Metastasis-directed radiotherapy without systemic therapy for oligometastatic clear-cell renal-cell carcinoma

Primary efficacy analysis of a single-arm, single-centre, phase 2 trial

Tang C, Sherry AD, Seo A, et al.
Lancet Oncology 2025; 26: 1289–99
 

The treatment of metastatic RCC: a historical perspective 1930s 1992 2001 2005–06 2018 2021–25 Early observations Spontaneous regression after nephrectomy (1–2% of cases) Lymphocyte infiltration Late recurrence (decades) Freed J Urol 1977 Fairlamb Cancer 1981 Miyao Urology 2011 Cytokine era HD IL-2 FDA approved 5–7% durable CR IFN-α: modest activity Median OS <1 year Placebo response 6% Elhilali BJU Int 2000 CN established SWOG 8949: OS 11.1 vs 8.1 mo EORTC 30947: OS 17 vs 7 mo Combined: 31% ↓ death Flanigan NEJM 2001 Mickisch Lancet 2001 TKI revolution Sunitinib: PFS 11 vs 5 mo vs IFN CN rates drop 31% → 15% Systemic-first begins Motzer NEJM 2007 IO era + CARMENA Nivo/ipi, pembro/axi OS 34–53 months CARMENA: CN not needed with sunitinib irAEs, cost, IV visits Méjean NEJM 2018 Motzer NEJM 2018 SBRT for RCC metastases is established Hannan et al. (Eur Urol Oncol 2022): n=23, 1-yr STFS 91% Tang feasibility cohort (Lancet Oncol 2021): n=30 Zhang et al. (IJROBP 2019): SBRT to defer systemic Tx Multiple retrospective series with favorable local control What was missing Prospective data large enough to estimate systemic therapy-free survival with a serial SBRT strategy in oligometastatic ccRCC off systemic therapy Local therapy era Systemic-first era Oligomet SBRT era RCC biology has always favored local approaches in selected patients This trial tests whether serial SBRT can meaningfully defer systemic therapy in the IO era
 The rise and fall of cytoreductive nephrectomy How local therapy became standard — then was challenged by better systemic options  The biological observations   Spontaneous regression 48 cases (Freed, J Urol 1977) 67 cases (Fairlamb, Cancer 1981) 6% in placebo arm (Elhilali 2000) Meta-analysis: 1.95% (Ghatalia 2016)  Immune infiltration Dense CD8+ T-cell infiltrate Th1/Tc1 polarization in TILs Clonal T-cell expansion at tumor Nakano Cancer Res 2001; Angevin IJC 1997  Indolent natural history 6.4% recur >10 yr post-nephrectomy Distant mets up to 26.6 years Case report: 25 yr latency Miyao Urol 2011; Kim Urol 2011; Xiao 2025 These observations — combined with RCC's resistance to chemotherapy — drove investigation of immune and local approaches   2001: Two landmark trials establish CN as standard of care  SWOG 8949 n=241; CN + IFN-α vs IFN-α alone Median OS: 11.1 vs 8.1 months (p=0.05) 1-year OS: 49.7% vs 36.8% Greatest benefit: good PS, lung-only mets Flanigan et al., NEJM 2001  EORTC 30947 n=85; CN + IFN-α vs IFN-α alone Median OS: 17 vs 7 months Time to progression: 5 vs 3 months Combined (J Urol 2004): 31% ↓ risk of death Mickisch et al., Lancet 2001   2018: CARMENA challenges the paradigm  CARMENA (Méjean, NEJM 2018) n=450; sunitinib alone non-inferior to CN + sunitinib Higher met burden than SWOG/EORTC: 43% intermediate, 57% poor risk  Result CN rates: ~31% → ~15% Role of upfront CN questioned  Note: CARMENA enrolled a higher-burden population. The favorable-biology oligometastatic subset — good PS, limited mets, long disease-free intervals — was not well-represented in this trial. The role of CN in this subset with modern IO remains an open question (PROBE trial, SWOG S1931). 

