Objectives: To identify the predictive tools which have emerged recently in the field of urothelial carcinomas.
Materials and methods: We performed a thorough MEDLINE literature review using a combination of the following keywords: urothelial carcinoma, transitional cell carcinoma, bladder, renal pelvis, ureter, predictive tools, predictive models and nomograms. We found 117 articles, but only the relevant reports were selected.
Results: The majority of available tools are prediction models, particularly nomograms. These models combine good performance accuracy with ease of use. They appear to be more accurate than risk grouping or tree modeling and are more suitable for clinicians than artificial intelligence. The most recent nomograms have been designed to be used in daily clinical practice and are even available as computer or smartphone applications. They focus on pathological outcomes or more frequently on survival statistics or recurrence risk after surgery. They provide an accurate prediction of disease evolution and may help clinicians to choose the most appropriate treatment option. However, these prediction tools still need to be validated and regularly utilized.
Conclusion: Predictive tools represent very helpful clinical decision-making aids but need to be validated in larger populations.