Stereotactic body radiotherapy plus sunitinib vs. sunitinib alone in treatment-naive oligometastatic renal cell carcinoma: A phase 2 clinical trial

Summary

Comparative data on stereotactic body radiotherapy (SBRT) in oligometastatic renal cell carcinoma (RCC) is lacking. This single-center, phase 2 non-randomized controlled trial enrolls patients with treatment-naive oligometastatic RCC. Patients choose to receive sunitinib alone (control arm, n = 24) or plus SBRT (SBRT arm, n = 24). The primary endpoint, objective response rate (83.3% vs. 29.2%; p < 0.001), is higher in the SBRT arm. With a median follow-up of 40.6 months, the 1-year local control rate is 91.5%. Progression-free survival (PFS) is significantly longer in the SBRT arm (17.3 vs. 6.3 months; p = 0.036). Post hoc multivariable analysis shows that SBRT significantly prolongs PFS (hazard ratio [HR], 0.42; 95% confidence interval [CI], 0.21–0.84; p = 0.015). Grade 3 toxicities are similar (54.2% vs. 50.0%, p = 0.773). The SBRT arm shows a trend toward better quality of life. These suggest potential benefit of SBRT in oligometastatic RCC, requiring further validation by phase 3 trials. The trial is registered at Chictr.org.cn (ChiCTR1800017136).

Graphical abstract

Highlights

• •The addition of SBRT to first-line sunitinib improves response in oligometastatic RCC
• •SBRT yields durable local control and prolongs progression-free survival
• •Toxicity remains comparable to sunitinib alone

Liu et al. report results from a controlled phase 2 trial demonstrating that stereotactic body radiotherapy added to first-line sunitinib improves response and progression-free survival in treatment-naive oligometastatic renal cell carcinoma. The addition of SBRT results in comparable toxicity to sunitinib alone and favorable quality-of-life trends.

Introduction

Immune checkpoint inhibitors have revolutionized the treatment of metastatic renal cell carcinoma (RCC) and reshaped the first-line therapeutic landscape.¹ However, with grade ≥3 toxicity exceeding 60% for most combinations,^2,3,4 their universal applicability is questioned. Oligometastatic RCC patients, characterized by a relatively indolent disease course, have shown promising responses to local therapy. This highlights the need to explore alternatives that achieve comparable efficacy at the early metastatic stage, while reserving immunotherapy combinations for cases of rapid progression to optimize therapeutic potential.

Increasing evidence indicates that stereotactic body radiotherapy (SBRT) provides durable local control with a favorable toxicity profile in RCC.⁵ Although SBRT has rapidly become the recommended treatment for oligometastatic RCC in current guidelines,⁶ the body of prospective evidence is limited to only a few trials.^7,8,9 These studies, with a single-arm design and largely inconsistent lines of systemic therapy at the time of SBRT, limit the robustness of the evidence. This significant gap underscores the need for controlled trials to clarify the clinical benefits and impact on patients’ quality of life, particularly in treatment-naive oligometastatic RCC. Therefore, this phase 2 trial aimed to investigate whether SBRT could improve the therapeutic efficacy of first-line sunitinib for treatment-naive oligometastatic RCC.

Results

Patient and treatment characteristics

Between August 21, 2018, and May 25, 2022, 48 patients were enrolled to either the SBRT (n = 24) or control (n = 24) arm (Figure 1). The baseline characteristics were generally balanced (Table 1). The median number of metastases was 2 per patient. The presence of bone metastasis was more frequent in the SBRT arm (66.7% vs. 29.2%; p = 0.009).

Figure 1.

CONSORT diagram

RT, radiotherapy

Table 1.

Baseline characteristics

	SBRT arm (n = 24, %)	Control arm (n = 24, %)	p value
Age, median (interquartile range)	54 (40–61)	59 (49–66)	0.079
Gender			0.763
Male	15 (62.5)	16 (66.7)	–
Female	9 (37.5)	8 (33.3)	–
IMDC risk factors			0.588
Favorable	4 (16.7)	7 (29.2)	–
Intermediate	14 (58.3)	12 (50.0)	–
Poor	6 (25.0)	5 (20.8)	–
Pathology			1.000
Clear cell	19 (79.2)	20 (83.3)	–
Non-clear cell	5 (20.8)	4 (16.7)	–
No. of metastases			0.131
1	11 (45.8)	6 (25.0)	–
2–5	13 (54.2)	18 (75.0)	–
Metastatic sites
Lung	2 (8.3)	10 (41.7)	0.008
Bone	16 (66.7)	7 (29.2)	0.009
Lymph node	5 (20.8)	9 (37.5)	0.204
Soft tissue	4 (16.7)	4 (16.7)	1.000

