The Radiobiology of Hypofractionation

If the α/β ratio is high (e.g. 10 Gy) for tumour clonogen killing, but low (e.g. 3 Gy) for late normal tissue complications, then delivering external beam radiotherapy in a large number (20-30) of small (≈2 Gy) dose fractions should yield the highest 'therapeutic ratio'; this is demonstrated via the linear-quadratic model of cell killing. However, this 'conventional wisdom' is increasingly being challenged, partly by the success of stereotactic body radiotherapy (SBRT) or stereotactic ablative radiotherapy (SABR) extreme hypofractionation regimens of three to five large fractions for early stage non-small cell lung cancer and partly by indications that for certain tumours (prostate, breast) the α/β ratio may be of the same order or even lower than that characterising late complications. It is shown how highly conformal dose delivery combined with quasi-parallel normal tissue behaviour (n close to 1) enables 'safe' hypofractionation; this can be predicted by the (α/β)eff concept for normal tissues. Recent analyses of the clinical outcomes of non-small cell lung cancer radiotherapy covering 'conventional' hyper- to extreme hypofractionation (stereotactic ablative radiotherapy) regimens are consistent with linear-quadratic radiobiology, even at the largest fraction sizes, despite there being theoretical reasons to expect 'LQ violation' above a certain dose. Impairment of re-oxygenation between fractions and the very high (α/β) for hypoxic cells can complicate the picture regarding the analysis of clinical outcomes; it has also been suggested that vascular damage may play a role for very large dose fractions. Finally, the link between high values of (α/β)eff and normal-tissue sparing for quasi-parallel normal tissues, thereby favouring hypofractionation, may be particularly important for proton therapy, but more generally, improved conformality, achieved by whatever technique, can be translated into individualisation of both prescription dose and fraction number via the 'isotoxic' (iso-normal tissue complication probability) concept.