Opening statements
Previous trials, particularly the UK START trials, have confirmed the feasibility and benefits of hypofractionation such that 40 Gy in 15 fractions was adopted as the UK standard schedule in 200950 and subsequently internationally102 but there is no reason why the concept of hypofractionation should be limited to 15 fractions. The UK FAST trial103 was planned in the early 2000s when 50 Gy in 25 fractions over 5 weeks was considered standard. The FAST trial compared 28.5 Gy in 5 fractions of 5.7 Gy or 30 Gy in 5 fractions of 6 Gy with 50 Gy in 25 fractions, with the 5-fraction regimens delivered once weekly over 5 weeks to reduce overall time as a variable. The primary endpoint was photographic assessment at 2 and 5 years. A low-risk patient population was deliberately chosen, excluding patients requiring a breast boost or systemic chemotherapy for instance. Two test groups were used, as in the START-pilot and START-A trials, to allow interpolation to estimate an isoeffective dose to control if necessary. The FAST trial identified a 5-fraction schedule estimated to be radiobiologically equivalent to the 25-fraction standard in terms of late NTE.
The 5-year results from the START trials,17,18 with the change in standard UK fractionation to 40 Gy/15 fractions (NICE 2009), the first results of the FAST trial,103 plus the acute toxicity and 2-year AE from a pilot study delivering 30 Gy in 5 fractions over 15 days26 gave impetus to the investigation of 1-week 5-fraction schedules in a phase III study. With these results the FAST-Forward Trial was developed to test two 5-fraction schedules over a week against the standard of 40 Gy in 15 fractions over 3 weeks, specifically to be non-inferior for local relapse and without an increase in late AE.
FAST-Forward local relapse
FAST-Forward is a non-inferiority trial with IBTR or local relapse as the primary endpoint. Based on START trial data18 and incorporating subsequent improvements in surgical technique and systemic therapy, it was anticipated that 5-year IBTR rate in the FAST-Forward control group given 40 Gy in 15 fractions would be 2%. Applying a non-inferiority margin of 1.6% at 5 years required a sample size of 4000 patients. This provided 80% power (one-sided α of 0·025 allowing for non-inferiority hypothesis and a simple Bonferroni correction taking into account comparisons between each test schedule and the control group) to exclude an absolute increase of 1·6% in 5-year IBTR incidence for a 5-fraction schedule compared with control. Estimated 5-year IBTR after 40 Gy in 15 fractions was 2.1% (95% CI 1.4 to 3.1), after 27 Gy was 1.7% (1.2 to 2.6) and after 26 Gy was 1.4% (0.2 to 2.2). The upper confidence limits (CI) excluded an increase in IBTR of > 1.6% after both 5-fraction schedules with p = 0.0022 and p = 0.00019 for non-inferiority of 27 Gy and 26 Gy schedules, respectively compared with 40 Gy in 15 fractions.
Therefore, it can be said emphatically that non-inferiority in terms of IBTR of 5-fraction schedules compared with 40 Gy in 15 fractions was demonstrated at 5 years’ follow-up for patients with early breast cancer, the majority of whom were treated by local tumour excision and sentinel node biopsy for node-negative disease. The lack of a detectable dose–response for local tumour control between 26 Gy and 27 Gy in 5 fractions is a potential limit to precision, but this feature reflects the shallowness of the dose–response curve for subclinical breast cancer around the 98% control level, so the −0.4% estimated difference in absolute levels of IBTR between 27 Gy and 26 Gy likely reflects random sampling variability in the IBTR rate and/or chance imbalances in unmeasured prognostic factors between test groups.
Longer follow-up is well recognised as having added value for RT trials. IBTR continues to present beyond 5 years, but trials of hypofractionation have so far reported almost identical HRs for IBTR between test and control groups at 5 and 10 years, the relevant metric for comparisons of effect. For example, START-B18 incidence rates of IBTR at 5/10 years after 50 Gy in 25 fractions versus 40 Gy in 15 fractions were 3.3%/5.2% and 1.9%/3.8%, respectively, reflecting crude HR of 0.72 (95% CI 0.43 to 1.21) and 0.70 (95% CI 0.46 to 1.07) at the respective time points that is barely changed between 5 and 10 years. The OCOG reported IBTR rates at 5/10 years after 50 Gy in 25 fractions versus 42.5 Gy in 16 fractions of 3.2%/6.7% and 2.8%/6.2%, respectively, giving an absolute difference of 0.4% (95% CI −1.5% to 2.4%) and 0.5% (95% CI −2.5% to 3.5%), respectively.104 This is reassuring but the 10-year IBTR rates will be reported when the data matures to address any remaining concerns over efficacy of five daily fractions.
Application of FAST-Forward to partial breast radiotherapy
IMPORT LOW tested 40 Gy/15 fractions over 3 weeks PBI compared with 40 Gy/15 fractions over 3 weeks to whole breast (control).12 This is the only PBI RCT worldwide where irradiated volume is the only variable and all other factors including dose/fractionation are constant. The reduced volume of breast tissue treated with PBI was associated with non-inferior local control and reduced late normal tissue toxicity. Given that we know that reduced volume (PBI) results in reduced late normal tissue toxicity for a constant dose/fractionation as per IMPORT LOW, then it is logical that 26 Gy/5 fractions is also adopted for PBI given that both trials used the same dose/fractionation for control. It would be counter-intuitive to state that 26 Gy/5 fractions can be used for whole breast RT with excellent local control and minimal toxicity, but not for PBI given that IMPORT LOW has shown (1) reduced volume is non-inferior for local control, (2) reduced volume has less toxicity, that is it would be illogical (and possibly unethical) to have a greater treatment burden for the lower risk group of patients eligible for PBI.
Use of hypofractionation in the post-mastectomy setting
Wang et al.105 reported on 810 patients in a single institution randomised non-inferiority trial of moderate hypofractionated RT post-mastectomy. All patients underwent axillary dissection and were at least four-node positive or T3–4, unless they received neoadjuvant chemotherapy in which case either clinical stage III or pathological axillary node-positive patients were eligible. The hypofractionated schedule was 43.5 Gy in 15 fractions over 3 weeks versus 50 Gy in 25 fractions over 5 weeks as standard. The RT target volume included the chest wall, level 3 axilla and SCF. The 5-year locoregional relapse rate was 8.3% (90% CI 5.8% to 10.7%) with the 15-fraction schedule and 8.1% (90% CI 5.4% to 10.6%) with the 25-fraction schedule. With a p < 0.0001 for non-inferiority they concluded that the hypofractionated regimen was non-inferior to standard.
