The SWIFT Model for Lichen Sclerosus Among Premenarchal Girls.

Abstract

Delay in diagnosis of childhood lichen sclerosus (LS) can be ameliorated with an efficient evaluation tool. We sought to create a useful prognostic tool for rapid and accurate risk stratification for LS in premenarchal girls. We conducted a retrospective chart review at a single institution of premenarchal girls presenting with vulvovaginal complaints at a specialty pediatric and adolescent gynecology clinic at a major academic center. Sixty-nine patients seen between July 2019 and September 2020 were used as a pilot study to create a model for LS based on 18 signs and symptoms. Accuracy of the pilot model was confirmed in a larger data set (additional 105 patients, seen between January 2017 and December 2020), and model parameters were refined through cluster-based analytics. Pilot study yielded 5 predictors for LS: soreness (S), whitening (W), urinary incontinence (I), fissures (F), and thickening of the clitoral hood (T)-SWIFT. The final refined model is given as log odds (LS) = -7 + 3·S + 17·W + 3·I + 3·F + 18·T. This model yielded a >97% accuracy in predicting LS among 174 unique patients (LS prevalence = 18%). The SWIFT model accurately predicts clinical diagnosis of LS in premenarchal girls. Replication in other patient populations is highly encouraged. Awareness of LS is paramount, and an efficient, accurate evaluation tool will prove invaluable in assuring timely diagnosis and treatment for premenarchal patients.