IMDC, International Metastatic Renal Cell Carcinoma Database Consortium

Fourteen patients (58.3%) in the control arm and 15 patients (62.5%) in the SBRT arm underwent sunitinib dose reduction due to adverse events (AEs) or at their own request. Three patients (12.5%) in the control arm and 3 (12.5%) in the SBRT arm discontinued sunitinib due to AEs. Two patients (8.3%) in the SBRT arm voluntarily stopped treatment after achieving complete remission. At last follow-up, 3 patients (12.5%) in the SBRT arm and 4 patients (16.7%) in the control arm were still on sunitinib treatment. SBRT was performed for 36 metastatic lesions, predominantly bone metastases (n = 22, 61.1%).

Clinical efficacy

The median follow-up was 40.6 months (interquartile range, 25.2–54.0). The sum of the largest diameter (SLD) change was depicted as the largest change from baseline (Figure 2A). Overall, 20 patients in the SBRT arm and 7 patients in the control arm achieved objective response. The objective response rate (ORR) was 83.3% in the SBRT arm vs. 29.2% in the control arm (p < 0.001). Complete response rates were 25.0% and 0%, respectively. The disease control rate (DCR) was 91.7% in the SBRT arm and 66.7% in the control arm (p = 0.072). After adjusting bone metastasis as the confounding factor (Figure S1A), SBRT was evaluated as the only independent predictor for ORR (adjusted odds ratio [aOR], 24.3, 95% confidence interval [CI], 3.89–152.78; p = 0.001; Table S2). In propensity-adjusted sensitivity analysis, the aOR for ORR was 97.0 (p = 0.001) favoring SBRT. Subgroup analysis for ORR was shown in Table S3.

Figure 2.

Clinical efficacy and health-related quality of life

(A) Waterfall plot of the best change in the SLD of lesions for all patients (n = 48). ∗One patient had a non-measurable lesion by RECIST v.1.1. Dashed lines indicate the ≥30% decrease and ≥20% increase in SLD.

(B) Freedom from local failure at the irradiated sites.

(C) Kaplan-Meier plot depicting progression-free survival in the SBRT arm vs. control arm.

(D) Kaplan-Meier plot depicting overall survival in the SBRT arm vs. control arm.

(E) Progression-free survival 2 among patients receiving upfront SBRT during OM (excluding 6 patients who received second-round SBRT for oligoprogressive lesions in SBRT arm, n = 18) vs. those at the time of OP (patients who received SBRT for oligoprogressive lesions in control arm, n = 12).

(F) Quality-of-life assessment measured by EuroQol 5 dimensions (EQ-5D) index and EQ-VAS per treatment arm at baseline, 3 months, and 1 year. Bars represent median, and error bars represent the interquartile range. Only statistically significant changes from baseline were marked.

SLD, sum of longest diameter; OP, oligoprogression; OM, oligometastasis; NE, not evaluable.

Thirty-nine (81.3%) patients developed disease progression. Five in-field recurrences occurred in the bones with 4 (80.0%) in the spine, 2 in lymph nodes, and 1 in kidney. The 1-year and post hoc 3-year local control rates after SBRT were 91.5% and 87.7%, respectively (Figure 2B). In the control arm, failure was most frequently due to progression at the original lesions (55.0%), while in the SBRT arm, new metastases (63.2%) were the predominant pattern of failure. The median progression-free survival (PFS) was 17.3 months (95% CI, 3.8–30.7) in the SBRT arm and 6.3 months (95% CI, 3.2–9.5) in the control arm (log rank p = 0.036; Figure 2C). After adjustment, the adjusted hazard ratio (HR) (aHR) for PFS was 0.42 (95% CI, 0.21–0.84; p = 0.015; Table S4). In propensity-adjusted sensitivity analysis, the aHR for PFS was 0.42 (95% CI, 0.20–0.90; p = 0.026) favoring SBRT. Subgroup analysis for PFS was shown in Table S5. At the time of last follow-up, 19 patients had died. The 3-year OS rates were 66.4% in the SBRT arm and 58.3% in the control arm (log rank p = 0.709; Figure 2D). At 48 months, the restricted mean survival time difference between the SBRT and control arms was 3.34 months (95% CI, −4.80 to 11.48; p = 0.422). Subgroup analysis for OS was shown in Table S6.

