Objective. The aim of our study was to evaluate the validity of the digital imaging fiber-optic transillumination (DIFOTI) method in comparison with clinical visual examination (CV) and bitewing (BW) radiography on detecting caries lesions on proximal surfaces of teeth. Materials and Methods. Proximal tooth surfaces of premolars and molars (n=2,103) of 91 voluntary university students aged from 18 to 30 years were examined with CV, BW radiography, and the DIFOTI method. Results. DIFOTI detected more initial and manifested caries lesions compared with CV and BW. Of the analyzed tooth surfaces, 69.8% were classified as sound by DIFOTI, 80.3% by BW, and 91.6% by CV. Initial caries lesions were found in 21.2% of the surfaces by DIFOTI, in 14.1% by BW, and in 6.2% by CV, whereas the proportions for manifested dental caries lesions were 9.0%, 5.6%, and 2.2%, respectively. The interexaminer agreement regarding the DIFOTI findings between an experienced clinician and a fifth-year dental student was high: κ=0.67 for initial and κ=0.91 for manifested caries lesions. Conclusions. The noninvasive DIFOTI method seems to offer a potential tool for everyday clinical practice. In clinical use, DIFOTI finds well even initial caries lesions on proximal surfaces, thus providing an instrument for detecting lesions potential for arresting as well as for monitoring the outcome after preventive measures.