5-Year Results of a Prospective Phase 2 Trial Evaluating 3-Week Hypofractionated Whole Breast Radiation Therapy Inclusive of a Sequential Boost

Abstract

Purpose:

To report 5-year outcomes of a phase 2 trial of hypofractionated whole breast irradiation (HF-WBI) completed in 3 weeks, inclusive of a sequential boost.

Methods and Materials:

Women with stage 0-IIIA breast cancer (ductal carcinoma in situ through T2N2a) were enrolled on a prospective, phase 2 trial of accelerated HF-WBI. We delivered a whole breast dose of 36.63 Gy in 11 fractions of 3.33 Gy, with an equivalent dose to the regional nodes when indicated, followed by a tumor bed boost of 13.32 Gy in 4 fractions of 3.33 Gy over a total of 15 treatment days. The primary endpoint was locoregional control; secondary endpoints included acute/late toxicity and physician-assessed and patient-reported breast cosmesis.

Results:

Between 2009 and 2017, we enrolled 150 patients, of whom 146 received the protocol treatment. Median age was 54 years (range, 33-82) and median follow-up was 62 months. Patients with higher-risk disease comprised 59% of the cohort, including features such as young age (33% ≤50 years), positive nodes (13%), triple-negative disease (11%), and treatment with regional nodal irradiation (11%) and/or neoadjuvant/adjuvant chemotherapy (36%). Five-year estimated locoregional and distant control were 97.7% (95% confidence interval [CI], 93.0%-99.3%) and 97.9% (95% CI, 93.6%-99.3%), respectively. Five-year breast cancer–specific and overall survival were 99.2% (95% CI, 94.6%-99.9%) and 97.3% (95% CI, 91.9%-99.1%), respectively. Acute/late grade 2 and 3 toxicities were observed in 30%/10% and 1%/3% of patients, respectively. There were no grade 4 or 5 toxicities. Physicians assessed breast cosmesis as good or excellent in 95% of patients; 85% of patients self-reported slight to no difference between the treated and untreated breast.

Conclusions:

Our phase 2 trial offers one of the shortest courses of HF-WBI; at 5 years of follow-up there continues to be excellent locoregional control and low toxicity with favorable cosmetic outcomes in a heterogeneous cohort of patients.

Introduction

Four large randomized phase 3 trials including over 7000 patients have reported long-term outcomes over the past decade and established hypofractionated whole breast irradiation (HF-WBI) as the new standard of care for early-stage breast cancer.^1–3 The Standardisation of Breast Radiotherapy (START) B and Ontario Cooperative Oncology Group (OCOG) trials found that HF-WBI fractionation schemes of 40 Gy in 15 fractions and 42.5 Gy in 16 fractions, respectively, resulted in equivalent tumor control and similar or improved breast cosmesis and late toxicity compared with conventionally fractionated whole breast irradiation (CF-WBI) regimens of 50 Gy in 25 fractions.^1,2 When the results of these trials are considered together with reduced costs and increased patient convenience,^4,5 HF-WBI is the clearly superior and cost-effective treatment choice. Both HF-WBI fractionation schemes were recently endorsed by the American Society for Radiation Oncology (ASTRO) as preferred dose regimens for patients not requiring regional nodal irradiation (RNI).⁶

However, although tumor bed boost has demonstrable local control benefit,^7,8 the HF-WBI trials did not incorporate standardized boosts—the START B trial allowed for an elective, nonhypofractionated boost of 10 Gy in 5 fractions,¹ and the OCOG trial did not allow any boost.² ASTRO consensus guidelines recommend including a boost in young or high-risk patients and omitting a boost in elderly patients with low-risk cancers, but otherwise guidelines suggest individualized decision making in the majority of women.⁶ In clinical practice, it is common to add a sequential boost of 10 to 16 Gy in 4 to 8 fractions after HF-WBI, resulting in total treatment courses that can be 4 weeks or longer. In a large registry study, 81% of patients undergoing HF-WBI received a boost, with the majority receiving 10 Gy in 4 or 5 fractions.⁹

