motion reproduction. ACL, anterior cruciate ligament.

(1) Data preprocessing: Publicly available knee MRI data sets from the Osteoarthritis Initiative (https://nda.nih.gov/oai/) and the 3D Sag PD Cube sequences from our 30 patients with ACL injury. These data sets were resampled to an isotropic resolution of 0.5 × 0.5 × 0.5 mm³ and normalized using Z-score standardization.

(2) Network training: A 5-fold cross-validation strategy was used. The objective function was a weighted combination of Dice loss and cross-entropy loss. The model was trained using 3D image patches of 128 × 128 × 128 voxels for 1000 epochs on an NVIDIA RTX 3090 GPU, with validation performed every 50 epochs to retain the best-performing model.

(3) Model inference: Automatic cartilage segmentation was performed on ACLR patient data. The cartilage probability maps were binarized using a 0.5 threshold, followed by 3D connected component analysis to remove noise regions with <50 voxels.

(4) Evaluation: Quantitative metrics, including Dice coefficient, Hausdorff distance, and mean surface distance, were computed. In addition, 2 radiologists (W.H.) with >5 years of experience independently conducted a blinded evaluation of the segmentation results. In cases of disagreement, consensus was reached through discussion, and the final result was considered the gold standard.

(5) Output: Segmentation results were converted into stereolithography-format 3D models using the Marching Cubes algorithm with a triangle mesh density of 0.5 mm, enabling the calculation of cartilage volume, mean thickness, and other morphological parameters.

All data processing was performed using Python 3.9.

Cartilage Morphological Measurement

In this study, MATLAB (R2022b, MathWorks) was used to perform 3D cartilage thickness analysis on the segmented medial and lateral tibial cartilage. First, we applied a 3D distance transformation (using the bwdist function) to the binarized cartilage masks. This computed the distance from each voxel to the nearest background voxel, creating a voxel-wise cartilage thickness map (unit: mm). Subsequently, a thickness heatmap was generated using a "hot" colormap to illustrate the spatial distribution of cartilage thickness, with red indicating thicker regions and blue indicating thinner areas (Figure 2).

Figure 2.

Three-dimensional reconstruction and quantitative analysis of tibial cartilage degeneration before and 3 years after ACLR. (A and B): Preoperative and 3-year postoperative 3D models of the tibia and tibial cartilage segmented using a deep learning-based approach. (C and D) 3D models and thickness heatmaps of the medial tibial plateau cartilage preoperatively and 3 years postoperatively. (E and F) 3D models and thickness heatmaps of the lateral tibial plateau cartilage preoperatively and 3 years postoperatively. (G) Cartilage wear heatmap illustrating the spatial distribution and severity of cartilage thinning, with colors ranging from blue to red indicating increasing degeneration. (H) Schematic illustration of the 5 predefined subregions used for regional cartilage analysis, adapted from previous studies. (I and J) Quantitative comparison of mean cartilage thickness changes within each subregion between the preoperative and 3-year postoperative conditions. ACLR, anterior cruciate ligament reconstruction; Post, postoperatively; Pre, preoperatively; 3D, 3-dimensional.

After the anatomical division method proposed by Wirth et al,^3,49 the tibial cartilage was subdivided into 5 regions (Figure 2H): anterior medial/lateral tibia (aMT/aLT), posterior medial/lateral tibia (pMT/pLT), external medial/lateral tibia (eMT/eLT), internal medial/lateral tibia (iMT/iLT), and central medial/lateral tibia (cMT/cLT). The mean cartilage thickness in each subregion was then calculated for subsequent statistical analyses.

Specimen Preparation

Eight fresh-frozen human knee specimens (aged 35-45 years, with an equal number of men and women) were obtained from the Department of Anatomy at Southern Medical University. All specimens were stored at −20°C and thawed overnight at room temperature before testing. Arthroscopic examinations were performed on each specimen to confirm the absence of osteoarthritis or ACL injuries. Autologous semitendinosus and gracilis tendon grafts were harvested via an anteromedial incision at the proximal tibia for ACLR. Skin and nonessential soft tissues were removed, but key stabilizing structures—including the quadriceps tendon, iliotibial band, joint capsule, cruciate ligaments, and collateral ligaments—were preserved to maintain structural integrity. The fibula was secured in its anatomical position relative to the tibia using 2.5-mm Kirschner wires. After establishing biomechanical protocols,^20,29 the femur and tibia were transected 15 cm proximal and distal to the joint line, respectively. Custom-made cylindrical fixtures were then mounted to the distal ends of the tibia-fibula complex using polymethyl methacrylate.

