Abstract
Background/Aim: The optimal treatment sequence after first-line nivolumab plus ipilimumab (Nivo+Ipi) for advanced renal cell carcinoma remains unclear, particularly for patients with early progressive disease (PD). We evaluated the effectiveness of second-line vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI) therapy after Nivo+Ipi according to early PD status.
Patients and Methods: This retrospective single-center study included patients with advanced renal cell carcinoma treated at Kyushu Cancer Center between September 2018 and January 2026. Among 54 patients who received systemic therapy, 40 were treated with first-line Nivo+Ipi. We analyzed 21 patients who subsequently received second-line VEGFR-TKI therapy (cabozantinib or axitinib). Early PD was defined as PD within 12 weeks of Nivo+Ipi initiation. Best response and progression-free survival (PFS) after second-line therapy were compared between early PD and non-early PD groups. A sensitivity analysis was performed among patients treated with cabozantinib.
Results: Of 21 patients receiving second-line therapy, 12 were classified as early PD and 9 as non-early PD. Best response differed between groups: early PD (partial response, 0/12; stable disease, 5/12; PD, 7/12) versus non-early PD (partial response, 4/9; stable disease, 3/9; PD, 2/9) (p=0.037). Median PFS after second-line VEGFR-TKI was 2.42 months [95% confidence interval (CI)=1.18-5.75] in the early PD group and 9.90 months (95%CI=3.45-10.59) in the non-early PD group (log-rank p=0.003). In the cabozantinib-restricted sensitivity analysis, median PFS was 2.53 months (95%CI=1.61-2.70) versus 18.67 months (95%CI=3.39-not estimable) (log-rank p<0.001).
Conclusion: Early PD after first-line Nivo+Ipi was associated with no objective response and markedly shorter PFS with second-line VEGFR-TKI. These findings suggest that early PD may identify a high-risk subgroup with limited benefit from standard second-line VEGFR-TKI treatment.
Introduction
Immune checkpoint inhibitor (ICI)-based combinations have reshaped first-line treatment of advanced renal cell carcinoma (RCC). Dual checkpoint blockade with nivolumab plus ipilimumab (Nivo+Ipi) provides durable responses and a long-term survival benefit in intermediate- and poor-risk disease; however, a subset of patients experiences early progression (1, 2). In parallel, ICI plus vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI) regimens, such as lenvatinib plus pembrolizumab and nivolumab plus cabozantinib, have demonstrated high response rates and improved outcomes compared with sunitinib in treatment-naïve advanced RCC (3, 4).
Despite these advances, evidence to guide treatment after failure of first-line ICI-based therapy remains limited, particularly for patients who do not derive early benefit from Nivo+Ipi (5-7). Post-ICI sequencing is challenging because much of the available evidence was generated in earlier treatment eras and is now applied across heterogeneous prior regimens (7). After progression on ICI-based therapy, VEGFR-TKIs remain a central therapeutic component, supported by pivotal trials such as METEOR (cabozantinib vs. everolimus) (8) and by accumulating real-world experience in the post-ICI setting (7, 9, 10). In contemporary practice, cabozantinib is among the most commonly selected VEGFR-TKI options after ICI-based combinations. Japanese real-world data from the Japanese Urological Oncology Group (JUOG) evaluating second-line cabozantinib after prior immunooncology combination therapy demonstrated clinically meaningful activity and suggested that discontinuation of first-line therapy because of progressive disease (PD) and the presence of liver metastasis were associated with poorer outcomes (9). Furthermore, in the randomized CONTACT-03 trial, continuing or adding another ICI after progression on prior ICI did not improve outcomes compared with cabozantinib alone, underscoring the importance of practical risk stratification rather than simply layering immune agents (11). Importantly, the post-ICI setting is not uniform: patients who fail dual checkpoint blockade early may represent a clinically fragile subgroup in whom the window to deliver effective subsequent therapy is short, making early identification and stratification highly relevant in daily practice.
From both biological and clinical perspectives, dual checkpoint blockade with Nivo+Ipi is expected to induce meaningful immune activation, and when effective, immunotherapy can yield durable disease control (1, 2). However, early progression suggests primary resistance and often coincides with rapid clinical deterioration (12, 13). In this context, a key sequencing-related concern in routine care is whether patients with early failure on Nivo+Ipi constitute a distinct high-risk subgroup with limited benefit from subsequent VEGFR-TKI therapy, and how such patients should be managed after early progression.
