The global incidence of bladder cancer is more than half a million diagnoses each year. Bladder cancer can be categorized into non-muscle-invasive bladder cancer (NMIBC), which accounts for ~75% of diagnoses, and muscle-invasive bladder cancer (MIBC). Up to 45% of patients with NMIBC develop disease progression to MIBC, which is associated with a poor outcome, highlighting a clinical need to identify these patients. Current risk stratification has a prognostic value, but relies solely on clinicopathological parameters that might not fully capture the complexity of disease progression. Molecular research has led to identification of multiple crucial players involved in NMIBC progression. Identified biomarkers of progression are related to cell cycle, MAPK pathways, apoptosis, tumour microenvironment, chromatin stability and DNA-damage response. However, none of these biomarkers has been prospectively validated. Reported gene signatures of progression do not improve NMIBC risk stratification. Molecular subtypes of NMIBC have improved our understanding of NMIBC progression, but these subtypes are currently unsuitable for clinical implementation owing to a lack of prospective validation, limited predictive value as a result of intratumour subtype heterogeneity, technical challenges, costs and turnaround time. Future steps include the development of consensus molecular NMIBC subtypes that might improve conventional clinicopathological risk stratification. Prospective implementation studies of biomarkers and the design of biomarker-guided clinical trials are required for the integration of molecular biomarkers into clinical practice.

膀胱癌的全球发病率每年超过 50 万例。膀胱癌可分为非肌层浸润性膀胱癌（NMIBC）和肌层浸润性膀胱癌（MIBC），前者约占诊断的 75%。高达 45% 的 NMIBC 患者病情进展为 MIBC，这与不良预后相关，这凸显了临床需要识别这些患者。目前的风险分层具有预后价值，但仅依赖于临床病理学参数，可能无法完全捕捉疾病进展的复杂性。分子研究已确定参与 NMIBC 进展的多个关键参与者。已确定的进展生物标志物与细胞周期、MAPK 途径、细胞凋亡、肿瘤微环境、染色质稳定性和 DNA 损伤反应有关。然而，这些生物标志物均未经过前瞻性验证。报告的进展基因特征并不能改善 NMIBC 风险分层。 NMIBC 的分子亚型提高了我们对 NMIBC 进展的理解，但由于缺乏前瞻性验证、肿瘤内亚型异质性导致的预测价值有限、技术挑战、成本和周转时间，这些亚型目前不适合临床实施。未来的步骤包括开发共识分子 NMIBC 亚型，这可能会改善传统的临床病理学风险分层。将分子生物标志物纳入临床实践需要生物标志物的前瞻性实施研究和生物标志物引导的临床试验的设计。