Dynamical Simulated Human Gait

Following the protocol described by Woo et al,⁵¹ the specimens were mounted on a 6 degree-of-freedom robotic testing system (Figure 1B). This system consists of a KUKA KR 120 R2500 Pro robot (KUKA Robotics) with a motion repeatability of ±0.06 mm, and an ATI Industrial Automation FT Delta 6-axis force/torque sensor, providing measurement accuracies of ±0.2 N for force and ±0.1 N·m for torque.^7,8

Knee kinematic control was achieved using a customized multitasking MATLAB program (MathWorks), which ensured high repeatability and operational reliability during testing.^34,35 Gait characteristics from ACLI, ACLD, and ACLR patients were obtained using an optical motion capture system and served as input trajectories for the robotic arm, thereby simulating physiologically realistic knee joint motion. To compensate for the loss of muscle force resulting from femoral resection and to maintain joint stability, simulated muscle loading was applied in accordance with established protocols.¹⁶ A customized pulley system was employed to apply a 100 N tensile force along the quadriceps tendon direction.³¹

Contact Stress Measurement

To assess contact mechanics, 1.5 cm long incisions were made beneath the anterior and posterior edges of both the medial and lateral menisci. After establishing protocols,^16,27 Tekscan pressure sensors (model 4010, Tekscan Inc) were carefully implanted beneath the menisci through these incisions. The sensor system comprises 2 independent sensing grids corresponding to the medial and lateral tibiofemoral compartments. Before implantation, the sensors were calibrated following the manufacturer's guidelines using the maximum axial load from the gait cycle. Calibration accuracy was maintained within ±25 N (approximately 5% of the calibration load), with a repeatability error <1%. Peak contact pressure analysis employed the same predefined tibial subregional framework as the MRI-based assessment.^3,49

Specimen Testing Conditions

Contact pressure parameters were measured sequentially under 3 conditions: (1) ACLI; (2) ACLD; and (3) single-bundle ACLR. The experimental procedure was as follows: initially, contact pressure data within the tibiofemoral joint were recorded under the ACLI condition; subsequently, the ACL was completely transected using a No. 11 scalpel to simulate ACLD. Finally, anatomical single-bundle ACLR was performed using quadrupled autografts of the semitendinosus and gracilis tendons, with graft diameters of 8 mm in 7 specimens and 7.5 mm in 1 specimen, in accordance with internationally recognized techniques and the guidelines of the American Academy of Orthopaedic Surgeons.^30,33 All surgical procedures were conducted by experienced orthopaedic surgeons (Y.Z.). After reconstruction, graft fixation and restoration of anterior tibial restraint were verified through standardized manual anterior tibial translation assessment.

Data collection points were set at key phases of the simulated gait cycle, including 2% (heel strike), 12% (loading response), 30% (mid-stance), 50% (terminal stance), 60% (pre-swing), 70% (initial swing), 75% (tibia vertical), 80% (mid-swing), and 100% (terminal swing).³² Mechanical distribution parameters within the joint (contact force and contact area) were obtained through repeated measurements, and a weighted algorithm was applied to calculate the contact center of stress (CCS).⁵² The algorithm for the weighted center of pressure is illustrated in Figure 3, A and B. The whole-gait-cycle contact center of stress (WCCS), derived from averaging phase-specific CCS coordinates across all 9 gait phases.

Figure 3.

Visualization of the weighted CCS across nine key phases of the gait cycle in knees with different ACL states. (A and B) Formula for calculating the weighted CCS, where x_i and y_i represent the coordinates of the i^th sensing element, S_i is the corresponding contact stress, and n is the total number of sensing elements. These calculations were performed separately for the medial and lateral compartments at each step of the gait simulation. (C and D) Left plots: CCS trajectories for intact ACL (green dots), ACLD (red dots), and ACLR (yellow dots) knees. Numbers 1 to 9 denote gait phases: 1 = heel strike; 2 = loading response; 3 = mid-stance; 4 = terminal stance; 5 = pre-swing; 6 = initial swing; 7 = tibial vertical; 8 = mid-swing; and 9 = terminal swing. Right plots: WCCS for each ACL state (ACL-intact: black star; ACLD: yellow star; ACLR: red star). ACL, anterior cruciate ligament; ACLD, anterior cruciate ligament deficiency; ACLR, anterior cruciate ligament reconstruction; CCS, contact center of stress; WCCS, whole-gait-cycle contact center of stress.

Statistical Analysis

All statistical analyses were performed using SPSS software (Version 26.0; IBM Corp) and GraphPad Prism (Version 9.0; GraphPad Software Inc). All statistical tests were 2-tailed, with significance set at P < .05. Normality of all continuous variables was assessed using the Shapiro-Wilk test, and homogeneity of variances was evaluated via the Levene test to determine appropriate statistical methods.

Comparisons of cartilage thickness between preoperative and 3-year postoperative time points were conducted using paired t tests if normality and homogeneity of variance assumptions were met; otherwise, paired Wilcoxon signed-rank tests were applied. For ex vivo measurements of tibiofemoral contact area under different ACL conditions (intact, deficient, reconstructed), 1-way analysis of variance with Bonferroni post hoc correction was used if data met normality and homogeneity assumptions; otherwise, the Friedman test with Bonferroni-adjusted pairwise comparisons was performed.

Additionally, to evaluate displacement changes in the WCCS in ACLD and ACLR states relative to the ACLI condition, differences between ACLD and ACLI and between ACLR and ACLI were calculated. Paired t tests were used for comparisons if normality assumptions were satisfied; otherwise, Wilcoxon signed-rank tests were applied.

Results

Cartilage Thickness Changes 3 Years After ACLR

Based on MRI data obtained preoperatively and at 3 years postoperatively in ACLR patients, we quantified the mean cartilage thickness in subregional compartments of the medial and lateral tibial plateaus (Table 1 and Figure 2).

