SUPREMO
ForPost-mastectomy breast, node-negative high-risk or 1-3 positive nodes
TL;DREditorial critique: chest-wall-only RT cut 10yr CW recurrence 1.1% vs 2.5% (HR 0.45) but RNI was prohibited.
The RT-relevant number is buried in SUPREMO's supplement: node-positive LRR 3.3% vs 4.8%, HR 0.51 (0.27-0.96), significant even with chest-wall-only fields and RNI prohibited. With supraclavicular coverage in 12% and IMN in under 2%, this trial never tested comprehensive PMRT, so it cannot settle the elective nodal decision.
In a post-mastectomy patient with 1-3 positive nodes staged by SLNB alone, this argues SUPREMO does not license PMRT omission; it says nothing about node-negative pts without adverse features, where EBCTCG also found no benefit.
The supplement carries the number the headline drops: node-positive LRR 3.3% vs 4.8%, HR 0.51 (0.27-0.96), achieved even with RNI prohibited, supraclavicular coverage in 12% and IMN in under 2%. SUPREMO therefore constrains chest-wall-only treatment, not the elective nodal volume decision.
Axillary staging gates how far SUPREMO travels: only 14% had SLNB alone, and roughly 30% of SLNB pts with 1-3 positive nodes harbour further nodes at completion ALND. Choosing ALND to justify skipping PMRT trades a 8% five-year lymphedema risk with SLNB alone for 25% with ALND alone.
Also covered Jul 7
9 details
ASO Perspectives editorial in Annals of Surgical Oncology, not new trial data. Two radiation oncologists re-read SUPREMO against the PMRT and RNI evidence base. PMRT after neoadjuvant therapy is explicitly out of scope.
The central claim is that SUPREMO tested chest wall alone, not PMRT: RNI was prohibited, supraclavicular nodes were covered in 12% of PMRT-arm pts (n=97), and internal mammary irradiation occurred in fewer than 2% across both arms. Non-UK centres could give RNI in the observation arm, and 12 control pts received supraclavicular RNI.
Only 25% were truly node-negative and would not be offered PMRT under current guidelines; the majority had N1 disease (1-3 nodes). 65% were hormone-receptor positive, TNBC was 10%, and only 14% had SLNB alone with the majority undergoing ALND.
SUPREMO's 10-year chest-wall recurrence fell from 2.5% to 1.1% (HR 0.45) with no gain in overall LRR, DFS or OS. The supplement carries the signal the headline drops: LRR 4.8% to 3.3%, HR 0.51 (95% CI 0.27-0.96) in node-positive pts, not in node-negative.
| Approach | 5yr lymphedema risk |
|---|---|
| SLNB alone | 8% |
| SLNB + RNI | 11% |
| ALND alone | 25% |
| ALND + RNI | 30% |
EBCTCG 2014 found PMRT cut 10-year LRR by 17.9% and 20-year breast cancer mortality by 8% in 1-3 node-positive women, persisting with a single positive node, with no node-negative benefit. MA-20 and EORTC 22922 both showed RNI benefit in 1-3 node-positive disease despite near-universal ALND, and a later EBCTCG RNI meta-analysis showed gains at 15 years including in contemporary systemic-therapy trials.
This is a single-perspective editorial from two radiation oncologists, so the framing selects evidence favouring comprehensive RNI, and the reader gets no independent re-analysis of SUPREMO's data. Its strongest number, the node-positive LRR HR 0.51, comes from a supplementary subgroup the trial did not power for.
The authors leave the genuinely open questions to trials in progress: MA.39 (Tailor-RT) for RNI omission at low recurrence score, T-Rex for RNI omission in hormone-sensitive disease with one to two macrometastatic SLNs. They also note the surgical corollary, that ALND should not be chosen to earn a PMRT omission, since ALND is the dominant lymphedema driver.
Editorial, no new data; contests the omission reading of SUPREMO on field design and population grounds, aligning with ASTRO 2025 and NCCN rather than the trial's public messaging.
- RNI omission in 1-3 node-positive pts staged by SLNB alone
- Whether low Oncotype score permits comprehensive RNI omission
- PMRT effect in triple-negative disease after mastectomy
📚 Sources · 📄 1 paper
The longer read
The argument here is a definitional one, and it is stronger than a typical editorial rebuttal because it does not dispute SUPREMO's numbers at all. It disputes what the trial's intervention was. If the PMRT arm prohibited regional nodal irradiation, covered supraclavicular nodes in 12% of pts and internal mammary nodes in fewer than 2%, then the trial is a test of chest-wall irradiation, and the evidence base it is being read against, the Danish and British Columbia trials, MA-20, EORTC 22922, and the EBCTCG pooled analyses, all rest on comprehensive fields. On that reading a null result for DFS and OS is the expected result rather than a surprising one, and the honest conclusion is that the comprehensive-PMRT question remains untested in a modern cohort rather than answered in the negative.
What should move a reader's confidence toward the authors is the internal consistency of the supplementary finding. Even with the incomplete field, LRR improved from 4.8% to 3.3% (HR 0.51, 95% CI 0.27-0.96) in node-positive pts and did not improve in node-negative pts. That is the pattern EBCTCG 2014 described, benefit concentrated in 1-3 node-positive disease and absent when nodes are clear, reproduced inside the trial now being cited for omission. A subgroup analysis buried in a supplement is fragile evidence on its own, and the confidence interval nearly touches unity, but it is a prespecified stratification variable rather than a fished-for subgroup, and its direction agrees with every prior dataset.
What should move confidence the other way is the direction of the authors' interests and the selectivity that follows. Both are radiation oncologists arguing that a trial suggesting less radiation actually supports more, and the TNBC subgroup is handled asymmetrically: worse outcomes with PMRT are attributed to underpowering, while the favourable node-positive LRR subgroup, from the same underpowered trial, is treated as confirmatory. Underpowering cuts both ways or it cuts neither. The editorial also does not seriously engage with the possibility that modern systemic therapy has lowered the absolute LRR floor enough that the historic proportional benefits translate into a much smaller absolute gain. A 2.5% ten-year chest-wall recurrence rate in the observation arm is remarkably low by the standards of the trials being invoked, and that fact does real work against the case for routine comprehensive coverage regardless of what the fields were.
The practical read differs by axillary surgery, and this is where the piece earns its place. In a cohort where only 14% had SLNB alone, SUPREMO cannot speak to the patient most clinics now see, a woman with one or two macrometastatic sentinel nodes and no completion dissection, where roughly 30% would harbour further nodal disease. The trial's population was staged aggressively enough to have removed much of the occult burden that RNI is meant to address, which weakens generalisation in the direction of omission, not toward it. The hormone-receptor-positive majority with only ten-year follow-up is a softer objection, since late recurrences after ten years are more often distant than locoregional, but the genomic gap is real and MA.39 and T-Rex are the correct places to resolve it. Until those report, the defensible position is that SUPREMO constrains the chest-wall question and leaves the nodal one open.