Occlusal cares diagnosis at non-cavitated, discoloured fissures remains difficult. Modern caries diagnostic devices have been poorly investigated under clinical conditions. The aim of this study was to compare the accuracy of the following methods in occlusal caries detection: (1) visual inspection using magnifying glasses (x2,5), (2) laser fluorescence measurement with KaVo DIAGNOdent device, (3) electric conductive measurement with ECM III, (4) electric conductive measurement with Cariometer CRM 800 (Gente 1999). A total of 381 sites on 192 occlusal surfaces from 40 patients were examined by two dentists using the four diagnostic methods mentioned above. If two or more methods indicated dentinal caries, a minimally invasive preparation was performed. For this reason 47 sites on 34 occlusal surfaces from 22 patients aged 5 to 30 years (mean 19 years) were selected. During stepwise preparation the presence of dentinal caries was determined visually as a gold standard for each of the examined locations within the fissure. ECM showed the best and the only significant predictive values concerning dentinal caries (sensitivity 0.94, specificity 0.62, significance 0.002; area under the ROC curve 0.75). CRM recordings as well as visual inspection and DIAGNOdent scores demonstrated decreasing predictive values (sensitivity/specificity/AUC 0.85/0.47/0.699, 0.69/0.62/0.661 and 0.79/0.38/0.589). Combinations of methods did not improve the results. The low specificity values obtained in this study are partially caused by the preselection of teeth with suspicious fissures, which correlates to the diagnostic needs in the clinical situation. In conclusion, among the four methods used, ECM demonstrated the best suitability for treatment decisions at non-cavitated, discoloured occlusal fissures. Further data should be collected to achieve higher significance. Schlagwörter: occlusal caries diagnostics in vivo, laser fluorescence