Are 5-Year Randomized Clinical Trial Results Sufficient for Implementation of Short-Course Whole Breast Radiation Therapy?

Introduction

Ten years ago, a woman with early-stage breast cancer would routinely undergo 6 to 7 weeks of daily “conventional” whole breast radiation.¹ In May 2020, 5-year results of the FAST-Forward randomized controlled trial confirmed equivalent safety and efficacy of a 1-week whole breast regimen.² Practical considerations in the setting of Coronavirus disease 2019 led to rapid uptake of this short course, especially for older patients, to better balance the benefit of radiation with risk of infection.³ And after a recent consensus discussion, the United Kingdom (UK) announced it would adopt FAST-Forward (26 Gy in 5 fractions) for all patients undergoing whole breast radiation.⁴ But are other countries, particularly the United States, ready to change the standard of care?

This is not the first time we are being faced with this consideration. In 2002, 5-year results from the Canadian randomized trial comparing conventional whole breast radiation to a hypofractionated (3-week) course were published.⁵ It was not, however, until 2008 that the National Comprehensive Cancer Network guidelines incorporated this treatment,⁶ shortly after a confirmatory trial from the United Kingdom published 5-year results.⁷ Additionally, in September 2008, 10-year follow-up of the Canadian trial results prompted a press release in the New York Times in which physicians at major academic centers reported that “up to half [of patients] choose” the shorter option.⁸ Once 10-year follow-up from the Canadian study was published in the New England Journal of Medicine in 2010,⁹ the American Society of Radiation Oncology released formal guidelines endorsing its use.¹⁰ Nonetheless, it was not until 2018 (nearly 17 years from the initial report in 2002) that it appeared that most women in the United States were actually receiving short-course radiation for breast cancer.¹¹

Hesitation among radiation oncologists to adopt novel short-course regimens for breast cancer before 10-year follow-up is due to multiple factors, including concern for late radiation toxicity as well as late recurrence related to the indolent nature of some estrogen-receptor-driven tumors.¹² Our group hypothesized that no significant changes in either toxicity or cancer recurrence have been observed in the interval between 5- and 10-years of follow-up in large phase 3 prospective randomized trials testing short-course regimens in breast cancer. Herein, we review the evidence to facilitate an informed discussion of the potential appropriateness of changing practice based on 5 years of follow-up data, specifically in the context of adopting “ultrahypofractionated” whole breast radiation.

Toxicity

Late toxicities of breast radiation are of particular concern due to the potential for adverse cosmesis and breast fibrosis to effect quality of life. It is worth recognizing that improving the therapeutic ratio (by reducing late toxicity) was the initial impetus for conventional fractionation. Importantly, the UK trials (both UK Standardization of Breast Radiotherapy [START] and FAST-Forward) tested different regimens to probe the α/β ratio of both breast cancer control and a number of normal tissue effects. For normal tissue effects, the investigators reported ratios in the range for 3.5 to 4.7 Gy.¹³ UK START B reported similar breast induration, while breast shrinkage, telangiectasia, and edema favored the 3-week regimen; these results did not change between 5 and 10 years of follow-up (Table 1). Cosmetic photographic assessments also favored the 3-week arm, which was apparent at 2 years and persisted through 5 years, but not further reported.⁷ In the Canadian trial, clinician-assessed cosmesis was not significantly different between the 2 arms at either 5 years (fair-poor: 22% vs 21% for 3-week vs 5-week, respectively) or 10 years (30% vs 29% for 3-week vs 5-week, respectively). The UK FAST trial (5 weekly fractions) recently published both 5- and 10-year results,¹⁴ which show no significant differences, and mirror 2-year cosmesis outcomes previously published, with generally low moderate-marked normal tissue effects in the breast.¹⁵

Table 1.

Comparison of 5-year and 10-year outcomes from randomized trials comparing different fractionation schemes for whole breast radiation.

Breast RT regimens tested	Clinical trial	Primary endpoint(s)	Adverse effects		Cancer recurrence
			5-y	10-y	5-y	10-y
Conventional vs hypofrationation	Canadian	5-y IBTR	Fair-poor cosmesis (NS) skin/subcutaneous (NS)	Fair-poor cosmesis (NS); skin/subcutaneous (NS)	NS	NS
	START B	5-y local-regional recurrence	Breast appearance* (NS) induration (NS) shrinkage, telangiectasia, edema (SS favors hypofrac) breast hardness, shrinkage, swelling† (NS)	Not reported induration (NS) shrinkage, telangiectasia, edema (SS favors hypofrac) Not reported	NS	NS
Hypofractionation vs 5 d (ultra-hypofractionation)	FAST	2-y photographic breast assessment	2-y mild-marked change* (NS)	Not reported	NS	NS
			Induration, shrinkage, telangiectasia, edema (NS)	Induration, shrinkage, telangiectasia, edema (NS)
	FAST-Forward	5-y IBTR; 5-y photographic breast assessment	Breast appearance*,† (NS) Induration, shrinkage, telangiectasia, edema (NS) Breast hardness† (NS) QLQ-BR23† (NS)	N/A	NS	NS

Adverse effects are clinician-assessed unless otherwise stated. Photographic assessments (*) and patient-reported outcomes (^†) are often not available beyond 5 years.

Abbreviations: IBTR = in-breast tumor recurrence; N/A = not applicable; NS = not significant; SS = statistically significant.

^*,†Means within the same row followed by different superscript letters differ significantly (P < 0.05).

