Cancer treatment planning is a complex and individualized process due to the variability of patient-specific factors, tumor characteristics, and evolving medical standards. Predicting the next step in diagnosis or therapy remains a significant challenge, due to the high variability and limited availability of structured medical data. Clinical Decision Support Systems (CDSS) offer a promising solution, with Case-Based Reasoning (CBR) standing out for its ability to provide interpretable and transparent recommendations. Unlike black-box machine learning models, CBR leverages past cases to generate predictions by analogy, aligning with the way clinicians naturally reason. In this work, we propose a CBR-based CDSS for skin cancer treatment that integrates medical taxonomies and patient-specific clinical features to predict the next treatment step. By focusing on both technical performance and real-world application in a medical setting, this study provides insights for the deployment of CBR-based systems in medical practice.