Artificial Intelligence-Driven Decision-Making for Knee Joint Manipulation Following Primary Total Knee Arthroplasty.
Stefano Ghirardelli, Kai Chun Augustine Chan, Pieralberto Valpiana et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review used literature on range-of-motion recovery and manipulation after primary total knee arthroplasty to train an artificial intelligence model and test decision scenarios. The model suggested considering manipulation under anesthesia at six weeks for knee flexion < 80 degrees and/or an extension deficit greater than 20 degrees, predicting mean gains of 26 degrees in flexion and three degrees in extension. Expected flexion gains fell to 17 degrees when manipulation was delayed beyond 90 days. The authors report that the framework aligns with published consensus but should support, not replace, clinical judgment.
Key findings
- At six weeks, the model indicated manipulation for flexion < 80 degrees and/or extension deficits greater than 20 degrees.
- Predicted mean improvements were 26 degrees in flexion and three degrees in extension.
- Delaying manipulation beyond 90 days reduced expected flexion gains to 17 degrees.
- For stiffness beyond three months, the model advised adding arthroscopic lysis of adhesions, particularly when flexion was less than 90 degrees.
Why it matters
Postoperative stiffness presents varied clinical scenarios, while evidence on thresholds and timing for manipulation remains limited. This model offers a structured way to relate postoperative motion and timing to expected outcomes. Its predictions may help frame decisions, but the evidence base and multifactorial nature of stiffness mean that model outputs require clinical interpretation.
What to keep in mind
The framework is based on literature described as limited in quality and heterogeneous; its recommendations are model-generated rather than established by a reported clinical trial.