BACKGROUND: A suspicious occlusal carious lesion (SOCL) can be defined as a lesion with no cavitation and no radiographic radiolucency but for which caries is suspected. The authors evaluated whether using a device changed the percentage of SOCLs that were opened surgically and, among those SOCLs that were opened, the proportion that had penetrated into dentin. METHODS: Eighty-two dentists participated. In phase 1 of the study, dentists identified approximately 20 SOCLs, obtained patient consent, and recorded information about the lesion, treatment or treatments, and depth, if opened. Dentists were then randomly assigned into 1 of 3 groups: no device, DIAGNOdent (KaVo), and Spectra (Air Techniques). In phase 2, dentists enrolled approximately 20 additional patients and recorded the same phase 1 information while using the assigned device to help make their treatment decisions. A mixed-model logistic regression was used to determine any differences after randomization in the proportion of lesions opened and, if opened, the proportion of lesions that penetrated into dentin. RESULTS: A total of 1,500 SOCLs were enrolled in each phase. No statistically significant difference was found in the change in proportion of lesions receiving invasive treatment from phase 1 to phase 2 across the 3 groups (P = .33) or in the change in proportion of percentage of opened lesions that extended into dentin (P = .31). CONCLUSION: Caries-detecting devices in the study did not change substantially dentists' decisions to intervene or the accuracy of the intervention decision in predicting lesion penetration into dentin. PRACTICAL IMPLICATIONS: The caries-detecting devices tested may not improve dentists' clinical decision making for SOCLs.