Resumen
Here we describe a serum metabolomics signature of bladder cancer coupled with a robust ensemble machine learning algorithm able to effectively discriminate patients with and without bladder cancer. This signature, if further confirmed and validated on a larger cohort, could represent a reliable screening test for this disease. Moreover, the signature was able to discriminate high- and low-grade cancers. The results represent an important clinical contribution since the prognosis of these conditions strongly depends on early detection and grading.