In the FAST-Forward Main Trial there were 264 patients who had had a mastectomy and only one of these had a local relapse. One hundred and seventy-three patients received chest wall only RT with one of the 5-fraction schedules with no local relapse reported. While this is too few events to make any firm statements the data raise no concerns regarding the efficacy or safety of the 5-fraction schedules. There is no logical reason why a 5-fraction schedule would not be as effective when irradiating the chest wall as opposed to the intact breast.
Breast cancer subgroups and local relapse
Commentators have suggested that hypofractionation may be less applicable in certain patient and pathological subgroups. The initial 2011 American Society for Radiation Oncology (ASTRO) evidence-based guidelines106 were unable to confirm agreement on hypofractionation in patients < 50 years. The small number of patients and the increased risk of IBTR at young age were cited, but no evidence of an adverse outcome by age after hypofractionation has been reported. A post hoc analysis of tumour grade in the OCOG trial104 suggested an interaction of grade and randomisation group, but subsequent central analysis of tumour blocks reported no trend for patients with high-grade tumours to be disadvantaged after 42.5 Gy in 16 fractions104 as compared to 50 Gy in 25 fractions. OCOG also found that tumour grade and molecular subtype did not predict response to hypofractionation. A subgroup meta-analysis of locoregional relapse was performed of the START-P, START-A and START-B trials in 5861 patients reporting 10-year results.107 Treatment effects of hypofractionation were not significantly different from 50 Gy in 25 fractions when examined by age, type of primary surgery, axillary node status, tumour grade, use of adjuvant chemotherapy or boost RT. The HR for locoregional control and the patient numbers for the combined hypofractionation schedules against 50 Gy in 25 fractions were 0.79 (95% CI 0.47 to 1.34) in 343 patients aged under 40 years, 0.88 (95% CI 0.60 to 1.28) in 1046 patients aged 40–49 years, 0.86 (95% CI 0.59 to 1.25) in 1272 patients with grade 3 tumours, 0.91 (95% CI 0.59 to 1.25) in 513 patients following mastectomy and 0.81 (95% CI 0.46 to 1.81) in 1480 patients after adjuvant chemotherapy. The 2018 ASTRO evidence-based guidelines subsequently approved hypofractionated breast RT with 40/42.5 Gy in 15/16 fractions over 3 weeks irrespective of age, tumour grade or receptor status.102
FAST-Forward reported non-inferior IBTR for both 5-fraction schedules and given the preceding arguments inferiority in any subgroups would not be expected. One thousand and five forty-five (37.8%) patients were in the high-risk category as defined by age < 50 years (604 patients), grade 3 tumour (1153 patients) or both, and this was a stratification factor at randomisation. Formal subgroup analysis was not performed due to the low number of IBTR events but descriptive data according to age, grade, and receptor status showed no signals of concern for the 5-fraction schedules. The use of boost and dose/fractionation, both declared prior to randomisation, were balanced between the three treatment groups so there is no risk of bias in dose intensity between trial groups.
Disease-free survival and overall survival
The EBCTCG meta-analysis reports 10-year relapse and 15-year mortality effect of RT after breast-conserving surgery.6 After breast-conserving surgery the results from 10,801 women showed that 15-year all-cause mortality was 3.0% (95% CI 0.6% to 5.4%; 2p = 0.03) better following RT; it was only 0.8% after 5 years. The trials included in the meta-analysis came from a different era when local relapse rates and risk of metastases were much higher than is currently the case. Thus, one would not anticipate any differences according to RT schedule after 5 years in FAST-Forward. Incidence of distant relapse, disease-free survival and overall survival were similar between the fractionation schedules, with no evidence of statistically significant differences between groups.
Normal tissue effects
For late NTE the dose–response curve is much steeper than for IBTR, enabling detection of clinically and statistically significant differences in event rates between 26 Gy and 27 Gy in 5 fractions. The 5-fraction schedule isotoxic for late NTE with 40 Gy in 15 fractions allows direct estimation of α/β for late NTE, which is consistent with values generated from our other trials. The α/β value of 3.7 Gy (0.3–7.1) for tumour control in FAST-Forward is similar to 3.5 Gy (1.2–5.7) estimated from the START pilot and START-A trials.108 Point estimates of α/β (assuming no effect of time) for late NTE in FAST-Forward scored by clinicians, patients and photographic assessments are closer to 2 Gy than the 3 Gy estimated in the earlier START and FAST trials, but 95% CI overlap for each endpoint in all trials. In FAST, 915 women were randomised after BCS for node-negative disease to 50 Gy in 25 fractions versus two dose levels of a 5-fraction regimen delivered once weekly, thereby ensuring complete repair between fractions and controlling for overall treatment time. The α/β value for change in photographic breast appearance in FAST was 2.6 Gy (1.4–3.7). There is uncertainty about biological processes, such as possible influence of a time factor in FAST-Forward. However, this does not interfere with clinical evaluation and thus decisions on implementation of FAST-Forward results in comparable patient groups.
In FAST-Forward, late NTE assessed by clinicians, patients and photographs were key secondary endpoints. On the basis of the similarity between NTEs for the 26 Gy in five daily fractions schedule and for the 40 Gy in 15 fractions schedule, 26 Gy in 5 fractions is the regimen recommended for clinical implementation. By ‘similar’ we mean that NTE were not statistically or clinically significantly different from 40 Gy in 15 fractions with respect to clinician- or patient-assessed NTE including external photographic assessments conducted blind to treatment allocation. In contrast, the 27 Gy 5-fraction schedule was significantly different to the 40 Gy standard for many late NTE and also to the 26 Gy schedule, confirming the sensitivity of trial outcome measures to a difference in dose intensity corresponding to 3 Gy in 2 Gy equivalents assuming α/β = 2 Gy (see below). The 27 Gy 5-fraction regimen exhibited late NTE rates of comparable magnitude to 50 Gy in 25 daily fractions, defended for so long as standard and still in use by many practitioners outside the UK. To provide some perspective for the late NTE after 5-fraction regimens, 40 Gy in 15 fractions is equivalent to about 46 Gy in 2 Gy fractions in terms of late NTE compared to 50 Gy in 25 fractions according to START trial outcomes.108 In FAST-Forward, the 26 Gy regimen is comparable to 47 Gy in 2 Gy equivalents in terms of late NTE.
