DBCG IMN2 NCT06549920
ForNode-positive early breast, upfront surgery + ALND, no neoadjuvant therapy
HR 0.85
95% CI 0.76-0.94, p=0.0016; 15y OS 65.0% vs 60.8%
TL;DRAdjusted OS HR 0.85 (0.76-0.94) with IMNI in node-positive breast cancer; 15y OS 65.0% vs 60.8% despite taxanes, trastuzumab, AIs and 3D RT.
The benefit held in the 1-3 node group, HR 0.85 (0.73-0.97), the population where guidelines still hedge on IMNI. It was delivered with free-breathing 3D wide tangents at 50 Gy/25 Fx, and dilution (20.4% incidental coverage left-sided) biases toward the null, so modern gated or VMAT coverage likely does no worse.
In node-positive breast cancer after upfront surgery and ALND, including 1-3 positive nodes, this supports covering the IMN in the regional field; it does not extend to pts treated after neoadjuvant therapy.
The 1-3 node HR 0.85 (0.73-0.97) addresses the group where IMN coverage is most contested, using 50 Gy/25 Fx free-breathing 3D tangents to ICS 1-4. Because incidental left-sided coverage (V90% 20.4%) dilutes the contrast, the benefit is likely understated with gated or VMAT coverage.
11 details 3 trials watching
Prospective nationwide population-based cohort across 6 Danish RT centres, N=4541, 2007-2014. Not randomized: IMNI allocated by tumour laterality per national guideline; ITT right vs left. Median follow-up 13.7y.
Macrometastatic node-positive breast cancer, all with ALND, mastectomy or BCS. Excluded neoadjuvant systemic therapy, bilateral BC, prior malignancy, recurrence before RT. Median age 59; 68.3% had 1-3 positive nodes.
EC x3 then docetaxel x3 when chemo indicated; AI postmenopausal, tamoxifen premenopausal; trastuzumab concurrent for HER2+. 96.2% of chemo pts received taxanes.
48 Gy/24 Fx before Jan 2009, 50 Gy/25 Fx after. IMN target ICS 1-4 with 3D conformal wide tangents, free breathing, plus level II-III, interpectoral and level IV in all pts.
Primary: OS. Secondary: breast cancer mortality and distant metastasis, with non-breast-cancer death as competing event.
15y death from ischemic or valvular heart disease 0.2% right vs 0.7% left. 15y secondary lung cancer death 1.8% vs 2.2%. No smoking or baseline cardiac data.
OS HR 0.85 sits with the EBCTCG regional RT RR 0.90 (0.84-0.96) and underpowered KROG 08-06 HR 0.87 (0.57-1.31). Magnitude matches DBCG IMN1, which treated pts before taxanes and 3D planning.
Left-sided controls got incidental IMN dose (median CTV V90% 20.4%) and 10.1% received IMNI, compressing separation. 1-3 node analysis was post hoc; no correction for multiple testing across subgroups.
Newer systemic therapy did not erase the IMNI gain. The medial/central ≥4 node HR 0.98 and ER-/HER2+ HR 1.49 are exploratory signals, not omission criteria; authors found no subgroup for safe omission.
| Endpoint | IMNI (n=2194) | No IMNI (n=2347) | Adjusted HR (95% CI), p |
|---|---|---|---|
| OS at 15y | 65.0% | 60.8% | 0.85 (0.76-0.94), p=0.0016 |
| Breast cancer mortality, 15y CI | 21.4% | 23.6% | 0.84 (0.74-0.95), p=0.0077 |
| Distant metastasis, 15y CI | 25.1% | 26.9% | 0.87 (0.78-0.98), p=0.026 |
| Subgroup | HR (95% CI) | 15y OS IMNI vs no IMNI |
|---|---|---|
| 1-3 LN (n=3100, post hoc) | 0.85 (0.73-0.97) | 72.3% vs 68.0% |
| 4-9 LN | 0.92 (0.75-1.12) | 55.3% vs 51.6% |
| ≥10 LN | 0.75 (0.59-0.96) | 39.2% vs 31.1% |
| Lateral ≥4 LN | 0.77 (0.64-0.93) | 53.8% vs 41.7% |
| Medial/central ≥4 LN | 0.98 (0.79-1.21) | 43.6% vs 50.4% |
| ER-/HER2+ | 1.49 (0.98-2.25) | 57.7% vs 68.2% |
Large prospective multicentre cohort, primary endpoint met with 13.7y f/u, but laterality allocation is not randomization. Aligns with EBCTCG and DBCG IMN1.
