A novel nomogram to predict psoriatic arthritis in patients with plaque psoriasis.

ABSTRACT

OBJECTIVE: To construct a predictive model for Psoriatic Arthritis (PsA) based on clinical and ultrasonic characteristics in patients with plaque psoriasis (PsP). PATIENTS AND METHODS: Demographic, clinical, and ultrasound data were collected from patients with PsP and PsA between May 2019 and December 2022. RESULTS: A total of 212 patients with PsP and 123 with PsA in the training cohort, whereas the validation cohort comprised 91 patients with PsP and 49 with PsA. The multivariate logistic regression identified nail psoriasis (odds ratio [OR] 1.88, 95% CI 1.07-3.29), synovitis (OR 18.23, 95% CI 4.04-82.33), enthesitis (OR 3.71, 95% CI 1.05-13.14), and bone erosion (OR 11.39, 95% CI 3.05-42.63) as effective predictors for PsA. The area under the curve was 0.750 (95% CI, 0.691-0.806) and 0.804 (95% CI, 0.723-0.886) for the training and validation cohorts, respectively. The Hosmer-Lemeshow goodness-of-fit test showed good consistency for both the training cohort (p  =  0.970) and the validation cohort (p  =  0.967). Calibration curves also indicated good calibration for both cohorts. The DCA revealed that the predictive model had good clinical utility. CONCLUSIONS: We have developed a quantitative, intuitive, and convenient predictive model based on nail psoriasis, synovitis, enthesitis, and bone erosion to assess the risk of PsA in patients with plaque psoriasis.