 

 

The IO era: effective systemic therapy — but do all oligometastatic patients need it? CheckMate 214 Nivo + ipi OS 47 mo (int/poor) 8-yr f/u: durable CRs Tannir, Ann Oncol 2024 KEYNOTE-426 Pembro + axitinib PFS 15.4 vs 11.1 mo OS benefit all risk groups Rini, NEJM 2019 CLEAR Pembro + lenvatinib PFS 23.9 vs 9.2 mo ORR 71% Choueiri, Lancet Oncol 2023 Treatment burden irAEs (colitis, hepatitis, pneumonitis, endocrine) Frequent IV visits High financial cost For oligometastatic patients: active surveillance showed systemic therapy can be deferred Rini et al. surveillance study (Lancet Oncol 2016) n=48; metastatic RCC observed without systemic Tx Median time to systemic therapy: 14.9 mo (95% CI 10.6–25.0) Median PFS: 9.4 mo; OS: 44.5 mo Could SBRT extend this interval? If surveillance alone defers Tx for 14.9 mo, can serial SBRT to oligometastatic sites extend this further while maintaining disease control? Prior evidence: SBRT for RCC metastases Tang feasibility (Lancet Oncol 2021) n=30; single-arm phase 2 1-yr STFS: 82% Demonstrated feasibility of SBRT-only strategy Hannan (Eur Urol Oncol 2022) n=23; systemic therapy-naïve 1-yr STFS: 91.3% 0% grade 3+ toxicity Favorable/intermediate IMDC Zhang (IJROBP 2019) Retrospective; SBRT used to defer systemic therapy in oligometastatic RCC UT Southwestern data What was missing: a prospective trial powered to estimate systemic therapy-free survival Small sample sizes, single-institution feasibility data, no prespecified STFS threshold The MD Anderson expansion cohort (n=121) was designed to fill this gap

Study design

 

  • Single-arm, single-centre, phase 2 trial (MD Anderson Cancer Center)
  • Enrolled July 2018 – May 2023; 121 patients
  • Co-primary endpoints:
    • Progression-free survival (RECIST 1.1) — per-protocol population
    • Systemic therapy-free survival (STFS) — intention-to-treat population
  • Prespecified threshold for success: median STFS ≥ 24 months
  • Benchmark: Rini et al. surveillance study — median time to systemic therapy 14.9 months
  • Sample size: 100 patients (84% power); increased to 120 for censoring

Eligibility criteria

 

  • Inclusion:
    • Age ≥ 18 years
    • Histologically confirmed renal-cell carcinoma
    • 1–5 sites of metastatic disease
    • ECOG performance status 0–2
    • Candidate for definitive radiotherapy to all disease sites
    • Adequate organ function (ANC ≥ 1000, Plt ≥ 50k, Hgb ≥ 9, Tbili ≤ 1.5, ALT/AST ≤ 3× ULN)
    • Off systemic therapy ≥ 1 month or systemic therapy-naïve
  • Exclusion:
    • Pregnancy
    • Comorbidities contraindicating RT (e.g. active scleroderma)
    • Diffuse metastatic processes (leptomeningeal, diffuse bone marrow, peritoneal carcinomatosis)
  • RECIST 1.1 measurable disease was not required

Trial schema

Enrollment: 1–5 metastatic lesions, ccRCC, ECOG 0–2
Staging: CT chest + MRI abd/pelvis, CT CAP, or PET-CT
↓
MDT to all sites of disease (SBRT preferred: ≤ 5 fx, ≥ 7 Gy/fx)
Surgery or IR ablation permitted if ≥ 1 site treated with RT
↓
Surveillance off systemic therapy
Imaging q12 weeks × 1 year, then q18 weeks
↓
Disease progression?
↓
Oligoprogression (≤ 3 new/progressing sites)
↓
Additional round of MDT
Reset surveillance schedule
Polymetastatic progression (> 3 sites)
↓
Initiate systemic therapy
Also if: RT toxicity precludes MDT, local progression, or physician/patient choice

Patient characteristics

 

  • 121 patients enrolled (ITT); 120 received RT (per-protocol)
  • Median age: 66 years (IQR 61–72); 74% male
  • 98% had prior nephrectomy; median time nephrectomy → enrollment: 71 months
  • Median time metastasis diagnosis → enrollment: 10 months (IQR 2–29)
  • ECOG PS: 0 (68%), 1 (29%), 2 (3%)
  • IMDC risk: Favorable 52%, Intermediate 47%, Poor 1%
  • Previous systemic therapy: 30% (most common: pazopanib n=11, nivo/ipi n=10)
  • Prior local therapy: Metastasectomy 32%, RT 16%, IR ablation 8%
  • Lesions treated (round 1): 1 lesion (59%), 2 lesions (31%), 3–5 lesions (10%)
  • Most common met sites: Lung (n=76), lymph nodes (n=16)
  • Median GTV of all irradiated sites: 5.7 cm³ (IQR 3.1–19.2)

Radiation and surveillance details

SBRT protocol

  • ≤ 5 fractions at ≥ 7 Gy/fx to all disease sites
  • Most common: 40–50 Gy / 4 fx (n=59)
  • Alternative RT if location precluded SBRT
  • Surgery/IR ablation if ≥ 1 site received RT

