Detection of Cisplatin Resistance in Bladder Cancer Using Machine Learning

Bladder Cancer (BC) is the ninth most common cancer worldwide, with cisplatin-based chemotherapy being a standard treatment. However, the emergence of cisplatin resistance poses significant challenges, contributing to high recurrence rates. This study investigates the mechanisms of cisplatin resistance in BC through the application of Machine Learning (ML) models, including Association Rule Mining, K-mode Clustering, and Decision Tree Classifier. By analyzing large-scale genomic data, we uncover the molecular pathways driving resistance. Specifically, our findings show that hypermethylation of CpG islands, such as those in UCSC_CPG_ISLANDS and GENOME_BUILD_37, leads to reduced cell accessibility, contributing to cisplatin resistance. The ML models highlight strong associations between hypermethylation and cell inaccessibility, with 99% confidence. This work provides valuable insights into the role of methylation patterns in BC recurrence and offers potential for personalized therapeutic strategies aimed at overcoming cisplatin resistance and improving clinical outcomes.