Accordingly, we aimed to clarify the clinical implications of early PD after first-line dual checkpoint blockade with Nivo+Ipi by evaluating the effectiveness of second-line VEGFR-TKI therapy in a real-world cohort. Specifically, we compared objective response and progression-free survival after second-line VEGFR-TKI between patients with early PD and those with non-early progression. We also performed a sensitivity analysis restricted to cabozantinib to mitigate potential confounding from imbalanced selection of second-line agents. By linking early failure of dual checkpoint blockade to both radiographic response patterns and time-to-event outcomes under subsequent VEGFR-TKI therapy, this study seeks to provide clinically actionable information to inform early treatment planning in routine practice.
Patients and Methods
Study design and patients. This retrospective, single-center study was conducted at the Kyushu Cancer Center. Between September 2018 and January 2026, 54 patients with advanced RCC, including those with metastatic disease, received systemic therapy. Of these, 40 were treated with first-line Nivo+Ipi and comprised the overall cohort (Figure 1). The primary analysis included patients who received second-line VEGFR-TKI therapy (cabozantinib or axitinib) after Nivo+Ipi. For patients who did not proceed to second-line therapy, the main reason was determined through chart review.
Patient flow diagram. Nivo, Nivolumab; Ipi, ipilimumab; Cabo, cabozantinib; Axi, axitinib; VEGFR-TKI, vascular endothelial growth factor receptor tyrosine kinase inhibitor.
Definitions and outcomes. Early PD was defined as PD documented within 12 weeks of initiating Nivo+Ipi, based on radiologic assessment according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Non-early PD was defined as PD occurring beyond 12 weeks. In routine practice, imaging assessments were generally performed at approximately 12-week intervals. Progression-free survival (PFS) for second-line therapy was measured from initiation of second-line VEGFR-TKI to radiologic progression or death from any cause, whichever occurred first. Patients without an event were censored at the time of last follow-up.
Data collection and treatment assessment. Data were extracted from the patients’ electronic medical records, including age, sex, Eastern Cooperative Oncology Group performance status (ECOG PS), prior nephrectomy, histologic subtype (clear cell vs. non-clear cell), International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk group at initiation of Nivo+Ipi, metastatic sites (including the liver and bone), and the reason for discontinuation of Nivo+Ipi (PD vs. adverse events). Details of second-line treatment included the agent used, starting dose (recorded as total daily dose, mg/day), best overall response (partial response, stable disease, or PD) according to RECIST version 1.1, and grade ≥3 adverse events during second-line therapy.
Statistical analysis and ethics. Continuous variables are presented as median (interquartile range) and were compared using the Mann-Whitney U test. Categorical variables are presented as number (%) and were compared using Fisher’s exact test. PFS was estimated using the Kaplan-Meier method and compared using the log-rank test. Given the small sample size, p-values were interpreted descriptively. A sensitivity analysis restricted to patients treated with cabozantinib was performed to mitigate potential confounding due to imbalanced selection of second-line agents between groups. Statistical analyses were conducted using EZR (version 1.68). The study protocol was approved by the institutional review board of the Kyushu Cancer Center (approval no. 2014-99), and the requirement for informed consent was waived using an opt-out approach because of the retrospective design.
Results
Patient flow diagram of second-line therapy after Nivo+Ipi. In total, 40 patients with advanced RCC were treated with Nivo+Ipi (Figure 1). At the data cut-off, 21 patients (52.5%) had received second-line VEGFR-TKI therapy: cabozantinib in 16 patients and axitinib in five. Among these 21 patients, 12 were classified as having early PD after Nivo+Ipi and nine as having non-early PD. In the early PD group, 11 patients received cabozantinib and one received axitinib. In the non-early PD group, five received cabozantinib and four received axitinib. Nineteen patients (47.5%) did not receive second-line therapy. The reasons were ongoing first-line therapy at the data cut-off (n=12), ineligibility for further systemic therapy (n=6), and patient refusal despite preserved performance status (n=1). Among the six patients deemed ineligible for second-line therapy, four experienced clinical deterioration following early PD, and two were unable to initiate subsequent therapy for medical reasons.
Patient characteristics. The baseline characteristics of the 21 patients who received second-line VEGFR-TKI after Nivo+Ipi are summarized in Table I. Their median age was 71 years (interquartile range=66-75), and 12 patients (57.1%) were male. Most had an ECOG PS of 0-1 (n=16, 76.2%), and 11 patients (52.4%) had undergone prior nephrectomy. Clear cell histology was observed in 14 patients (66.7%), whereas seven (33.3%) had non-clear cell histology. At initiation of Nivo+Ipi, 16 patients (76.2%) were classified as IMDC intermediate risk and five (23.8%) as poor risk. Liver and bone metastases were present in five (23.8%) and 13 (61.9%) patients, respectively. Nivo+Ipi was discontinued because of PD in 19 patients (90.5%) and because of adverse events in two (9.5%).
Baseline characteristics of patients receiving second-line vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI) after nivolumab plus ipilimumab.
