Quadriceps force and medio-laterally directed joint force during knee flexion in a personalized patellofemoral joint model

TL;DR

This study uses MRI data to create personalized knee models and analyzes how anatomical differences affect patellofemoral joint forces during a simulated squat.

Contribution

The study introduces a method to integrate MRI-derived knee geometries into musculoskeletal simulations for analyzing patellofemoral joint loading.

Findings

Mediolateral patellofemoral joint reaction force increased with knee flexion, peaking between 7 and 67 N/kg BW.

A model with Dejour type B trochlear dysplasia showed the highest lateral loading and quadriceps force.

The method enables in-silico analysis of anatomical risk factors for patellofemoral instability.

Abstract

Patellofemoral instability (PFI) is a multifactorial condition influenced by complex interactions between anatomical structures and dynamic stabilizers. Accurate assessment of mediolateral patellofemoral joint loading remains challenging, particularly during physiologically relevant joint angles simulated under controlled kinematic conditions. Anatomy-based individualized musculoskeletal models provide an in-silico approach to investigate these biomechanical parameters, such as the loading patterns. The purpose of this study was to develop individualized musculoskeletal knee models based on MRI-derived anatomical data from cadaveric specimens and to quantify the mediolateral component of the patellofemoral joint reaction force during a standardized, computer-driven squat simulation. Quadriceps muscle forces were estimated to assess the demand on dynamic stabilizers.

: MRI scans from four cadaveric lower limbs were segmented to reconstruct the distal femur, proximal tibia, and patella. The resulting STL bone models of the femur, tibia, and patella were aligned to TLEM2 templates using anatomical landmarks in CATIA V5 and integrated into the AnyBody Modeling System via custom AnyScript routines. A predefined, computer generated squat motion consisting of asymmetric flexion–extension cycle between 0° and 90° at a constant, scripted angular velocity of 60°/s was applied using a kinematic driver. The original bodyweight-level loading conditions from the AnyBody squat model were retained; no external weights or EMG data were included. Patellofemoral joint reaction forces and quadriceps muscle forces were estimated using inverse dynamics calculations.

The mediolateral patellofemoral joint reaction force increased with knee flexion in all models, reaching specimen-specific peaks ranging from 7 N/kg BW to 67 N/kg BW (maximum of 24 ± 25 N/kg BW). The model exhibiting a pronounced supratrochlear bony prominence (Dejour type B trochlear dysplasia) showed the highest lateral loading and quadriceps force. Muscle forces are reported as magnitudes.

This study demonstrates that MRI-derived, individualized knee geometries can be integrated into a musculoskeletal simulation framework. This allows the investigation of how anatomical variation affects patellofemoral joint loading during a controlled, computer-generated squat motion. By integrating individualized knee anatomy into a validated simulation framework, our approach enables the in-silico analysis of anatomical risk factors—such as trochlear dysplasia and increased TT–TG distance—that are clinically relevant for patellofemoral instability. This method may support future biomechanical investigations and preoperative planning by providing reproducible, anatomy-driven insights into joint mechanics.

Figures