Bladder cancer is one of the most common malignant diseases involving the urinary system. Accurate prediction of the disease course and outcome is crucial for choosing an appropriate treatment strategy in these patients. Currently, there are several prognostic models for predicting non-muscle invasive bladder cancer outcomes. The scoring systems developed by the European Organization for Research and Treatment of Cancer (EORTC) and the Spanish Urological Club for Oncological Treatment (CUETO) are the most widely used prognostic models for bladder cancer. Despite the undeniable merits of these scales, they need to be supplemented. Since the prognostic score has a direct impact on the treatment strategy, intensity and costs of postoperative follow-up, and outcome, its accuracy should be higher than it is now. Identifying the additional parameters that would increase the robustness of these models is one of the major challenges for researchers.The molecular and genetic characteristics of the tumor, that can be estimated after the first surgery, are probably the best candidates for this role. The main limitation of these prognostic models lies in the fact that they assess only morphological properties of the tumor, while the most important molecular characteristics are neglected. These scoring systems do not evaluate clinical factors, concomitant diseases, and iatrogenic complications occurring during the treatment of relapses. The assessment of molecular mechanisms and clinical characteristics underlying the development of non-muscle-invasive bladder cancer as well as identification of key molecular markers, that could complement the currently existing risk assessment models, are the most important goals for researchers dealing with bladder cancer. It will significantly improve predictive capabilities of these models, ensuring the choice of an optimal treatment strategy.
Editorial Comment from Dr Todenhöfer and Dr Schwentner to Recurrence and progression in patients with non-muscle invasive bladder cancer: Prognostic models including multicolor fluorescencein situhybridization molecular grading
