Comparison of photothermal radiometry and modulated luminescence, intraoral radiography, and cone beam computed tomography for detection of natural caries under restorations.

Abstract

The aim of this ex vivo study was to measure the sensitivity, specificity, and receiver operating characteristic (ROC) area under the curve (AUC) of a caries diagnostic system based on photothermal radiometry and modulated luminescence (PTR/LUM) and compare them with the values for digital intraoral radiography (IR) and cone beam computed tomography (CBCT) in detecting recurrent decay. Class 2 composite restorations were prepared on 70 proximal surfaces: 35 with caries and 35 without caries. The gingival floor of the restored surfaces was assessed for caries under the restorations using each of the 3 modalities. Statistical calculations and analysis were performed using the R statistical computing environment. The average scores for sensitivity among the 6 observers were 0.89 for PTR/LUM, 0.38 for IR, and 0.40 for CBCT. Sensitivity for PTR/LUM was significantly greater than sensitivity for IR and CBCT. Average scores for specificity were 0.83, 0.80, and 0.70 for PTR/LUM, IR, and CBCT, respectively. CBCT had significantly lower specificity. The AUC was 0.65 for IR and 0.59 for CBCT, which were significantly different. PTR/LUM had moderate intraobserver agreement. PTR/LUM, which involves non-ionizing radiation, can serve as a sensitive adjunct in early caries detection and monitoring.