Table 1

Mean Thickness Changes Within the 5 Subregions^a

Subregion	Medial, mm				Lateral, mm
Subregion	ACLR-Pre	ACLR-Post	P	Difference	ACLR-Pre	ACLR-Post	P	Difference
Center part cMT/cLT	3.58 ± 0.36	3.23 ± 0.33	.027	10	4.35 ± 0.42	3.94 ± 0.43	.031	9
Anterior part, aMT/aLT	2.89 ± 0.27	2.65 ± 0.25	.618	8	3 ± 0.29	2.75 ± 0.31	.371	8
Internal part, iMT/iLT	2.46 ± 0.26	1.95 ± 0.2	.019	21	3.16 ± 0.28	2.82 ± 0.29	.461	11
Posterior part, pMT/pLT	2.8 ± 0.33	2.44 ± 0.21	.041	13	2.9 ± 0.33	2.51 ± 0.27	.029	14
External part, eMT/eLT	3.44 ± 0.36	3.23 ± 0.34	.414	6	2.55 ± 0.26	2.17 ± 0.22	.035	15

Data are presented as mean ± SD or %. Boldface P values indicate statistical significance. ACLR, anterior cruciate ligament reconstruction; aMT/aLT, anterior medial/lateral tibia; cMT/cLT, central medial/lateral tibia; eMT/eLT, external medial/lateral tibia; iMT/iLT, internal medial/lateral tibia; Post, postoperatively; Pre, preoperatively; pMT/pLT, posterior medial/lateral tibia.

On the medial tibial plateau, cartilage thickness showed varying degrees of reduction at 3 years postoperatively compared with baseline, with significant changes observed in the cMT, iMT, and pMT subregions (Figure 2J). Specifically, cMT thickness decreased from 3.58 ± 0.36 mm preoperatively to 3.23 ± 0.33 mm postoperatively (P = .027), representing a mean reduction of 10%. iMT thickness declined from 2.46 ± 0.26 mm to 1.95 ± 0.20 mm (P = .019), a 21% reduction, while pMT thickness decreased from 2.80 ± 0.33 mm to 2.44 ± 0.21 mm (P = .041), corresponding to a 13% decrease. Although reductions were also observed in the aMT and eMT regions, these changes were not statistically significant (P > .05).

On the lateral tibial plateau, cartilage thickness also generally decreased, with several subregions exhibiting statistically significant differences (Figure 2I). cLT thickness declined from 4.35 ± 0.42 mm to 3.94 ± 0.43 mm (P = .031), reflecting a 9% reduction. pLT thickness dropped from 2.90 ± 0.33 mm to 2.51 ± 0.27 mm (P = .029), a 14% reduction, and eLT thickness decreased from 2.55 ± 0.36 mm to 2.17 ± 0.32 mm (P = .035), corresponding to a 15% decrease. No statistically significant changes were found in the aLT and iLT regions (P > .05).

Changes in Anterior-Posterior and Medial-Lateral Displacement of the WCCS

As shown in Figure 3, C and D, the weighted contact centers of stress of the knee joint were calculated at 9 selected gait phases across the 3 ACL conditions. Subsequently, a secondary weighting was applied to derive the WCCS. The quantitative displacements of the stress centers in ACLD and ACLR conditions relative to the intact ACL are detailed in Table 2.

Table 2

Relative Translation of WCCS in AP and ML Direction for ACLD and ACLR Knee^a

Direction	Medial, mm			Lateral, mm
Direction	ACLI-ACLR	ACLI-ACLD	P	ACLI-ACLR	ACLI-ACLD	P
AP	5.29 ± 3.11	9.32 ± 4.15	.036	4.73 ± 2.78	11.24 ± 5.58	.001
ML	–0.49 ± 0.25	1.28 ± 0.67	.422	–3.55 ± 1.1	5.18 ± 1.98	.021

Data are presented as mean ± SD. Boldface P values indicate statistical significance. ACLD, anterior cruciate ligament deficient; ACLI, anterior cruciate ligament intact; ACLR, anterior cruciate ligament reconstruction; AP, anterior-posterior; ML, medial-lateral; WCCS, whole-gait-cycle contact center.

On the medial tibial plateau, the WCCS in the ACLD condition shifted posteriorly by 9.32 ± 4.15 mm compared with the ACLI state. Although ACLR knee significantly improved this displacement (P = .036), a posterior shift of 5.29 ± 3.11 mm remained. In the mediolateral direction, the WCCS shifted medially by 1.28 ± 0.67 mm in the ACLD condition. After reconstruction, a lateral shift of 0.49 ± 0.25 mm was observed relative to the ACLI state. However, this change was not statistically significant.

On the lateral tibial plateau, the WCCS in the ACLD knee shifted posteriorly by 11.24 ± 5.58 mm compared with the ACLI condition. ACLR knee significantly reduced this displacement (P = .001), although a residual posterior shift of 4.73 ± 2.78 mm persisted. In the mediolateral direction, the ACLD knee showed a medial shift of 5.18 ± 1.98 mm. After reconstruction, a significant lateral displacement of 3.55 ± 1.10 mm was observed relative to the ACLI state (P = .021).

Tibiofemoral Contact Area Variation Across Gait Phases in Intact, ACLD, and ACLR Knees

The robotic gait simulation revealed significant differences in tibiofemoral contact area throughout the gait cycle among the ACLI, ACLD, and ACLR knees, as summarized in Table 3 and Figure 4.