Early NTE are much less responsive to fraction size than late NTE, the contribution of total dose to early NTE being relatively more important. FAST-Forward offers a good example in that breast erythema was less intense and also settled a fortnight earlier after 5-fraction than 15-fraction schedules.3 In this context, the milder erythema was due to 26 and 27 Gy total dose levels much more than to fraction sizes of 5.2 Gy and 5.4 Gy. Acute reactions were also milder in both 5-fraction schedules (total doses 27 and 30 Gy) of the FAST trial than the 50 Gy schedule.26
Induration is a key late NTE that is expected to increase with the passage of time irrespective of radiation schedule. Other factors contributing to breast appearance include fat necrosis and oedema, particularly in the early years. Appendix 2, Table 56 reproduces results already presented in Chapter 5, but shows clinical and patient assessments together, for the longitudinal FAST-Forward moderate or marked assessments that equate to fibrosis (including clinician-assessed distortion, shrinkage and induration outside the tumour bed) as well as patient-assessed breast appearance change (smaller and harder or firmer). The ORs do not tell the whole story, the absolute level of events also being very important. For example, for breast shrinkage, the most frequent of any clinician-assessed effects, the rate at 5 years was 5.1% in the 40 Gy and 6.3% in the 26 Gy schedule. Also, it should be noted that, for all clinician-assessed events documented in the moderate/marked categories, most were moderate rather than marked in severity. For the patient assessments, in which there was a baseline post-surgery but pre-radiotherapy assessment, the number of longitudinal events is generally not much higher than that post-surgery baseline. This is likely to be the same event rather than a distinct event. These late effect results are very important in terms of giving patients realistic expectations. With regard to increasing late NTE with time, stability of the HR at later time points is clinically relevant as discussed earlier for IBTR. In START-B the breast shrinkage HR for test to control arm was 0.83 (95% CI 0.66 to 1.04) at 5 years and 0.80 (95% CI 0.67 to 0.96) at 10 years. For 5 fractions we are now able to investigate the FAST 5- and 10-year results noting the same principle but with confidence limits which are wider given the smaller number of patients compared to START.23 For breast shrinkage the test to control arm HR at 5 and 10 years, respectively was 2.03 (95% CI 1.15 to 3.58) and 1.83 (95% CI 0.88 to 3.81) for the 30 Gy schedule and 1.20 (95% CI 0.63 to 2.27) and 1.83 (95% CI 0.88 to 3.81) for 28.5 Gy schedule. For ‘any NTE’ the test to control arm at 5 and 10 years, respectively was 2.40 (95% CI 1.45 to 3.97) and 2.03 (95% CI 1.06 to 3.89) for the 30 Gy schedule and 1.32 (95% CI 0.75 to 2.34) and 1.61 (95% CI 0.81 to 3.18) for 28.5 Gy schedule. The principle of the relative difference between test and control arm changing little with time can therefore be applied to FAST-Forward, again noting the low absolute levels of marked and moderate events.
The Danish-led HYPO trial of 1864 patients was designed to determine whether 40 Gy in 15 fractions would not increase in duration of the breast at 3 years compared to 50 Gy in 25 fractions. Results confirmed no increase in induration with moderate hypofractionation.109 Breast size was a stratification at randomisation and they reported that induration at 3 years was significantly increased for large compared with small-breasted patients but that at no time point was the hypofractionated schedule worse than control for this NTE. Shaitelman et al.110 conducted a randomised non-inferiority trial due to low uptake of hypofractionated whole breast RT in the United States, the low uptake being attributed to concerns of safety for patients receiving a tumour bed boost, chemotherapy or having large breast size. The standard arm was 50 Gy in 25 daily fractions with a 10–14 Gy boost in 5–7 fractions and hypofractionation was 42.56 Gy in 16 daily fractions with a 10–12.5 Gy boost in 4–5 fractions. Primarily they investigated the proportion of patients with adverse patient-reported cosmetic outcome at 3 years. One hundred and six patients (36.9%) of the 287 patients were defined as large breast size with a bra cup size of at least D. Adverse patient-reported cosmetic outcome was 5.4% lower (8.2% versus 13.6%, p = 0.002 for inferiority) in the hypofractionated group overall and 18.6% lower (90% upper confidence limit 8% lower) for large breasted patients. They conclude that this offers strong reassurance for hypofractionation not compromising cosmetic outcome based on large breast size.
Tsang et al.111 looked at dose heterogeneity with regard to the FAST trial and the risk of ‘triple trouble’.112 Three hundred and ninety full CT-planning data sets were reviewed for patients where there was a baseline and 2-year photographic assessment, the primary endpoint of FAST. The two 5-fraction groups were combined for analysis and there was no significant difference between these and control for breast volume or for patient tumour and treatment characteristics from the whole FAST population. Multiple logistic regression analyses show that after adjusting for breast size (and surgical deficit) there was no evidence of late NTE associated with dose inhomogeneity using various definitions of hotspots. The effect of inhomogeneity was not significantly different for any of the dosimetric parameters between control and 5-fraction schedules. In FAST-Forward the α/β estimate for any clinician-assessed moderate or marked NTE was barely different unadjusted or when adjusted for breast size, using whole breast planning treatment volume as the proxy for breast size. The same lack of difference was found with photographic assessment and breast size. We can conclude that breast size is an established factor for increased NTE following breast RT but that hypofractionation, including 5-fraction schedules, is not an additional concern for larger-breasted patients.
What about other organs at risk? The heart is often mentioned and, given that very long follow-up is required to detect late cardiac events, is cited as a reason to wait for longer-term data before adopting hypofractionation as standard of care. However, there is no specific reason to expect an increased cardiac sensitivity to hypofractionation. Darby et al. have shown that there is no safe dose to the heart and therefore the effort is to reduce or eliminate cardiac dose.113 At this early stage, after imaging and further investigation, excluding cases confirmed not to be radiotherapy-related, for left-sided RT there are six cases of ischaemic heart disease in the 40 Gy arm and three cases in the 26 Gy arm.