When stratified by early PD status, the early PD group more frequently had an ECOG PS of ≥2 compared with the non-early PD group (41.7% vs. 0.0%; p=0.045). Similarly, IMDC poor risk at initiation of Nivo+Ipi was observed only in the early PD group (41.7% vs. 0.0%; p=0.045). Other baseline variables–including age, sex, prior nephrectomy, histology, and the presence of liver or bone metastases–were broadly comparable between groups (Table I).
Second-line treatment and response. Second-line treatment exposure and outcomes are summarized in Table II. Among the 21 patients who received second-line VEGFR-TKI therapy, cabozantinib was administered to 16 (76.2%) and axitinib to five (23.8%). The distribution of second-line regimens differed numerically between groups (cabozantinib: 91.7% in the early PD group vs. 55.6% in the non-early PD group; p=0.119). Regarding starting dose (total daily dose), cabozantinib was initiated at 40 mg/day in nine patients (42.9%) and 20 mg/day in seven (33.3%), whereas axitinib was initiated at 10 mg/day in four patients (19.0%) and 6 mg/day in one (4.8%).
Second-line vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI) regimen, starting dose, best response, and adverse events.
Best response to second-line therapy according to RECIST differed between groups (p=0.037). In the early PD group, no patient achieved a partial response (0/12); stable disease was observed in five patients (41.7%) and PD in seven (58.3%). By contrast, in the non-early PD group, four patients achieved a partial response (44.4%), three had stable disease (33.3%), and two had PD (22.2%). Grade ≥3 adverse events during second-line VEGFR-TKI therapy occurred in four patients overall (19.0%), all within the early PD group (33.3% vs. 0.0% in the non-early PD group; p=0.104).
PFS after second-line VEGFR-TKI. With 20 progression or death events among 21 patients, PFS after initiation of second-line VEGFR-TKI was significantly shorter in the early PD group than in the non-early PD group (Figure 2). Median PFS was 2.42 months [95% confidence interval (CI)=1.18-5.75] in the early PD group and 9.90 months (95%CI=3.45-10.59) in the non-early PD group (log-rank p=0.003).
Progression-free survival after initiation of second-line vascular endothelial growth factor receptor tyrosine kinase inhibitor.
Sensitivity analysis restricted to cabozantinib. To address potential confounding arising from the imbalanced selection of second-line agents between groups, we performed a sensitivity analysis restricted to patients who received second-line cabozantinib (n=16; early PD, n=11; non-early PD, n=5) (Figure 3). The median PFS was 2.53 months (95%CI=1.61-2.70) in the early PD group and 18.67 months (95%CI=3.39-not estimable) in the non-early PD group (log-rank p<0.001).
Progression-free survival after initiation of second-line cabozantinib (sensitivity analysis). PD: Progressive disease.
Discussion
In this real-world cohort of patients treated with first-line dual checkpoint blockade with Nivo+Ipi, early PD was strongly associated with poor outcomes following subsequent VEGFR-TKI therapy. Patients classified as having early PD achieved no objective responses and experienced markedly shorter PFS after second-line VEGFR-TKI compared with those with non-early progression. Importantly, this association persisted in a sensitivity analysis restricted to cabozantinib, supporting the robustness of our findings despite potential confounding from imbalanced selection of second-line agents.
These results are clinically relevant in the current era, in which multiple ICI-based combinations are available in the first-line setting, yet direct comparative evidence remains limited (5-7). Notably, no head-to-head prospective trials have compared an immunotherapy doublet (e.g., Nivo+Ipi) with an ICI-VEGFR-TKI regimen; available comparisons are largely indirect or derived from real-world datasets (14). Although a randomized comparison tailored to specific clinical questions–including optimal sequencing–would be ideal, such trials are challenging to conduct in a rapidly evolving therapeutic landscape. In this context, pragmatic real-world stratifiers that inform post-ICI decision-making retain clear clinical value.
Post-ICI sequencing remains challenging because the evidence base supporting VEGFR-TKIs was established largely in earlier treatment eras and is now extrapolated to patients who progress after ICI-based combinations (7). Cabozantinib has longstanding evidence of activity in previously treated RCC, as demonstrated in METEOR (8). More recently, Japanese real-world data from JUOG showed clinically meaningful outcomes with second-line cabozantinib after first-line immuno-oncology combinations and identified discontinuation because of PD and the presence of liver metastasis as adverse factors for PFS (9). Additional real-world evidence supports the feasibility and activity of cabozantinib following ICI failure (10). In our cohort, the early PD group more frequently had an ECOG PS of ≥2 and IMDC poor risk at initiation of Nivo+Ipi, consistent with a clinically fragile and biologically aggressive subset that may be less likely to benefit from standard post-ICI VEGFR-TKI strategies.