Oligoprogression occurred in 14 patients (58.3%) in the SBRT arm and 17 patients (70.8%) in the control arm. Following oligoprogression, 6 patients in the SBRT arm (20.8%) and 12 patients in the control arm (50.0%) received SBRT directed at the progressive sites. The 3-year PFS2 was 61.1% in the group irradiated during the oligometastatic phase and 16.7% in the group delivered at the time of oligoprogression (log rank p = 0.026; Figure 2E).

Safety

All patients experienced at least one AE, and no grade 5 event was reported (Table 2). One grade 4 AE occurred in the SBRT arm. A patient developed chronic kidney disease following sunitinib and thoracic spine SBRT. Five years later, the patient experienced acute exacerbation of chronic kidney disease triggered by COVID-19 infection, leading to dialysis. The rates of grade 3 AEs were similar between the two arms (54.2% vs. 50.0% in the SBRT vs. control arm, p = 0.773). Leukopenia (66.7% vs. 75.0% in the SBRT vs. control arm) and elevated creatinine levels (62.5% in both arms) were the most common AEs. Grade 3 non-hematological toxicity related to SBRT included radiation enteritis (n = 1) and radiation pneumonitis (n = 1).

Table 2.

Treatment-related adverse events

Type of Adverse Event	SBRT arm (n = 24)		Control arm (n = 24)
	Any grade No. (%)	Grade 3-4 No. (%)	Any grade No. (%)	Grade 3-4 No. (%)
Total	24 (100)	13 (54.2)	24 (100)	12 (50.0)
Fatigue	8 (33.3)	3 (12.5)	7 (29.2)	3 (12.5)
Hypertension	3 (12.5)	0 (0)	8 (33.3)	1 (4.2)
Increased alanine aminotransferase	8 (33.3)	1 (4.2)	5 (20.8)	1 (4.2)
Increased aspartate aminotransferase	10 (41.7)	1 (4.2)	6 (25.0)	0 (0)
Creatinine increased	15 (62.5)	1 (4.2)	15 (62.5)	0 (0)
Palmar-plantar erythrodysesthesia syndrome	10 (41.7)	0 (0)	11 (45.8)	0 (0)
Radiation enteritis	1 (4.2)	1 (4.2)	N/A	N/A
Hemorrhoidal hemorrhage	1 (4.2)	0 (0)	0 (0)	0 (0)
Leukopenia	16 (66.7)	3 (12.5)	18 (75.0)	2 (8.3)
Thrombocytopenia	8 (33.3)	3 (12.5)	12 (50.0)	5 (20.8)
Anemia	8 (33.3)	1 (4.2)	12 (50.0)	0 (0)
Mucositis oral	11 (45.8)	0 (0)	9 (37.5)	0 (0)
Diarrhea	6 (25.0)	0 (0)	2 (8.3)	0 (0)
Vomiting	5 (20.8)	0 (0)	1 (4.2)	0 (0)
Proteinuria	9 (37.5)	1 (4.2)	10 (41.7)	0 (0)
Hypothyroidism	10 (41.7)	0 (0)	6 (25.0)	0 (0)
Wound infection	1 (4.2)	1 (4.2)	0 (0)	0 (0)
Epistaxis	1 (4.2)	0 (0)	0 (0)	0 (0)
Adrenal insufficiency	1 (4.2)	0 (0)	0 (0)	0 (0)
Rash	2 (8.3)	0 (0)	0 (0)	0 (0)
Heart failure	1 (4.2)	0 (0)	1 (4.2)	0 (0)
Radiation pneumonitis	1 (4.2)	1 (4.2)	N/A	N/A
Coma	0 (0)	0 (0)	1 (4.2)	1 (4.2)
Cough	0 (0)	0 (0)	1 (4.2)	0 (0)
Hematuria	0 (0)	0 (0)	2 (8.3)	0 (0)
Increased bilirubin	0 (0)	0 (0)	2 (8.3)	0 (0)
Colitis	1 (4.2)	0 (0)	0 (0)	0 (0)
Fall	1 (4.2)	1 (4.2)	0 (0)	0 (0)
Hyperuricemia	10 (41.7)	0 (0)	13 (54.2)	0 (0)
Chronic kidney disease	1 (4.2)	1 (4.2)	0 (0)	0 (0)

N/A, not applicable.

