Development of a Prediction Model for Patients at Risk of Incidental Skin Cancer: A Multicentre Prospective Study

Abstract

In the absence of guidelines recommending routine total-body skin examination, patient concern remains the main factor in seeking consultation regarding suspicion of skin cancer. This study explores gaps in patients’ understanding of malignant skin lesions, through the factors associated with incidental skin cancer. Included patients had a confirmed histological diagnosis of basal cell carci­noma, squamous cell carcinoma or melanoma. Tumour characteristics, patient demographics and other risk factors related to the development of skin cancer were obtained from each participant. The main measure was incidental skin cancer detection, using both binary logistic regression and Chi-squared Automatic Interaction Detection (CHAID) algorithm. Of the total tumours, 26.6% were detected incidentally. The following variables: male sex, living alone, long-axis diameter, tumour location, symptoms and time of disease evolution were independent predictors of incidental skin cancer. According to the CHAID algorithm, the most significant risk factor for incidental skin cancer was the absence of symptoms at diagnosis.

Keywords

• early diagnosis
• incidental findings
• projections and predictions
• skin cancer