A key practical insight from our data is that early failure is not only prognostic but may also limit the delivery of subsequent therapy. Although 19 of 40 patients did not proceed to second-line treatment, most were still receiving first-line therapy at the data cut-off (n=12) or declined treatment despite preserved performance status (n=1). The critical operational finding was that six patients were unable to initiate second-line therapy because of clinical deterioration following early progression (n=4) or other medical reasons (n=2). These findings underscore that in patients with suspected early PD, timely reassessment and proactive planning for subsequent therapy are essential to avoid missing the window during which systemic treatment remains feasible.
Biologically, early PD likely reflects primary resistance to immune checkpoint blockade arising through multiple, non-mutually exclusive mechanisms, including insufficient antigen recognition and priming, impaired T-cell trafficking or infiltration, and dominant immunosuppressive programs within the tumor microenvironment (12, 13). These frameworks help reconcile the apparent paradox that dual checkpoint blockade can confer durable benefit when effective, yet a subset of patients progresses early: durable immune control characterizes responders, whereas primary resistance may manifest before meaningful immune-mediated tumor control is established (1, 2, 12, 13).
Another practical question is whether VEGFR-TKI monotherapy administered immediately after Nivo+Ipi can approximate the benefit observed with upfront ICI-VEGFR-TKI combinations. In the first-line setting, ICI-VEGFR-TKI regimens have demonstrated high response rates and survival advantages over sunitinib (3, 4, 15). However, the clinical context at the time of VEGFR-TKI initiation differs substantially between upfront combination therapy and salvage treatment after early PD. Tumor burden, performance status, and organ function may deteriorate rapidly, potentially limiting the benefit achievable even with an active agent. The persistence of the early PD disadvantage in our cabozantinib-only sensitivity analysis further suggests that the observed difference is not solely attributable to the specific second-line agent selected.
Finally, our findings align with emerging evidence that empirically adding another ICI after progression on prior ICI is unlikely to overcome resistance in unselected patients. In CONTACT-03, atezolizumab plus cabozantinib did not improve outcomes compared with cabozantinib alone and was associated with increased toxicity, indicating that simply “stacking” immunotherapies is unlikely to be an effective strategy after ICI failure (11). This underscores the need for practical risk stratification and rational sequencing in the post-ICI setting (7). In this context, our data suggest that early PD after dual checkpoint blockade with Nivo+Ipi identifies a high-risk subgroup with limited benefit from subsequent VEGFR-TKI therapy.
This study has limitations inherent to its retrospective, single-center design and small sample size. Treatment selection was not randomized, and baseline imbalances between groups were present; however, the direction and magnitude of the association were consistent in a cabozantinib-only sensitivity analysis. Imaging was generally performed at approximately 12-week intervals, and response assessments were based on routine clinical practice rather than centralized review. In addition, our analysis focused on patients who received second-line therapy; thus, the overall clinical burden of early failure may be underestimated, as some patients became ineligible for further systemic treatment.
Conclusion
In this real-world cohort, early PD after first-line dual checkpoint blockade with Nivo+Ipi was associated with markedly poorer outcomes following second-line VEGFR-TKI therapy. Patients with early PD achieved no objective responses and experienced substantially shorter PFS than those with non-early progression; this difference persisted in a sensitivity analysis restricted to cabozantinib. These findings suggest that early PD may function as a practical high-risk marker to inform timely treatment planning after Nivo+Ipi.
Acknowledgements
The Authors would like to thank Angela Morben, DVM, ELS, from Edanz (https://jp.edanz.com/ac) for professional English language editing of this manuscript.
Footnotes
Authors’ Contributions
Conceptualization, N.F., J.T., A.T., Y.S., M.N., and T.N.; methodology, N.F., J.T., A.T., Y.S., M.N., and T.N.; software, N.F. and T.N.; validation, N.F., J.T., A.T., Y.S., M.N., and T.N.; formal analysis, N.F. and T.N.; investigation, N.F., J.T., A.T., Y.S., M.N., and T.N.; resources, N.F., M.N., and T.N.; data curation, N.F., J.T., A.T., Y.S., M.N., and T.N.; writing—original draft preparation, N.F., M.N., and T.N.; writing—review and editing, N.F., M.N., and T.N.; visualization, N.F., J.T., A.T., Y.S., M.N., and T.N.; supervision, T.N.; project administration, T.N.; funding acquisition, not applicable. All Authors have read and agreed to the published version of the manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest in relation to this study.
Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.
Funding
This research received no external funding.
- Received March 4, 2026.
- Revision received March 20, 2026.
- Accepted March 27, 2026.
- Copyright © 2026 The Author(s). Published by the International Institute of Anticancer Research.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.