Automated analysis of knee joint alignment using detailed angular values in long leg radiographs based on deep...

Area of Science:

• Orthopedics
• Medical Imaging
• Artificial Intelligence

Background:

• Lower limb malalignment necessitates precise evaluation for effective correction.
• Accurate quantification of mechanical tibiofemoral angle (mTFA), mechanical lateral distal femoral angle (mLDFA), medial proximal tibial angle (MPTA), and joint line convergence angle (JLCA) is crucial.

Purpose of the Study:

• To develop an automated system for evaluating lower limb alignment using deep learning on full-length weight-bearing radiographs.
• To quantify mTFA, mLDFA, MPTA, and JLCA automatically.

Main Methods:

• A retrospective study involving 404 radiographs for algorithm development and 30 for external validation.
• Performance evaluation using Dice Similarity Coefficient (DSC) for segmentation and Intraclass Correlation Coefficient (ICC) for alignment parameters.
• Comparison of automated measurement time against manual measurement.

Main Results:

• The segmentation algorithm showed excellent agreement with manual segmentation (average similarity: 89-97%).
• Internal validation demonstrated good to very good agreement for alignment parameters (ICC: 0.7213-0.9865).
• External validation showed good to very good interobserver correlations (ICC: 0.7126-0.9695), with automated measurements being 3.44 times faster.

Conclusions:

• A deep learning-based automated measurement algorithm accurately quantifies lower limb alignment from radiographs.
• The automated system offers a faster and reliable alternative to manual measurements for clinical use.