Evidence-based evaluation of three imaging methods for the temporomandibular disc.

OBJECTIVES To review the literature and evaluate the evidence for the diagnostic outcome of arthrography, computed tomography (CT) and magnetic resonance imaging (MRI) in the assessment of temporomandibular joint (TMJ) disc position. METHODS A literature search from 1978 to 1994 was conducted. Data on sensitivity, specificity, predictive values and likelihood ratios were collected and, if not available, calculated for each imaging method. Measures of observer performance were also recorded. RESULTS In total, more than 400 publications were retrieved on imaging of TMJ disc position, 219 on arthrography, 99 on CT and 147 on MRI, with a marked decrease in recent years. The majority of the publications (54%) presented series of patients. The diagnostic outcome cold be obtained from only 7% of the studies and the observer performance from only two of these. Arthrography had the highest diagnostic outcome for the diagnosis of anterior disc position; sensitivity 0.90, specificity 0.80, positive predictive value 0.88 and negative predictive value 0.82. Corresponding figures for CT were 0.66, 0.68, 0.66 and 0.74 and for MRI 0.86, 0.63, 0.67 and 0.83. The diagnostic outcomes expressed as the likelihood ratios for positive test outcome were 4.5 for arthrography, 2.3 for MRI and 2.1 for CT. The outcomes in diagnosing sideways and rotational displacements were higher for MRI (sensitivity 0.81, specificity 0.87, positive predictive value 0.82, negative predictive value 0.88) than for arthrography (0.64, 0.83, 0.70, 0.79). The likelihood ratios were 6.2 for MRI and 3.8 for arthrography. The interobserver performance ranged between moderate to substantial for arthrography and was almost perfect for MRI. CONCLUSIONS Based on current evidence on the diagnostic outcome, MRI seems to be the method of choice for diagnosing TMJ disc position. Arthrography has a higher diagnostic outcome for anterior disc position but the disadvantage of being an invasive method. We suggest that the quality of the evidence should be improved and that an analysis of the impact of the imaging methods on patient treatment should be performed.