We may be living in an era of fast-moving molecular medicine, but historical influences exert profound, and often justified, effects on everyday practices. Radiotherapy fractionation represents a good example, since dose-limiting toxicities of new radiotherapy regimens take decades to mature, setting limits on the pace of change. Radiotherapy for carcinomas of the head and neck, bronchus, and uterine cervix has been a most important influence on radiotherapy practices as a whole. Squamous carcinomas are, on average, relatively fast-growing neoplasms best treated to high total doses (>65 Gy) delivered in multiple fractions ⩽2.0 Gy [Reference Stuschke and Thames1,Reference Overgaard, Hansen and Specht2]. In an attempt to overcome the compensatory tumour proliferation between schedules delivering five fractions per week over 7 weeks or more, six daily 2.0 Gy fractions per week or twice daily fractions of 1.8 Gy have been tested [Reference Beck-Bornholdt, Dubben and Liertz-Petersen3,Reference Bourhis, Overgaard and Audry4]. Post-operative radiotherapy for early breast cancer has also traditionally used fractions of 1.8–2.0 Gy to a total dose of around 50 Gy, albeit without a perceived need to shorten treatment time (so-called accelerated fractionation).

There is no doubt that 50 Gy in 25 fractions of 2.0 Gy is effective in the adjuvant therapy of early breast cancer, current levels of local tumour control after breast conservation surgery or mastectomy being >95% at 5 years [Reference Bentzen, Agrawal and Aird5,Reference Bentzen, Agrawal and Aird6]. However, colleagues in the UK and Canada have long believed that a lower total dose delivered in fewer, larger fractions can be at least as safe and effective as this standard schedule [Reference Patterson7,Reference Peters8]. For example, the Christie Hospital in Manchester, UK, introduced the use of 40 Gy in 15 fractions of 2.67 Gy over 3 weeks, with satisfactory results in terms of late adverse effects and local tumour control [Reference Magee, Stewart and Swindell9].

The responses of normal and malignant tissues to changes in fraction size are non-linear and are well described using a linear quadratic model [Reference Jones, Dale and Deehan10]. Reductions in total dose are needed to take account of the increasing effect of larger fractions on normal tissues. The main controversy focuses on whether a single hypofractionated regimen can be identified that is at least equivalent to 50 Gy in 25 fractions in every clinically relevant respect, including a range of late adverse effects. It is certain that some forms of hypofractionation are unsuitable for treating the axilla and supraclavicular fossa by virtue of the sensitivity of brachial plexus to fraction size. However, it appears that schedules using fractions >2.0 Gy can be identified, which are equivalent in all clinically relevant aspects of late normal tissue damage to patients undergoing breast radiotherapy. Assuming this to be the case, the only other uncertainty is whether breast cancer responds to change in fraction size in a similar way to the critical late responding normal tissues. This remains controversial in the view of some experts, given that squamous carcinomas are known to be less responsive to fraction size than the dose-limiting late normal tissue reactions (atrophy and fibrosis) [Reference Bentzen, Ruifrok and Thames11].

Several prospective randomised trials have compared 50 Gy in 25 fractions with a lower total dose delivered in fewer, larger fractions, a practice referred to as hypofractionation (see Table 1). In the recently published UK Start Trial A, 2236 women with early breast cancer at 17 centres in the UK were randomly assigned after primary surgery to receive 50 Gy in 25 fractions of 2.0 Gy (749 women) or 41.6 Gy in 13 fractions of 3.2 Gy (750 women) or 39 Gy in 13 fractions of 3.0 Gy (737 women) [Reference Bentzen, Agrawal and Aird5]. All regimens were given over 5 weeks, and women were eligible if they were aged over 18, did not have immediate surgical reconstruction, and were available for follow-up. The end-points of the trial were tumour relapse, defined as the reappearance of cancer at irradiated sites, the effect of the regimen on normal tissues, and quality of life. At a median follow-up of 5.1 years, rates of tumour relapse at 5 years were very similar in all treatment groups: 3.6% after 50 Gy, 3.5% after 41.6 Gy, and 5.2% after 39 Gy. Photographic and patient self-assessments suggested lower rates of late adverse effects after 39 Gy than after 50 Gy. The conclusions were that 41.6 Gy in 13 fractions over 5 weeks offers similar rates of tumour control and normal tissue damage as the international standard fractionation schedule of 50 Gy in 25 fractions. These results are entirely consistent with the hypothesis under test viz. that breast cancer is as sensitive to fraction size as the late reacting normal tissues (if as insensitive as squamous carcinomas, the larger reduction in total dose, from 50 to 41.6 Gy, would lead to higher local tumour relapse rates with this 13-fraction schedule).