- Does IMNI benefit persist with adjuvant immunotherapy or ADCs?
- Is regional RT omission safe after neoadjuvant therapy with pCR? 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 · regional nodal XRT vs breast/chestwall only post chemon=100 · primary completion 2028-06 · RNI de-escalation in cN1 → ypN0 after neoadjuvant therapyrecruiting Axillary Management in Breast Cancer Patients With Needle Biopsy Proven Nodal Metastases After Neoadjuvant Chemotherapy Phase NAn=1900 · primary completion 2030-02 · omits axillary RT if SLNB node-neg after NACT; DFS
- Is ER-/HER2+ subtype a true predictor of IMNI harm?
📚 Sources · 📄 1 paper
The longer read
The central worry this cohort answers is whether better systemic therapy has absorbed the survival contribution of internal mammary node irradiation. The EBCTCG meta-analysis enrolled trials initiated before 2009, and DBCG IMN1 treated patients before taxanes and CT-based planning, so a skeptic could reasonably argue that the historical benefit belongs to an era of weaker drugs and cruder fields. DBCG IMN2 undercuts that argument directly: with taxane-based chemotherapy, aromatase inhibitors and trastuzumab, the OS hazard ratio of 0.85 is essentially the same point estimate as the EBCTCG rate ratio of 0.90 and the KROG 08-06 hazard ratio of 0.87. The Korean trial's null result reads better as an underpowered trial than as contrary evidence, and here the three estimates converge.
The design deserves careful weighting in both directions. Allocation by laterality is not randomization, but it is a close approximation: tumour side is unlikely to track prognosis, and the baseline table is balanced across nearly every variable. The residual concern is that laterality is not neutral for RT harm, since left-sided patients carry cardiac dose. That confounder works against the right-sided IMNI arm only if left-sided patients are healthier, which is implausible, and the heart-disease death rates (0.2% vs 0.7%) are small against an absolute OS difference of 4.2% at 15 years. The stronger methodological point runs the other way. Median IMN coverage in the left-sided controls was 20.4%, and 10.1% of them were deliberately treated, so the comparison is partial IMNI versus substantial incidental dose, not IMNI versus none. Dilution of this kind biases toward the null, which makes the observed effect a floor rather than a ceiling.
For the radiation oncologist, the decision this moves is the 1-3 node patient, where practice varies most. The HR of 0.85 (0.73-0.97) in that group is post hoc, and should be read as consistency with the overall result rather than an independent finding. It is still meaningful that the effect did not shrink as nodal burden fell, because the usual argument for omitting IMNI in N1 disease is that the absolute gain becomes too small to justify cardiac and lung dose. Here the 15-year cardiac death rates were low even with free-breathing wide tangents, and left-sided mean heart dose was 2.3 Gy.
Two subgroup signals invite overinterpretation. The medial/central, 4+ node HR of 0.98 inverts the long-held assumption that medial tumours benefit most, and the ER-/HER2+ HR of 1.49 suggests possible harm. Neither survives a multiplicity lens, the interaction test for location was non-significant, and NSABP B-51 did not show harm in HER2+ disease, albeit post-neoadjuvant. They are reasons to watch, not reasons to omit.
What this does not settle is the neoadjuvant population, which was excluded, and the IO and ADC era, whose effect on nodal RT benefit is unknown. It would be wrong if laterality carried an unmeasured prognostic imbalance of roughly the size of the effect, which the balanced baseline makes unlikely.