PURPOSE: Evaluation of the prognostic significance of four different scoring systems in a real-world cohort of patients with metastatic urothelial carcinoma (mUC) or renal cell carcinoma (mRCC) undergoing immunotherapy (IO). METHODS: For 120 patients with mUC (n = 67) and mRCC (n = 53) who received IO between July 2016 and December 2020 at the tertiary Urological University Medical Centre Mannheim, the following scores were recorded at pre-treatment baseline: modified Glasgow prognostic score (mGPS), systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR), neutrophil-to-eosinophil ratio (NER). Overall survival (time between the beginning of IO until the patients' death or last contact) was determined for every patient. RESULTS: Kaplan-Meier analyses revealed that high baseline mGPS, SII (> 979) and NLR (> 3) were associated with poor overall survival (OS) (p < 0.05). Cox proportional hazards regression analyses showed that baseline mGPS and NLR had a significant independent prognostic influence on OS (p < 0.05), of which mGPS had a greater significance (p < 0.001, mGPS Score 2 vs. Score 0: HR 4.1, 95% CI 1.9-8.8). Although a high baseline NER (63.9) was associated with poor OS, it did not reach statistical significance. Baseline NER was also not identified as a significant score in the regression analyses. CONCLUSION: mGPS, SII and NLR are scoring systems that are easy to record in routine clinical practice. As they provide good prediction of OS in patients with mUC and mRCC under IO, they may allow identification of patients at high-risk and monitor them more cautiously in addition to imaging.