Table 1 Summary of prospective randomised trials in early breast cancer testing radiotherapy fractions larger than 2.0 Gy.

In Start Trial B, 2215 women with early breast cancer at 23 centres in the UK were randomly assigned after primary surgery to receive 50 Gy in 25 fractions of 2·0 Gy over 5 weeks (1105 women) or 40 Gy in 15 fractions of 2.67 Gy over 3 weeks (1110 women) [Reference Bentzen, Agrawal and Aird6]. The eligibility for the trial and its measured end-points were the same as for Start Trial A. At a median follow-up of 6.0 years, the rate of tumour relapse at five years was very similar in both groups: 2.2% after 40 Gy and 3.3% after 50 Gy. Both photographic and patient self-assessments suggested lower rates of late adverse effects after 40 Gy than after 50 Gy. The conclusions were that after surgery for early breast cancer, a radiotherapy schedule delivering 40 Gy in 15 fractions over 3 weeks appears to offer local regional tumour control and rates of late normal tissue effects at least as good as the accepted international standard of 50 Gy in 25 fractions over 5 weeks. The 5-year results of a Canadian trial evaluating a 16-fraction regimen are entirely consistent with those of the START Trials [Reference Whelan, MacKenzie and Julian12].

The outcomes of >7000 women participating in well-designed randomised trials testing hypofractionation in early breast cancer are strongly supportive of the hypothesis that hypofractionation can be both safe and effective for breast cancer. Residual uncertainties focus on the period of follow-up required before comparisons of late adverse effects and local tumour control are reliable enough to change practice. Some commentators argue for longer follow-up, on the basis that the very late-onset damage to the heart, for example, may be more sensitive to changes in fraction size than other late-reacting healthy tissues [Reference Bartelink and Arriagada13]. Such concerns are likely to be misplaced, although it is true that 20+ years of follow-up will be needed to establish this. There is no evidence that the fractionation sensitivity of tissues change in the decades following radiotherapy, most systematically studied in human skin, for example [Reference Turesson14]. Meanwhile, heart exposure is undesirable whatever fractionation schedule is used, and this organ should be shielded whenever possible.

Fifteen or 16 fractions is unlikely to represent the limits of hypofractionation. The UK FAST Trial is testing two dose levels of a five-fraction regimen delivered in 5 weeks against 50 Gy in 25 fractions with 900 women in follow-up (see Table 2). Meanwhile, a pilot trial (N = 30) testing 30 Gy in five fractions over 15 days in terms of early and late adverse effects is also in follow-up phase, with favourable results at 2 years [Reference Martin, Mannino and Rostom15]. Future proposals include testing a five-fraction schedule delivered over 5 days, which we consider a realisable research objective. At the time of writing, discussion focuses on whether 50 Gy in 25 fractions or 40 Gy in 15 fractions should represent the control comparator schedule. In conclusion, evidence is building that modest degrees of hypofractionation are both safe and effective for women with early breast cancer. There appear to be no disadvantages to this approach, which is simpler for the patient and health services. Future trials evaluating shortened treatment times may reveal advantages for tumour control. Meanwhile, research into the molecular determinants of fractionation sensitivity focuses on differences in DNA repair processes in slowly and rapidly proliferating tissues. The goals of this research are two-fold: first, to identify biomarkers that predict tumour sensitivity to fraction size; and second, to identify molecular targets for interventions that selectively modulate fractionation sensitivity and enhance treatment outcome. In conclusion, after decades of resistance to hypofractionation in curative radiotherapy, this approach is gaining rapid acceptance in selected tumour types, including breast cancer.

Table 2 Schema of UK FAST Trial (N = 900; in follow-up phase).