Ten-Year Results of FAST: A Randomized Controlled Trial of 5-Fraction Whole-Breast Radiotherapy for Early Breast Cancer

PURPOSE Previous studies of hypofractionated adjuvant whole-breast radiotherapy for early breast cancer established a 15- or 16-fraction (fr) regimen as standard. The FAST Trial (CRUKE/04/015) evaluated normal tissue effects (NTE) and disease outcomes after 5-fr regimens. Ten-year results are presented. METHODS Women ≥ 50 years of age with low-risk invasive breast carcinoma (pT1-2 pN0) were randomly assigned to 50 Gy/25 fr (5 weeks) or 30 or 28.5 Gy in 5 once-weekly fr of 6.0 or 5.7 Gy. The primary end point was change in photographic breast appearance at 2 and 5 years; secondary end points were physician assessments of NTE and local tumor control. Odds ratios (ORs) from longitudinal analyses compared regimens. RESULTS A total of 915 women were recruited from 18 UK centers (2004-2007). Five-year photographs were available for 615/862 (71%) eligible patients. ORs for change in photographic breast appearance were 1.64 (95% CI, 1.08 to 2.49; P = .019) for 30 Gy and 1.10 (95% CI, 0.70 to 1.71; P = .686) for 28.5 Gy versus 50 Gy. α/β estimate for photographic end point was 2.7 Gy (95% CI, 1.5 to 3.9 Gy), giving a 5-fr schedule of 28 Gy (95% CI, 26 to 30 Gy) estimated to be isoeffective with 50 Gy/25 fr. ORs for any moderate/marked physician-assessed breast NTE (shrinkage, induration, telangiectasia, edema) were 2.12 (95% CI, 1.55 to 2.89; P < .001) for 30 Gy and 1.22 (95% CI, 0.87 to 1.72; P = .248) for 28.5 Gy versus 50 Gy. With 9.9 years median follow-up, 11 ipsilateral breast cancer events (50 Gy: 3; 30 Gy: 4; 28.5 Gy: 4) and 96 deaths (50 Gy: 30; 30 Gy: 33; 28.5 Gy: 33) have occurred. CONCLUSION At 10 years, there was no significant difference in NTE rates after 28.5 Gy/5 fr compared with 50 Gy/25 fr, but NTE were higher after 30 Gy/5 fr. Results confirm the published 3-year findings that a once-weekly 5-fr schedule of whole-breast radiotherapy can be identified that appears to be radiobiologically comparable for NTE to a conventionally fractionated regimen.


INTRODUCTION
Ten-year results of 4 randomized trials totaling . 7,000 patients confirm the safety and efficacy of hypofractionated radiotherapy after primary surgery for early breast cancer. [1][2][3][4] The UK START-B and Ontario trials established 15-and 16-fraction schedules as new standards of care delivered over 21-22 days. [5][6][7] Sensitivity to fraction size was tested in the START pilot and START-A trials by controlling for treatment time, generating an a/b estimate of 3.5 Gy (95% CI, 1.2 to 5.7) for tumor control, comparable to that for late adverse effects. 2,4,8 Fifteen-or 16-fraction regimens are unlikely to represent the clinical limits of hypofractionation, and 3-year adverse effects of 5-fraction schedules in the UK FAST trial were reported in 2011. 9 In FAST, 5.7 or 6.0 Gy once weekly were tested against 50 Gy in 25 fractions, the standard of care at the time. The explanatory trial design allowed interpolation between 2 5-fraction schedules that suggested a schedule equivalent to 50 Gy in 25 fractions in terms of late adverse effects. Five fractions of 5.7 and 6.0 Gy were predicted to be radiobiologically equivalent to 25 fractions of 2.0 Gy, assuming a/b values of 3.0 and 4.0 Gy for late normal tissue responses and tumor control, respectively. 10 At a median follow-up of 3 years, 28.5 Gy in 5 fractions was comparable to 50 Gy in 25 fractions and milder than 30 Gy in 5 fractions in terms of adverse effects in the breast. 9 This manuscript presents the 5-year results for change in photographic breast appearance and physician assessments of breast normal tissue effects (NTE) up to 10 years after radiotherapy, as well as breast cancer disease events.