Health-related quality of life

All patients completed the baseline health-related quality of life (HRQoL) assessments. Forty-six patients (96%) completed the questionnaires at 3 months, and the number fell to 39 (81%) at 1 year. The baseline EuroQol (EQ)-index (0.937 vs. 1.000, p = 0.001), EuroQol Visual Analogue Scale (EQ-VAS) (83 vs. 90, p = 0.002), and Verbal Rating Scale (VRS) (1 vs. 0, p = 0.002) scores were inferior in the SBRT vs. the control arm. In addition, the proportion of patients reporting pain of any grade was nearly four times higher in the SBRT arm than in the control arm (63% vs. 17%, p = 0.003). The control arm experienced a decline in HRQoL at 3 months (3-month vs. baseline EQ-index 0.937 vs. 1.000, p = 0.001; EQ-VAS 85 vs. 90, p = 0.001; VRS 1 vs. 0, p = 0.001), returning to baseline levels at 1 year. The HRQoL parameters of the SBRT arm remained relatively stable at 3 months and showed significant improvements in all aspects at 1 year (1-year vs. baseline EQ-index 1.000 vs. 0.937, p = 0.004; EQ-VAS, 97 vs. 83, p = 0.002; VRS 0 vs. 1, p = 0.001) (Figure 2F). The baseline difference in HRQoL parameters between the two arms was no longer evident at 1 year. Change-from-baseline analysis showed similar trend (Data S1).

Post hoc biomarker analysis

A total of 29 tumor and 26 blood samples with successful DNA libraries were collected. Identity-by-descent analysis yielded 24 genetically matched pairs that met the quality criteria and were, therefore, included in the final analysis. The most frequently mutated genes were VHL (50%), GYPB (33%), RBM38 (25%), BAP1 (21%), NLGN4X (21%), and PARP4 (21%). Kyoto Encyclopedia of Genes and Genomes enrichment analysis indicated that the PI3K-Akt (71%), RAS (63%), and mTOR (54%) pathways were the most frequently altered cancer-related pathways (Figure 3A). These genes exhibited ethnic and geographic heterogeneity in mutation frequencies among different cohorts compared to prior studies (Figure S2).^{10,11,12,13,14} Patients with mutate-type BAP1 or PARP4 were associated with significantly lower ORR (33.3% vs. 80.0%, p = 0.036) and DCR (55.6% vs. 100.0%, p = 0.012). Additionally, patients with wild-type BAP1 and PARP4 showed better PFS than those with mutate-type BAP1 or PARP4 (HR, 0.20; 95% CI, 0.07–0.55; p = 0.002; Figures 3B and 3C). In the mutated subgroup, the median PFS was 11.5 months in the SBRT arm and 3.5 months in the control arm (log rank p = 0.062; Figure 3D), while in the wild subgroup, the median PFS was 29.6 months in the SBRT arm and 19.1 months in the control arm (log rank p = 0.996; Figure 3E). Immunohistochemistry demonstrated similar trend (Figure S3).

Figure 3.

Post hoc biomarker analysis based on WES

(A) Oncoplot depicting clinical features and somatic alterations of 24 patients.

(B) PFS according to BAP1 and PARP4 mutation status.

(C) OS according to BAP1 and PARP4 mutation status.

(D) PFS stratified by treatment arms for those with mutate-type BAP1 or PARP4 subgroup.

(E) PFS stratified by treatment arms for those with wild-type BAP1 and PARP4 subgroup.

WES, whole-exome sequencing; PFS, progression-free survival; OS, overall survival.

Discussion

The current study shows that for treatment-naive oligometastatic RCC, the addition of SBRT to first-line sunitinib provided significant benefits of ORR and DCR. A significant improvement in PFS was also observed in favor of the combined modality therapy, with no OS advantage at present. Additionally, SBRT plus sunitinib combination had a comparable toxicity profile to sunitinib alone, while improving the HRQoL. Controlled comparative data evaluating SBRT plus first-line sunitinib vs. sunitinib monotherapy in oligometastatic RCC have not been previously reported.

