Radiation decisions after neoadjuvant chemotherapy and mastectomy
TL;DRNo randomised post-NAC PMRT data: retrospective MDACC/NSABP series put ypN+ LRR without RT at ≥15%, cT1–T2 ypN0 <10%, supporting response-adapted PMRT.
NSABP B-18NSABP B-27RAPCHEM
The omission case rests on small denominators: NSABP pCR + ypN0 (n=68) had 5.9% 8-yr LRF without PMRT, MDACC cT1–2 ypN0 (n=19) 5%. Clinical T3N0 converting to ypN0 is the contested cell (MDACC 14% vs 2%, p=0.06; NSABP 6.2% and 11.8%), so chest wall and nodal RT stay individualized there.
In clinical T1–T2 disease that is ypN0 after NAC and mastectomy, these data support discussing PMRT omission; they do not extend to clinical T3–T4, N2–N3, or ypN+ disease, where retrospective series favor PMRT.
The omission case rests on small denominators: NSABP pCR + ypN0 (n=68) had 5.9% 8-yr LRF without PMRT, MDACC cT1–2 ypN0 (n=19) 5%. Clinical T3N0 converting to ypN0 is the contested cell (MDACC 14% vs 2%, p=0.06), so chest wall and nodal RT stay individualized there.
12 details 2 trials watching
Lancet Oncol 2012 Personal View from MD Anderson radiation and surgical oncology authors. Medline and Embase searched to October 2011; inflammatory breast cancer excluded.
Upfront-mastectomy PMRT benefit comes from the Danish and Canadian randomised trials and the EBCTCG meta-analysis, all predating routine NAC. Pathological extent after NAC carries higher risk: with ≥4 involved nodes, MDACC 5-yr LRR was 53% vs 23% neoadjuvant vs upfront surgery (p<0.001).
MDACC series overlap across publications and span eras of evolving staging, imaging, chemo and surgery. NSABP pts were earlier stage than MDACC pts: only 35% had clinical nodal involvement at diagnosis.
Authors propose response-adapted PMRT: LRR at least 15% without RT in ypN+ supports PMRT, <10% in cT1–T2 ypN0 supports omission. Clinical stage III with pCR is unsettled, with MDACC and Le Scodan pointing in opposite directions.
| Subgroup | No PMRT | PMRT | p |
|---|---|---|---|
| Clinical T3 | 22% | 8% | 0.002 |
| Clinical T4 | 46% | 15% | <0.0001 |
| Clinical N2–3 | 40% | 12% | <0.0001 |
| Clinical N1 | 14% | 9% | 0.062 |
| Clinical N0 | 23% | 10% | 0.014 |
| Path tumour 2.1–5.0 cm | 31% | 14% | 0.002 |
| Path tumour >5.0 cm | 52% | 13% | 0.001 |
| ≥4 nodes at resection | 59% | 16% | <0.0001 |
| Stratum | n | 5-yr LRR |
|---|---|---|
| T1–2, ypN0 | 19 | 5% |
| T3–4, ypN0 | 23 | 34% |
| T1–2, ypN+ | 42 | 13% |
| T3–4, ypN+ | 64 | 36% |
| Cohort | Endpoint | No PMRT | PMRT | p |
|---|---|---|---|---|
| MDACC cStage III + pCR | 10-yr LRR | 33% | 7.3% | 0.04 |
| Le Scodan stage II, ypN0 | 5-yr LR-RFS | 92.9% | 97.4% | >0.4 |
| Le Scodan stage III, ypN0 | 5-yr LR-RFS | 94.7% | 90.9% | >0.2 |
- Can PMRT be safely omitted in cT1–T2 cN+ converting to ypN0? active Standard or Comprehensive Radiation Therapy in Treating Patients With Early-Stage Breast Cancer Previously Treated With Chemotherapy and Surgery Phase NAn=1636 · primary completion 2023-09 · phase 3 randomised chest wall/nodal RT after NAC + surgeryn=100 · primary completion 2028-06 · response-adapted RNI de-escalation, cT1-3 N1 to pN0 post-NAT
- PMRT benefit in clinical T3N0 that is ypN0 after NAC
- Does PMRT benefit clinical stage III with pCR?
📚 Sources · 📄 1 paper
The longer read
Published in 2012, this Personal View is best read as the hypothesis document for response-adapted PMRT rather than an evidence base for it. Its central argument is that the upfront-surgery rules (four or more involved nodes, pathological stage III) cannot be transplanted to the post-NAC setting, because chemotherapy changes what the pathology report means. The strongest evidence offered is the MDACC comparison of neoadjuvant and upfront cohorts treated without RT, where four or more involved nodes carried 53% vs 23% 5-yr LRR. Residual nodal disease after NAC marks chemoresistant biology, not just burden, so the same nodal count signals a worse tumour.
The addition argument is stronger than the omission argument, and the asymmetry matters. The case for PMRT in clinical T3–T4, N2–N3 and ypN+ disease rests on large absolute LRR differences in a 676-patient MDACC series, consistent in direction with the Danish and Canadian randomised trials and the EBCTCG overview in the upfront setting. The authors note the irradiated groups carried more adverse features, which would bias against finding a benefit rather than toward it. The case for omission rests on small denominators: 68 NSABP pts with pCR and ypN0, 19 MDACC pts in the cT1–T2 ypN0 stratum, 30 stage II pCR pts with no events in either arm. Low event rates in small groups leave wide uncertainty, and the low-risk NSABP cohorts were earlier stage than the MDACC experience, so part of what defines the low-risk group is who entered the trials.
The contested cells are clinical stage III with pCR and clinical T3N0 converting to ypN0. MDACC and Le Scodan disagree on the first, and the disagreement cannot be resolved from these data: both are retrospective with few events, and the Le Scodan adjusted HR of 0.37 has a 95% CI of 0.09–1.61. For cT3N0 ypN0, MDACC reports 14% vs 2% (p=0.06), while NSABP no-RT rates are 11.8% with residual breast disease and 6.2% with pCR. The gap between institutions in untreated cT3N0 LRR is the tell: baseline risk differs across referral populations by more than the effect being estimated, which is exactly the situation in which retrospective comparison misleads.
For a radiation oncologist the durable contribution is the two-axis framework, clinical stage at presentation plus nodal response, and the insistence on documenting pre-NAC nodal extent with ultrasound of all regional basins and FNA so the later decision has an anchor. What the review cannot address is the modern regimen. The cohorts received doxorubicin-based NAC, and the authors flag that trastuzumab-based therapy converted 74% of cN+ pts to ypN0 in one series, enlarging precisely the group whose omission data are thinnest. It also addresses whether to irradiate, not what to irradiate: chest wall vs regional nodal coverage after nodal conversion is never separated.
The authors' own conclusion, that RAPCHEM and a randomised Alliance proposal are needed before omission becomes routine, is the correct read. The omission recommendation would be wrong if ypN0 status after effective NAC under-calls residual nodal disease, or if biology-driven systemic therapy shifts baseline LRR in ways that make these historical no-RT rates unrepresentative.