onc brain

The longer read

The value of this series is that it measures the untreated baseline. Most of what the field believed about high-risk node-negative disease after mastectomy came from cohorts treated in the 1980s to early 2000s, when far fewer pts received adjuvant systemic therapy and HER2-directed therapy was not routine. The authors report 86% systemic therapy use against the 38% to 77% they cite for older series, and a crude LRR of 4.7% against historical figures as high as 20% in pts with multiple features. If the baseline has fallen that far, any proportional benefit from PMRT now buys a much smaller absolute gain, and that is the real argument of the paper.

The comparison with the nodal irradiation trials makes the same point from a different direction. EORTC 22922 and MA.20 are the evidence usually invoked to justify regional RT in high-risk N0 disease, yet node-negative pts were a minority of MA.20 and EORTC enrolled mostly breast-conserving surgery. The regional recurrence here with no RT at all, 2.6%, falls below the 4.2% the source quotes for the EORTC 22922 control arm. Cross-series comparison of this kind is weak, since definitions, follow-up and populations differ, but it does suggest the trial-derived benefit should not be ported directly onto a modern mastectomy population with negative SLNs on careful serial sectioning.

The methodological limits should move confidence in specific ways. First, this is a cohort with no RT arm, so it cannot say what PMRT would have prevented; it can only say what happens without it. Second, the headline negative finding, that counting risk factors does not stratify LRR, rests on a log-rank test across groups where the highest-risk stratum had only 32 pts. The crude rate there was 9.4% with a lower 5-year LRR-free bound of 0.71, which is not reassurance; it is an unresolved estimate. A reader should treat "features do not matter" as "features could not be shown to matter in this sample." Third, the risk score weights five heterogeneous features equally, which can dilute a genuine signal from any one of them. Tumor size, the one variable analyzed continuously, was the one that came through (HR 1.70), and that may partly reflect the greater statistical power of a continuous term.

The pattern of failure deserves more attention than the abstract gives it. Eight of 22 isolated LRRs were axilla-only, and 7 of those 8 followed SLNB alone. The authors suspect false-negative SLNB. That is a surgical staging issue, and chest wall-directed RT would not address it, while comprehensive nodal RT might. This is a reminder that a low aggregate LRR can hide a failure mode that one specific RT field would target.

The systemic therapy association (HR 0.40) is real-world and confounded by indication, and a third of pts who failed had not received recommended systemic therapy. For an RT decision, the practical read is that the low baseline applies to pts actually receiving indicated systemic treatment. The paper supports not referring T1-T2N0 pts for PMRT on the strength of LVI, age, grade, multifocality or medial location alone. It leaves open T2 tumors carrying several features, and those are the pts SUPREMO's node-negative subset is positioned to inform.