Table 3

Contact Areas (in cm²) of the Intact, ACLD, and ACLR Knee in Response to 9 Phases of the Gait Cycle^a

Gait Cycle Phases, %	Medial						Lateral
Gait Cycle Phases, %	ACLI	ACLD	ACLR	P1	P2	P3	ACLI	ACLD	ACLR	P1	P2	P3
2	2.1 ± 0.9	1.5 ± 0.8	3.1 ± 0.7	.255	.012	.008	1.5 ± 0.72	1.5 ± 0.8	2.3 ± 1.02	.567	.031	.123
12	2.7 ± 0.7	2.3 ± 0.9	3.3 ± 0.7	.203	.023	.021	2.4 ± 1.14	2.3 ± 0.9	2.6 ± 1.08	.263	.27	.02
30	2.9 ± 1.2	2.3 ± 0.7	3 ± 0.8	.232	.763	.019	2 ± 1.2	2.3 ± 0.7	2.4 ± 1.08	.698	.374	.001
50	3.3 ± 1.1	2.5 ± 1	3.1 ± 0.6	.161	.575	.025	3.3 ± 1.14	2.5 ± 1	3.8 ± 1.02	.033	.093	.001
60	2.7 ± 1.1	2.3 ± 1.1	2.8 ± 1.1	.078	.7	.03	2.9 ± 0.9	2.3 ± 1.1	2.8 ± 0.78	.005	.824	.004
70	2.2 ± 1.2	1.9 ± 1.2	2.5 ± 1.1	.109	.169	.021	2.2 ± 0.72	1.9 ± 1.2	2.2 ± 0.72	.12	.575	.166
75	1.9 ± 0.8	1.7 ± 0.9	2.2 ± 0.8	.212	.039	.012	2 ± 0.6	1.7 ± 0.9	2.4 ± 0.6	.353	.233	.002
87	1.9 ± 0.9	2 ± 1.1	2.8 ± 0.7	.783	.017	.027	1.5 ± 0.66	2 ± 1.1	2.3 ± 0.9	.746	.034	.02
100	1.7 ± 0.8	1.5 ± 0.8	3 ± 0.8	.715	.007	.009	1.3 ± 0.66	1.5 ± 0.8	2.4 ± 1.08	.291	.025	.036

Data are presented as the mean ± SD. Boldface P values indicate statistical significance. ACLD, anterior cruciate ligament deficient; ACLI, anterior cruciate ligament intact; ACLR, anterior cruciate ligament reconstruction.

P1: Intact vs ACLD; P2: Intact vs ACLR; P3: ACLD vs ACLR.

Figure 4.

Comparison of gait contact mechanics across 3 ACL states. (A and B) Contact area variation across gait phases (ACLI vs ACLD vs ACLR). (C and D) WCCS deviations of ACLD and ACLR from the ACLI baseline are quantified in the AP and ML directions. ACL, anterior cruciate ligament; ACLD, anterior cruciate ligament deficient; ACLI, anterior cruciate ligament intact; ACLR, anterior cruciate ligament reconstruction; AP, anterior-posterior; ML, medial-lateral; WCCS, whole-gait-cycle contact center.

On the medial tibial plateau, the contact area under ACLD conditions consistently differed from that of the ACLI knee across the entire gait cycle (P < .05), as shown in Figure 4A. Although the ACLR condition partially resembled the ACLI knee at certain phases, significant deviations were still observed at several time points, including 2% (P = .012), 12% (P = .023), 75% (P = .039), 87% (P = .017), and 100% (P = .007), indicating incomplete restoration of joint loading after reconstruction. Moreover, when compared with the ACLD condition, the ACLR knee demonstrated significantly larger contact areas at most gait phases: 2% (P = .008), 12% (P = .021), 30% (P = .019), 50% (P = .025), 60% (P = .030), 70% (P = .021), 75% (P = .012), 87% (P = .027), and 100% (P = .009).

On the lateral tibial plateau, the overall trend was similar (Figure 5B). The ACLD condition showed only minor differences compared with the ACLI knee, with significant reductions in contact area occurring primarily at 50% (P = .033) and 60% (P = .005) of the gait cycle. After reconstruction, the ACLR knee exhibited significantly increased contact area at key gait phases, such as 2% (P = .031), 87% (P = .034), and 100% (P = .025). Furthermore, ACLD and ACLR knees differed significantly at multiple time points, including 12% (P = .020), 30% (P = .001), 50% (P = .001), 60% (P = .004), 75% (P = .002), 87% (P = .020), and 100% (P = .036), underscoring the substantial impact of ACLD and its reconstruction on lateral joint loading patterns.

Figure 5.

Subregional tibial peak contact pressure during the gait cycle under different ACL conditions. ACL, anterior cruciate ligament. ACLD, anterior cruciate ligament deficient; ACLI, anterior cruciate ligament intact; ACLR, anterior cruciate ligament reconstruction. (A) Anterior lateral tibial region (aLT). (B) Posterior lateral tibial region (pLT). (C) Central lateral tibial region (cLT). (D) Internal lateral tibial region (iLT). (E) External lateral tibial region (eLT). (F) Anterior medial tibial region (aMT). (G) Posterior medial tibial region (pMT). (H) Central medial tibial region (cMT). (I) Internal medial tibial region (iMT).

