Assessment of temporomandibular joint disease by panoramic radiography: reliability and validity in relation to...

To assess the reliability and validity of panoramic radiography for TMJ evaluation in comparison with tomography.

One TMD specialist assessed the panoramic radiographs of 50 TMD patients and 20 non-TMD subjects on two occasions. A third evaluation was made after calibration with an oral radiologist who also assessed the radiographs twice and the patients' corresponding tomograms. Each TMJ was assessed for five features.

Absence of TMJ disease varied between 56 and 87% for the different features and the two observers on the panoramic radiographs. Condylar flattening was the most frequent positive finding. Mean intra-observer agreement varied between 0.34 and 0.79. Mean inter-observer agreement varied between 0.22 and 0.65 and increased after calibration. Agreement was worse for the temporal component. Pathological findings were more frequent on the tomograms and varied between 6 and 19%. Mean sensitivity varied between 0.00 and 0.60 and mean specificity between 0.59 and 0.95. Mean PPV ranged from 0.25 to 0.90 and NPV from 0.02 to 0.22.

Evaluation of bony changes on the condyles has acceptable reliability and specificity but low sensitivity, whereas the temporal component has low reliability and accuracy. Positive findings often correspond to disease while negative findings do not exclude it. If bony abnormality is suspected in the TMJ, and the panoramic radiograph is negative, tomography may be indicated.