Explainable Ensemble Machine Learning Framework for Lung Cancer Status Classification: A Retrospective Cross-sectional Study

Abstract

Non-muscle-invasive bladder cancer (NMIBC) is characterized by a high recurrence rate and a substantial long-term surveillance burden. However, current detection and surveillance strategies rely largely on invasive cystoscopy and urine cytology, which are limited by patient discomfort, suboptimal adherence, and insufficient sensitivity, particularly for low-grade disease. This mini review examines the role of liquid biopsy in the detection, risk stratification, and surveillance of NMIBC. It summarizes current evidence on circulating tumor DNA, circulating tumor cells, exosomes and extracellular vesicles, urinary tumor DNA, and DNA methylation and protein biomarkers in blood or urine. For detection, urine-based molecular assays complement cystoscopy and cytology in the noninvasive assessment of patients with suspected bladder cancer. For risk stratification, molecular detection of residual disease and longitudinal biomarker changes help identify patients at increased risk of recurrence or progression. During surveillance, urine-based assays support earlier recurrence detection and, in selected settings, help reduce the frequency of cystoscopy. Overall, liquid biopsy is a promising minimally invasive complement to conventional NMIBC management although technical standardization and prospective clinical validation are required before routine implementation.