It has also been acknowledged that these landmark trials, although practice changing, are unlikely to have tested the limits of safe and effective hypofractionation.¹⁰ In the START pilot trial, which was designed to test fractionation sensitivity, the α/β ratio was estimated to be 4.0 Gy for tumor control³ and 2.3 to 5.1 Gy for various endpoints of late breast toxicity,¹¹ in agreement with previous clinical and experimental data.^12–16 The similarity of these α/β estimates suggests that dose fractionation does not preferentially spare normal tissue over tumor; therefore, using even larger fraction sizes should result in shorter but equally safe and effective treatment schedules.

In 2009, we initiated a prospective phase 2 trial investigating a more condensed HF-WBI dose regimen completed in 3 weeks inclusive of a sequential boost. Furthermore, we designed the trial to include a heterogenous cohort of women, including those with more advanced disease requiring RNI, and sought to use modern breast techniques including field-in-field, 3-dimensional conformal radiation treatment, and deep inspiration breath hold, which had not been used in the previous HF-WBI trials. Previous report of an initial cohort showed excellent rates of tumor control, toxicity, and breast cosmesis.¹⁷ In this planned expansion cohort, we sought to evaluate 5-year outcomes in a larger number of patients with high-risk disease.

Methods and Materials

Study design and patient eligibility

We conducted a prospective, single-arm phase 2 trial of accelerated HF-WBI delivering a whole breast dose of 36.63 Gy in 11 daily fractions of 3.33 Gy, with an equivalent dose to the regional nodes if indicated, followed by a mandatory tumor bed boost of 13.32 Gy in 4 daily fractions of 3.33 Gy over a total of 15 treatment days. Detailed methods of the trial have been previously reported.¹⁷ In brief, women ≥18 years old with American Joint Committee on Cancer 7th edition pathologic stage 0 ductal carcinoma in situ (DCIS; Tis) or stage IA-IIIA invasive breast cancer (T0-2, N0-N2a, M0) who underwent lumpectomy with negative margins (no tumor on ink) were eligible. Axillary staging with axillary dissection or sentinel node biopsy was required for invasive disease; neoadjuvant and/or adjuvant systemic therapy were allowed at the discretion of medical oncology. The trial is registered with ClinicalTrials.gov (NCT00909909). The study and informed written consent were approved by our institutional review board.

Radiobiological rationale of dose regimen

The linear quadratic model¹⁸ was used to ensure that the proposed fractionation scheme would result in equivalent probabilities of tumor control and toxicity compared with conventionally fractionated regimens. Using an α/β ratio of 4.0 Gy for tumor control,³ the equivalent dose in 2 Gy fractions (EQD2) of the whole breast dose (36.63 Gy in 11 fractions) is 44.7 Gy and the EQD2 of the boost dose (13.32 Gy in 4 fractions) is 16.3 Gy, for a total tumoricidal EQD2 of 61.0 Gy. Using a range of α/β ratios of 2.0 to 11.0 Gy to further model the acute and late fraction sensitivities of normal tissue gives whole breast EQD2 values of 40.4 to 48.8 Gy.^11,14 Altogether, these EQD2s are lower than conventional doses for normal tissue responses while providing equivalent tumor dose, and they are comparable to those of the hypofractionated regimens used in the START B and OCOG trials (Table 1).

Table 1.