Gait-Dependent Changes in Subregional Tibial Contact Pressure

During walking, peak tibial contact pressure demonstrated clear phase-dependent and region-specific patterns across the gait cycle. In the anterior tibial regions (Figure 5, A and F), higher contact pressures were predominantly observed before 60% of the gait cycle. Compared with ACLI knees, ACLD knees exhibited significantly reduced peak contact pressure in the anterior tibia during these gait phases. After ACLR, anterior tibial contact pressure showed partial recovery at selected gait intervals. However, overall pressure magnitudes remained lower than those observed in ACLI knees.

In contrast, elevated contact pressures in the posterior tibial regions (Figure 5, B and G) and central tibial regions (Figure 5, C and H) were primarily concentrated between 30% and 75% of the gait cycle. Within these phases, ACLD knees demonstrated significantly increased peak contact pressures in the pLT, pMT, and cLT compared with ACLI knees. After ACLR, peak contact pressures in these regions were reduced relative to the ACLD condition but remained significantly higher than those in ACLI knees at multiple gait phases.

In the internal tibial regions (iLT and iMT; Figure 5, D and I), peak contact pressure did not differ substantially between ACLD and ACLI knees, whereas ACLR knees demonstrated significantly increased pressure, particularly within the iMT region. In contrast, the external tibial regions (eLT and eMT; Figure 5, E and J) exhibited relatively low contact pressure throughout the gait cycle, with no significant between-group differences, although absolute pressure values tended to be higher in ACLD and ACLR knees compared with ACLI knees.

Overall, ACLD was associated with a marked redistribution of tibial contact pressure. ACLR partially attenuated these abnormal loading patterns but did not fully restore the physiological spatiotemporal distribution of contact pressure observed in ACLI knees. Notably, these subregional contact mechanics findings were consistent with the imaging-based observations reported in this study.

Discussion

This study integrated dynamic gait biomechanics and 3D MRI-based quantitative analysis to reveal, for the first time, the regional characteristics of cartilage degeneration after ACLR. At the 3-year postoperative follow-up, MRI data demonstrated site-specific thinning of the femorotibial cartilage, with the most pronounced degeneration observed in the medial compartment (inner region, −21%; posterior region, −13%) and the lateral compartment (posterior region, −14%; outer region, −15%). Moreover, an in vitro 6 degree-of-freedom gait simulation further elucidated the underlying mechanical mechanisms, demonstrating a persistent posterior shift of the WCCS post-ACLR (5.29 ± 3.11 mm medially, 4.73 ± 2.78 mm laterally), accompanied by abnormal increases in contact area during the early stance and late swing phases of the gait cycle. The results of this study may clarify the relationship between dynamic mechanical forces and region-specific cartilage degeneration.

Some studies have consistently reported progressive increases in T1ρ/T2 values within 1 to 2 years after ACLR,^25,4546,53 reflecting sustained disturbances in cartilage matrix metabolism. However, a study by Li et al,²² which simultaneously assessed biochemical and morphological changes, reported significant increases in T2 values 1 year after ACLR but no statistically significant changes in cartilage thickness. Rather than representing conflicting findings, these observations likely reflect different stages of cartilage degeneration along a temporal continuum. Early biochemical alterations detected by T1ρ and T2 mapping may precede overt structural damage, whereas measurable cartilage thinning may emerge only at later follow-up. Accordingly, the 3-year postoperative thinning observed in the present study may represent a downstream morphological manifestation of earlier compositional degeneration of cartilage.

Consistent with this temporal progression, studies incorporating longer-term morphological assessments have reported findings in line with our results. For instance, Zhang et al⁵³ observed significant reductions in medial and lateral tibial cartilage volume at 24 and 60 months after ACLR, compared with 6 months postoperatively. Similarly, Su et al⁴⁵ demonstrated that the posterior region of the lateral tibial cartilage exhibited significant thinning relative to control participants 2 years after ACLR. Collectively, these findings confirm that early degeneration occurs in the tibial cartilage within 3 years after ACLR, with 10% to 21% loss of thickness, predominantly in the middle and posterior regions.

Although ACLR knees exhibit significantly restored anterior-posterior laxity compared with ACLD knees, a posterior shift of the contact pressure center (approximately 5 mm) persists during gait when compared with ACLI knees. This discrepancy may be attributed to incomplete biomechanical restoration. While anatomically positioned single- or double-bundle ACLRs can partially restore anterior stability under static loading, the reconstructed grafts often fail to replicate the native ACL's nonlinear viscoelastic properties, time-dependent behavior, neuromuscular control, and synergy with other joint structures during dynamic activities, such as gait.^11,14,44 Another possible explanation is the quadriceps-dominant anterior tibial traction effect during knee extension. Our results indicate that the phases with significantly increased intra-articular contact area in ACLR knees—early stance and late swing—coincide with knee extension during gait. During these phases, intense quadriceps contraction induces anterior tibial translation, contributing to the posterior shift of the contact pressure center.⁴² Insufficient recovery or delayed activation of the hamstrings after reconstruction may further diminish their capacity to resist anterior tibial translation, thereby exacerbating this displacement.¹⁴

Two potential explanations exist for the observed regional cartilage degeneration after ACLR, particularly the thinning in the central and posterior regions. On the one hand, the sustained posterior shift of the contact center after ACLR concentrates joint loading on the posterior tibial plateau, exposing cartilage to prolonged nonphysiological stress and accelerating degeneration. Previous studies support this hypothesis. For example, Amano et al² found a positive correlation between anterior tibial translation and T1ρ/T2 values in the posterolateral region. Chu et al⁹ reported that residual anterior translation was associated with increased relaxation times in the posterior horn of the meniscus. On the other hand, the medial-lateral shift of the contact center on the lateral tibial plateau—from medial displacement in ACLD knees to lateral displacement after ACLR—suggests excessive external rotational coupling.^21,48 This rotational change may alter load trajectories, potentially exacerbating cartilage wear in the lateral peripheral region.

