Second-line TKI therapy following first-line immune checkpoint inhibitor combinations in metastatic RCC: A systematic review and meta-analysis of safety outcomes.

Abstract

e16561 Background: Sequential tyrosine kinase inhibitor (TKI) therapy following immune checkpoint inhibitor (ICI)-based first-line treatment is standard practice in metastatic renal cell carcinoma (mRCC), yet the cumulative toxicity burden remains poorly characterized. We conducted the first meta-analysis to quantify toxicity with second-line TKIs following first-line ICI-based combinations. Methods: A systematic search of PubMed, EMBASE, the Cochrane Library, and MEDLINE through January 2026 identified studies reporting toxicity outcomes (Grade ≥3 adverse events per CTCAE, serious adverse events, treatment discontinuation) of second-line TKIs in mRCC patients previously treated with ICI-based combinations. A single-arm meta-analysis used the Freeman-Tukey double arcsine transformation with the DerSimonian-Laird random-effects model. Subgroup analyses by TKI type and prior ICI regimen, sensitivity analysis restricted to prospective studies, and publication bias assessment using funnel plots and Egger's regression test were performed. Results: Ten studies (1,093 patients; 2 phase 3 RCTs, 2 phase 2 trials, 6 retrospective cohorts) were included. The pooled incidence of Grade ≥3 adverse events was 45.4% (95% CI: 35.2%-55.8%; I² = 87.3%; 7 studies, n = 832), serious adverse events 30.0% (95% CI: 23.1%-37.3%; I² = 62.5%; 4 studies, n = 523), and treatment discontinuation 12.9% (95% CI: 2.5%-29.8%; I² = 93.9%; 5 studies, n = 486). Subgroup analysis by TKI type showed Grade ≥3 AE rates of 48% for cabozantinib (3 studies, n = 368), 60% for tivozanib (1 study, n = 171), 63% for axitinib (1 study, n = 40), and 30% for various TKIs (5 studies, n = 443), with substantial between-subgroup heterogeneity (I² = 95.2%, p0.0001). By prior ICI regimen, Grade ≥3 AE rates were approximately 40% following IO-IO, 45% following IO-TKI, and 48% following mixed regimens. Sensitivity analysis restricted to prospective studies confirmed robustness (pooled 45.1%; 95% CI: 40.0%-50.2%; I² = 44.6%, p = 0.165). Egger's test showed no significant publication bias for Grade ≥3 AEs (p = 0.893) or discontinuation (p = 0.652), with marginal asymmetry for SAEs (p = 0.087). Conclusions: Approximately 45% of mRCC patients receiving second-line TKI after first-line ICI-based combinations experience Grade ≥3 adverse events. Toxicity varies by TKI agent (cabozantinib ~48%, tivozanib ~60%, axitinib ~63%), while prior ICI regimen has modest impact. Despite substantial toxicity, low discontinuation rates (~13%) suggest most patients tolerate therapy with appropriate management. These findings support individualized TKI selection and proactive toxicity monitoring post-ICI. Future prospective studies with standardized reporting and head-to-head comparisons are needed to optimize sequencing and develop predictive toxicity biomarkers.