While prior single-arm studies have consistently reported improved tumor response (ORRs of 66%–76%) and PFS (median 16–23 months) with SBRT in oligometastatic RCC,^7,8,15 these findings are limited to short-term endpoints and predominant inclusion of patients with lung metastases and International Metastatic RCC Database Consortium (IMDC) favorable or intermediate risk. In comparison, our trial enrolled a broader and less-selected population with poorer prognostic features. This included mixed histology; a higher proportion of bone, lymph node, and soft tissue metastases; as well as patients with IMDC poor risk, better reflecting real-world disease. The ORR in our control arm was consistent with prior findings,^16,17 supporting the representativeness of our cohort, while the slightly shorter PFS may reflect the inclusion of non-clear cell histology. We found that SBRT significantly improved ORR and PFS, like in other solid tumors.^18,19,20,21 This benefit should be confirmed by phase 3 trials in the era of immunotherapy. Current metastasis-directed therapy for oligometastatic RCC includes SBRT alone, sequential SBRT, and SBRT combined with systemic therapy. SBRT alone and sequential SBRT showed promising efficacy in IMDC favorable or intermediate group, with median PFS of 22.7 and 32.0 months and 1-year PFS of 64% and 82.6%, respectively.^7,9,22 SBRT combined with systemic therapy reported a median PFS of 15.6 months and 1-year PFS of 60%,⁸ similar to our findings (median PFS was 17.3 months and 1-year PFS was 66.7%). Definitive comparison between strategies remains limited by small cohorts, variability in systemic treatments, and heterogeneity across studies.

Notably, with the longest follow-up time, our study provides comparative prospective data on OS following SBRT in oligometastatic RCC. The absence of a significant OS benefit in the overall population underscores the importance of careful patient selection for local therapy even in oligometastasis. Prior studies have shown that IMDC poor-risk patients derive little or no benefit from cytoreductive nephrectomy or metastasectomy.^23,24 In subgroup analyses, IMDC risk group was the only variable with a significant interaction across ORR, PFS, and OS, suggesting a possible differential treatment effect. However, these post hoc findings may be underpowered and require further validation.

Current consensus recommends temporary interruption of tyrosine kinase inhibitors (TKIs) before and after SBRT, largely based on findings in liver cancer.²⁵ In RCC, however, quite a few studies have observed acceptable toxicity combining SBRT and TKIs.²⁶ Prospective single-arm studies in RCC have reported grade ≥3 AE rates of 4%–10% with SBRT alone,^7,9 0%–5% when TKIs were withheld,^27,28 and 0%–7% with immune checkpoint inhibitors.^8,29 The rate of SBRT-related grade ≥3 toxicity with concurrent sunitinib in our study was comparable to prior findings (8%), suggesting acceptable safety of this combination. One grade 4 AE in the SBRT arm was attributed to prolonged sunitinib exposure, highlighting the need for awareness of cumulative toxicity in long-term RCC survivors. Combining SBRT with systemic therapy followed by response-adapted treatment adjustment may balance efficacy and toxicity. Although HRQoL improvements generally trended toward benefit with SBRT in our study, the results need validation in more balanced cohorts.

The timing of metastasis-directed SBRT in oligometastatic RCC remains uncertain. Two strategies are commonly used: upfront SBRT during the oligometastatic phase and deferred SBRT at oligoprogression. Reported median PFS ranges from 16 to 23 months with upfront SBRT^7,8,9 and 9 to 11 months with deferred SBRT.^27,28 In our post hoc analysis of treatment-naive oligometastatic RCC, upfront SBRT significantly improved PFS2 compared to deferred treatment until oligoprogression. While further validation is needed, these findings support the concept that early treatment of all visible metastases may reduce systemic tumor burden, enhance antitumor immunity, and prevent reseeding from untreated lesions,³⁰ as observed in both RCC^31,32 and prostate cancer.¹⁹

In conclusion, our trial showed that the addition of SBRT to first-line sunitinib can safely improve ORR and PFS in patients with oligometastatic RCC, suggesting potential therapeutic benefit. Ongoing randomized phase 3 trials, including SOAR (NCT05863351) and STROKER (NCT06726421), will validate these findings.

Limitations of the study

The current study has some limitations. The major limitation was the non-randomized design due to recruitment challenges. While sensitivity analysis yielded estimates directionally aligned with the primary results, residual confounding cannot be fully excluded given the non-randomized design, limited sample size, baseline imbalance, and sparse event distribution. Moreover, the systemic therapy used was sunitinib, which is no longer a standard in the immunotherapy era. The magnitude of benefit from SBRT and optimal sequencing with systemic therapy in oligometastatic RCC—particularly in the context of contemporary immunotherapy-based systemic therapy—remains uncertain and awaits confirmation from larger trials.

