Possibilities of ultrasound diagnostics and the use of artificial neural network to assess the morphology and size of the articular disc of the temporomandibular joint; [Возможности ультразвуковой диагностики и применения искусственной нейронной сети для оценки морфологии и размеров суставного диска височно-нижнечелюстного сустава]

To date, the «gold standard» for diagnosing internal disorders of the temporomandibular joint (TMJ) is magnetic resonance imaging, but ultrasonography allows to accurately assess the presence of anterior displacement of the articular disc, as well as takes less time, requires less expensive equipment. The aim of this work was to develop and train an artificial neural network (ANN) to describe the shape and size of the TMJ articular disc on static images during ultrasonography. Materials and methods. A total of 300 TMJ ultrasounds were performed. The TMJ ultrasound was performed on high-resolution equipment Philips HD 11XE (USA) using a 12 MHz linear transducer. The ViSurgery computer program (Russia) was used to construct the ANN. Results. Following dimensions of the articular disk were obtained: width of the anterior thickening — 2.09±0.69 mm, posterior thickening — 1.6±0.55 mm, middle part — 1.98±0.69 mm, which does not contradict the literature data. Conclusion. The obtained results indicate that it is possible to use ANN for visualization of the articular disc of TMJ on ultrasonograms and to measure its size with an average accuracy of 0.4—0.5 mm. © 2024 Clinical Dentistry LLC. All rights reserved.