The most frequent specialist referral we have seen is to lymphoedema clinics for breast lymphoedema, 90 (6.6%) following 40 Gy and 106 (7.7%) after 26 Gy. Breast oedema is predominantly an early side effect which we have seen settling such that at 5 years the moderate/marked incidence on clinician-assessment is nine (0.7%) patients after 40 Gy and 17 (1.7%) patients after 26 Gy with none at all in 94% and 93%, respectively. These rates are low and not significantly different. These low rates are also encouraging for patients.
Tumour control and radiobiology
In a review of the linear-quadratic model and implications for practice, Brand and Yarnold present FAST-Forward as an example of a trial evaluating of 5-fraction hypofractionated accelerated RT.114 To make sense of FAST-Forward in terms of fraction size effects, the START trials offer a good entry point. The START-P/-A trials (1986–2003) each compared 50 Gy in 25 fractions over 5 weeks (control) with two test dose levels of a 13-fraction regimen over 5 weeks (5 fractions per fortnight). By controlling for time-related effects, repopulation being the likeliest mechanism in tumours, and assuming complete repair of sub-lethal damage between fractions in all groups, an unconfounded estimate of sensitivity to fraction size is possible. This simply involves identifying the total dose in 13 fractions matching the IBTR rate in the control group by interpolation between test dose levels if needed. The START-A test dose level of 41.6 Gy in 13 fractions of 3.2 Gy was the closest, from which the α/β estimate of 3.5 Gy (95% CI 1.2 to 5.7) was derived based on a 10-year total of 349 IBTR events in 3646 women. The reduction from 50 Gy to 41.6 Gy (around 8 Gy) total dose needed to match antitumour effects of 13 fractions of 3.2 Gy and 25 fractions of 2.0 Gy is a vivid example of increasing fraction size sensitivity at play.
To our knowledge, START-P/-A generated the only direct clinical estimate of α/β for a cancer, others being based on non-randomised or randomised comparisons that do not control for one or more variables, especially time. START-B is a good example of the latter, testing 50 Gy in 25 fractions over 5 weeks against 40 Gy in 15 fractions of 2.7 Gy over 3 weeks. Applying the α/β = 3.5 generated by START-P/-A, the equivalent total dose in 2.0 Gy fractions (EQD2/3.5) of the 3-week schedule is only 45 Gy, yet based on 95 IBTR events in 2215 patients (4.3%), the test schedule was non-inferior to 50 Gy (HR 0.77, 95% CI 0.51 to 1.16; p = 0.21). In fact, the point estimate 10-year IBTR rate of the 3-week regimen was about 1% superior 50 Gy.
A post hoc analysis asked the question ‘If this difference is real, what would it tell us about the impact of treatment time?’ We know that in laryngeal carcinomas at least 0.5 Gy/day can be ‘wasted’ compensating for accelerated repopulation from the fourth week of treatment onwards, first described by Withers in patients treated with primary RT and confirmed by Overgaard et al. in a randomised clinical trial comparing 60 Gy in 30 fractions delivered five versus six times per week.115,116 Breast cancers have relatively low mitotic rates at presentation, but they might be in an accelerated phase of repopulation by the time RT starts several weeks or months after primary surgery and/or cytotoxic chemotherapies. In the context of the START-B result, the post hoc analysis estimated 0.6 Gy/day (95% CI 0.1 to 1.8; p = 0.02) ‘wasted’ dose in control group patients during weeks 4 and 5. If true, it implies roughly 14 × 0.6 = 8 Gy of the control regimen (50 Gy) ‘wasted’, leaving an EQD2/3.5 of 42 Gy against 45 Gy for the 3-week schedule. The HYPO trial offers an independent test of START-B regimens in a comparable group of patients, in whom the 9-year risk of locoregional relapse was 3.3% (95% CI 2.0 to 5.0) in the 50 Gy in 25 fractions group compared to 3.0% (95% CI 1.9 to 4.5) in the 40 Gy in 15 fractions group (risk difference 20.3%, 95% CI 22.3 to 1.7),8 a result almost identical to START-B.
What have we seen in FAST-Forward? The trial generated an α/β estimate for IBTR of 3.7 Gy (95% CI 0.3 to 7.1), the CI reflecting very low IBTR rates. The analysis plan did not incorporate a hypothetical time correction, so the α/β estimate of 3.7 Gy necessarily incorporates a putative time effect by contributing to the efficacy of the shorter test regimens. Regardless of whether or not there is a time effect, the clinically effective EQD2/3.7 of 26 Gy in 5 fractions is 41 Gy in 2 Gy fractions, see Table 3. The difference between this EQD and the 45 Gy of 40 Gy in 15 fractions at such high levels of local control may be just too small to detect clinically, but the most important clinical conclusion that can be robustly drawn is that the 5-fraction regimen has demonstrated non-inferiority in relation to the pre-defined ≤ 1.6% excess IBTR boundary set in the protocol. Questions have been raised whether 26 Gy in 5 fractions has any antitumour effect at all.109 With 5-year IBTR rate of 2.1% (95% CI 1.4% to 3.1%) after 40 Gy in 15 fractions, the rate would be expected to be about 6% at 5 years and perhaps 10% at 10 years without any RT according to systematic overviews of RT effects.6 The observed 5-year IBTR rates after 26 Gy in 5 fractions are hardly consistent with an absence of effect.
To take the 0.6 Gy/day ‘wasted’ dose further a hypothesis could be generated with regard to 26 Gy in 5 daily fractions. If the ‘wasted dose’ continues down to a week, 40 Gy in 15 fractions is 14 days longer and an EQD2/3.7 of 37 Gy is calculated as opposed to an EQD2/3.7 of 41 Gy for 26 Gy in 5 fractions. The same calculation for 50 Gy in 25 daily fractions gives an EQD2/3.7 of 33 Gy. These are all hypothesis-generating.
