OBJECTIVES: To compare the implant size (width and length) planned with digital panoramic radiographs, cone beam computed tomography (CBCT)-generated panoramic views, or CBCT cross-sectional images, in four implant systems. MATERIAL AND METHODS: Seventy-one patients with a total of 103 implant sites in the upper premolar and/or lower molar regions were examined with digital panoramic radiography (D-PAN) and (CBCT). A metal ball 5 mm in diameter was placed in the edentulous area for the D-PAN. CBCT data sets were reformatted to a 10-mm thick CBCT panoramic view (CBCT-pan) and 1-mm cross-sections (CBCT-cross). Measurements were performed in the images using dedicated software. All images were displayed on a monitor and assessed by three observers who outlined a dental implant by placing four reference points in the site of the implant-to-be. Differences in width and length of the implant-to-be from the three modalities were analyzed. The implant size selected in the CBCT-cross images was then compared to that selected in the other two modalities (D-PAN and CBCT-pan) for each of the implant systems separately. RESULTS: The implant-to-be (average measurements among observers) was narrower when measured in CBCT-cross compared with both D-PAN and CBCT-Pan. For premolar sites, the width also differed significantly between D-PAN and CBCT-pan modalities. The implant-to-be was also significantly shorter when recorded in CBCT-cross than in D-PAN. In premolar sites, there were no significant differences in implant length among the three image modalities. It mattered very little for the change in implant step sizes whether CBCT-cross was compared to D-PAN or CBCT-pan images. CONCLUSION: Our results show that the selected implant size differs when planned on panoramic or cross-section CBCT images. In most cases, implant size measured in cross-section images was narrower and shorter than implant size measured in a panoramic image or CBCT-based panoramic view