Resource availability

Lead contact

Further information and resource requests should be directed to and will be fulfilled by the lead contact, Li-Ru He (helir@sysucc.org.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •De-identified individual patient-level data required to reanalyze are available from the lead contact upon request. The genomic variation data reported in this paper have been deposited in the Genome Variation Map (GVM) in National Genomics Data Center, Beijing Institute of Genomics, Chinese Academy of Sciences and China National Center for Bioinformation, under accession number GVM001326.
• •No custom code was generated in this study. Statistical analyses were performed using standard R packages as described below.
• •Any additional information required to reanalyze the data reported in this paper is available from the corresponding author upon reasonable request.

Acknowledgments

This study is supported by grants from the National Natural Science Foundation of China grant no. 82373210 to L.-R.H., grant no. 82472655 to P.D., and grant no. 82303922 to Y.-L.P. Presented as a poster at the 2023 American Society of Clinical Oncology Annual Meeting, Chicago, IL, June 2–6, 2023.

Author contributions

Conception and design, Y.L., F.-J.Z., P.D., and L.-R.H.; acquisition of data, Y.L., X.-Y.Z., W.-S.W., R.-Q.L., Q.-W.P., Y.-J.Z., Y.-L.P., P.D., and L.-R.H.; analysis and interpretation of data, Y.L., X.-Y.Z., W.-S.W., R.-Q.L., Q.-W.P., J.-B.L., Y.-L.P., F.-J.Z., P.D., and L.-R.H.; drafting of the manuscript, Y.L., X.-Y.Z., W.-S.W., R.-Q.L., Q.-W.P., Y.-J.Z., J.-B.L., Y.-L.P., T.-X.H., Z.-L.Z., H.H., S.-J.G., J.-M.G., F.-J.Z., P.D., and L.-R.H.; critical revision of the manuscript for important intellectual content, L.-R.H. and P.D.; statistical analysis, Y.L., X.-Y.Z., W.-S.W., R.-Q.L., Q.-W.P., J.-B.L., Y.-L.P., F.-J.Z., P.D., and L.-R.H.; obtaining funding, Y.-L.P., P.D., and L.-R.H.; administrative, technical, or material support, F.-J.Z.; supervision, F.-J.Z., P.D., and L.-R.H.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-PARP4	Abcam	Cat# ab24110; RRID: AB_447876
Mouse monoclonal anti-BAP1	Santa Cruz Biotechnology	Cat# sc-28383; RRID: AB_626723
Biological samples
RCC patient tumor tissue	This study	N/A
RCC patient blood sample	This study	N/A
Deposited data
Patient data	This study	Chictr.org.cn ChiCTR1800017136
Genomic variation data	This study	Genome Variation Map: GVM001326
Software and algorithms
R (version 4.1.0)	R foundation	https://www.r-project.org
SPSS (version 25.0)	IBM	https://www.ibm.com/spss

Experimental model and study participant details

Human subjects

The trial included 48 patients (24 in each arm), all of Han Chinese ethnicity and East Asian ancestry. Key inclusion criteria were as follows: age ≥18 years; histologically confirmed RCC; primary tumor resected with cytoreductive, radical, or partial nephrectomy; radiologically confirmed distant metastases; and either no prior systemic therapy or ongoing first-line sunitinib for less than 3 months without response evaluation. Eligible patients were required to have at least one measurable metastatic lesion per RECIST 1.1 or MD Anderson criteria for bone metastases,³³ and no more than five metastatic lesions measuring 1–6 cm in diameter, with at least one lesion ≥3 cm in cases with lung-only metastases. Key exclusion criteria included liver or brain metastases, >5 metastatic lesions, any lesion >6 cm, prior systemic therapy other than sunitinib, or radiologically confirmed progression during prior first-line sunitinib treatment. The trial was approved by the Ethics Committee of Sun Yat-sen University Cancer Center (approval number: B2018-066-03), in accordance with the Declaration of Helsinki. All patients provided written informed consent. Both male and female patients were included. Gender distribution is reported in Table 1. No pre-specified gender-based subgroup analyses were conducted. The full eligibility criteria are listed in trial protocol (Methods S1).

Method details

Study design

This is an open-label, non-randomized, controlled phase II trial, starting in July 2018. Since SBRT was not endorsed by guidelines for oligometastatic RCC at the time our study was initiated, non-randomized controlled design was adopted considering both ethical and practical barriers to randomization. Patients were enrolled into the control arm (sunitinib alone) or SBRT arm (SBRT plus sunitinib) based on their own selection.

