Artificial Intelligence vs. Human Experts in Temporomandibular Joint MRI Interpretation: A Systematic Review

Area of Science:

• Radiology
• Artificial Intelligence
• Medical Imaging

Background:

• Magnetic resonance imaging (MRI) is the gold standard for evaluating temporomandibular joint (TMJ) disorders.
• Artificial intelligence (AI) and deep learning offer potential advancements in medical image analysis.
• Increasing imaging demands necessitate efficient and accurate diagnostic tools.

Purpose of the Study:

• To systematically review and compare the diagnostic performance of AI models against human experts in TMJ MRI analysis.
• To assess the potential of AI as an adjunct tool for radiologists interpreting TMJ MRIs.
• To evaluate AI's effectiveness in identifying specific TMJ pathologies.

Main Methods:

• Systematic review adhering to PRISMA 2020 guidelines, registered in PROSPERO.
• Comprehensive search of major scientific databases (PubMed, ScienceDirect, etc.) for studies from 2020-2026.
• Inclusion of studies comparing AI/ML/DL models against human expert interpretation for TMJ MRI, extracting sensitivity, specificity, accuracy, and AUC.

Main Results:

• Five retrospective studies (118-1474 patients) evaluated AI for TMJ disorders like disc displacement and osteoarthritis.
• AI models demonstrated strong discriminative performance (AUC 0.79-0.98), achieving accuracy comparable to experienced radiologists.
• AI showed higher specificity and similar overall accuracy, while human experts often had higher sensitivity; AI performance neared expert levels for osteoarthritis.

Conclusions:

• AI achieves diagnostic performance comparable to experienced clinicians in TMJ MRI interpretation.
• AI shows significant promise as a decision-support tool for TMJ disorder assessment.
• Further rigorous validation is required before AI can be fully integrated as a complementary tool to expert radiological assessment.