*4.3. Statistical Methods*

Correlations between the mRNA of *CXCL9*, *PD1*, *PD-L1*, *KRT5*, *KRT20* and *KI67* and clinicopathological data were calculated using Spearman's bivariate correlation. Optimized cut-off values for dichotomizing each marker with respect to survival were defined using Youden's index on the receiver operating characteristic (ROC). Detailed information about the calculated optimal cut-off values, the associated area under the ROC curve and internal validation using bootstrapping are provided in Tables S1 and S2. Following standard practice in retrospective survival analysis, the common time point zero for all patients was the date of the first TURB. The associations of mRNA with recurrence-free survival (RFS), overall survival (OS) and cancer-specific survival (CSS) were determined by univariate (Kaplan–Meier analysis and Cox's regression hazard models) and multivariate (Cox's regression hazard models, adjusted for age and the molecular parameters PD1, PD-L1 and CXCL9) analyses. A *p*-value < 0.05 was considered statistically significant. Statistical analyses were performed with the SPSS 21.0 software package (SPSS Inc., Chicago, IL, USA) and R V3.2.1 (The R foundation for statistical computing, Vienna, Austria).

#### **5. Conclusions**

Altogether, we confirmed that high *PD-L1* mRNA is associated with increased DSS and RFS. Furthermore, we demonstrated for the first time that *CXCL9* mRNA is associated with a longer OS, DSS and RFS. Associations with RFS were also identified or further pinpointed to special groups, including the younger age group (*CXCL9*, *PD1*), the high *KRT5* or high *KRT20* group (*CXCL9*, *PD-L1*), the high *KI67* group (*CXCL9*, *PD-L1*) or the no instillation group (*CXCL9*, *PD-L1*).

An increased mRNA for *PD1*, *PD-L1* and *CXCL9* being associated with a better prognosis may mirror the host–tumor interaction. In this way, we suggest that the increased mRNA levels of all three genes may reflect the immune response of the host.

Our finding of associations between these immune markers and prognosis may aid in future therapeutic options and decisions.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/10/2794/s1. Table S1: Optimized Ct cutoff values and internal validation and Table S2: Area under the ROC curve and internal validation.

**Author Contributions:** D.S., H.T., S.W., R.M.W. and B.K. designed the study. D.S., J.K., S.W., V.W., R.S., A.H. and B.W. acquired the clinical samples and patient information. A.H. and M.E. performed the pathological review of all cases. J.K. and A.N. performed qRT-PCR experiments. H.T., S.W., D.S. and J.K. performed statistical analyses, and H.T., S.W., J.K., D.S., M.E. prepared the tables and figures. H.T., S.W., D.S., B.W., M.E. and A.H. wrote the main manuscript. All authors reviewed the manuscript and approved the final version of the manuscript.

**Funding:** This study was funded by the ELAN Fund (ELAN 18¨C08-18¨C1-Sikic) and was supported by the Interdisciplinary Center for Clinical Research (IZKF) at the University Hospital of the Friedrich-Alexander University Erlangen-Nuremberg. We thank the Rudolf und Irmgard Kleinknecht-Stiftung for supporting H.T., and the Johannes und Frieda Marohn-Stiftung and the Wilhelm Sander-Stiftung for supporting S.W. and H.T.

**Acknowledgments:** The present work was performed in (partial) fulfillment of the requirements for obtaining the degree "Dr. med." (M.D.) of the Friedrich-Alexander-Universität Erlangen-Nürnberg, Medizinische Fakultät for Jennifer Kubon. The authors thank Johannes Breyer (University of Regensburg) and Philipp Erben (Heidelberg University) for helpful discussion. We thank American Journal Experts for editing the manuscript. The authors also acknowledge support from Deutsche Forschungsgemeinschaft and Friedrich-Alexander-Universität Erlangen-Nürnberg within the funding program Open Access Publishing.

**Conflicts of Interest:** The authors declare that there are no financial and/or nonfinancial conflicts of interest.
