Aesthetic outcomes are a primary concern for adult orthodontic patients. Clinicians often rely on soft tissue simulation software, such as Dolphin Imaging (D-Tech) and WebCeph (D-Tech), to manage patient expectations. This study aimed to compare the accuracy of these two tools in predicting soft tissue profile changes in non-growing skeletal Class II patients treated with maxillary first premolar extractions.

A retrospective comparative study was conducted on patients treated by the same practitioner with bilateral maxillary first premolar extractions and maximum anchorage. From a private-practice database of 2535 patients treated between January 2015 and January 2025, 34 met the inclusion criteria after eligibility screening. Pre- and post-treatment lateral cephalometric radiographs and profile photographs were used to generate simulations. Predicted cephalometric measurements from D-Tech and W-Tech were compared with each other and with the actual post-treatment results. Statistical analyses included paired t -tests, one-sample t -tests, Bland-Altman plots, and linear regression.

Thirty-four patients (10 males, 24 females; 35.7 ± 3.5 years) were analysed. Both tools showed a significant discrepancy in the nasolabial angle versus actual results ( P < 0.001). D-Tech differed significantly in the subnasale to Holdaway-line distance (mean difference 0.81 mm [95% CI 0.03 to 1.60]; P = 0.042), while W-Tech differed in upper and lower lip position relative to the True Vertical Line (LL–TVL P = 0.004; UL–TVL P = 0.006). For the remaining parameters, both tools showed good agreement; D-Tech showed a proportional bias for lip–TVL position (greater underestimation at larger values), whereas W-Tech overestimated lower lip thickness.

Both Dolphin Imaging and WebCeph provide clinically applicable predictions for most soft tissue parameters. Several statistically significant lip-position discrepancies were within the range generally considered clinically imperceptible (= 1.5-2.0 mm), whereas the nasolabial angle showed a discrepancy that is both statistically and clinically relevant. Predictions should therefore be interpreted with caution, particularly for parameters prone to systematic bias.