EQD2

		EQD2 (Gy)
Normal tissue responses	α/β ratio (Gy)	Conventional 50 Gy/25 fractions	START B 40 Gy/15 fractions	OCOG 42.5 Gy/16 fractions	Current study 36.63 Gy/11 fractions
Fibrosis	2.0	50	46.8	49.6	48.8
Contracture	3.5	50	44.9	47.7	45.5
Telangiectasia	4.0	50	44.5	47.2	44.7
Desquamation	11.0	50	42.1	44.7	40.4
Locoregional tumor control	α/β ratio (Gy)	WBI + boost 10 Gy/5 fractions	WBI + boost 10 Gy/5 fractions*	WBI + boost 10 Gy/4 fractions*	WBI + boost 13.32 Gy/4 fractions
Tumor	4.0	60	54.5	58.1	61.0

Abbreviations: EQD2 = equivalent dose in 2 Gy fractions; WBI = whole breast irradiation.

^*Boost of 10 Gy in 5 fractions is commonly delivered when using Standardisation of Breast Radiotherapy (START) B hypofractionated regimen, although elective on the trial. Boost of 10 Gy in 4 fractions is commonly delivered when using Ontario Cooperative Oncology Group (OCOG) hypofractionated regimen, although not allowed on the trial.

Radiation treatment planning and techniques

Simulation with computed tomography was mandatory and done in either the supine or prone position at the treating physician’s discretion. Radiation treatment began within 21 to 63 days of last surgical intervention or last cycle of chemotherapy. The breast target volume was defined clinically with radiopaque markers and/or contoured to create a breast clinical target volume. Standard whole-breast tangential fields were used, and dose to nonbreast structures was limited according to standard of care. On beams-eye views, the tangent beams could not include more than 3 cm of lung at any level. The heart was required to be excluded from the primary beam, generally using deep inspiration breath hold in left-sided cases and/or multileaf collimator blocking as appropriate. RNI was administered with a standard supraclavicular field and posterior axillary boost if needed. Nodal volumes including the axillary and supraclavicular nodes were contoured with or without the internal mammary nodes as deemed clinically appropriate. The tumor bed was visualized on computed tomography and contoured with a 1- to 2-cm margin to create a boost planning target volume.

All patients were treated with forward-planned, field-in field 3-dimensional conformal radiation treatment using ≥6 MV photon beams with or without the addition of electrons of any energy. Tangents were prescribed to a point 1.5 cm anterior to the posterior edge of the fields at midseparation or to a point one-third of the distance from this point to the skin. Supraclavicular dose was prescribed to a depth of 3 cm, or to the deepest point of the contoured volumes. Boost dose was prescribed to an isodose line that completely covered the planning target volume. The dose was required to be within 90% to 115% of prescription dose in the breast clinical target volume and below 107% in the supraclavicular volume to constrain brachial plexus dose. Treatment plans including beams-eye views, axial planes, skin renderings, and dose-volume histograms were reviewed to ensure appropriate dose distributions.

Outcomes and statistical analysis

Our primary endpoint was locoregional control; secondary endpoints included acute and late toxicity, physician-assessed breast cosmesis, and patient-reported outcomes. Acute toxicities were scored using Common Terminology Criteria for Adverse Events version 4.0. Late toxicities were scored using the Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer scale. Cosmesis was scored by the treating physician using the Harvard Cosmesis Scale. Patient-reported outcomes were assessed using the validated Breast Cancer Treatment Outcomes Scale (BCTOS), which asks patients to report the degree of difference between treated and untreated breasts in the areas of cosmesis, functional status, and breast pain.¹⁹ Interim analysis of 83 patients excluded a locoregional control rate of ≤88%, grade 3+ breast toxicities of ≥25%, and nonbreast toxicities of ≥5%¹⁷; therefore, 67 additional patients were enrolled in a planned expansion cohort for a total planned sample size of 150 patients.

All time intervals were calculated from the date of diagnosis. Kaplan-Meier estimates of the survival and recurrence-free distributions were calculated, and the 5-year relapse and survival rates with 95% confidence intervals (CIs) were estimated. Tests were declared statistically significant if the calculated P value was ≤.05. All tests used 2-sided P values. Statistical analysis was performed with SAS statistical software (version 9.3, SAS Institute).

