APBI-IMRT-Florence Technique Update (VMAT)
ForEarly breast cancer selected for PBI, postoperative
TL;DR30 Gy/5 consecutive-fraction VMAT PBI, N=50: G1-only toxicity (acute 32%, late 16%) at median 4.5 yr f/u, no G>1 events.
The PTV-to-ipsilateral-breast volume ratio predicts ipsilateral breast V15Gy (R2 .911), so a planner can tell before optimizing whether a pt will meet the V15Gy ≤50% constraint. Consecutive daily fractions gave numerically more G1 toxicity than the trial's alternate-day schedule (acute 32% vs 21.1%, late 16% vs 4.5%), none above G1.
In early breast cancer pts selected for PBI, this supports 30 Gy/5 on consecutive days as an alternative to every-other-day delivery; it does not extend to pts whose plan misses the acceptable OAR constraints, who go to WBI.
The PTV-to-ipsilateral-breast volume ratio predicts ipsilateral breast V15Gy (R2 .911), which screens PBI candidates before planning. Consecutive daily 30 Gy/5 carried more G1 toxicity than alternate days (acute 32% vs 21.1%, late 16% vs 4.5%), all G1, so the schedule change costs little.
13 details
Single-center (University of Florence) report of 50 consecutive pts, 39 left and 11 right-sided, with dosimetric analysis and RTOG/EORTC toxicity scoring. No comparator arm. Median follow-up 4.5 yr (mean 3.4; range 0.6-5).
30 Gy in 5 consecutive fractions via partial-arc VMAT, free breathing, replacing five-field step-and-shoot IMRT on alternate days. CTV = clips + 1.5 cm, PTV = CTV + 0.5 cm, PTV up to 4 mm into lung and cropped 3 mm from skin. A plan that misses the acceptable constraints goes to WBI.
Acute any-grade G1 toxicity in 32% (dermatitis 28%), late G1 in 16% (fibrosis 14%). No G>1 event. Physician cosmesis excellent 84%, good 14%, fair 2%.
The parent APBI-IMRT-Florence phase 3 trial showed no significant IBTR or survival difference vs WBI at 10 yr, with better toxicity and cosmesis. Against that trial this series reports more G1 toxicity (acute 32% vs 21.1%, late 16% vs 4.5%). FAST-Forward 26 Gy/5 WBI reported moderate or marked normal tissue effects in 11.9% at 5 yr but has not been tested within a PBI phase 3.
Median follow-up of 4.5 yr is hard to square with VMAT adopted in the past 3 yr. Some pts had a 1-day weekend break, and the lung, heart and contralateral breast dosimetry comes from free-breathing plans at one center only.
The useful output is a planning template: the constraint set plus the volume-ratio predictor let a center screen candidates before optimizing. The authors note VMAT and inverse planning may be unavailable in low-income settings or hard to get insurance-authorized in the US, and cite planning work suggesting 3D-CRT is a feasible alternative.
| Parameter | Original protocol | Current practice |
|---|---|---|
| Delivery | 5 coplanar 6-MV step-and-shoot IMRT fields | Partial-arc VMAT, pt-specific gantry angles |
| Schedule | 30 Gy/5 fx, nonconsecutive once-daily | 30 Gy/5 fx, consecutive days |
| Structure | Original | Updated ideal | Accepted deviation |
|---|---|---|---|
| PTV | V28.5Gy = 100% · Dmax ≤105% · Dmin ≥28 Gy | V28.5Gy ≥98% · V31.5Gy ≤5% · Dmax ≤110% | V28.5Gy ≥95% · V31.5Gy ≤10% |
| Ipsilateral uninvolved breast | V15Gy ≤50% | V15Gy ≤50% | V15Gy ≤60% |
| Ipsilateral lung | V10Gy ≤20% | V10Gy ≤20% | not listed |
| Contralateral lung | V5Gy ≤10% | V5Gy ≤10% | not listed |
| Contralateral breast | Dmax ≤1 Gy | Dmax ≤1 Gy | Dmax ≤2 Gy |
| Heart | V3Gy ≤10% | V3Gy ≤10% | not listed |
| Endpoint | This series (consecutive) | Florence trial (alternate days) |
|---|---|---|
| Acute toxicity, any grade | 32% | 21.1% |
| Early late toxicity, any grade | 16% | 4.5% |
Single-center consecutive series, 50 pts, no comparator arm; toxicity and dosimetry only, no local control. Consecutive-day schedule never tested within the parent phase 3.
- Does consecutive-day delivery preserve the parent trial's 10-yr IBTR?
- Does the G1 toxicity excess vs alternate-day delivery persist with longer f/u?
- Can 3D-CRT meet the updated Florence constraints?
📚 Sources · 📄 1 paper
The longer read
This is a technique paper, not an efficacy paper, and it should be read that way. Its clinical weight comes from the parent APBI-IMRT-Florence phase 3 trial, which reported no significant difference in ipsilateral breast tumor recurrence or survival vs WBI at 10 years, with better toxicity and cosmesis in the PBI arm. This report adds the move of that schedule into current practice: partial-arc VMAT in place of five-field step-and-shoot IMRT, and consecutive daily fractions in place of the alternate-day schedule the trial used. Neither change was tested inside the trial, so the oncologic results carry over only on an assumption: that dose, volumes and constraints drive the efficacy, and the delivery method and schedule do not.
The schedule change is the part a radiation oncologist should weigh most. The trial chose alternate days as a hedge against toxicity, and the authors' own comparison suggests removing that hedge had a cost: any-grade acute toxicity of 32% here vs 21.1% in the trial, and early late toxicity of 16% vs 4.5%. Every event was grade 1, and physician-rated cosmesis was good or excellent in 98%, so the absolute cost is small. Still, the direction fits a per-day dose effect, and a 50-patient series at this follow-up cannot rule out a late fibrosis or cosmesis gap that shows up only later. The external support the authors cite for consecutive delivery is the prone Shah et al series, which mixed alternate-day and consecutive treatment, so it is supportive rather than a direct test.
The FAST-Forward comparison is where the argument is strongest, and also where it needs the most care. FAST-Forward established 26 Gy in 5 fractions as noninferior to 40 Gy in 15 for whole-breast treatment, with moderate or marked clinician-assessed normal tissue effects in 11.9% of 26 Gy pts at 5 years. The two series grade toxicity on different scales (RTOG/EORTC grade here, clinician-assessed breast effects there), so their numbers should not be compared directly. A fair reading is that both 30 Gy/5 PBI and 26 Gy/5 WBI are defensible one-week options, and that the choice between them rests on target volume and pt selection. No head-to-head toxicity result exists.
For a planning team, the most portable finding is the near-linear relationship between the PTV-to-ipsilateral-breast volume ratio and ipsilateral breast V15Gy (R2 .911). It turns the V15Gy ≤50% constraint into a screen before planning starts: a large cavity in a small breast predicts a failed plan before anyone optimizes it, and the protocol's answer to a failed plan is WBI. That makes selection explicit. It also means the favorable toxicity belongs to pts whose anatomy already met the constraint. The relationship comes from one center's contours and free-breathing VMAT plans, so a center using prone positioning, different margins or 3D-CRT should check it against its own plans first.
The report leaves several questions open. It gives no recurrence data for this cohort, so it cannot show that consecutive delivery keeps the trial's 10-year local control. It cannot show whether the small toxicity excess grows with longer follow-up. And it cannot show that the results hold outside a center with VMAT access. The authors raise that last point themselves: inverse-planned IMRT and VMAT may be unavailable or hard to get authorized, and they cite planning work in which 3D-CRT met coverage goals with somewhat worse homogeneity.