MR imaging of the temporomandibular joint

Using deep learning to predict temporomandibular joint disc perforation based on magnetic resonance imaging

The goal of this study was to develop a deep learning-based algorithm to predict temporomandibular joint (TMJ) disc perforation based on the findings of magnetic resonance imaging (MRI) and to validate its performance through comparison with previously reported results. The study objects were obtained by reviewing medical records from January 2005 to June 2018. 299 joints from 289 patients were divided into perforated and non-perforated groups based on the existence of disc perforation confirmed during surgery. Experienced observers interpreted the TMJ MRI images to extract features. Data containing those features were applied to build and validate prediction models using random forest and multilayer perceptron (MLP) techniques, the latter using the Keras framework, a recent deep learning architecture. The area under the receiver operating characteristic (ROC) curve (AUC) was used to compare the performances of the models. MLP produced the best performance (AUC 0.940), followed by random forest (AUC 0.918) and disc shape alone (AUC 0.791). The MLP and random forest were also superior to previously reported results using MRI (AUC 0.808) and MRI-based nomogram (AUC 0.889). Implementing deep learning showed superior performance in predicting disc perforation in TMJ compared to conventional methods and previous reports.

Imaging Approach to Temporomandibular Joint Disorders

Internal derangement is the most common temporomandibular joint disorder. Degenerative osteoarthritis and trauma are next in frequency. Less common pathology includes rheumatoid arthritis, synovial chondromatosis, calcium pyrophosphate dehydrate deposition disease, pigmented villonodular synovitis, tumors, infection, and osteonecrosis. We provide a systematic approach to facilitate interpretation based on major anatomic structures: disc-attachments, joint space, condyle, and lateral pterygoid muscle. Relevant graphic anatomy and state of the art imaging are discussed in correlation with current clinical and therapeutic highlights of pathologic entities affecting the joint.

Sensitivity of quantitative UTE MRI to the biomechanical property of the temporomandibular joint disc

PURPOSE

To quantify MR properties of discs from cadaveric human temporomandibular joints (TMJ) using quantitative conventional and ultrashort time-to-echo magnetic resonance imaging (UTE MRI) techniques and to corroborate regional variation in the MR properties with that of biomechanical indentation stiffness.

METHODS

This study was exempt from the institutional review board approval. Cadaveric (four donors, two females, 74 ± 10.7 years) TMJs were sliced (n = 14 slices total) sagittally and imaged using quantitative techniques of conventional spin echo T2 (SE T2), UTE T2*, and UTE T1rho. The discs were then subjected to biomechanical indentation testing, which is performed by compressing the tissue with the blunt end of a small solid cylinder. Regional variations in MR and indentation stiffness were correlated. TMJ of a healthy volunteer was also imaged to show in vivo feasibility.

RESULTS

Using the ME SE T2 and the UTE T1rho techniques, a significant (each p < 0.0001) inverse relation between MR and indentation stiffness properties was observed for the data in the lower range of stiffness. However, the strength of correlation was significantly higher (p < 0.05) for UTE T1rho (R(2) = 0.42) than SE T2 (R(2) = 0.19) or UTE T2* (R(2) = 0.02, p = 0.1) techniques.

CONCLUSION

The UTE T1rho technique, applicable in vivo, facilitated quantitative evaluation of TMJ discs and showed a high sensitivity to biomechanical softening of the TMJ discs. With additional work, the technique may become a useful surrogate measure for loss of biomechanical integrity of TMJ discs reflecting degeneration.

MR imaging of the temporomandibular joint

MR imaging allows detailed evaluation of temporomandibular (TMJ) anatomy because of its inherent tissue contrast and high resolution. Joint biomechanics can be assessed through imaging patients in the closed and open jaw positions. Despite the accuracy of MR imaging in detecting disc position, results must be interpreted together with clinical findings, because an anteriorly displaced disc can be seen in up to one-third of asymptomatic patients, and a normal disc position can be seen in up to one-quarter of symptomatic patients. Interpretation of MR imaging requires knowledge of the normal anatomy and an understanding of normal and abnormal biomechanics.

The radiologic evaluation of tinnitus

Tinnitus ("ringing in the ears") is a prevalent symptom in the general population, and often brings patients to medical attention. Many causes of tinnitus are evident radiographically. The most frequently-encountered causes of tinnitus are discussed, and imaging recommendations are provided.