In vitro validation of near-infrared transillumination at 780 nm for the detection of caries on proximal surfaces

Lederer, Alexander; Kunzelmann, Karl-Heinz; Heck, Katrin; Hickel, Reinhard und Litzenburger, Friederike (2019): In vitro validation of near-infrared transillumination at 780 nm for the detection of caries on proximal surfaces. In: Clinical Oral Investigations, Bd. 23, Nr. 11: S. 3933-3940

Abstract

Objectives The aim of the present in vitro study was to determine the ability of Diagnocam to detect caries at an early stage and to compare it with digital radiography. Materials and methods One hundred twenty proximal surfaces composed equally of sound and decayed surfaces were evaluated by assessing images captured with Diagnocam (Kavo, Biberach, Germany) and digital radiography (DR). All images were assessed twice by two calibrated dentists with a minimum interval of 2 weeks between examinations. The results were compared with mu CT scans. Results Inter-rater reliability showed nearly perfect agreement;a high intra-rater reliability was calculated. Spearman's rank correlation coefficients showed a strong correlation of Diagnocam and mu CT (0.82). DR and mu CT achieved a slightly lower correlation (0.73). The surfaces were categorized into sound surfaces, enamel lesions, and dentin lesions to determine intraclass correlation coefficients (ICCs), sensitivity and specificity. ICCs between mu CT and DR ranged from 0.20 to 0.63. ICCs between Diagnocam and DR were higher in all categories and ranged from 0.56 to 0.83. Sensitivity, specificity, and accuracy of Diagnocam achieved mostly higher values than DR. In detecting enamel lesions, sensitivity was 0.36 for DR and 0.59 for the DC. The areas under the receiver operating characteristic (ROC) curves of Diagnocam were larger than those of DR in all categories. Conclusions Diagnocam is more capable of detecting initial proximal lesions than DR and also has a higher sensitivity for dentin lesions. However, caries progression in dentin cannot be reliably determined with Diagnocam.

• Dokumententyp: Zeitschriftenartikel
• Fakultät: Medizin
• Themengebiete: 600 Technik, Medizin, angewandte Wissenschaften > 610 Medizin und Gesundheit
• ISSN: 1432-6981
• Sprache: Englisch