Our findings have 3 important clinical implications. First, surgical and rehabilitation strategies should aim to improve dynamic biomechanical restoration of the knee during functional activities, rather than focusing solely on static stability. Second, rehabilitation protocols should target critical gait phases. This includes enhancing quadriceps-hamstring co-contraction to control anterior tibial translation during the early stance phase (0%-12%) and employing eccentric training of the gluteus medius during the terminal swing phase (75%-100%) to mitigate posterior loading induced by femoral internal rotation. Third, monitoring strategies should incorporate posterior tibial cartilage thickness as a sensitive biomarker to enable early intervention for patients with significant thinning.

This study has several limitations. First, cadaveric knee models cannot replicate the neuromuscular control present in living subjects. To approximate physiological loading, calibrated weights were applied during the gait cycle to simulate quadriceps tension.^7,8 Second, although the 3-year follow-up period was adequate to detect early degenerative changes, it may not fully capture the long-term progression of post-ACLR osteoarthritis. We plan to conduct an extended longitudinal follow-up with this patient cohort. Third, we did not control for potential confounding factors, such as differences in rehabilitation adherence and patient activity levels. In addition, only hamstring tendon autografts were evaluated; thus, the findings may not be generalizable to other graft choices, including bone–-patellar tendon-bone or quadriceps tendon autografts. Future studies comparing graft choices under dynamic loading conditions are warranted.

Conclusion

Our findings demonstrate a temporal and spatial link between the posterior shift of the WCCS after ACLR and the degeneration of the posterior tibial cartilage. The integrated framework of dynamic loading and structural evolution presented in this study not only provides a novel standard for the early detection of osteoarthritis but also opens new avenues for targeted mechanical interventions to halt cartilage degeneration.

Footnotes

Submitted September 13, 2025; accepted May 6, 2026.

The study was endorsed by the Key Research and Development Program of Guangzhou (No. 2023B01J0022), the National Natural Science Foundation of China (No. 12472327), and the Shanghai Municipal Education Commission, Artificial Intelligence-Driven Research and Applications in Sports Biomechanics—Precision Modeling, Key Point Identification, and Sports Injury Diagnosis.

Ethical approval was obtained from the Ethics Committee of Guangdong Provincial People's Hospital (No. 2019-226H-1).

Authors’ Contribution

Jinpeng Lin, Rongshan Chen, and Yuan Yan made substantial contributions to research design, data acquisition, and data analysis, and drafted the manuscript. Xiaolong Zeng and Wenhan Huang were the major contributors to data acquisition and analysis. Tsung-Yuan Tsai helped in data collection and the manuscript evaluation. Chunlin Deng and Shaobai Wang supervised data processing and analysis and critically revised the manuscript. Yu Zhang supervised the study's development, helped with study design and manuscript evaluation, and served as the corresponding author.

ORCID iDs

Jinpeng Lin

Xiaolong Zeng

Wenhan Huang

References

Aga

Risberg

Fagerland

, et al. No difference in the KOOS quality of life subscore between anatomic double-bundle and anatomic single-bundle anterior cruciate ligament reconstruction of the knee: a prospective randomized controlled trial with 2 years’ follow-up. Am J Sports Med. 2018;46(10):2341-2354.

Amano

Pedoia

Souza

CB.

Persistent biomechanical alterations after ACL reconstruction are associated with early cartilage matrix changes detected by quantitative MR. Orthop J Sports Med. 2016;4(4):2325967116644421.

Andreisek

White

Sussman

, et al. Quantitative MR imaging evaluation of the cartilage thickness and subchondral bone area in patients with ACL-reconstructions 7 years after surgery. Osteoarthritis Cartilage. 2009;17(7):871-878.

Aubonnet

Ramos

Recenti

, et al. Toward new assessment of knee cartilage degeneration. Cartilage. 2023;14(3):351-374.

Bustin

Witschey

WRT

van Heeswijk

Cochet

Stuber

Magnetic resonance myocardial T1ρ mapping: technical overview, challenges, emerging developments, and clinical applications. J Cardiovasc Magn Reson. 2023;25(1):34.

Chastain

Wach

Pekmezian

, et al. ACL transection results in a posterior shift and increased velocity of contact on the medial tibial plateau. J Biomech. 2022;144:111335.

Cheng

Dimitriou

Yao

, et al. Outperformance of combined artificial anterolateral ligament and ACL reconstruction compared with isolated artificial ACL reconstruction in knees with anterolateral structure and ACL deficiency: a biomechanical analysis. Orthop J Sports Med. 2025;13(2):23259671241309270.

Cheng

Yao

Dimitriou

Jiang

Yang

Tsai

TY.

The 45° and 60° of sagittal femoral tunnel placement in anterior cruciate ligament reconstruction provide similar knee stability. Knee Surg Sports Traumatol Arthrosc. 2024;32(11):3031-3038.

