SARON NCT02417662
ForSynchronous oligometastatic AGA-negative NSCLC, 1-5 mets, non-progressing on 1L SACT
HR 0.99
95% CI 0.63-1.55, p=0.95; mOS 31.0 vs 28.0 mo, not met
TL;DRConsolidation RT added to 1L SACT: mOS 31.0 vs 28.0mo (HR 0.99, p=0.95) despite mPFS 13.9 vs 8.9mo (HR 0.64) in synchronous oligometastatic AGA-negative NSCLC.
Pattern of failure is the RT read: 12-mo local failure fell 39.5% to 13.0%, but new distant mets ran 36.2% vs 32.3%. Comprehensive RT controls what it treats without changing systemic seeding, so consolidation in AGA-negative oligometastatic NSCLC is a PFS and local-control decision, not a survival claim.
In synchronous oligometastatic AGA-negative NSCLC stable or responding on 1L SACT, this supports consolidation RT for local control and PFS when lesions threaten morbidity; it does not support counselling a survival gain, and it does not address EGFR/ALK-positive disease.
12-mo local failure fell 39.5% to 13.0% with radical RT to primary ± nodes plus SABR to all OM sites, while new distant mets were unchanged (36.2% vs 32.3%). Consolidation is justified on local control and PFS; dose and fractionation are not reported in source, so technique transfer is unverified.
The OS null (HR 0.99) arose on a mixed backbone: 40.0% chemo, 15.0% IO, 45.0% chemo-IO, with IO added to control mid-trial. Whether consolidation adds survival on uniform 1L chemo-IO remains untested, so referral for RT is a PFS decision, not a survival one.
10 details 2 trials watching
Multicentre randomised (1:1) phase II, N=140 across 22 UK centres, Oct 2016 to Jul 2023. Stratified by SACT, histology, N-stage, number of metastases, brain mets and site. Median follow-up 37.4 mo.
Synchronous oligometastatic AGA-negative NSCLC, 1-5 metastases in ≤3 organs, all sites amenable to SABR/SRS, randomised after initial SACT with CR, PR or SD. Median age 68, 58.6% adenocarcinoma, 52.9% N2/3, 62.9% a single OM site.
SACT was 40.0% chemo, 15.0% IO and 45.0% chemo-IO; the control arm changed to include IO in January 2020. Median SACT cycles was 4 in both arms, with maintenance permitted.
Radical RT to the lung primary ± locoregional nodes plus SABR to all OM sites. 62/70 (88.6%) received RT: 51 conventional and 11 SABR to the primary, 54 SABR to ≥1 OM site. Dose and fractionation not reported in source.
OS. Major secondary: PFS. Also reported: cumulative incidence of local failure and new distant metastasis, AEs, EORTC QLQ-C30/LC13.
G≥3 AE rate 28.6% in both arms; 6 pts had a G3 event possibly RT-related. QoL similar between arms.
The Gomez LCT phase II reported OS 41.2 vs 17.0 mo, but in 49 pts before routine IO. SABR-COMET supported the oligometastatic paradigm with few lung primaries. SARON is the larger NSCLC-specific randomised test and does not reproduce the OS benefit.
Systemic backbone shifted mid-accrual, so the RT effect on a uniform chemo-IO backbone is not isolated. Post-progression therapy, including salvage RT in the control arm, is not reported in source and could compress OS.
RT did what it was aimed at: local failure fell and PFS improved, while new distant metastasis was unchanged. The OS null points to systemic relapse as the limiting event, so lesion count alone selects poorly for the pts in whom ablation alters natural history.
| Endpoint | SACT+RT | SACT | Effect |
|---|---|---|---|
| mOS (primary) | 31.0 mo | 28.0 mo | HR 0.99 (95% CI 0.63-1.55), p=0.95 |
| mPFS | 13.9 mo | 8.9 mo | HR 0.64 (95% CI 0.43-0.95), p=0.03 |
| 12-mo cumulative incidence, local failure | 13.0% | 39.5% | p<0.01 |
| 12-mo cumulative incidence, new distant mets | 36.2% | 32.3% | p=0.28 |
| G≥3 AE rate | 28.6% | 28.6% | n/a |
CONSORT flow
Randomised, completed analysis with OS as stated primary and null result, diverging from prior small LCT trial OS signal. Phase II size leaves OS CI wide.