Patients
FAST is a multicenter, phase III randomized controlled trial. Full details of trial design, eligibility criteria, radiotherapy planning and delivery, and study procedures have been presented previously (Protocol, online only). 9 Eligible patients were women having invasive early breast cancer age $ 50 years, pathologic tumor size , 3 cm, axillary node negative, breast-conserving surgery with complete microscopic resection, and whole-breast radiotherapy. Patients requiring mastectomy, lymphatic radiotherapy, tumor bed boost, or cytotoxic therapy were ineligible.
Patients were randomly assigned (1:1:1) to receive 50 Gy in 25 fractions of 2.0 Gy, 30 Gy in 5 once-weekly fractions of 6.0 Gy, or 28.5 Gy in 5 once-weekly fractions of 5.7 Gy. Random assignment was performed by telephone or fax from the recruiting center to the Clinical Trials and Statistics Unit, Institute of Cancer Research, London. Computergenerated random permuted blocks stratified by participating center were used. Treatment allocation was not blinded because of the nature of the intervention.
All patients provided written informed consent. FAST (CRUKE/04/015) was approved by the national South-West Multicentre Research Ethics Committee (04/MRE06/17) and the local ethics committees of participating centers. FAST was sponsored by The Institute of Cancer Research and is registered as an International Standard Randomized Controlled Trial (ISRCTN62488883).

Radiotherapy
Patients lay supine on an inclined plane in a position that remained unchanged during imaging/simulation and treatment, verified by orthogonal laser beams. Clinical target volume included soft tissues of the whole breast down to deep fascia but not including underlying muscle, ribcage, overlying skin, or excision scar. Planning target volume included the entire breast with 1-cm margins to palpable breast tissue. Medial and lateral borders did not normally extend beyond the anterior midline or the midaxilla. Margins were reduced in selected patients if the tumor bed did not encroach, to exclude or reduce the volume of heart and/or lung within the high-dose volume. The deep margin extended down to the deep fascia.
Transverse cross-sections of the patient were taken through the center of the planning target volume; a minimum of 5 slices was recommended, spaced appropriately. Sixteen out of 18 centers used full-dose compensation with computerized tomography; others used optical outlining devices capturing the central external contour supplemented by 2 additional outlines collected 1 cm inside the superior field border and 1 cm superior to the inframammary fold. 11 The maximum thickness of lung included in the tangential field was 2 cm; cardiac shielding used multileaf collimator (MLC) or other technique. The dose distribution across the target volume was modified to ensure homogeneity within ICRU50/62 guidelines. 12 Doses were prescribed to the reference point at/near the center of the target volume. Maximum and minimum doses were # 10% of doses on the central plane after full dose compensation; where full dose compensation was not possible, maximum doses in the superior plane and plane through the inframammary fold were recorded. Three main dose compensation methods were used to improve dose homogeneity: (1) physical breast compensators, (2) simple forward-planned intensity-modulated radiation therapy (IMRT) MLC segment fields/field-in-field technique, and (3) inverse-planned IMRT MLC segment fields. 13

Outcome Assessment
The primary end point was change in photographic breast appearance. Secondary end points were physician assessments of radiation-induced breast changes and ipsilateral disease in the breast (relapse or new primary).
Photographs were taken at baseline and 2 and 5 years after radiotherapy. Change in photographic breast appearance compared with the postsurgical (preradiotherapy) baseline was scored on a qualitative 3-point scale (no, mild, or marked change), on the basis of changes in size, shrinkage, and shape. Patients were ineligible for additional photographic assessments after breast reconstructive

