Purpose <p>Immune checkpoint plus tyrosine kinase inhibition (IO + TKI) has emerged as the first-line therapy in metastatic renal cell carcinoma (RCC), but no biomarker can predict its efficacy. Thymidine kinase 1 (TK1) is closely associated with immune evasion in tumors.</p> Methods <p>Metastatic RCC patients treated by IO + TKI were enrolled from two cohorts (ZS-MRCC, <i>n</i> = 45; Javelin-101, <i>n</i> = 726). High-risk localized RCC were also enrolled (ZS-HRRCC, <i>n</i> = 40). TK1 was assessed by RNA-sequencing in all cohorts, and the immune contexture was assessed by flow cytometry and immunohistochemistry.</p> Results <p>Higher TK1 expression was found in patients resistant to IO + TKI therapy (<i>p</i> = 0.025). High-TK1 group showed poor progression-free survival (PFS) in both the ZS-MRCC cohort (<i>P</i> = 0.008) and the Javelin-101 cohort (<i>P</i> = 0.036). By multivariate Cox regression, high-TK1 was determined as an independent factor for poor PFS (hazard ratio (HR) = 3.855, <i>P</i> = 0.002). High-TK1 expression was associated with decreased granzyme B⁺ CD8⁺ T cells (ρ=-0.22, <i>P</i> = 0.18), increased PD1⁺ CD4⁺ T cells (ρ = 0.33, <i>P</i> = 0.04), increased PDL1⁺ macrophages (ρ = 0.45, <i>P</i> &lt; 0.001), and increased regulatory T cells (ρ = 0.35, <i>P</i> = 0.03). A novel random forest (RF) risk score was built by machine learning based on TK1 and immunologic parameters. Combined IO + TKI therapy surpassed sunitinib monotherapy in the low RF risk score group (HR = 0.158, <i>P</i> &lt; 0.001), but was inferior to sunitinib in the high RF risk score group (HR, 2.195, <i>P</i> &lt; 0.001).</p> Conclusion <p>High-TK1 expression could be a potential indicator for therapeutic resistance, poor PFS and immune evasion in metastatic RCC under IO + TKI therapy. The novel RF risk score may help stratify patients for IO + TKI therapy.</p>