Results

Baseline characteristics

We enrolled 150 women between June 2009 and December 2017, of whom 146 completed the protocol treatment and were included in the analysis. Baseline characteristics of the study cohort are given in Table 2. Median follow-up was 62 months (range, 5-121 months). Median age at diagnosis was 54 years, with 48 patients (32.9%) ≤50 years old. We enrolled a substantial portion of higher-risk patients, including 19 (13.0%) with pathologically positive nodes, 16 (11.0%) with triple-negative breast cancer, 16 (11.0%) who required RNI, and 52 (35.6%) who required chemotherapy. Neoadjuvant chemotherapy was given in 17 patients (11.6%) who had up to clinical T3N1 disease; 5 of these patients had a pathologically complete response at the time of lumpectomy. The cohort had a mean central axis separation of 22.1 cm (range, 8.9-31.2 cm); 4 patients were treated in the prone position with smaller average separation. Mean maximum point dose in the breast was 107% of prescription dose (range, 104%-113%) with 3 patients having hotspots ≥110%.

Table 2.

Baseline characteristics

Characteristic	Total
Patients who received protocol treatment, N	146
Follow-up
Median (range), mo	62 (5-121)
Age at diagnosis
Median (range), y	54 (33-82)
≤50 y, n (%)	48 (32.9)
Breast laterality, n (%)
Left	69 (47.3)
Right	77 (52.7)
Histology, n (%)
Ductal carcinoma in situ	29 (19.9)
Invasive ductal carcinoma	106 (72.6)
Invasive lobular carcinoma	11 (7.5)
Tumor size
Median (range), cm	1.0 (0.0-5.0)
AJCC pathologic T stage, n (%)
Tis	29 (19.9)
T0	5 (3.4)
T1	96 (65.8)
T2	16 (11.0)
AJCC pathologic N stage, n (%)
N0	127 (87.0)
N1	18 (12.3)
N2	1 (0.7)
Receptor status, n (%)
Estrogen receptor-positive	115 (78.8)
Progesterone receptor-positive	109 (74.7)
HER2/neu-amplified	14 (9.6)
Triple-negative breast cancer	16 (11.0)
Radiation target volume, n (%)
Whole breast alone	130 (89.0)
Whole breast + regional nodes	16 (11.0)
Systemic treatment, n (%)
Hormone therapy	108 (74.0)
Chemotherapy	52 (35.6)
Neoadjuvant chemotherapy	17 (11.6)
Adjuvant chemotherapy	35 (24.0)

Abbreviation: AJCC = American Joint Committee on Cancer.

Tumor control

In 117 patients with invasive disease, there were 3 local recurrences (2.6%; 2 invasive and 1 DCIS). In 29 patients with DCIS, there were 2 DCIS recurrences (6.9%). There were no regional recurrences. Five-year estimated locoregional control for all patients was 97.7% (95% CI, 93.0%-99.3%) (Fig. 1a). There were 5 distant recurrences, all in patients with invasive disease (4.3%). Five-year estimated distant control was 97.9% (95% CI, 93.6%-99.3%) (Fig. 1b). Seven patients experienced contralateral breast tumor recurrence. One patient developed a secondary malignancy (angiosarcoma) in the treated breast approximately 4 years after radiation therapy (RT); the malignancy was treated with mastectomy but continued to progress. There were 3 deaths related to breast cancer for a 5-year estimated breast cancer–specific survival of 99.2% (95% CI, 94.6%-99.9%) (Fig. 2a). Five-year estimated overall survival was 97.3% (95% CI, 91.9%-99.1%) (Fig. 2b).

Fig. 1.

Kaplan-Meier analysis with 5-year estimates for locoregional (a) and distant (b) control.

Fig. 2.

Kaplan-Meier analysis with 5-year estimates for breast cancer-specific (a) and overall (b) survival.

