BackgroundReliable caries detection is a cornerstone in the modern caries treatment schema. This study aimed to evaluate adopting traditional and new caries detection methods by third-year dental students.MethodsFifty-seven students were given lectures on caries detection, after which they evaluated 27 extracted carious teeth using traditional clinical assessment (CE), Nyvad’s, and ICDAS methods. On three teeth they also performed DIAGNOdent pen® (LF) scanning. Histological scores of the sectioned teeth (ICDAS, LF) and activity estimations of the lesions by the supervisors were used as golden standards (Nyvad, CE). For the ICDAS method , sensitivity and specificity were calculated using dentine caries (D3) as a cut-off point. Mean ICC and kappa values were calculated to evaluate interexaminer agreement for all lesions and methods. Spearman’s correlation coefficient evaluated LF scanning.ResultsICDAS method presented good sensitivity (0.78) and specificity (0.87). The inter-examiner agreement for different methods was fair or good (CE ICC = 0.69, κ = 0.53; Nyvad’s method ICC = 0.68, κ = 0.48, ICDAS ICC = 0.66, κ = 0.47). Variation in LF values was the greatest with lesions extending to middle third of dentin. In that case, the Spearman’s correlation coefficient was also the weakest.ConclusionsTo follow the guidelines by the European Core Curriculum on Cariology, the third year dental students are introduced to methods for detecting lesion depth and assessing lesion activity as well as using new caries detection methods. Their performance in estimating lesion depth is good, and fair to good in estimating lesion activity even after basic training only.