Treatment details

  • Median GTV: 5.7 cm³ (IQR 3.1–19.2)
  • Round 1: 1 lesion (59%), 2 (31%), 3–5 (10%)
  • Most common sites: lung (n=76), LN (n=16)
  • 59% received ≥ 2 MDT rounds; up to 6 total
  • Only 7% local progression in irradiated sites

Surveillance protocol

  • Imaging q12 weeks × 1 year, then q18 weeks
  • AE review + CBC with differential at each visit
  • RECIST 1.1 with central radiology review
  • Schedule resets after each MDT round

Oligoprogression → more MDT

  • ≤ 3 new or progressing metastases
  • Additional round of MDT, reset surveillance

Triggers for systemic therapy

  • Polymetastatic progression (>3 sites)
  • RT toxicity precluding further MDT
  • Local progression after RT
  • Physician or patient choice

51/121 (42%) ultimately started systemic Tx

Primary outcomes

Median follow-up: 36.3 months (IQR 26.5–51.1)

Systemic therapy-free survival (co-primary endpoint, intention-to-treat, n=121)

Time from enrollment to initiation of systemic therapy or death from ccRCC

Median 34.0 months (95% CI 28.3–54.1)
1-year rate 87.5% (95% CI 80.0–92.2)
Prespecified target Median ≥ 24 months — exceeded ✓ (lower 95% CI also exceeded)
Events 51 (42%) started systemic therapy; 2 (2%) died from disease
Benchmark Rini et al. surveillance: median 14.9 months (95% CI 10.6–25.0)

Progression-free survival (co-primary endpoint, per-protocol, n=120)

Time from enrollment to RECIST 1.1 progression, clinical progression, or death from any cause

Median 17.7 months (95% CI 14.9–22.4)
1-year rate 67.5% (95% CI 58.3–75.1)
Events 86/120 (72%) — 73 RECIST progression, 7 clinical progression, 6 death
Comparators Rini surveillance: 9.4 mo; RAPPORT (MDT + pembrolizumab): 15.6 mo

Secondary outcomes and safety

Overall survival

Median OS Not reached
1-year OS 96.7% (95% CI 91.4–98.7)
3-year OS (post hoc) 86.5% (95% CI 77.5–92.1)
Deaths 23 total: 14 disease, 9 other causes

Patterns of progression

Freedom from new lesions (12-mo) 72.7%
Median time to new lesion 22.7 months
First event: new lesions 57% of patients
First event: local progression Only 7% — high local control

Toxicity (per-protocol, n=120)

Grade ≥ 2 AEs 25 patients (21%)
Grade 3–4 AEs 8 patients (7%)
Treatment-related deaths 0 (none)
Most common grade 3 Pain (4 events), leukocytosis (2)
Grade 4 event Hyperglycemia (RT to pancreas → insulin-dependent)
Grade 2 pneumonitis 5 patients (4%)

Key safety takeaway: 7% grade 3+ with no treatment-related deaths — compare to IO combination toxicity (grade 3+ in 46–73% across CheckMate 214, KEYNOTE-426, CLEAR) and metastasectomy complications (45.7% overall, Meyer J Urol 2017)

 treatment course of all patients

Exploratory: ctDNA molecular residual disease

 

  • Assay: Tumor-informed, patient-specific ctDNA panels via WGS (up to 2000 somatic variants per patient; Precise MRD, Myriad Genetics)
  • 87/89 patients had successful panel creation (98%); 149/168 plasma samples analyzed (89%)
  • Baseline MRD: 47/78 patients (60%) were ctDNA MRD-positive
    • MRD+ vs MRD−: median STFS 28.1 vs 54.1 months (HR 2.75, p=0.0064)
  • 3-month MRD (landmark analysis): 37/71 patients (52%) MRD-positive
    • MRD+ vs MRD−: median STFS 19.1 vs 49.9 months (HR 4.42, p<0.0001)
  • 25% of baseline MRD+ patients converted to MRD− at 3 months
  • Multivariable model: ctDNA MRD status remained independently prognostic after adjusting for # lesions and prior systemic therapy lines (all p<0.05)

ctDNA MRD and systemic therapy-free survival

A) Baseline

ctDNA MRD+ vs MRD− and STFS (HR 2.75, p=0.0064)


