Prediction accuracy of Dolphin software for soft-tissue profile in Class I patients undergoing fixed orthodontic treatment

•Dolphin software can successfully predict the result of fixed orthodontic treatment in borderline extraction/nonextraction cases.•In extraction treatment, Dolphin simulated photographs had more-protruded lips than did the real photographs.•In nonextraction treatment, Dolphin simulated photographs showed a more inferiorly positioned lower lip than did the real photographs.•Differences between the simulated prediction photographs and the real photographs were small and clinically insignificant. Simulation and prediction can facilitate the decision between an extraction versus nonextraction treatment plan. This study aimed to assess the accuracy of Dolphin Imaging software in predicting profile changes in class I borderline cases. In this retrospective study, class I borderline patients (i.e., both extraction and nonextraction treatment plans were possible for them) aged 15–35 years were enrolled. All of the cephalometric analyses were done by Dolphin Imaging software, version 11.8 Premium. The initial cephalograms were superimposed on initial photographs. The final values for the 7 angular and linear landmarks of the upper and lower incisors were extracted from post-treatment cephalograms and inserted in the “Goals” tab of the Dolphin Imaging software. Post-treatment images of patients were simulated. Final post-treatment photographs were superimposed on the simulated pictures. The differences between the real and simulated pictures were calculated in relation to the reference lines. The P value was set at 0.05. A total of 70 patients (36 with tooth extraction, and 34 without tooth extraction) were included. In the extraction group, the horizontal position of both lips was predicted to be significantly more protruded than it actually was, whereas in the nonextraction group, the only difference between the simulated and actual results was that the vertical position of the lower lip, which was simulated as being more inferior than it actually was. Despite the statistically significant differences between the real and simulated pictures, the differences were small and clinically insignificant. Prediction via Dolphin Imaging software can be an appropriate guide in extraction–nonextraction borderline cases.