Radiotherapy details

For SBRT, the gross tumor volume (GTV) was defined as all visible tumor by imaging. Clinical target volume was created for spinal metastases according to International Spine Radiosurgery Consortium Consensus Guidelines for Target Volume Definition in Spinal Stereotactic Radiosurgery (by Cox et al.). For patients receiving decompression and fixation surgery for spine, delineation followed Consensus Contouring Guidelines for Post-Operative Stereotactic Body Radiation Therapy (SBRT) for Metastatic Solid Tumor Malignancies to the Spine (by Redmond et al.). Margins depend on the site irradiated, with 3mm margins for bony lesions,^34,35 5–10 mm margins for lymph nodes and lung lesions. The prescribing dose for SBRT ranged from 18 to 50Gy in 1–5 fractions, depending on the size and location of lesion. Dose constraints for organs at risk (OARs) were in accordance with the recommendations of American Association of Physicists in Medicine Task Group 101, and dose constrains for OARs prioritized over target coverage. SBRT was conducted either daily or every other day. Cone-beam CT was mandatory before each fraction.

Systemic therapy and follow-up

In both arms, patients began sunitinib (50 mg daily) for 4 weeks, followed by 2 weeks off. For grade 3 toxicity, sunitinib was paused and restarted at 50 mg (2 weeks on/1 week off) after improvement, with dose reductions of 12.5 mg for recurrence until discontinuation. In the SBRT arm, SBRT was mandated for all metastatic lesions within 2 months of starting sunitinib, which was uninterrupted during SBRT. Follow-ups occurred every 3 months, including adverse events (AEs), lab tests (hematology, chemistry, urinalysis), and imaging (chest CT, abdominal/pelvic CT or MRI). Bone scans or PET-CT were performed as needed. Post-progression treatment was determined by the oncologist.

Outcomes

Tumor response was evaluated according to the MDA criteria for bone lesions and the RECIST 1.1 criteria for other lesions. The primary endpoint was objective response rate (ORR). Based on an estimated ORR of 40% in the control arm and 80% in the SBRT arm, an enrollment of 46 patients would provide a two-sided α of 0.05 and a power of 0.80. Assuming a dropout rate of 5%, 48 patients were needed. Progression-free survival (PFS) was a secondary endpoint, defined as the time from enrollment to first disease progression or any-cause death or to the last follow-up for censored patients. Other secondary endpoints included overall survival (OS), disease control rate (DCR) based on best overall response, 1-year local control rate in the irradiated site measured from SBRT completion, health-related quality of life (HRQoL) assessed with the EuroQol 5 dimensions 3 level (EQ-5D-3L) instrument, EuroQol Visual Analogue Scale (EQ-VAS), and Verbal Rating Scale (VRS). AEs during enrollment were graded according to the CTCAE 5.0.

Post-hoc molecular analysis

A total of 35 tumor samples collected from enrolled patients, and 6 samples were precluded from library preparation due to tissue degradation. Of the 29 blood samples collected, 3 samples were excluded due to contamination. DNA extraction was performed using the TIANamp Genomic DNA Kit (Tiangen) and qualified by Nanodrop spectrophotometer and Qubit 3.0 Flurometer. VAHTS Universal Plus DNA Library Prep Kit for Illumina V2 was used for Library preparation, followed by hybridization capture using the method of the SureSelect XT (Agilent SureSelect XT Reagent kit) target enrichment system. After quantification by Qubit 3.0, the library was sequenced on the MGI platform (DNBSEQ-T7). To ensure the matched tumor-normal sample pairs were derived from the same patient, we performed identity-by-descent (IBD) analysis using the generated variant data. Only tumor-normal pairs demonstrating IBD proportions >0.9 were considered genetically matched and thus qualified for subsequent somatic mutation analysis. Genes with somatic mutations were evaluated enrichment in the Kyoto Encyclopedia of Genes and Genomes (KEGG) using ClusterProfiler package (version 4.10.1) to explore their biological significance. Immunohistochemistry (IHC) was performed for gene mutations significantly associated with prognosis (anti-PARP4, ab24110 Abcam, 1:1600; anti-BAP1, sc-28383 Santa Cruz Biotechnology, 1:200). The results were evaluated by two board-certified uropathologists. For PARP4 expression, IHC intensity was scored as follows: negative staining, 0; weak staining, 1; moderate staining, 2; and strong staining, 3. Percentage scores were categorized into quartiles. The intensity and percentage scores assigned to each sample were multiplied by a final score of 0 points to 12 points.³⁶ Expression levels were categorized based on median values. While BAP1 expression was evaluated using a binary scoring system that classified tumors as either BAP1 positive or BAP1 negative based on tumor cell nuclear immunoreactivity.³⁷