Chu

Williams

West

, et al. Quantitative magnetic resonance imaging UTE-T2* mapping of cartilage and meniscus healing after anatomic anterior cruciate ligament reconstruction. Am J Sports Med. 2014;42(8):1847-1856.

10.

Cuzzolin

Previtali

Zaffagnini

Deabate

Candrian

Filardo

Anterior cruciate ligament reconstruction versus nonoperative treatment: better function and less secondary meniscectomies but no difference in knee osteoarthritis: a meta-analysis. Cartilage. 2021;13(suppl 1):S1658-S1670.

11.

Escamilla

Macleod

Wilk

Paulos

Andrews

JR.

Anterior cruciate ligament strain and tensile forces for weight-bearing and non-weight-bearing exercises: a guide to exercise selection. J Orthop Sports Phys Ther. 2012;42(3):208-220.

12.

Gao

Peng

Wang

Deng

Zhang

Liu

Cartilage T2 mapping-based radiomics in knee osteoarthritis research: status, progress and future outlook. Eur J Radiol. 2024;181:111826.

13.

Hewett

Di Stasi

Myer

GD.

Current concepts for injury prevention in athletes after anterior cruciate ligament reconstruction. Am J Sports Med. 2013;41(1):216-224.

14.

Hewett

Myer

Zazulak

BT.

Hamstrings to quadriceps peak torque ratios diverge between sexes with increasing isokinetic angular velocity. J Sci Med Sport. 2008;11(5):452-459.

15.

Howe

Thompson

Teeter

, et al. Early degenerative changes are different between partial and complete anterior cruciate ligament injury and associate with joint instability in a skeletally immature porcine model. Osteoarthritis Cartilage. 2025;33(2):302-312.

16.

Huang

Ong

Man

Liu

Lau

Yung

PS.

Posterior tibial loading results in significant increase of peak contact pressure in the patellofemoral joint during anterior cruciate ligament reconstruction: a cadaveric study. Am J Sports Med. 2021;49(5):1286-1295.

17.

Imhauser

Mauro

Choi

, et al. Abnormal tibiofemoral contact stress and its association with altered kinematics after center-center anterior cruciate ligament reconstruction: an in vitro study. Am J Sports Med. 2013;41(4):815-825.

18.

Isensee

Jaeger

Kohl

SAA

Petersen

Maier-Hein

KH.

NnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods. 2021;18(2):203-211.

19.

Keays

Newcombe

Bullock-Saxton

Bullock

Keays

AC.

Factors involved in the development of osteoarthritis after anterior cruciate ligament surgery. Am J Sports Med. 2010;38(3):455-463.

20.

Koh

Zimmerman

Patel

Ren

Zhang

L-Q.

Tibiofemoral contact mechanics with horizontal cleavage tears and treatment of the lateral meniscus in the human knee: an in vitro cadaver study. Clin Orthop Relat Res. 2018;476(11):2262-2270.

21.

Lai

Chen

Huang

, et al. Whether patients with anterior cruciate ligament reconstruction walking at a fast speed show more kinematic asymmetries? Orthop Surg. 2024;16(4):864-872.

22.

Tao

Hua

Chen

Quantitative magnetic resonance imaging assessment of cartilage status: a comparison between young men with and without anterior cruciate ligament reconstruction. Arthroscopy. 2013;29(12):2012-2019.

23.

Cheng

Lin

, et al. Quantitative MRI using T1ρ and T2 in human osteoarthritic cartilage specimens: correlation with biochemical measurements and histology. Magn Reson Imaging. 2011;29(3):324-334.

24.

Han

Busse

Majumdar

: in vivo T(1rho) mapping in cartilage using 3D magnetization-prepared angle-modulated partitioned k-space spoiled gradient echo snapshots (3D MAPSS). Magn Reson Med. 2008;59(2):298-307.

25.

Kuo

Theologis

, et al. Cartilage in anterior cruciate ligament-reconstructed knees: MR imaging T1{rho} and T2–initial experience with 1-year follow-up. Radiology. 2011;258(2):505-514.

26.

Lie

Risberg

Storheim

Engebretsen

Øiestad

BE.

What's the rate of knee osteoarthritis 10 years after anterior cruciate ligament injury? An updated systematic review. Br J Sports Med. 2019;53(18):1162-1167.

27.

Liu

Wang

, et al. Tibiofemoral joint contact area and stress after single-bundle anterior cruciate ligament reconstruction with transtibial versus anteromedial portal drilling techniques. J Orthop Surg Res. 2018;13(1):247.

28.

Logan

Aman

Kemler

Storaci

Dornan

LaPrade

RF.

Influence of medial meniscus bucket-handle repair in setting of anterior cruciate ligament reconstruction on tibiofemoral contact mechanics: a biomechanical study. Arthroscopy. 2019;35(8):2412-2420.

29.

Mae

Shino

Miyama

, et al. Single- versus two-femoral socket anterior cruciate ligament reconstruction technique: biomechanical analysis using a robotic simulator. Arthroscopy. 2001;17(7):708-716.

30.

Mao

Marshall

Price

, et al. Notchplasty alters knee biomechanics after anatomic ACL reconstruction. Knee Surg Sports Traumatol Arthrosc. 2020;28(2):614-621.

31.

Markolf

O'Neill

Jackson

McAllister

DR.

