270 Development and validation of a simple model to predict the risk of keratinocyte carcinoma on screening total body skin examination

Abstract

Keratinocyte carcinoma (KC) is the most common cancer in the U.S., and diagnostic delay is associated with increased morbidity and treatment costs. Although data suggests screening total body skin examination (TBSE) detects incidental skin cancers, there are no guidelines for providers to triage patients to TBSE. Prior predictive models of KC risk are cumbersome to implement clinically. We sought to develop a simple model to predict the risk of KC on TBSE. We used data from a previously published prospective, multicenter international prospective study that examined rates of skin cancer detection via TBSE in patients who presented to dermatology clinic for unrelated focused chief complaints. Multivariable analysis identified age, prior history of KC, and skin phototype as significant risk factors. These risk factors were incorporated into a points system that exhibited a ROC curve with high discrimination (c-statistic 0.82). The model was then validated in an American cohort of patients; those with a positive score were 10-fold more likely to have KC on TBSE (p < 0.001). We propose the acronym CAP (cancer history, age, phototype) as a simple mnemonic for this model which may be used clinically to increase the yield of TBSE by dermatologists in diagnosing KC and guide primary care referrls for TBSE for KC detection.