Accelerated MRI in temporomandibular joints using AI-assisted compressed sensing technique: a feasibility study

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Temporomandibular joint (TMJ) disorders are common, often causing orofacial pain that limits patient tolerance for lengthy MRI examinations.
• Accelerated MRI techniques are crucial for improving patient comfort and optimizing clinical workflow for TMJ imaging.

Purpose of the Study:

• To evaluate the feasibility of artificial intelligence-assisted compressed sensing (ACS) for accelerated temporomandibular joint (TMJ) MRI.
• To compare the performance of ACS MRI with parallel imaging (PI) and standard (STD) MRI protocols in terms of image quality, quantitative metrics, and diagnostic agreement.

Main Methods:

• Prospective enrollment of 51 participants with TMJ symptoms.
• Bilateral TMJ imaging using ACS (6:08 min), PI (10:57 min), and STD (13:28 min) protocols.
• Qualitative assessment of image quality and structure visibility; quantitative analysis of SNR and CNR; evaluation of diagnostic agreement for joint effusion and disc displacement.

Main Results:

• ACS protocol demonstrated significantly higher overall image quality and structure visibility than the STD protocol (p < 0.05), and comparable quality to the PI protocol.
• Quantitative analysis revealed significantly higher SNR and CNR for the ACS protocol compared to STD (p < 0.05), with SNR comparable to PI.
• Good to excellent inter-protocol and inter-observer agreement (κ = 0.699–1.000) was achieved for diagnosing TMJ abnormalities.

Conclusions:

• Accelerated MRI with ACS technique is feasible for TMJ imaging, significantly reducing scan time.
• ACS provides superior or equivalent image quality and diagnostic performance compared to PI and STD protocols.
• The ACS technique enhances patient comfort and optimizes the clinical MRI workflow for TMJ disorders.