CONTEXT Key Objective
To test the reduction in total dose of adjuvant whole-breast radiotherapy delivered by 5 once-weekly fractions needed to match the late adverse effects of a standard 25-fraction schedule. Knowledge Generated A once-weekly 5-fraction schedule of 28 Gy is estimated to be radiobiologically equivalent to 50 Gy in 25 fractions in terms of late adverse effects at 10 years of follow-up. Relevance a/b estimates for late adverse effects are consistent with historical estimates of fraction size sensitivity in patients prescribed adjuvant whole-breast radiotherapy and can be used to inform additional trials of accelerated hypofractionation. Five-year results from the UK FAST-Forward trial have confirmed the efficacy of a 5-fraction schedule delivered in 1 week in terms of local tumor control. Our findings from the FAST trial may be relevant to the needs of patients who are unable to comply with or gain access to standard 25-, 16-, 15-or 5-fraction schedules, and in whom once-weekly treatment is preferable.
surgery and after additional ipsilateral disease. All photographs were scored by at least 2 observers blind to patient identity and treatment allocation following procedures established in the START Trials 14 (Appendix Figure A1, online only). Because a number of years had elapsed since the scoring of the 2-year photographs for the previous publication, 9 these were rescored along with the 5-year photographs to ensure consistency of assessment criteria (Appendix Table A1, online only). Breast size and surgical deficit were assessed from the baseline photographs using a qualitative 3-point scale (small, medium, large), with surgical deficit expressed relative to the contralateral breast size.
Late-onset NTE in the breast (shrinkage, induration, telangiectasia, edema) were assessed by physicians at annual follow-up and graded on a 4-point scale for the treated breast relative to the contralateral breast (none, a little, quite a bit, or very much; interpreted as none, mild, moderate or marked). Incidence of symptomatic rib fracture, symptomatic lung fibrosis, and ischemic heart disease was recorded. Physicians were not blinded to randomized treatment allocation. No patient-reported outcomes were assessed within the FAST trial. Clinical assessments of acute skin toxicity have been previously reported. 9 Ipsilateral disease was defined as a malignancy (invasive or ductal carcinoma in situ) presenting anywhere in the ipsilateral breast parenchyma and/or overlying skin, whether considered ipsilateral breast relapse or new primary tumor. Data on first regional relapse (axilla, supraclavicular fossa, and internal mammary chain), distant metastases, new primary cancer, and death were also collected.

Statistical Considerations
Using START pilot trial results, 2 an average 2-year rate of mild or marked change in photographic breast appearance for the test groups of 20% was assumed, allowing a sample size of 900 to detect a 10% difference in the prevalence of change in photographic breast appearance between test dose levels with 90% power, 2-sided a 5 0.05, allowing for 10% loss to follow-up/unevaluable. The trial was not statistically powered to test for differences in local tumor control.
Scores for change in photographic breast appearance at 2 and 5 years were modeled using generalized estimating equations (GEE). 15 Mild and marked categories were combined, because marked change was rare. Pairwise comparisons of mild/marked change between schedules were described by odds ratios (ORs, with 95% CI) obtained from the GEE models and the Wald test.
Cross-sectional analyses of physician-assessed breast NTE at 5 and 10 years compared frequencies of moderate/ marked effects versus none/mild between pairs of schedules using risk ratios and risk differences (with 95% CI), and Fisher's exact test. Longitudinal analyses of moderate/marked physician-assessed NTE (v none/mild) used GEE models including all annual assessments, comparing schedules across the whole follow-up period using OR (with 95% CI) and the Wald test; a term representing years of follow-up was included, enabling time trends to be modeled. Survival analysis methods analyzed time to first moderate/marked physician-assessed NTE, including Kaplan-Meier plots and estimates of cumulative incidence rates. Hazard ratios (HRs, with 95% CI) were obtained from Cox proportional hazards regression, and schedules were compared using the log-rank test. Inconsistencies between the GEE and Cox models for some end points appeared to be due to more patients in the 28.5-Gy group having only 1 event, which has a greater influence on the time-to-event analysis (where only 1 event is needed) compared with the longitudinal models including all events over follow-up.
Kaplan-Meier estimates (with 95% CI) of 5-and 10-year cumulative incidence of ipsilateral disease in the breast were calculated, and HR (with 95% CI) compared schedules obtained from Cox proportional hazards regression, with patients censored at date of distant metastases, new primary cancer (contralateral breast or nonbreast), death, or date of last follow-up.
Five-and 10-year cumulative incidence rates of moderate/ marked NTE in the breast were higher for 30 Gy compared with 50 Gy, with statistically significant differences for any NTE in the breast, breast shrinkage, breast induration, and    breast edema (Fig 1; Appendix Table A3, online only). Cumulative incidence rates of any moderate/marked NTE in the breast and breast induration were significantly higher for 28.5 Gy versus 50 Gy.
Modeling all annual physician assessments over follow-up, rates of moderate/marked effects were statistically significantly higher for 30 Gy compared with 50 Gy (OR for any breast NTE, 2.12; 95% CI, 1.55 to 2.89; P , .001), but with no significant difference between 28.5 Gy and 50 Gy (OR, 1.22; 95% CI, 0.87 to 1.72; P 5 .248; Table 3). Statistically significant differences between the test schedules were found for breast shrinkage, telangiectasia, and breast edema, with higher rates for 30 Gy compared with 28.5 Gy. The prevalence of breast shrinkage and telangiectasia increased over time, with a decline in breast edema (Fig 1).
Change in photographic breast appearance gave an unadjusted a/b estimate of 2.7 Gy (95% CI, 1.5 to 3.9 Gy); adjusting for breast size and surgical deficit made little difference (Table 4). Using an a/b of 2.7 Gy, the isoeffect doses expressed in 2.0-Gy equivalents for 30 and 28.5 Gy in 5 fractions were approximately 56 and 51 Gy, respectively, and the once-weekly 5-fraction schedule estimated to be isoeffective with 50 Gy/25 fractions was 28 Gy (95% CI, 26 to 30 Gy). Estimates of a/b for physician-assessed NTE were consistent with the photographic end point (Table 4).
A total of 123 patients (13.4%) were referred to a specialist for radiotherapy-related adverse effects, most frequently lymphedema, with similar rates between the schedules (Appendix Table A4, online only). Symptomatic rib fracture was reported for 11 patients (1.2%), symptomatic lung fibrosis for 8 (0.9%), and ischemic heart disease for 17 (1.9%), including 7 cases in patients treated for left-sided breast cancer (Appendix Table A5, online only).

