Evaluating Anterior Cruciate Ligament Risk of Injury Using a Graph Convolutional Autoencoder
Authors: Athanasios Babouras, Paul A. Martineau, Thomas Fevens
Organizations: Department of Surgical and Interventional Sciences, McGill University, Montreal, QC, Canada · Division of Orthopaedic Surgery, McGill University, Montreal, QC, Canada · Department of Computer Science and Software Engineering, Concordia University, Montreal, QC, Canada
Current solutions for assessing the risk of anterior cruciate ligament (ACL) injury typically rely on expensive motion-capture labs that require substantial resource investment or subjective evaluations of specific athletic stress tests by experts. The Drop Vertical Jump (DVJ) and the Single Leg Hop (SLH) are two examples of such tests typically used to track progress in ACL rehabilitation. In this paper, we develop an unsupervised learning model that runs entirely on a smartphone for two tasks, based on an autoencoder architecture with multiple layers of graph convolutional networks, which utilizes pose estimation data to determine a person's injury risk status. Our model accurately predicts athletes at high risk of injury for both athletic stress tests, with average sensitivities and specificities of 78% and 78% for the DVJ, and 75% and 74% for the SLH. Our findings support the development of a widely accessible smartphone solution for real-time screening of ACL injury risk.
Figures & tables
Age (years)
Height (m)
Weight (kg)
BMI (kg/m 2 )
Female
20±2
1.69±0.07
66.84±8.30
23.49±2.73
Male
21±2
1.82±0.07
86.51±15.54
25.95±4.08
Table 1: Mean ± Standard Deviation for Participant Demographics
Figure 2
Figure 3: Model Architecture Overview
Joint
Weight
Joint
Weight
Spine Base
0.213
Right Wrist
0.010
Spine Mid
0.142
Right Hand
0.006
Spine Shoulder
0.028
Right Hand Tip
0.003
Neck
0.081
Right Thumb
0.003
Head
0.069
Left Hip
0.071
Left Shoulder
0.028
Left Knee
0.100
Table 2: Joint Weights for Center of Gravity Calculation
Configuration
Sens.
Spec.
F1
All Joints
0.88
0.71
0.62
Lower Extremities + Trunk
0.83
0.69
0.59
Lower Extremities + Trunk + CoG
0.83
0.81
0.69
Table 3: Performance Metrics for Different Joint Sets
Run
Drop Vertical Jump (DVJ)
Single Leg Hop (SLH)
Sensitivity
Specificity
F1 Score
Sensitivity
Specificity
F1 Score
1
83%
76%
0.65
76%
70%
0.50
2
83%
81%
0.69
76%
77%
0.55
3
71%
83%
0.68
76%
75%
0.53
4
76%
82%
0.68
73%
74%
0.52
5
78%
68%
0.58
73%
76%
0.53
Table 4: Repeated Cross-Validation Results for DVJ and SLH (5 Runs)
Method
Sensitivity
Specificity
Zazulak et al. [ 25 ]
0.90
0.56
Bates et al. [ 3 ]
0.78
0.83
Corban et al. [ 7 ]
0.80
0.74
Ours
0.78
0.78
Table 5: Sensitivity and Specificity of Injury Prediction Using the DVJ
Figure 8Figure 9
Figure 8: Drop Vertical Jump Angle Importance Heatmap
Figure 9: Single Leg Hop Angle Importance Heatmap
Variable
Explanation
KneeLeftCoronal
Angle of the left knee in the coronal plane.
KneeLeftSagittal
Angle of the left knee in the sagittal plane.
KneeRightCoronal
Angle of the right knee in the coronal plane.
KneeRightSagittal
Angle of the right knee in the sagittal plane.
PelvisTilt_Y
Angle measuring pelvic obliquity.
PelvisTilt_Z
Angle measuring pelvic rotation.
Table 6: Descriptions of Biomechanical Parameters Used for Figures 8 and 9
One of the most common reasons for anterior cruciate ligament (ACL) reconstruction failure is femoral tunnel malpositioning (ACL footprint center and tunnel orientation). Such failures may lead to the development of meniscal pathology and osteoarthritis. Accurate ACL femoral footprint identification is therefore essential for precise tunnel placement, restoration of the native knee joint mechanics, post-surgical knee joint health and prevention of graft failure. Recent advances in artificial intelligence (AI) bring new opportunities to improve image-guided orthopedic surgery. However, at present, existing AI research focuses primarily on ACL segmentation and rupture classification based on pre- and post-operative magnetic resonance (MR) images. Identification of the ACL footprint center using deep learning methods has not been thoroughly researched. Thus, the purpose of this study is to explore 3D deep learning models for ACL femoral footprint identification directly from 3D MR images. Two comprehensive 3D deep learning architectures were developed: a 3D graph convolutional neural network-based geometric model applied to 3D femoral meshes; and a 3D landmark-enhanced identification model based on 3D MR images. A total of 4883 right and 3087 left knee image sets were used from a publicly available database. Eighty percent (80%) were applied to model generation, and twenty percent (20%) were preserved for model testing. Both models achieved excellent performance; however, the image-based method outperformed the model-based method (average error of 2.1mm vs 2.8 mm). Thus, 3D deep learning provides a feasible clinical approach for ACL footprint localization and has the potential to improve ACL reconstruction footprint accuracy.
Ruida Cheng, Gabriel Gibson, Ali Uneri +2
Scientific Application Services, Center for Information Technology, NIH, Bethesda, MD · Rehabilitation Medicine Department, Clinical Center, NIH, Bethesda, MD · Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD +1
Accident anticipation aims to recognize anomalous driving cues before a crash while avoiding false alarms during normal driving. Existing approaches typically formulate this task as binary classification, focusing on whether an accident will occur rather than when it will occur. We propose PRE-ACT, a framework that models accident risk as a continuously evolving signal that increases as the crash approaches. By explicitly enforcing temporal ordering and distance-to-accident awareness, our method progressively raises risk while suppressing premature alarms, leading to significant improvements on MM-AU subsets and Nexar. We further introduce a Separation Score to evaluate the global behavior of predicted risk curves beyond local temporal windows. Code and visualizations are available at https://github.com/giddyyupp/PRE-ACT.
Samet Hicsonmez, Eray Çakar, Nermin Samet +1
Independent Researcher, Paris, France · Department of Computer Engineering, KUIS AI, Koç University, Istanbul, Turkey · Valeo.ai, Paris, France
Accident anticipation aims to predict impending collisions from dashcam videos and trigger early alerts. Existing methods rely on binary supervision with manually annotated "anomaly onset" frames, which are subjective and inconsistent, leading to inaccurate risk estimation. In contrast, we propose RiskProp, a novel collision-anchored self-supervised risk propagation paradigm for early accident anticipation, which removes the need for anomaly onset annotations and leverages only the reliably annotated collision frame. RiskProp models temporal risk evolution through two observation-driven losses: first, since future frames contain more definitive evidence of an impending accident, we introduce a future-frame regularization loss that uses the model's next-frame prediction as a soft target to supervise the current frame, enabling backward propagation of risk signals; second, inspired by the empirical trend of rising risk before accidents, we design an adaptive monotonic constraint to encourage a non-decreasing progression over time. Experiments on CAP and Nexar demonstrate that RiskProp achieves state-of-the-art performance and produces smoother, more discriminative risk curves, improving both early anticipation and interpretability.
Yiyang Zou, Tianhao Zhao, Peilun Xiao +9
School of Computer Science, Wuhan University · Zhongguancun Academy, Beijing, China · Didi Chuxing +1