Normal tissue effects and radiobiology
Turning to late AE, meticulous data generated in human skin are consistent with minimal measurable effect of time on probability of late NTE; Turesson reported a tiny time effect for telangiectasia associated with complete absence of mitotic figures in capillary endothelium on serial skin biopsies over many weeks of RT, the lack of mitoses excluding repopulation as a mechanism.66 The effect was thought more likely to represent a very slow component of repair decaying with a T1/2 of around 40 days.66 The same post hoc analysis of time in START-B described above examined late AE as a negative control, yielding an estimate of 0.14 Gy/day (95% CI −0.09 to 0.34 Gy/day; p = 0.29) for change in photographic breast appearance, consistent with the absence of a time effect for AE between 3 and 5 weeks of breast RT.117
The reason for going into this amount of detail is that FAST-Forward test dose levels of 27 and 26 Gy assumed firstly no clinically significant time effect for late AE between 1 and 3 weeks, secondly complete sublethal damage repair between fractions and thirdly an α/β of 2.8 Gy for late NTE, the last assumption based on the combined estimates of α/β in START-A and FAST. On this basis, the relative EQD of the FAST-Forward schedules to 50 Gy in 25 fractions were calculated (see Table 31, Chapter 5).
Although the 5-fraction regimen predicted to be isoeffective with 40 Gy in 15 fractions was 27 Gy the observed iso-effect at 5 years was closer to 26 Gy, suggesting a slightly lower α/β value, see Table 3. The α/β point estimates are all around 2 Gy, corresponding to EQD2/2 of 46.8 Gy for 26 Gy in 5 fractions compared to EQD2/2.8 = 45.6 Gy for 40 Gy in 15 fractions, the latter using α/β = 2.8, a combined estimate based on START-A and FAST. The CIs of α/β values generated for a wide range of AE in the FAST-Forward Trial are all consistent with the α/β values of all late AE generated by FAST and START-P/-A. One interpretation, and statistically speaking the most likely, is that they are all internally consistent with each other, as reflected by wide CIs associated with individual point estimates. The implication is that for mean α/β values for specific late AE endpoints are the same across all trials and fraction sizes in the range 2.0–6.0 Gy. A combined estimate based on all trials would likely be somewhere between 2.0 Gy and 2.5 Gy.
In conclusion, the slightly reduced α/β values in FAST and FAST-Forward relative to START-P/-A might be real, in which case, one potential explanation includes a time effect related to a slow component of repair. These somewhat esoteric considerations should not obscure the all-important clinical conclusion that 26 Gy in five daily fractions offers patients comparable AE rates and non-inferior IBT to 40 Gy in 15 fractions.
Resource implications
The economic analysis demonstrated that 5-fraction RT is expected to be cost-effective. When compared to 15 fractions, 5 fraction RT is expected to result in a reduction in direct treatment costs from £2315 to £982 per person treated. Total costs are also expected to reduce from £31,640 to £29,638, where these resource savings are primarily due to reductions in direct treatment costs. The implications of these resource savings for the health system are complex and depend on how the extra capacity is used. The same number of individuals may be treated with lower RT capacity, freeing up these resources for other uses. The additional capacity may also be used to drive quality, for instance increasing the use of cardiac sparing breath hold technique where that is not fully utilised for breast cancer patients. The capacity and technical advances may be utilised for cancers other than breast cancer.
COVID-19
Dissemination of the Main Trial primary results was expedited due to the COVID-19 pandemic. In March 2020 emergency guidelines for RT were published by the NICE118 and the RCR created an online resource of clinical advisory documents containing risk mitigation approaches for RT.119 The aim was to reduce footfall in RT departments and hospitals to reduce person-to-person contact to minimise the spread of COVID-19. The breast cancer RCR guidance cited FAST-Forward 3-year NTE outcomes, recommending the 26 Gy/5 fractions schedule for adjuvant treatment of early breast cancer, stating that the 5-year local relapse publication was imminent.51 This schedule was rapidly adopted nationwide, since FAST-Forward involved most UK RT centres which therefore had the required quality assurance measures already in place. From the National Radiotherapy Dataset, 60.6% of breast RT courses used 26 Gy/5 fractions in April 2020 versus only 0.2% in April 2019.96
Breast consensus/guidelines and moving forward
The UK held a breast fractionation consensus meeting under the auspices of the Royal College of Radiologists and included an in-depth look at the FAST-Forward data. Lewis et al.120 explain the Royal College of Radiologists consensus statement program and the explicit methodology that is used. There was very strong support to offer 26 Gy in 5 fractions over 1 week for whole breast, chest wall and partial breast RT. Where there had been a mastectomy and breast reconstruction there was strong support to consider 26 Gy in 5 fractions over a week. Using the 10-year FAST data there was very strong support to consider 28.5 Gy in 5 fractions over 5 weeks instead of 26 Gy in 5 fractions over 1 week for patients with significant co-morbidities and/or frailty that make daily RT difficult. For patients requiring a boost there was strong support to offer 26 Gy in 5 fractions whole breast RT plus either a sequential normofractionated boost or a hypofractionated boost (delivered in no more than 5 fractions RCR Consensus Guidance, 2016) or 15 fraction simultaneous integrated boost (SIB), for example 48 Gy to boost volume and 40 Gy to rest of breast all over 3 weeks.
European Society for Radiotherapy and Oncology (ESTRO) Advisory Committee in Radiation Oncology Practice (ACROP) consensus recommendations on patient selection and dose/fractionation for external beam radiation therapy in early breast cancer were developed due to a lack of agreement persisting in the radiation oncology community on radiation dose/fractionation/volume selection in early-stage breast cancer. These were published in 2022.121 Most of the evidence was on moderate hypofractionation (15–16 fraction regimens) but the FAST-Forward Trial on ‘ultra-(5-fraction-) hypofractionation’ was included. Concentrating on the statements on 26 Gy in 5 fractions in a week there was consensus that whole breast or chest wall irradiation without breast reconstruction can be offered as (1) standard-of-care or (2) within a randomised controlled trial or prospective registration cohort. There was strong consensus that moderate hypofractionation (40 Gy in 15 fractions) and ultra-hypofractionation (26–30 Gy in 5 fractions) represent acceptable schedules for external beam partial breast irradiation.
The St Gallen 2021 International consensus guidelines, produced bi-annually, cover local and systemic therapies for early breast cancer.122 Regarding 5-fractions, FAST and FAST-Forward schedules, the panel did not endorse these as standard treatment as yet but stated that there was growing interest.122 The National Comprehensive Cancer Network clinical practice guidelines in oncology (NCCN guidelines123), probably the most used in the USA, recommend either 25–28 fractions or preferably hypofractionation with 15/16 fractions for whole breast RT. For the chest wall 25–28 fractions are recommended. The current guidelines thus show a mixed picture with the 5-year FAST-Forward results not enough to produce universal change.
