Assessment of sensitivity and specificity of clinical evaluation of cast restoration marginal accuracy compared to stereomicroscopy

Abstract

One of the most important criteria for the clinical acceptability and longevity of cast restorations is marginal adaptation. However, there is limited information on the accuracy of commonly practiced evaluation techniques. This study assessed the effect of preparation design on marginal adaptation. In addition, sensitivity and specificity of clinical evaluations of marginal adaptations of cast restorations were assessed and compared to stereomicroscopy. Three Ivorine molar teeth of differing designs were prepared: (1) a complete crown preparation with a chamfer finish line, (2) a complete crown preparation with a buccal shoulder and beveled finish line, and (3) a three-quarter crown preparation with proximal boxes and beveled finish line. Twenty-four castings (n = 8) were prepared (Ney technique casting metal). Twelve circumferential sites were identified for examination on each casting using a stereomicroscope, which was interpreted using software (Bioquant 98). The clinical examinations included explorer examination to detect marginal gap by tactile feel, followed by examination with a disclosing media (GC FitChecker). Chi-square tests of independence were used to evaluate the effect of preparation design and compare agreement between examination methods for detection of marginal gap size of greater than or equal to 30 microm (alpha=.05). The overall marginal adaptation of preparation designs was not significantly different. Results of explorer examination at 30-microm sensitivity and specificity were 46% and 97%, respectively. Similar values for disclosing media at 30 microm were 12% sensitivity and 97% specificity. The overall percent agreement with the stereomicroscope using the explorer was 58.3%, with 23.6% correct acceptance rate and 34.7% correct rejection rate. With respect to detection of marginal fit, the combination of disclosing material with explorer data did not yield significantly improved results (59.4%), as compared to explorer alone. A receiver operator characteristic curve yielded satisfactory sensitivity (0.8) and specificity (0.9) at a gap size value of 124 microm. Within the limitations of this study, the preparation design did not significantly affect the marginal adaptation. Clinical detection, with similar sensitivity and specificity as the stereomicroscope, occurred at marginal openings greater than or equal to 124 microm. Therefore, commonly used clinical evaluation techniques using an explorer and disclosing media may be inadequate for assessments of marginal accuracy.