DISCUSSION
The FAST Trial tested once-weekly 5-fraction schedules of whole-breast radiotherapy in terms of late NTE against a standard regimen of 50 Gy in 25 fractions. Patient eligibility focused on factors associated with a low absolute risk of local tumor relapse, as experienced by an older patient age group with early-stage pathologically node-negative disease.
Change in photographic breast appearance was the primary end point of late NTE as in the START trials, because breast appearance after breast cancer treatment is of importance to women, and photographs allow external assessors to control for baseline surgical deficit and to score postradiotherapy changes blind to treatment allocation. 14 Marked change in photographic breast appearance in the FAST trial was rare. The low rates of change recorded in the FAST trial after 50 Gy incorporate the benefits of 3-dimensional dosimetry compared with 2-dimensional dosimetry used in the START and Ontario trials, as well as fewer women with large breast size included in FAST. Physician assessments, although not blinded to allocated treatment and hence potentially subject to bias, nevertheless provide a valuable assessment of late NTE from a different perspective to the photographs, and both sets of results contribute to the overall evidence from the trial. Annual physician assessments identified few moderate or marked effects over 10 years. The prevalence of breast shrinkage and telangiectasia increased over follow-up in FAST, as shown in other studies, 4,16 whereas breast edema declined, consistent with patient-reported outcomes of the IMPORT LOW trial of partial breast radiotherapy. 17 Incident cases of ischemic heart disease were rare, but longer follow-up is required to adequately monitor cardiac risk after breast radiotherapy.
The a/b estimates from FAST are consistent with the 10-year analysis of the START-A trial, which generated estimates around 3-4 Gy for late NTE in the breast. 4 This consistency supports the validity of the linear-quadratic model for fraction sizes as high as 5.0-6.0 Gy. However, fractionation sensitivity might be slightly higher (a/b value slightly lower) than predicted by the model because of much lower rates of moist desquamation and later consequential late skin damage when larger fractions are used. Rates of patchy/confluent moist desquamation in the FAST trial after 50.0 Gy, 30.0 Gy, and 28.5 Gy were 11.7%, 2.7%, and 2.8%, respectively (including only 1 confluent case), confirming the well-established insensitivity of early-reacting self-renewal tissues to fraction size and the importance of total dose. 9,18 The FAST trial was not powered for formal statistical comparison of local tumor control; the 10-year cumulative incidence estimate was 1.3%, in keeping with the low-risk population for which the trial was designed. The extremely low number of local tumor events reflects the patient demographics, tumor characteristics, careful attention to microscopic excision margins, the use of adjuvant endocrine therapy, and high-quality radiotherapy. Deaths from other causes were the most frequent consequential event.
The    Any NTE in the breast includes shrinkage, induration, telangiectasia, and edema. c Lower limit truncated at 0. the 26 Gy 5-fraction schedule compared with 40 Gy in 15 fractions, and is already considered standard in many UK radiotherapy departments. 19 A substudy within FAST-Forward tests the same dose schedules as the main trial in patients who also require radiotherapy to the axilla and/or supraclavicular fossa.
In conclusion, the FAST trial identifies a 5-fraction schedule estimated to be radiobiologically equivalent to the 25-fraction standard in terms of late NTE. Identification of a 5-fraction schedule equivalent with respect to tumor control is being evaluated in the UK FAST-Forward trial. Although not powered for tumor control, the FAST trial suggests that for patients at low risk of relapse and for whom daily visits over 3 or 5 weeks are not possible because of frailty or comorbidities, 28 Gy in 5 fractions as a once-weekly schedule might be an appropriate alternative to no treatment.