Quantification and statistical analysis

Primary and secondary endpoint analyses

All analyses followed the intention-to-treat principle. Categorical variables were presented as frequency (%) and analyzed using the Chi-square test or Fisher’s exact test, as appropriate. Continuous variables were depicted as median interquartile range (IQR) and analyzed with the Mann-Whitney U test. PFS, OS, and local control rates were calculated using Kaplan-Meier methods and compared via log rank tests. Treatment effects were reported as hazard ratios (HRs) with 95% confidential interval (95% CI) from Cox models for PFS and as differences in Restricted Mean Survival Time for overall survival (OS), based on proportional hazards evaluation. EQ-5D-3L scores were converted to EQ-Index values using a Chinese value set.³⁸ HRQoL parameters were compared using the Wilcoxon rank-sum test and signed-rank test, with Bonferroni correction for within-group comparisons (adjusted α = 0.017). A post hoc analysis of 3-year local control rate and PFS2 (defined as the time from enrollment to progression on second-line treatment or death)³⁹ was performed to compare the survival between patients only receiving upfront SBRT to all oligometastatic lesions versus those receiving SBRT only after oligoprogression. Post-hoc estimates of adjusted odds ratios (aORs) and hazard ratios (aHRs) were conducted for ORR and PFS. Post-hoc subgroup and biomarker analyses were performed for survival outcomes, and change-from-baseline analysis was performed for HRQoL. Sensitivity analyses were conducted to account for selection bias. All analyses were performed using R (v4.1.0) and SPSS (v25.0). with two-sided p < 0.05 considered significant.

Post-hoc and exploratory analyses

To explore the durability of SBRT, post-hoc 3-year local control rate was calculated using the Kaplan–Meier method. The 3-year local control rate was defined as time from SBRT completion to local progression of the irradiated sites. Post-hoc subgroup analysis was performed for ORR, PFS and OS, and unadjusted HRs were calculated for each subgroup. Subgroups included age, histology, IMDC risk groups, number of lesions and bone metastasis. Change-from-baseline analysis was performed for EQ-5D index and EQ-5D VAS. Change-from-baseline (Δ) values were calculated as follow-up score minus baseline score at each time point and compared using Mann-Whitney U test.

To account for potential confounding, post-hoc univariate and multivariable logistic regression models were conducted for ORR. Adjusted odds ratios (aORs) were estimated adjusting for histology and bone metastasis (Data S1). Post-hoc univariate and multivariable Cox regression models were performed for PFS, and adjusted hazard rates (aHRs) were estimated incorporating IMDC risk groups, number of lesions, bone metastasis and baseline EQ-5D index.

Sensitivity analyses were performed including two directed acyclic graphs (DAGs) based analyses and propensity-adjusted sensitivity analysis. DAGs were constructed with online software (Dagitty.net; Figure S1). The construction was based on published data and empirical causal assumptions. For ORR, bone metastasis, non–clear cell histology, and multiple metastases have been associated with poorer response to sunitinib,^40,41,42 while bone metastasis and socioeconomic factors (Data S1) may influence patients’ likelihood of receiving radiotherapy.^43,44,45 For PFS, bone metastasis, IMDC risk group, histology, number of metastases, and baseline HRQoL (EQ-5D) have been reported to correlate with prognosis in RCC,^41,46,47,48 and similarly, bone metastasis and socioeconomic factors were retained due to their influence on treatment selection.^43,44,45 The propensity-adjusted sensitivity analysis was performed accounting for confounders of treatment selection. Propensity scores were generated using a logistic regression model including variables potentially influencing treatment selection (e.g., education level, socioeconomic status, EQ-5D index, and presence of bone metastasis). The propensity score was then incorporated into multivariable logistic and Cox regression models to obtain aORs (for ORR) and aHRs (for PFS).

Additional resources

The trial was registered at www.chictr.org.cn (ChiCTR1800017136).

Published: April 20, 2026