- Does consolidation RT add OS on a uniform chemo-IO backbone? active Nivolumab and Ipilimumab With or Without Local Consolidation Therapy in Treating Patients With Stage IV Non-Small Cell Lung Cancer Phase 3n=339 · primary completion 2027-03 · phase 3 nivo/ipi ± local consolidation, stage IV NSCLCrecruiting First-line Immunotherapy-based Standard of Care and Local Ablative Treatments for Oligometastatic Non-small Cell Lung Cancer Patients. Phase 3n=124 · primary completion 2030-02 · phase 3 1L IO ± chemo with vs without radical local tx
- Can ctDNA or systemic response select pts who gain OS from consolidation?
- How much did control-arm salvage RT compress the OS difference?
📚 Sources · 📄 2 papers
Abstract
Abstract
The longer read
SARON is the first NSCLC-specific randomised trial of this size to test consolidation RT in synchronous oligometastatic disease with OS as the stated endpoint, and its answer is split. The local-therapy hypothesis holds where RT acts: cumulative incidence of local failure within 12 months was 13.0% vs 39.5%, and mPFS 13.9 vs 8.9 mo (HR 0.64). It does not hold where the oligometastatic argument has always made its larger claim. OS was 31.0 vs 28.0 mo with an HR of 0.99, and new distant metastases appeared at 36.2% vs 32.3%. The trial treated visible disease well and left the rate of new seeding untouched.
That pattern is the most informative part of the result, more than the OS point estimate. A survival benefit from consolidation requires either that the treated lesions were the main source of later dissemination, or that preventing local progression spares pts the morbidity and treatment breaks that shorten survival. The distant-metastasis curves argue against the first mechanism in this population. Most pts who progressed did so at new sites that no RT field would have covered.
Against prior data, the contrast with the Gomez LCT trial is the obvious one. That study's OS of 41.2 vs 17.0 mo came from 49 pts in a pre-IO setting and has carried a lot of the enthusiasm for comprehensive consolidation. A small trial with a large OS effect is exactly the result most likely to shrink on replication, and SARON is a reasonable test of that. Still, SARON is not a definitive refutation. With 140 pts in a phase II design the OS confidence interval runs from 0.63 to 1.55, which is compatible with a clinically meaningful benefit as well as with harm. The fair statement is that SARON gives no evidence of an OS effect, not that it proves there is none.
Two design features should temper confidence in the null as well. First, the systemic backbone changed during accrual: only 60% received any IO, and the control arm moved to IO in January 2020. If consolidation is most valuable in pts whose micrometastatic disease is controlled by durable checkpoint response, a chemo-heavy cohort is the wrong place to find that. Second, post-progression therapy is not reported in the source. Control-arm pts who progressed locally may have received salvage RT, which would convert an upfront-versus-deferred comparison into something close to a timing question and push OS together.
For the RT reader the practical consequence is a change in the counselling frame rather than the abandonment of consolidation. Toxicity was not increased overall (G≥3 28.6% both arms, 6 pts with G3 events possibly RT-related) and QoL was similar, so local control and PFS remain defensible goals, particularly for a primary or nodal mass likely to cause symptoms. What SARON removes is the license to describe comprehensive RT as life-prolonging in unselected AGA-negative oligometastatic NSCLC. The trial would be wrong in its OS conclusion if a uniformly chemo-IO-treated, response-selected cohort, or one enriched by ctDNA clearance, showed a survival effect that the heterogeneous SARON population diluted. That is the question the next trial should be built around, rather than another count-based eligibility rule.