Abstract: Objective: To study a non-invasive mathematical model for predicting the effect of dual-wavelength pulsed dye laser on port-wine stains (PWS). Methods:The data of 91 patients with port-wine stains treated with dual-wavelength pulsed dye laser were retrospectively analyzed and  were divided into modeling group (59 patients) and verification group (32 patients). In the modeling group, using single-factor and multi-factor methods to analyze independent related factors which affect the effect of PWS. The mathematical model PWS index was established through logistic regression. Finally, areas under the receiver operating characteristics (AUROC), sensitivity, and specificity were used to evaluate the predictive ability of the model. Results: The dermoscopic feature of PWS and skin thickness under B-ultrasound were independent factors on the effect of dye laser treatment on PWS on the face and neck by single-factor and multi-factor methods (P<0.05). Mathematical model PWS index=-5.899 +［-1.846 (dermoscopy type 1)/2.017 (dermoscopy type 2)］ + 6.808 * B-ultrasonic thickness. The AUROC (95% CI) for the effect of dye laser treatment on PWS in the modeling group and verification group were 0.952 (95% CI: 0.895, 1.01) and 0.902 (95% CI: 0.785, 0.102). Conclusion: The dermoscopic feature of PWS and skin thickness under B-ultrasound are independent factors that predict the efficacy of dual-wavelength pulsed dye laser trentment on PWS on the face and neck. The mathematical model PWS index established based on the two parameters exhibits significant prediction of efficacy in both the modeling group and the verification group.

Key words: port-wine stains, pulsed dye laser, none-invasive model