Effects of applied quadriceps and hamstrings muscle loads on forces in the anterior and posterior cruciate ligaments. Am J Sports Med. 2004;32(5):1144-1149.

32.

Moltedo

Baček

Verstraten

Rodriguez-Guerrero

Vanderborght

Lefeber

Powered ankle-foot orthoses: the effects of the assistance on healthy and impaired users while walking. J Neuroeng Rehabil. 2018;15(1):86.

33.

Morimoto

Ferretti

Ekdahl

Smolinski

FH.

Tibiofemoral joint contact area and pressure after single- and double-bundle anterior cruciate ligament reconstruction. Arthroscopy. 2009;25(1):62-69.

34.

Nakamura

Linde

Marshall

, et al. Arthroscopic centralization restores residual knee laxity in ACL-reconstructed knee with a lateral meniscus defect. Knee Surg Sports Traumatol Arthrosc. 2019;27(11):3699-3704.

35.

Papageorgiou

Gil

Kanamori

Fenwick

Woo

FH.

The biomechanical interdependence between the anterior cruciate ligament replacement graft and the medial meniscus. Am J Sports Med. 2001;29(2):226-231.

36.

Pfeiffer

Spang

Nissman

, et al. Gait mechanics and T1ρ MRI of tibiofemoral cartilage 6 months after ACL reconstruction. Med Sci Sports Exerc. 2019;51(4):630-639.

37.

Porter

Shadbolt

"Anatomic" single-bundle anterior cruciate ligament reconstruction reduces both anterior translation and internal rotation during the pivot shift. Am J Sports Med. 2014;42(12):2948-2954.

38.

Potter

Jain

Black

Fung

Lyman

Cartilage injury after acute, isolated anterior cruciate ligament tear: immediate and longitudinal effect with clinical/MRI follow-up. Am J Sports Med. 2012;40(2):276-285.

39.

Sanders

Maradit Kremers

Bryan

, et al. Incidence of and factors associated with the decision to undergo anterior cruciate ligament reconstruction 1 to 10 years after injury. Am J Sports Med. 2016;44(6):1558-1564.

40.

Sanders

Maradit Kremers

Bryan

, et al. Incidence of anterior cruciate ligament tears and reconstruction: a 21-year population-based study. Am J Sports Med. 2016;44(6):1502-1507.

41.

Sasho

Katsuragi

Yamaguchi

, et al. Associations of three-dimensional T1 rho MR mapping and three-dimensional T2 mapping with macroscopic and histologic grading as a biomarker for early articular degeneration of knee cartilage. Clin Rheumatol. 2017;36(9):2109-2119.

42.

Simonson

Piussi

Högberg

, et al. Effect of quadriceps and hamstring strength relative to body weight on risk of a second ACL injury: a cohort study of 835 patients who returned to sport after ACL reconstruction. Orthop J Sports Med. 2023;11(4):23259671231157386.

43.

Slater

Hart

Kelly

Kuenze

CM.

Progressive changes in walking kinematics and kinetics after anterior cruciate ligament injury and reconstruction: a review and meta-analysis. J Athl Train. 2017;52(9):847-860.

44.

Solomonow

Krogsgaard

. Sensorimotor control of knee stability. A review. Scand J Med Sci Sports. 2001;11(2):64-80.

45.

Hilton

Nardo

, et al. Cartilage morphology and T1ρ and T2 quantification in ACL-reconstructed knees: a 2-year follow-up. Osteoarthritis Cartilage. 2013;21(8):1058-1067.

46.

Ushio

Okazaki

Osaki

, et al. Degenerative changes in cartilage likely occur in the medial compartment after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 2019;27(11):3567-3574.

47.

van Melick

van Cingel

Brooijmans

, et al. Evidence-based clinical practice update: practice guidelines for anterior cruciate ligament rehabilitation based on a systematic review and multidisciplinary consensus. Br J Sports Med. 2016;50(24):1506-1515.

48.

Williams

Erhart-Hledik

Asay

, et al. Patient-reported outcomes and knee mechanics correlate with patellofemoral deep cartilage UTE-T2* 2 years after anterior cruciate ligament reconstruction. Am J Sports Med. 2021;49(3):675-683.

49.

Wirth

Eckstein

A technique for regional analysis of femorotibial cartilage thickness based on quantitative magnetic resonance imaging. IEEE Trans Med Imaging. 2008;27(6):737-744.

50.

Wirth

Ladel

Maschek

Wisser

Eckstein

Roemer

Quantitative measurement of cartilage morphology in osteoarthritis: current knowledge and future directions. Skeletal Radiol. 2023;52(11):2107-2122.

51.

Woo

Kanamori

Zeminski

Yagi

Papageorgiou

FH.

The effectiveness of reconstruction of the anterior cruciate ligament with hamstrings and patellar tendon. A cadaveric study comparing anterior tibial and rotational loads. J Bone Joint Surg Am. 2002;84(6):907-914.

52.

Yoo

Papannagari

Park

DeFrate

Gill

The effect of anterior cruciate ligament reconstruction on knee joint kinematics under simulated muscle loads. Am J Sports Med. 2005;33(2):240-246.

53.

Zhang

Liu

Sun

Xie

Chen

Knee cartilage change within 5 years after ACLR using hamstring tendons with preserved tibial-insertion: a prospective randomized controlled study based on magnetic resonance imaging. J Clin Med. 2022;11(20).