B) 3-month

ctDNA MRD+ vs MRD− and STFS (HR 4.42, p<0.0001)]

ctDNA as a treatment sufficiency signal after SBRT The change in ctDNA — not baseline alone — reveals whether local therapy was enough NSCLC — CURB ccRCC — Tang 2025 Breast — CURB TUMOR oligoprogressive sites 1–5 metastatic sites visible + occult micromets SBRT to all visible sites all ablated ablated — 25% convert MRD− visible gone, occult persists ctDNA SBRT arm — ctDNA clears (p=0.02) baseline 8 weeks VAF values reported graphically only SOC only — no change (p=0.11) baseline 8 weeks MRD+ → MRD− converters (25%) baseline 22.5 ppm median 3 months undetectable STFS 50 mo (vs 19 mo if persistent) Persistent MRD+ (75%) baseline 22.5 ppm 3 months 18.5 ppm SBRT arm — no change (p=0.56) baseline 8 weeks VAF values reported graphically only SOC only — no change (p=0.56) baseline 8 weeks OUTCOME SBRT works PFS 10.0 vs 2.2 mo (HR 0.41) Only 14% new distant mets Response stratifies MRD−: STFS 50 mo → defer Tx MRD+: STFS 19 mo → start Tx SBRT fails PFS 4.4 vs 4.2 mo (no benefit) 61% new distant mets What does ctDNA change tell the clinician? "Treatment was enough" ctDNA fell → SBRT ablated all disease-driving clones → Continue current therapy "Was it enough?" — check ctDNA MRD− at 3 mo → defer systemic Tx MRD+ at 3 mo → start Tx sooner → ctDNA-guided decision "Treatment was never enough" ctDNA unchanged → occult disease persists beyond SBRT reach → Prioritize systemic therapy Tang ccRCC trial advance: ctDNA moved from correlative biomarker to treatment sufficiency signal MRD− patients safely stayed off systemic therapy for a median of 50 months (22.5 ppm → undetectable) Active tumor Ablated Occult micromet ctDNA before → after Note: CURB reports VAF change graphically (exact values not published). Tang reports tumor fraction in ppm. Bar sizes are schematic; assays use different scales.

Prognostic factors for STFS

 

  • Prespecified baseline clinical factors associated with STFS:
    • Number of metastatic lesions at enrollment: HR 1.53 per lesion (95% CI 1.14–2.07, p=0.0044)
    • Number of previous lines of systemic therapy: HR 1.50 per line (95% CI 1.11–2.04, p=0.0078)
  • ctDNA MRD status independently prognostic on multivariable analysis (all p<0.05 with lesion count and prior systemic therapy)
  • Factors associated with baseline MRD positivity (not adjusted for multiple testing):
    • IMDC risk group (p=0.0034)
    • Bone metastases (p=0.038)
    • Total irradiated GTV (p=0.0045)
  • Notably: factors associated with MRD status were not the same as those associated with STFS, suggesting independent prognostic value
SBRT instead of systemic therapy: not all histologies are equal Willmann et al., JAMA Netw Open 2025 — meta-analysis of SBRT without upfront systemic therapy (29 studies, n=2,074) Pooled 1–2 yr systemic therapy-free survival by histology RCC 87% (76–95) n=53; I²=0% Prostate 78% (67–87) n=292; I²=72% Gynecological 66% (59–73) n=178; 1 study Sarcoma 56% (30–80) n=16; 1 study Mixed tumors 47% (18–78) n=445; I²=98% Pooled 70% (57–81) n=984; 13 studies; I²=93% Histology explains 35.7% of heterogeneity (R²). RCC subgroup: Tang 2021 feasibility (n=30) + Hannan 2022 (n=23). Note: Tang 2025 expansion cohort (n=121, median STFS 34 mo) postdates this meta-analysis and is not included. Two key histologies not separately analyzed in the meta-analysis NSCLC — favorable CURB (Tsai, Lancet 2024): randomized, oligoprogressive SBRT + SOC: PFS 10.0 vs 2.2 mo (HR 0.41, p=0.004) SBRT cleared ctDNA (p=0.02); only 14% new distant mets Also: Gomez (JCO 2019): PFS 14.2 vs 4.4 mo (HR 0.30) Tsai Lancet 2024; Gomez JCO 2019; Palma Lancet 2019 Breast — negative CURB (same trial): SBRT did NOT clear ctDNA (p=0.56) PFS 4.4 vs 4.2 mo (no benefit); 61% new distant mets NRG-BR002 (n=125): also negative; phase 3 canceled EXTEND breast basket: negative. Consistent across trials. Tsai Lancet 2024; Chmura ASCO 2022 A spectrum of benefit — biology determines the value of SBRT alone Breast Mixed / Sarcoma NSCLC / Prostate RCC ← Less benefit from SBRT alone Greater STFS with SBRT alone → Tumor biology determines whether SBRT alone can meaningfully defer systemic therapy

Prostate cancer: the most mature MDT evidence — and model for RCC

Two key questions answered by randomized data: (1) Does MDT beat observation? (2) Should MDT be combined with systemic therapy?

