Organs at Risk Radiation Dose Constraints: 2025 Update
TL;DRSFRO 2025 OAR dose-constraint update: normofractionated, moderate hypofractionated and ablative tables across skull base, H&N, lung, breast, oesophagus, liver.
START BFAST-FORWARDRTCMIENDOMETRE
The practical shift is coverage of ablative regimens per fraction number (1, 3, 4, 5, 8) alongside normofractionation, plus cardiac substructures (LAD, left ventricle) and a proposed dorsal vagal complex OAR. Numbers are tiered optimal vs mandatory, so a plan that exceeds an optimal value is a documented conversation, not a violation.
Practical shift is the per-fraction-number ablative columns (1, 3, 4, 5, 8) sitting alongside normofractionation in one grid, plus cardiac substructures (LAD, left ventricle) as constrained organs. The optimal versus mandatory tiering is what changes planning behaviour: it names which values are negotiable against target coverage.
12 details
SFRO consensus update, part of the RecoRad 4th edition, covering adult external beam radiotherapy and brachytherapy practice. Constraints were compiled from international consensus guidelines and clinical trial data, then validated by national reference experts rather than derived from a new dataset.
Tables are indexed by tumour site and by fractionation: normofractionation, moderate hypofractionation, accelerated hyperfractionation (lung), and ablative regimens split by fraction number (1, 3, 4, 5, 8). Dmax is approximated by D2 % in normofractionation and D0.035 cm3 in stereotactic plans; several hypofractionated entries are given as EQD2.
No clinical endpoint. The deliverable is a set of dose-volume objectives intended for direct implementation into treatment planning systems as standardized clinical objectives.
Constraints are tiered as optimal versus mandatory, with ALARA used where no numeric threshold is offered (parotid in the brain table, lacrymal gland under moderate hypofractionation). The authors require that pts be informed of the risk when a constraint is exceeded.
Breast constraints are anchored to the schedules of START B (40 Gy in 15) and FAST-FORWARD (26 Gy in 5); liver constraints are built on the ASTRO clinical practice guideline across 3, 5 and ≥ 20 fractions. This is a national harmonization layer over those sources, not a competing evidence synthesis.
The authors state most constraints still rest on 3D conformal-era data, which sits awkwardly with IMRT's larger low-dose bath and with ablative dose gradients. Biological modifiers of risk (individual radiosensitivity, genomic profile, concurrent immunotherapy) are not accounted for, and dose to immune organs at risk is named as unresolved.
The value here is harmonization, not novelty: the stated problem is the multiplicity of circulating constraint sets and the absence of a single reference. The optimal/mandatory tiering is the part that changes daily behaviour, because it tells a planner which values are negotiable against target coverage and which are not.
- Validity of 3D conformal-derived constraints under IMRT and SBRT
- Dose thresholds for immune organs at risk with concurrent immunotherapy
- Dorsal vagal complex as an OAR for radiation-induced nausea
📚 Sources · 📄 1 paper
The longer read
The honest way to read this document is as an act of standardization rather than of evidence generation, and that is its main claim on a reader's attention. Anyone planning an ablative case has watched constraint tables proliferate: one number from the trial protocol, another from AAPM TG-101, another from the local physics group, another remembered from residency. The SFRO's answer is a single indexed grid, site by site and fraction number by fraction number, that a physics team can load into the planning system as clinical objectives. That is a real contribution even though it settles no clinical question, and it should be judged on internal consistency and usability rather than on whether any individual threshold is correct.
The authors themselves supply the most important caveat, and it deserves more weight than a reader skimming the tables will give it. Most constraints trace back to three-dimensional conformal data. IMRT redistributes dose rather than removing it, trading high-dose exposure to a small volume for low-dose exposure to a larger one, and a constraint calibrated in the conformal era encodes a dose distribution that no longer exists. The problem compounds in stereotactic regimens, where the gradient is steeper still and where the linear-quadratic model that underwrites the EQD2 conversions is on its least secure footing. The document uses EQD2 freely for the hypofractionated columns, which is standard practice and also the step where the most uncertainty is silently absorbed.
Two structural choices are worth flagging because they change how the tables should be used. The optimal versus mandatory tiering is the more consequential one. A brainstem Dmax of 54 Gy labelled optimal against a surface Dmax of 60 Gy labelled mandatory tells a planner precisely where the negotiation with target coverage happens, which is the information a bare threshold withholds. The second is the appearance of cardiac substructures, the left anterior descending artery and the left ventricle, as constrained organs in their own right. That reflects a genuine shift in the breast and oesophageal literature toward substructure dosimetry, and it moves the contouring workload before it moves any outcome. The proposed dorsal vagal complex sits a tier below this, offered as a hypothesis about radiation-induced nausea rather than a validated constraint.
What the document does not attempt is any grading of the evidence behind individual numbers, and the authors say so explicitly. A constraint drawn from a randomized trial's protocol and one drawn from a single retrospective series appear in the same cell format with the same authority. For a reader who wants to know how hard to fight for a particular value, that flattening is the central limitation, and it is why this belongs on the planning workstation rather than in a debate about whether a specific threshold is right. The related gap is biological: individual radiosensitivity, genomic profile and concurrent systemic therapy all modify risk, and none of them enter a dose-volume histogram. The authors name dose to immune organs at risk as the open question that combined radiotherapy and immunotherapy will force, which is a fair read of where the next round of constraints has to come from.