Radiobiology of High Dose per Fraction

1.What the LQ model says

The linear-quadratic (LQ) model describes cell survival as a function of dose d through two terms: a linear component (αd, single-hit lethal damage) and a quadratic component (βd², accumulated sublethal damage). The α/β ratio (high (~10 Gy) for early-responding tissues and most tumors, low (~2–3 Gy) for late-responding normal tissues) captures why fractionation spares late-responding tissue: small fractions sit on the shallow part of the survival curve for low-α/β tissue. The model, and derived quantities like the biologically effective dose (BED), are the backbone of conventional radiotherapy planning.

2.Where it strains at radiosurgical doses

Most of the data underpinning the LQ model come from in-vitro survival curves at doses far below those used in radiosurgery. Extrapolated to 15–24 Gy single fractions, the LQ model's continuously bending curve tends to over-predict cell kill (and conversely can mis-estimate normal-tissue effect). Kirkpatrick and colleagues argued prominently that the LQ model is inappropriate for the high-dose-per-fraction regime, while Brown, Brenner and others countered that the LQ model remains well validated to about 10 Gy per fraction and arguably usable to about 18 Gy and that invoking new mechanisms is often unnecessary. The honest position is that this is genuinely contested: the LQ model is a useful approximation that becomes progressively less reliable as dose per fraction rises, and clinicians should treat BED comparisons across very different fractionations with caution rather than as exact equivalences.

3.Vascular and stromal injury

A defining feature of ablative single-fraction biology is endothelial and microvascular injury. Experimental work by Garcia-Barros and colleagues linked high single-dose tumor response to endothelial apoptosis through the acid sphingomyelinase/ceramide pathway, with effects becoming more visible at large dose per fraction. The finding helped explain why vascular and stromal response may contribute to AVM obliteration and tumor control, but its clinical dominance and any sharp dose threshold remain debated. The safest teaching point is that classical clonogenic killing remains central while vascular, stromal, and immune effects may add biology not fully represented by a simple LQ calculation.

4.Reoxygenation, repopulation, and the immune question

Single-fraction treatment forgoes reoxygenation between fractions, so a hypoxic subpopulation may be relatively protected: one rationale some cite for a few-fraction regimen in larger or hypoxic targets. With a single session there is no meaningful tumor repopulation during treatment, removing the accelerated-repopulation concern of protracted courses. High-dose radiation also produces immunomodulatory effects (release of tumor antigens, interferon signaling, and the much-discussed but clinically inconsistent abscopal effect); these are an active research area in combination with immunotherapy and are covered, for brain metastases, on the disease pages rather than asserted as established practice here.

5.How doses are actually chosen

Because the models are imperfect at these doses, radiosurgical prescriptions are anchored empirically. The single most cited safety anchor is RTOG 90-05, which established maximum tolerated single-fraction doses by tumor diameter: 24 Gy for ≤ 20 mm, 18 Gy for 21–30 mm, and 15 Gy for 31–40 mm, in previously irradiated patients, and demonstrated that larger targets carry steeply higher toxicity. Disease-specific marginal doses (for example, 12–13 Gy for vestibular schwannoma, or 16–25 Gy depending on AVM size and eloquence) come from outcome series, not from BED calculation. When a target is too large or too close to a critical structure to treat safely in one fraction, the response is hypofractionation (typically 3–5 fractions), which trades some of the single-fraction vascular effect for improved normal-tissue tolerance: a pragmatic, evidence-guided compromise rather than a model-derived certainty. Normal-tissue tolerances for these regimens are compiled in HyTEC and AAPM TG-101 and applied on the planning and constraints page.