Anatomically Consistent TMJ Disc Segmentation via Semantic Anchoring and Clinical Priors

2026-06-19

AI Summary

This study addresses the challenges in segmenting the temporomandibular joint (TMJ) disc in MRI, which suffers from small volume, low contrast, and high morphological variability, leading to fragmented and anatomically inconsistent results that compromise clinical reliability. To overcome these limitations, the authors propose the TISC framework, which innovatively integrates Prototype Semantic Anchoring (PSA) with Clinical Metadata-guided Point-level Refinement (C-MPR). PSA enables robust disc localization within the MedDINOv3 feature space, while C-MPR leverages clinical indicators—such as limited mouth opening—to refine segmentation boundaries. Additionally, multi-slice feature aggregation enhances inter-slice contextual consistency. Evaluated on 2,488 MRI scans from 1,300 patients, the method achieves up to a 4.96-point improvement in Dice score over strong baselines, significantly enhancing both anatomical accuracy and clinical credibility of TMJ disc segmentation.

Abstract

Segmenting the temporomandibular joint (TMJ) disc from MRI is essential for accurate diagnosis of internal derangement, yet it remains unreliable in practice due to its small size, low contrast, and morphological variability. Existing methods, primarily adapted from general segmentation architectures, often produce fragmented or anatomically inconsistent masks, leading to unstable measurements of disc position and shape for downstream diagnosis. To address these challenges, we propose TISC, a TMJ disc segmentation framework that integrates semantic anchoring with clinical metadata-guided boundary refinement. The framework first establishes robust disc localization in the foundation model feature space via a Prototypical Semantic Anchoring (PSA) module that aggregates adjacent-slice MedDINOv3 features and derives a prototype-driven similarity map. It then performs targeted boundary refinement through a Clinical-Metadata Point Refinement (C-MPR) module, with point-wise predictions modulated by Mouth Open Limitation (MOL), a clinical indicator associated with disc displacement without reduction. On a large-scale cohort of 2,488 PD MRI volumes from 1,300 patients, our method achieves up to a 4.96 Dice improvement over strong baselines across diverse architectures, delivering more anatomically coherent and clinically reliable TMJ disc segmentation.

Problem

• TMJ disc segmentation
• anatomical consistency
• MRI
• internal derangement
• morphological variability

Innovation

• semantic anchoring
• clinical priors
• TMJ disc segmentation
• prototype-driven similarity
• boundary refinement