Reducing uncertainty when using knee-specific finite element models by assessing the effect of input parameters.

Area of Science:

• Biomechanics
• Computational modeling
• Orthopaedic research

Background:

• Knee-specific factors influencing contact mechanics are not well understood.
• Finite Element (FE) models are valuable for studying knee joint contact mechanics.
• Input parameter uncertainty in FE models leads to a range of output values.

Purpose of the Study:

• To quantify the reduction in output uncertainty in human knee FE models.
• To assess the impact of using clinically measurable inputs (patient-derived inputs, PDIs) on FE model output variability.
• To investigate how fixing specific input parameters affects the range of FE model outputs.

Main Methods:

• A statistically augmented FE approach was applied to three human cadaveric knees.
• Two simulation conditions were used: varying all inputs versus fixing clinically measurable inputs (PDIs).
• Model output uncertainty was compared between conditions with all variables free and with PDIs fixed.

Main Results:

• Fixing body mass index and meniscal-bony insertion points reduced FE model output uncertainty by 20-60%.
• The degree of uncertainty reduction varied significantly between individual knees.
• Knees with greater anterior-posterior translation during gait showed larger uncertainty reductions when PDIs were used.

Conclusions:

• Clinically measurable inputs can significantly reduce uncertainty in human knee FE models.
• This approach is a foundational step towards developing patient-specific FE models for knee joint analysis.
• FE models incorporating PDIs hold promise for more accurate clinical applications in knee biomechanics.