The aim of this study was to compare the validity of the measurements of the laser fluorescence device, KaVo DIAGNOdent((R)), with the result of polarized light microscopy in the detection of occlusal fissure caries in extracted third molars. Ten impacted, surgically removed, and 25 extracted third molars with macroscopically intact occlusal surfaces were selected. The DIAGNOdent measurements of the occlusal test site were recorded by two observers at intervals of 2 days. The teeth were then sectioned at the specified test sites for histological examinations. Prepared specimens were evaluated under the polarized light microscopy and all images were scored with the caries classification of D1 (sound and fissure lesion in the half of the outer enamel), D2 (enamel decay) and D3 (dentin decay) level (gold standard). The kappa value for the inter-observer repeatability was calculated and the value 0.83 for the first measurements and 0.67 for the second measurement were obtained,respectively. Inter-observer correct classification percentages were 88.5 and 77.1 for the first and second measurements, respectively. The kappa value for intra-observer repeatability was 0.79 for the first observer and 0.75 for the second observer. Intra-observer correct classification percentage values were 85.7 and 82.8 for the first and second observer, respectively. Value of specificity for the detection of enamel caries at D1 level was 0.74 and sensitivity values at D2 and D3 levels were 0.66 and 1.00, respectively. The present study indicates that the DIAGNOdent((R)) provides not only almost perfect agreement but also sufficient repeatability at D1, D2, D3 levels and better specificity at D1 level as well as lower sensitivity at D2 level and excellent sensitivity at D3 level.