Global health
Breast cancer is the commonest cancer worldwide, both in terms of incidence and prevalence and is the leading cause of death from any cancer in women with around 2.3 million cases and 700,000 deaths in 2020.124 It is predicted that by 2040 there will be around 3 million cases annually with the majority occurring in low- and middle-income countries (LMICs).125 In addition, 70% of the predicted 1 million deaths will also be in lower-income settings if there is no intervention in the current mortality trends.125
Radiotherapy is an important component of the multimodality treatment of breast cancer and contributes to improving both local control and survival.126,127 However, RT infrastructure is severely under resourced in many low- and middle-income countries with some regions having no access to RT at all.128 Compounding this is the relative lack of uptake of hypofractionated RT in LMICs compared with higher-income countries.129 Furthermore, these regions also have very high rates of patients who do not initiate or abandon breast cancer treatment for a variety of reasons.130
Five-fraction RT over 1 week would go some way to helping this disparity. Mature results of the FAST-Forward Nodal Sub-Study will be particularly relevant, given that many people in LMICs present with more advanced disease. This shortened fractionation would help alleviate health system costs and limit the potential financial toxicity for patients who would have received 3–5 weeks of RT. Access to and completion of good quality multimodality treatment, including RT would prevent premature breast cancer deaths currently seen in many LMICs.130
Other trials and areas of future research and collaboration
FAST-Forward provides the only published randomised controlled trial evidence for 1-week 5 fractions schedules. There are several other studies either in set-up or open to recruitment evaluating the 26 Gy 1-week schedule in early breast cancer; effectively almost a decade behind FAST-Forward. However, it will be many years until any results of these are known, and FAST-Forward has more patients enrolled than all of the other studies combined. The Indian HYPORT-Adjuvant trial opened in 2019 with a recruitment target ~2100 (NCT03788213) and randomises patients requiring adjuvant RT to the breast/chest wall ± regional lymph nodes to 40 Gy/15 fraction over 3 weeks versus 26 Gy/5 fractions in 1 week. The use of a SIB (dose of 8 Gy and 6 Gy, respectively) is allowed in patients who have undergone breast conservation. Another Indian trial HYPART; NCT04472845) is testing 26 Gy/5 fractions in 1 week versus 34 Gy/10 fractions in 2 weeks in high-risk patients; the trial opened in March 2020 with a target sample size of 1018. The Canadian RHEAL trial (NCT04228991) has a similar design to the FAST-Forward Nodal Sub-Study; the target sample size is 588 patients and the estimated trial completion date is end of 2027. There are four other studies involving between 36 and 100 patients investigating the 26 Gy/5 fractions schedule that are currently recruiting and will have no long-term data for many years. The UK is therefore best placed to maximise the knowledge already gained from the testing of 5 fractions schedules.
FAST-Forward will contribute to future collaborations such as the planned EBCTCG meta-analysis breast RT fractionation trials. Another potential collaboration is the CONFLUENCE project. This National Cancer Institute (USA) funded initiative aims to use genome-wide association studies (GWAS) to improve knowledge of breast cancer genetics. This research resource of different races/ethnicities will include at least 300,000 cases of both breast cancer and controls. The confluence of existing GWAS and new genome-wide genotyping data generated through this project will achieve this aim.
Limitations of the trial
The boost policy for each patient (yes or no) and dose prescription (10 Gy in five daily fractions or 16 Gy in eight daily fractions) had to be declared to the CTSU ahead of randomisation. In the planning stage of the trial we debated the sequential normofraction boost against a hypofractionated boost. There was logic in using a hypofractionated boost for a trial of hypofractionation. To limit variables it was decided to continue with the established schedules for boost at the time. Boost is therefore treated as a randomly distributed variable that impacts (randomly) across all groups and contributes, like excision volume, breast size, smoking and genetic factors, to likelihood of adverse events. The reason for retaining sequential boost was to avoid having to commit to synchronous boost dose levels considered most likely to be isoeffective with control before knowing the outcome of the trial. That involves making an assumption about the likely α/β value (testing two values to cover upper and lower estimates would have made things very complex and prone to error). The limitation is that we used standard-of-care 5- or 8-fraction boost schedules in the Main Trial and not hypofractionated boosts. The boost policy was changed via an amendment for the Nodal Sub-Study to allow a hypofractionated boost in no more than five fractions. The trial will not contribute to information on a hypofractionated boost.
Confirmation that the 5 fractions schedule is as safe as 15 fractions for patients also receiving axillary RT is more important than ever as a higher proportion of node-positive patients who require axillary treatment are now having nodal RT rather than surgery than when the FAST-Forward Nodal Sub-Study was conceived. This shift in practice followed publication of the AMAROS study which showed non-inferiority of nodal RT to axillary nodal dissection, with halved lymphoedema rates in the irradiated group.30 Contemporary practice means there is also increasing use of neoadjuvant chemotherapy in patients with bulky or involved axillary nodes at presentation, especially in the HER-2-positive and triple-negative breast cancer sub-types, where high complete pathological responses are observed. The need for subsequent axillary surgery in these patients is being questioned, and is diminishing, being replaced by axillary RT or no treatment (NIHR ATNEC trial; NCT04109079). Lymphoedema (highlighted in a Living With and Beyond Cancer (LWBC) priority setting workshop) is the focus of the FAST-Forward Nodal Sub-Study which provides the best opportunity to quantify these effects and further funding will supplement the shorter current follow-up currently available since it started midway through the Main Trial.
For some, the long-term results are essential to change practice, long-term meaning at least 10 years. While the START trial’s 5-year outcomes were published in 2008 and 10-year outcomes in 2013, international guidelines were slow to adopt the 15 fractions schedule that has been the UK standard of care since 2009. US guidelines were updated to encompass broader adoption of hypofractionation for breast cancer only in 2018,102 and a number of countries still use the historic standard 25 fractions schedule. Lack of long-term safety data is often cited as a reason for remaining cautious regarding adoption of new standards of RT treatment. To understand this caution in adoption it is worth emphasising what long-term follow-up means for RT trials, and the continued relevance and importance of reporting on outcomes, up to and even beyond 10 years. In the START-B trial of 15 fractions versus 25 fractions moderate/marked breast shrinkage increased from 11.4% at 5 years to 26.2% at 10 years after 40 Gy/15 fractions. Outcomes for local relapse were 1.9% and 3.8%, respectively. Similar increases in prevalence of outcomes after 5 years were seen in other trials.104,109 Other trials have continued to report outcomes to 15 years.131,132
There is caution regarding effectiveness of the 5 fractions schedule in higher-risk patients, as mentioned above.133,134 There are no data to support this concern but the lack of events per biological subgroup at 5 years is the argument used. Again, as stated above there are concerns over late toxicities emerging with longer-term follow-up, such as cardiac and respiratory toxicity, that are not obviously apparent now. Some are concerned that early adoption may lead to long-term irreversible changes for patients and are sticking to their current 15–25 fractions schedules. Particular emphasis has been placed on statistically rather than clinically significant values, such as a 1–2% increase in breast hardening outside the tumour bed with the 1-week schedule. We expect the cumulative effects of long-term toxicity and relapses to go up over time, but do not expect the relative differences between the groups to change. However, we need to demonstrate this.
