Cross-Modal Feature Alignment for Automated Diagnosis of Temporomandibular Joint Disorders Using Medical Imaging

Authors

  • Amie L. Mulvany School of Computing, Clemson University, Clemson, SC, USA. Author

Keywords:

temporomandibular joint disorders; cross-modal alignment; medical image analysis; contrastive learning; uncertainty calibration; multimodal diagnosis

Abstract

Temporomandibular joint disorders (TMDs) comprise heterogeneous disc, bone, and inflammatory conditions whose imaging appearance varies across magnetic resonance imaging, computed tomography or cone-beam computed tomography, and three-dimensional intraoral scans. Automated diagnosis is difficult because no single modality captures both soft-tissue and osseous pathology, images are acquired with different geometry and contrast, and labelled cohorts are small and unevenly annotated. This manuscript develops a cross-modal feature alignment framework for TMD diagnosis that treats images from different modalities as partial observations of a shared latent disease state. The contribution is a methodological architecture: modality-specific encoders map each input into a latent representation; a shared projection space is trained with contrastive and registration-consistency objectives; anatomically homologous structures, including the condyle, glenoid fossa, articular eminence, disc, and dental occlusion, provide anchors; and an attention-weighted fusion layer produces multi-label diagnostic probabilities with uncertainty estimates. Diagnostic inference is formulated as latent-variable evidence integration, and alignment strength, registration accuracy, missing-modality behavior, and calibration are analyzed as coupled design choices. The analysis shows why latent alignment is preferable to early pixel-level fusion and late score averaging for heterogeneous TMJ imaging, identifies failure modes such as over-alignment, registration drift, and shortcut learning, and specifies patient-level evaluation and calibration criteria needed for clinical interpretation. The framework provides a reproducible design for combining complementary imaging evidence without treating any modality as a substitute for another.

References

1. Schiffman, E., Ohrbach, R., Truelove, E., Look, J., Anderson, G., Goulet, J. P., List, T., & International RDC/TMD Consortium Network. (2014). Diagnostic criteria for temporomandibular disorders (DC/TMD) for clinical and research applications: Recommendations of the International RDC/TMD Consortium Network and Orofacial Pain Special Interest Group. Journal of Oral & Facial Pain and Headache, 28(1), 6–27.

2. Dworkin, S. F., & LeResche, L. (1992). Research diagnostic criteria for temporomandibular disorders: Review, criteria, examinations and specifications, critique. Journal of Craniomandibular Disorders, 6(4), 301–355.

3. Tasaki, M. M., & Westesson, P. L. (1993). Temporomandibular joint: Diagnostic accuracy with sagittal and coronal MR imaging. Radiology, 186(3), 723–729.

4. de Leeuw, R., & Klasser, G. D. (Eds.). (2018). Orofacial pain: Guidelines for assessment, diagnosis, and management (6th ed.). Quintessence Publishing.

5. Baltrušaitis, T., Ahuja, C., & Morency, L.-P. (2019). Multimodal machine learning: A survey and taxonomy. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(2), 423–443.

6. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770–778).

7. Ronneberger, O., Fischer, P., & Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015 (pp. 234–241). Springer.

8. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. In Advances in Neural Information Processing Systems (Vol. 30, pp. 5998–6008).

9. Balakrishnan, G., Zhao, A., Sabuncu, M. R., Guttag, J., & Dalca, A. V. (2019). VoxelMorph: A learning framework for deformable medical image registration. IEEE Transactions on Medical Imaging, 38(8), 1788–1800.

10. Avants, B. B., Epstein, C. L., Grossman, M., & Gee, J. C. (2008). Symmetric diffeomorphic image registration with cross-correlation: Evaluating automated labeling of elderly and neurodegenerative brain. Medical Image Analysis, 12(1), 26–41.

11. Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020). A simple framework for contrastive learning of visual representations. In Proceedings of the 37th International Conference on Machine Learning (pp. 1597–1607).

12. Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., Krueger, G., & Sutskever, I. (2021). Learning transferable visual models from natural language supervision. In Proceedings of the 38th International Conference on Machine Learning (pp. 8748–8763).

13. Liu, S., Jin, J., Li, W., & Yan, X. (2026). Segmentation-Guided Semi-Supervised Landmark Detection on Intraoral 3D Scans. Preprints. https://doi.org/10.20944/preprints202609.1722.v1

14. Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017). Grad-CAM: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE International Conference on Computer Vision (pp. 618–626).

15. Gal, Y., & Ghahramani, Z. (2016). Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In Proceedings of the 33rd International Conference on Machine Learning (pp. 1050–1059).

16. Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2017). Simple and scalable predictive uncertainty estimation using deep ensembles. In Advances in Neural Information Processing Systems (Vol. 30, pp. 6402–6413).

17. Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., & Bengio, Y. (2014). Generative adversarial nets. In Advances in Neural Information Processing Systems (Vol. 27, pp. 2672–2680).

18. Kingma, D. P., & Welling, M. (2014). Auto-encoding variational Bayes. In Proceedings of the International Conference on Learning Representations.

19. Hu, J., Shen, L., & Sun, G. (2018). Squeeze-and-excitation networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 7132–7141).

20. Oord, A. van den, Li, Y., & Vinyals, O. (2018). Representation learning with contrastive predictive coding. arXiv. https://arxiv.org/abs/1807.03748

Downloads

Published

2026-10-15