Q1: MDT alone vs observation

STOMP
Ost, JCO 2018; 5-yr ASCO GU 2020		 Median ADT-free survival: 21 vs 13 mo
5-yr ADT-free: 34% vs 8% (HR 0.57)
5-yr CRPC-free: 76% vs 53%
ORIOLE
Phillips, JAMA Oncol 2020		 6-mo progression: 19% vs 61% (p=0.005)
PFS: not reached vs 5.8 mo (HR 0.30)
Pooled analysis
Deek, JCO 2022		 PFS: 11.9 vs 5.9 mo (HR 0.44, p<0.001)
15–20% durable PFS beyond 4–5 yrs
High-risk mutations (ATM/BRCA/Rb1/TP53): HR 0.05

Q2: MDT + ADT vs ADT alone

EXTEND
Tang, JAMA Oncol 2023		 PFS: NR vs 15.8 mo (HR 0.25, p<0.001)
Eugonadal PFS: NR vs 6.1 mo (HR 0.32, p=0.03)
MDT + intermittent ADT preserves testosterone
RADIOSA
Marvaso, Lancet Oncol 2025		 PFS: 32.2 vs 15.1 mo (HR 0.43, p=0.001)
SBRT + 6-mo ADT vs SBRT alone
First RCT showing ADT adds to SBRT
WOLVERINE
X-MET, ASCO GU 2025		 Pooled IPD from 5 RCTs (n=472)
48-mo OS: 87% vs 75% (HR 0.64, p=0.057)
First signal of OS benefit with MDT

The parallel to RCC: Both prostate and RCC have indolent biology in the oligometastatic subset. Both show SBRT can defer systemic therapy. Prostate has randomized data; RCC now has the Tang expansion cohort (n=121, STFS 34 mo). The next step for RCC is what prostate already has — randomized trials (SOAR, ASTROs).

Key difference: In prostate, the systemic therapy being deferred is ADT — with known QoL effects (sexual dysfunction, metabolic syndrome, osteoporosis). In RCC, the deferred therapy is IO combinations — with different but significant toxicity.

Conclusions

 

  • Serial MDT without systemic therapy in oligometastatic ccRCC achieved a median STFS of 34.0 months, exceeding the 24-month prespecified threshold
  • Favorable PFS (median 17.7 months)
  • OS (median not reached; 3-year OS 86.5%)
  • Compared favorably to surveillance alone (Rini: PFS 9.4 mo) and MDT + pembrolizumab (RAPPORT: PFS 15.6 mo, OS 20 mo)
  • Modest toxicity profile: grade ≥ 3 in only 7%; no treatment-related deaths
  • 59% of patients received ≥ 2 rounds of MDT, demonstrating feasibility of serial approach
  • ctDNA MRD status at baseline and 3 months was an independent prognostic biomarker for STFS
  • Number of metastatic lesions and prior systemic therapy lines also prognostic
  • MDT can enable meaningful treatment de-escalation, potentially improving QoL and reducing costs

Strengths and limitations

 

  • Strengths:
    • Largest prospective trial of serial MDT without systemic therapy for oligometastatic ccRCC (n=121)
    • Longest follow-up in this setting (median 36.3 months)
    • Central radiology review (RECIST 1.1) by subspecialized radiologist
    • Novel integration of second-generation ultrasensitive ctDNA assay
    • 60% baseline ctDNA detection rate (vs 25% with first-generation assays)
    • Prespecified endpoints, futility analyses, and exploratory biomarker analyses
  • Limitations:
    • Single-arm, single-institution — no randomized comparator
    • No quality-of-life data collected
    • Systemic therapy initiation partly physician/patient discretion
    • RECIST 1.1 developed for systemic therapy assessment — may not optimally capture oligomet MDT response
    • ctDNA analysis exploratory and in a subset only; tissue procurement biases
    • Predominantly favorable/intermediate IMDC risk and predominantly lung metastases — limits generalizability

Discussion points

 

 

  1. How should radiation oncologists select patients for MDT-only strategies? Should ctDNA MRD testing be integrated into routine practice, and at what time points?
  2. What are the implications for non-clear-cell histologies, which have fewer effective systemic therapy options? Could MDT have an even greater role? 
  3. In what other histologies can metastasis directed therapy be used to delay systemic therapy vs used to maintain efficacy of systemic therapy?