Purpose: The aim of the study was to determine the ability of digital sensors (CMOS and CCD sensors) and D and F-speed films to detect cavitated and non-cavitated enamel caries lesions at different exposure conditions compared to a gold standard.
Methods: 100 extracted human molars and premolars were selected and mounted in a block between two neighbouring teeth. Sensors or films were exposed with voltages of 60 or 70 kVp at varying times. Three observers assessed each approximal site independently. Lesion depth was rated according to an anatomical 5-point scale (0 = no lesion to 4 = lesion reaching inner half of dentin). Serial sections of resin-embedded teeth were prepared. Gold-standard scores were established by consensus based on histological sectioning. A carious lesion was present at scores of 1 and higher. Statistical evaluation (sensitivity, specificity and receiver-operating curves), was based on caries-free surfaces and those presenting enamel caries (n=116).
Results: The ROC-curves had Area Under the Curve values (Az) from 0.50 (F-speed, 70 kVp, 0.20s) to 0.58 (CCD 60 kVp, 0.08 s). The detection percentage of cavitated lesions was generally higher (0-52%, depending on technique and observer) than that of non-cavitated lesions (3-32%). The CMOS sensor showed Az values comparable to the CCD sensors but required higher exposure times. There was no significant difference between 60 and 70 kVp.
Clinical significance: Digital radiography using CMOS sensors is at least comparable to conventional radiography for the detection of enamel lesions and neither system was sufficiently effective in the detection of approximal enamel lesions.
Original languageEnglish
Pages (from-to)74-78
Number of pages4
JournalAmerican Journal of Dentistry
Issue number2
Publication statusPublished - 1 Apr 2011

    Research areas

  • digital radiography, receiver operating characterics, dental caries

ID: 2038429