Long-term follow-up of patients in cancer trials is recognised as essential to understanding recovery from treatment and its long-term effects and limitations as well as developing optimal management strategies for patients. Patients have shown high compliance with long-term follow-up questionnaires, and this will provide important additional data in the planned 10-year outcomes study. This is particularly pertinent for RT where AE continue to occur and progress years after treatment completion. Trial follow-up is essential in order to measure long-term outcomes since in the UK most breast cancer patients are now discharged from routine clinical follow-up within 12 months. Long-term safety data from FAST-Forward, only available with continued trial follow-up, will be invaluable for future patients in order to support fully informed consent and shared decision-making regarding treatment. The UK health data systems do not routinely collect the type of data that is required to assess breast-related radiation toxicity. Even serious toxicity such as lung/heart damage and second malignancies, would be difficult to routinely and accurately attribute in a comprehensive and timely manner. In the earliest studies of breast RT the benefits of reducing relapse were at least counterbalanced if not outweighed by long-term excess of deaths from cardiac causes. Although this is avoided or reduced by heart-sparing techniques in the modern era, we need to be sure that there is no excess of these rare events with the 5 fractions schedule. Although relapses and deaths could be collected from the UK databases, they are not always complete, potentially indistinguishable from new primary breast cancer (contralateral) and are subject to a time lag in availability. With relapse rates so low, it is important to collect as accurately and comprehensively as possible all of these occurrences. National breast RT consent forms developed by the RCR state expected rates of long-term risks after 15 fractions and will need to be updated for 5 fractions in future.119
Survivorship issues are especially relevant for breast cancer; an estimated 600,000 people are alive in the UK after a diagnosis of breast cancer currently, predicted to rise to 1.2 million by 2030.135 Given the increasing incidence and low local relapse rates with increasing survival rates in breast cancer the issue of survivorship is magnified. If 50% of patients undergo RT and a side effect was 10% more prevalent then by 2030 60,000 people will be affected. In the recent debate held at the San Antonio breast conference (December 2021), it was argued by some that in the long term, if the 5 fractions regimen led to more rare but expensive complications this could be detrimental. The FAST-Forward Trialists do not believe this will be the case, but it is vital to be able to evidence this with long-term data.
Evaluation of progressive long-term and late side effects occurring many years after cancer treatment are highlighted in the James Lind Alliance Top 10 priorities for LWBC.136 Data collection beyond 5 years in FAST-Forward is the only available source to provide robust estimates of these effects for 5 fractions schedules contributing significantly to worldwide knowledge in the field of breast RT extreme hypofractionation. Collection of PRO finished at 5 years in our previous UK trials; the planned collection of 10-year data in FAST-Forward will enable a unique comparison of patient and clinician assessments of AE and wider survivorship issues at this time point. Reports published since START-B have illustrated that this is particularly important given that patients and clinicians have different perspectives of symptoms with patients’ ratings of symptomology consistently higher than ratings by clinicians.137
Future plans
The National Institute for Health and Care Research Health Technology Assessment Programme (NIHR HTA) has approved new funding for follow-up in the Main Trial to 10 years, and to at least 5 years in the Nodal Sub-Study. This proposal contained details for crucial long-term follow-up of breast cancer patients and was developed by the TMG including patient views via a strong PPI partnership. In Q2 2021 hospitals were asked to share a flyer with FAST-Forward patients giving them the chance to work with the trial team to optimise the future plans for the trial. Given that participants were on annual follow-up it was reassuring to have 146 responses within a month. Due to the larger number of people responding than had been anticipated a two-stage consultation was established, firstly to complete an online survey and secondly the establishment of focus groups to gauge participant opinion. In spite of a time lag in sending out the survey due to the need to get appropriate approvals, 82 responses were received (56% response rate). Of these, 39 also attended one of two focus group meetings held in Q4 2021. Survey respondents appeared to be representative for both geographical region and treatment allocation 58/82 (71%) having received hypofractionated treatment (expected 2/3rds).
Participants were unanimous in endorsing the continuation of follow-up to 10 years and largely supportive of the pragmatic approach to reduce burden of data collection at sites during year 8 and 9 while focussing on getting complete data from face-to-face clinical visits at year 10. In the focus groups the emerging themes were the reassurance provided by the annual visits and the recognition that side effects continue to emerge in later years. Some felt that they had signed up for 10 years and that this should be honoured if possible. NHS resource constraints were also noted.
They also confirmed their support for collecting patient views on their health and treatment at 10 years using a PRO questionnaire. Opinions were sought on the aspects of quality of life, toxicity and well-being that should be included. Concerns about continued changes in the treated breast, pain and physical functioning were highlighted but also fear of relapse, psychological wellbeing, fatigue and impact of other life events and comorbidities (previously reported in.138 The views expressed led the trial team to seek out a variation to the original questionnaires to extend data collection to cover survivorship issues and to offer the PRO assessment to all patients not just those who had previously been in the PRO sub-study.
Forty-six Main Trial sites opted to participate in the PRO sub-study, which was offered to all of their eligible patients (1796 consented to PRO). All Nodal Sub-Study patients were asked to consent to PRO collection since the primary endpoint is patient-assessed arm/hand swelling. We propose to offer all Main Trial patients still in follow-up involvement in the PRO study at 10 years, as this would greatly increase the number evaluable for PRO assessment. The trial provided a unique opportunity to capture extremely valuable PRO assessments of the long-term effects of living with and beyond breast cancer and capturing longer-term outcomes of hypofractionated RT. In reality due to timings this assessment would be collected between follow-up years 10 and 11. Given the chronic nature of RT-related effects this is considered a pragmatic solution and is endorsed by our participant involvement groups which had been retained as an ongoing patient representative collaborative resource.
