International Delphi consensus: pelvic SABR re-irradiation
ForPelvic recurrence after prior pelvic RT, SABR re-irradiation candidates
TL;DR66% of statements reached consensus among 21-23 international radiation oncologists; interval since prior RT, lesion number/size and cumulative OAR limits stayed unresolved.
The transferable piece is the agreed technique: 30-37.5 Gy in 5-6 fx or 21-27 Gy in 3 fx on alternate days, IMRT, daily volumetric verification. Luminal OAR contact split the panel at 50% while direct invasion was a 90% exclusion, so contact cases remain judgment calls without a consensus fallback.
In a pt with pelvic recurrence after prior pelvic RT, no direct luminal OAR invasion, and WHO PS 2 or better, this supports SABR re-irradiation as an MDT option alongside exenteration; it does not settle the interval since prior RT or cumulative OAR limits.
The agreed technique is the transferable piece: 30-37.5 Gy in 5-6 fx or 21-27 Gy in 3 fx on alternate days, IMRT, daily volumetric verification. Luminal OAR contact split the panel at 50% versus a 90% exclusion for direct invasion, leaving contact cases without consensus backing.
9 details
Three-round international Delphi; Round 1 opened 27/10/2020, Round 3 closed 22/03/2021. 44 statements after Round 1, 29 reached consensus in Round 2, 15 re-presented in Round 3.
Panel of radiation/clinical oncologists who had published on pelvic SABR re-irradiation or were judged international experts; one per research group. 23 of 41 invited joined Round 1; median 10 cases/yr per participant (range 10-100). Subspecialty interest GU 19 (83%), lower GI 11 (48%), gyn 8 (35%).
Agreed schedules: 30-37.5 Gy in 5-6 fx or 21-27 Gy in 3 fx, alternate days. Agreement also covered IMRT-level conformity and daily online volumetric or fiducial verification. The 140% PTV point-max cap reached only 73%; 1 participant favoured 115-125%.
66% of statements reached consensus, covering definition, patient selection, set-up, delineation, planning and delivery. No consensus on minimum interval since prior RT (43% for 12 mo), maximum lesion number/size, or cumulative OAR constraints.
| Statement | Agreement | Consensus |
|---|---|---|
| SABR re-RT not usually appropriate with direct luminal OAR invasion | 90% | Yes |
| May not be appropriate with luminal OAR contact (no invasion) | 50% | No |
| Minimum 12 mo since prior pelvic RT | 43% | No |
| PTV point max dose not to exceed 140% | 73% | No |
Sits beside the ESTRO ACROP Delphi on salvage SBRT for intraprostatic relapse, which addresses a narrower prostate-only scenario. A thoracic re-irradiation Delphi hit the same wall: no agreed quantified dose overlap to define re-irradiation.
The panel is GU-dominant, so transfer to gyn and rectal recurrences rests on fewer voices. Optimal OAR constraints were borrowed from de novo AAPM TG-101 limits used cumulatively, not from re-irradiation outcome data.
The selection and technique framework is usable now. The unresolved items (interval, lesion count/size, cumulative constraints with recovery) are the ones that set the dose actually deliverable, so the core dosimetric decision remains case-by-case.
Delphi panel output: agreement percentages measure expert opinion, not outcomes. Frames practice where no randomised re-irradiation data exist; key dosimetric questions explicitly unresolved.
- Minimum safe interval between pelvic RT courses
- Cumulative OAR EQD2 limits allowing for tissue recovery
- Maximum lesion number and size for pelvic SABR re-irradiation
📚 Sources · 📄 1 paper
The longer read
This document is most useful where it agrees and most informative where it does not. The areas of agreement are largely procedural: who belongs in the MDT, how to immobilise and prepare bladder and rectum, which imaging to fuse, which OARs to contour, and that IMRT-level conformity and daily volumetric verification are expected. Almost nobody practicing pelvic SABR re-irradiation would argue with any of that, and the high agreement reflects how uncontroversial it is. The agreed fractionation range, 30-37.5 Gy in 5-6 fractions or 21-27 Gy in 3 fractions on alternate days, is the closest the paper comes to a dosimetric anchor, and it gives departments a defensible default to point to.
The failures to agree carry more information. The panel could not settle a minimum interval since prior RT (43% backed 12 months), a cap on lesion number or size, or cumulative OAR constraints. Those are exactly the parameters that decide whether a meaningful re-irradiation dose can be delivered at all. The comments on the interval statement are telling: several participants argued that previously delivered OAR dose and primary disease type matter more than elapsed time, which amounts to saying the question is posed on the wrong axis. Similarly, direct luminal OAR invasion drew 90% agreement as a contraindication, yet contact without invasion split the panel at 50%, with lower total dose, compromised PTV coverage, or close intra/inter-fraction monitoring offered as workarounds. Contact is the common clinical scenario, so the gap sits where the reader most needs guidance.
The constraint problem reflects the evidence base. The proposed optimal constraints were the AAPM TG-101 de novo 5-fraction limits applied cumulatively, and the less conservative published approaches cited in Table 4 come from small retrospective series: Abusaris et al reported no grade 3+ toxicity at median 15 months follow-up, and Smith et al one grade 3 pain event at median 24.5 months. Short follow-up in re-irradiated bowel and sacral plexus is precisely where late events would hide, so the reluctance to endorse higher cumulative limits with assumed tissue recovery is defensible rather than timid.
Panel composition should temper how a reader applies this outside prostate. Most participants listed a GU interest, fewer lower GI, fewer still gyn, and the organisers were UK-based. A gyn reader facing a sidewall recurrence after prior brachytherapy is working in territory the panel covered thinly, and the brachytherapy overlap statement did not settle either. Response was also modest, with 23 of 41 invited joining Round 1, so the panel is a self-selected subset of an already selected expert pool.
What would make this obsolete is prospective outcome data with recorded cumulative OAR doses, which the panel itself endorsed collecting. Until then, the practical read is that the technique framework and patient-selection gates are a reasonable floor, and the cumulative constraint and interval decisions remain a per-case dosimetric judgment that no consensus currently backs.