To further determine time to detection of subtle differences in late breast toxicity with adjuvant external beam breast radiation, trials of partial breast irradiation (PBI) and tumor bed boost radiation are additionally informative. The Canadian RAPID trial detected a significant difference in cosmesis between accelerated PBI and whole breast radiation by 3 years, with absolute differences appearing stable by 5 years (3-, 5-, and 7-year absolute differences of 11.3%, 16.5%, and 17.7%, respectively).¹⁶ The Florence PBI trial also reported significant differences in clinician-assessed cosmesis at both 5- and 10-year follow-up.^17,18 Of note, although RAPID and Florence made different overall conclusions about PBI, likely due to differences in dose, fractionation, and technique, neither overturned initial 5 year results at 10 years of follow-up. Finally, the European Organization for Research and Treatment of Cancer (EORTC) 22881–10882 boost trial provides multiple longitudinal reports with up to 20 years of follow-up. In this case, a significant difference in breast fibrosis was not reported until 10 years (moderate-severe: 28% boost vs 13% no boost),¹⁹ beyond which absolute differences stabilized (20-year moderate-severe: 30% boost vs 15% no boost).²⁰ Again in this trial, though, a significant difference in breast cosmesis was first detected as early as 3 years (fair-poor: 29% boost vs 14% no boost),²¹ solidifying early the overall conclusion that boost radiation would affect long-term breast toxicity.

Regarding additional late toxicities, START B noted ischemic heart disease, symptomatic lung fibrosis, and symptomatic rib fracture to be rare and no different at 5 and 10 years. Arm lymphedema, although a low risk when treating the breast alone, shows a similar pattern of no difference at 5 or 10 years in the subset of patients undergoing regional nodal irradiation on START trials.^13,22 Furthermore, the largest retrospective series confirms median time to lymphedema caused by radiation is within 4 years.²³

Cancer recurrence

When departing from a widely accepted standard of care, it is critical to confirm that oncologic efficacy is not compromised. This issue rose to prominence in trials of short-course breast radiation before the underlying radiobiologic α/β ratio was fully defined. As described above, the START trials sought to answer this question, revealing an α/β ratio of 3.5 for breast cancer locoregional relapse, and demonstrating that short-course radiation did not compromise local control.^7,24 The 10-year report was consistent with earlier results, showing no decrement in disease control in the intervening years.¹³ In the Canadian trial, 5-year locoregional recurrence-free survival was wholly consistent between 5 and 10 years. One exception was increased recurrence among patients with high-grade tumors seen on an unplanned subset analysis at 10-year follow-up. However, a subsequent study-level meta-analyses confirmed this to be a spurious result.²⁵

Although not comparing 2 fractionation schemes, the Canadian MA.20 and European EORTC 22922 are informative for evaluating locoregional recurrence at multiple timepoints.^26,27 The trials studied similar populations and were reported concurrently in 2015 with 10-year outcomes showing that, for certain patients, regional nodal irradiation improved systemic disease control and disease-free survival, broadening its use in current practice. These results were, notably, foreshadowed by 5-year reports; a late-breaking abstract at the 2011 Annual Meeting of the American Society of Clinical Oncology reported improvement in distant disease-free survival (hazard ratio = 0.64, P = .002) and disease-free survival (hazard ratio = 0.68, P = .003),²⁸ commensurate with later publication.

Discussion

There is limited evidence in the external beam breast radiation therapy literature to suggest that 5-year results stand to be materially overturned upon longer follow-up. Although breast fibrosis is a slowly progressive phenomenon, differences in breast cosmesis appear early, within 2 to 3 years, and stabilize shortly thereafter, representing an early marker of differential breast toxicity. To this end, with no significant difference in late breast toxicity by 5 years of follow-up on UK FAST-Forward, which used comprehensive assessment (including both patient-reported and blinded photographic assessment by 3 masked observers), we anticipate that additional followup is unlikely to alter the clinical calculus surrounding its implementation. Nonetheless, given trends favoring 40 Gy/15 compared with 26 Gy/5 (10.6% vs 12.2%, respectively for any breast or chest wall event), we still cannot rule out this possibility. An informed discussion with patients is therefore warranted, weighing theoretical long-term risks with potential short-term benefits of fewer treatment visits such as reduced fatigue,^29,30 trouble meeting family needs,³⁰ and financial toxicity.³¹ Additionally, although not designed for statistical testing, FAST Forward also reported acute dermatitis to be numerically lower grade and resolving sooner with the shorter regimen.³²

We recognize that the sole inclusion of major landmark studies in this report may have survivorship bias or publication bias, and the absence of a systematic review or meta-analysis is a methodological limitation. We therefore invite counterexamples whereby 5-year published results of external beam breast radiation therapy were wholly overturned by findings at 10 years of follow-up, to ensure a balanced and informed discussion.

In light of the present review, we submit that 5-year results demonstrating significant clinical utility should not be dismissed outright solely on the basis of absent 10-year outcomes, which represent a traditional benchmark. With the average interval for translation of early trial results into widespread clinical practice extending nearly 17 years,³³ it is important to explore opportunities such as this to expedite strategies to reduce the evidence to practice gap. And with 10-year results from UK FAST already available, the current scenario is not dissimilar from that seen in 2008, when many informed patients began to opt for shorter regimens.⁸ In the United Kingdom, this regimen has already become standard,³⁴ with population-based data confirming 60% of eligible patients receiving it as of June 2020.³⁵ We therefore propose consideration of a “phased implementation” strategy, whereby those patients most likely to benefit and least likely to be harmed by an intervention are the first in which the new treatment is used. A national consensus statement to this effect would support physicians in routine practice and ensure equal access to this new patient-centered treatment option.