Our participant focus groups have already demonstrated the huge value of engaging with, listening to and actively involving patients with lived experience of breast cancer. The focus groups may assist with other PPI issues that arise during the trial as the focus group members are keen to continue working with us.
Research recommendations
Future research with regard to hypofractionation of 26 Gy/5 fractions in a week would include confirmation of our results in other geographic populations. The HYPORT-adjuvant trial is open to recruitment in India (NCT03788213). This is a two-arm multicentre RCT aiming to recruit 2100 patients. The control arm is 40 Gy/15 fractions over 3 weeks and is being compared to 26 Gy/5 fractions in a week. The inclusion criteria are described as ‘non-restrictive’ to enable the results to be applicable to an Indian population. The OCOG group in Canada is recruiting to the RHEAL trial of locoregional RT also with a control arm of 40 Gy/15 fractions over 3 weeks compared to 26 Gy/5 fractions in a week (ClinicalTrials.gov Identifier: NCT04228991). There are other trials in the planning phase.
The appropriate boost dose, ideally delivered using a SIB, needs to be determined. The HYPORT trial is using either a sequential or SIB boost for both the trial and test arms of the study. This is a topic that would be appropriate for further study to build on both the IMPORT High and FAST-Forward Trials. For instance the 15-fraction SIB dose of 48 Gy/15 fractions from IMPORT High as the standard schedule could be compared with two 5FR SIB doses over a week.
The FAST-Forward Nodal Sub-Study will provide information on axillary RT in 5 fractions. Evidence on the safety of a hypofractionated approach in patients requiring IMC RT also needs to be gathered. There will be some data from centres that used 5 fractions at the height of the COVID-19 pandemic but trial-based evidence is required. Both the HYPORT and RHEAL trials include patients having IMC RT.
Is 5-fraction hypofractionation the limit for external beam RT, might two fractions be an option? We are unaware of any trials examining this question.
Equality, diversity and inclusion for study participants
The study recruited from nearly 100 centres (RT and referring centres) across the whole of the UK. The inclusion criteria were very broad to allow easy access to the study regardless of where the patient lived. The study was attractive to many patients because of the usual reasons to help with research, but also because of the 2/3 chance of a 1-week RT schedule, thus completing treatment more quickly and with less expense and incursion. The trial included all the subtypes of breast cancer, which were broadly represented as per the incidence that they occur in the UK. The age of patients in FAST-Forward ranged from 25 to 90 years with 16.4% aged 70+ years. Men were eligible although being a rare disease in men, only 12 were entered. Ethnicity data has been collected (for those who donated a blood sample for genetic research) and will be reported with the 10-year results.
The study is in the follow-up stage, so no more patients will be recruited. However, extension of the PRO sub-study to all available Main Trial participants will provide a broader, all-encompassing view of patient outcomes at 10 years.
Patient and public involvement
Patient advocate involvement has been integral to ICR-CTSU breast RT trials for over 25 years as demonstrated by the involvement of the RAGE consortium in the START trials. Such involvement is from trial concept to protocol review and membership of the TMG, the patient advocates are a constant guide to ensure that the trialists are acting on patients’ best interests and answering the questions they want answered. Patient advocates have also been integrally involved in discussions at TMG meetings regarding the trade-off between toxicity and efficacy and in advising on content of PRO questionnaires. Other essential discussions with patient advocates have included differentiating between what is statistically significant with what is clinically relevant to patients. Issues such as, does an increase incidence of 1% of breast hardening at 5 years matter to patients, with a trade-off of a week’s treatment with generally low toxicity?
In November 2021 around 80 FAST-Forward Trial participants completed an online questionnaire seeking their feedback on plans for further follow-up in the trial including content of a PRO questionnaire at 10 years. Over 40 of these patients then participated in participant involvement group discussions organised by ICR-CTSU in December 2021 that further explored patients’ views on long-term follow-up in the trial and issues important to them at 10 years following cancer treatment. Outputs from the online survey and the discussion meetings have informed the future plans of FAST-Forward.
Conclusions
Extremely low rates of IBTR and of clinically assessed moderate/marked late NTE can be attributed to improvements in all diagnostic and treatment modalities and to the commitment of patients to early diagnosis and randomised trials. The 26 Gy dose level has been shown to be comparable to 40 Gy in 15 fractions in terms of patient-assessed NTE, clinician-assessed NTE, and NTE photographic change in breast appearance, and to NTE expected after 46–48 Gy in 2 Gy fractions.
Beyond its safety and effectiveness, the 26 Gy FAST-Forward schedule is convenient and less expensive for patients and for health services. The 26 Gy FAST-Forward schedule reduces the estimated health-care fiscal cost of breast RT by over 50%. The 5-fraction regimen reduces the machine time required for breast RT patients, thus improving patient access for other groups of cancer patients within the NHS.
The reduced number of patient attendances in a 5-fraction schedule places less burden on patients and the carers. In some geographically large countries, where travel times to RT centres are prohibitive, fewer women choose breast conservation for early breast cancer. Reducing the travel and personal financial burden may allow more women to choose breast conservation rather than mastectomy.
In response to global warming, it is important that healthcare systems respond in an environmentally conscious manner. Reducing the number of patient visits to only 5 fractions will reduce the carbon footprint for patient travel. Furthermore, the impact of a 5-fraction regimen on improving access times for other cancer patients will lead to more efficient use of linear accelerators and hence reduce carbon footprint for each RT centre.
It is also likely to be safe for patients requiring regional RT, an approach currently under formal evaluation in our randomised FAST-Forward Nodal Sub-Study comparing 40 Gy in 15 fractions and 26 Gy in 5 fractions. Areas of future research with regard to hypofractionation of 5 fractions would include confirmation of our results in other geographic populations, trials are both ongoing and in set-up. The appropriate boost dose, ideally delivered using a SIB, needs to be determined. Evidence on the safety of a hypofractionated approach in patients requiring IMC RT is being trialled. Is 5-fraction hypofractionation the limit for external beam RT, might two fractions be an option? FAST-Forward can and is influencing both practice and future research.