An Examination of Temporomandibular Joint Disc Displacement through Magnetic Resonance Imaging by Integrating Artificial Intelligence: Preliminary Findings

• Published in: Medicina (Kaunas, Lithuania) Vol. 60; no. 9; p. 1396
• Main Authors: Almășan, Oana, Mureșanu, Sorana, Hedeșiu, Petra, Cotor, Andrei, Băciuț, Mihaela, Roman, Raluca, TEAM Project Group, TEAM Project Group
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.09.2024; MDPI
• Subjects: Adult; Algorithms; Artificial Intelligence; Computer-aided medical diagnosis; Datasets; deep learning; Diagnosis; disc displacement; Female; Humans; Image processing; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Male; Methods; Middle Aged; temporomandibular joint; Temporomandibular Joint Disc - diagnostic imaging; Temporomandibular joint disorders; Temporomandibular Joint Disorders - diagnostic imaging; Romania; deep learning; disc displacement; magnetic resonance imaging; temporomandibular joint; artificial intelligence
• ISSN: 1648-9144; 1010-660X; 1648-9144
• DOI: 10.3390/medicina60091396

Background and Objectives: This research was aimed at constructing a complete automated temporomandibular joint disc position identification system that could assist with magnetic resonance imaging disc displacement diagnosis on oblique sagittal and oblique coronal images. Materials and Methods: The study included fifty subjects with magnetic resonance imaging scans of the temporomandibular joint. Oblique sagittal and coronal sections of the magnetic resonance imaging scans were analyzed. Investigations were performed on the right and left coronal images with a closed mouth, as well as right and left sagittal images with closed and open mouths. Three hundred sagittal and coronal images were employed to train the artificial intelligence algorithm. Results: The accuracy ratio of the completely computerized articular disc identification method was 81%. Conclusions: An automated and accurate evaluation of temporomandibular joint disc position was developed by using both oblique sagittal and oblique coronal magnetic resonance imaging images.