Actinic Keratosis (AK) is skin damage lesions on forearms, face, ears, bald scalp, back of hands, and lower legs. If untreated, AK could deteriorate and lead to keratinocyte carcinoma (KC). Patients with AK have a high risk of contracting this non-melanoma skin cancer. This study aims to develop a predictive model to identify high-risk AK-patients for KC occurrence.

The dermatology researchers studied 1169 patients with AK in the Netherlands. They added known risk factors for KC, such as genes, lifestyle, and phenotype (physical characteristics). The data analysis method using backward stepwise selection helped to predict the outcomes. The effect of each predictor got analyzed, and the impact of all other variables underwent corrections simultaneously.

Around 176 patients, or 15.1%, developed KC after a median follow-up of 1.8 years. The risk of KC increased where AK got localized in upper body extremities. The risk was lower in patients who consumed coffee or had AK localized in the head. The presence of 4 to 9 AKs, or 10 or more AKs, also considerably increased the risk. 

These predictors were efficient in calculating KC risk in AK patients.

The study’s internal validation of the model showed a moderate discriminative ability. There are unknown predictor variables for KC occurrence in this patient group.

Ref: https://onlinelibrary.wiley.com/doi/10.1111/bjd.19355