AFFILIATIONS SUPPORT
The funding source provided peer-reviewed approval for the trial but had no role in study design, collection, analysis, interpretation of data, or writing of the report.

Methods
Photographs were scored by a team comprising 2-3 observers (2 consultant clinical oncologists including J.R.Y., Chief Investigator of FAST Trial, and a research manager in the chief investigator's research team). Each scoring session began with a review of photographs previously scored, followed by scoring of the new photographs; sessions generally lasted for 1 day. As previously described, 12 observers conferred and agreed on a score by consensus. The same processes were followed for the 5-year photographs as for the original 2-year photograph scoring, with one change of personnel (clinical oncologist).
The categories of mild and marked change were assessed qualitatively, as it was not possible to quantitatively measure breast shrinkage from the photographs. Examples of no change and marked change in photographic breast appearance are shown in Appendix Figure A1.

Update of 2-Year Results
Three additional 2-year photographs were scored since the 2011 publication, 9 taking the total at year 2 to 732.
When the year 5 photographs were scored, it was noted that the overall prevalence of mild and marked changes was unexpectedly lower than reported at 2 years in the 2011 publication. Marked changes in particular would not be expected to reverse, except for some patients with marked breast edema. Because there was no objective measure used, such as a quantitative measurement of breast shrinkage, for example, it is considered more likely that perceptions of radiotherapyrelated changes changed over the long time period since the 2-year photographs were originally scored, causing discrepancies between the published 2-year results and the 5-year results reported here.
Hence it was decided to rescore all 2-year photographs originally scored as mild or marked change, together with a random sample of those originally scored as no change.
Overall, 472 paired scores (original and rescore) were available for year 2. The number of scores expected to agree by chance were calculated, along with the weighted k statistic to test agreement between the pairs. There were fewer mild and marked changes in the rescores at year 2 (Appendix Table A1). The number of pairs of scores expected to agree by chance were 227 (no change), 26 (mild), 0.8 (marked) (ie, observed agreement for mild and marked changes was higher than would be expected by chance; weighted k, 0.46; "moderate" agreement; SE 0.03).
In summary, although the observed agreement between the original scores and the rescores for mild and marked changes was higher than would be expected by chance, it was decided to use the rescores for all analyses presented in this article, as the level of agreement overall was only moderate. The number of patients with mild/marked change in photographic breast appearance at 2 years reported here is less than in the 2011 article, but conclusions regarding differences between the fractionation schedules are as before. 9 No. randomly assigned

(n = 301)
Did not receive allocated treatment (n = 4) Ineligible due to second primary (n = 1) Limited arm movement so standard superior border not used (n = 1) Prescribed tumor bed boost after random assignment (n = 1) Standard treatment given due to dose inhomogeneity for allocated treatment (n = 1)