Acute and late toxicity

In the entire cohort, grade 2 acute toxicity was observed in 44 patients (30.1%), consisting mostly of radiation dermatitis; 1 patient (0.7%) developed grade 3 radiation dermatitis. Grade 2 late toxicity was observed in 12 of 119 patients with at least 2 years of follow-up (10.1%) and consisted of fibrosis and fat necrosis. Three patients (2.5%) developed grade 3 late toxicity, including fibrosis, contracture, and seroma formation. The contracture developed after the patient underwent elective breast reduction, and the seroma developed after the patient experienced a local recurrence and opted for repeat breast-conserving surgery with partial breast reirradiation (45 Gy in 30 fractions delivered twice daily), eventually forming a non-healing wound; neither toxicity was related to the experimental fractionation. There were no grade 4 or 5 toxicities, and there were no heart or lung toxicities or findings of brachial plexopathy.

Physician-assessed cosmesis and patient-reported outcomes

Breast cosmesis was assessed by the treating physician as good or excellent in 59 of 62 patients (95.2%) who had a cosmesis evaluation completed at least 2 years after treatment completion (median of 38 months). Cosmesis was assessed as fair in 3 (4.8%) and poor in 0 (0%) patients.

Forty-one patients completed the BCTOS at least 3 years after treatment completion (median of 44 months). On the cosmetic subscale, 35 (85.4%) reported slight to no difference between the treated and untreated breast, 5 (12.2%) reported moderate difference, and 1 (2.4%) reported a large difference. On the functional subscale, 39 (97.5%) reported slight to no difference, 1 (2.5%) reported moderate difference, and 0 (0%) reported a large difference. On the breast pain subscale, 31 (77.5%) reported slight to no difference, 8 (20.0%) reported moderate difference, and 1 (2.5%) reported a large difference.

Discussion

We conducted a phase 2 trial evaluating one of the shortest courses of daily HF-WBI, delivering 36.63 Gy in 11 fractions to the whole breast followed by a boost of 13.32 Gy in 4 fractions over a total of 3 weeks. We found excellent disease control, low rates of acute and late toxicity, and favorable cosmetic outcomes at 5 years of follow-up, extending our previous findings in an initial cohort.¹⁷ Our 5-year locoregional failure rate of 2.3% is comparable to 2.2% and 2.8% found in the HF-WBI arms of the START B and OCOG trials, respectively,^20,21 although our study included a higher-risk population than either trial. Our 5-year distant relapse rate of 2.1% compares favorably with 7.6% found in the START B HF-WBI arm²⁰ (distant failure rate was not reported in the OCOG trial) and potentially reflects greater usage of chemotherapy and RNI in our trial as well as the advent of more effective taxane-based chemotherapy.^22,23 Furthermore, although there are differences in the grading scales used for late toxicity and cosmesis, our results compare favorably with outcomes of both trials, which may reflect the benefits of improved dose homogeneity with modern RT techniques.^24,25 Our patient-reported outcomes were also similar to previous BCTOS data indicating that 20%, 4%, and 18% of patients reported moderate to large differences on the cosmetic, functional, and breast pain subscales, respectively.¹⁹

We found acute grade 2 and 3 toxicity in 30% and 1% of patients, respectively. These rates are consistent with the results of a randomized trial from MD Anderson that, using the same Common Terminology Criteria for Adverse Events criteria, found grade 2 and 3 toxicity rates of 47% and 0% for patients undergoing HF-WBI versus 73% and 5% for patients undergoing CF-WBI.²⁶ The low acute toxicity rates in our trial and other HF-WBI regimens can be explained by radiobiological principles; namely, the reduction in total dose that is required to achieve equivalence for low α/β outcomes (breast tumor and late toxicity) necessitates a larger reduction in equivalent dose for high α/β outcomes (acute toxicity), thus sparing more rapidly proliferating tissues such as skin. For example, the EQD2 of our whole breast dose for fibrosis is 48.8 Gy, but only 40.4 Gy for desquamation (Table 1).

Our study population included women with locally advanced disease and high-risk features including young age, triple-negative disease, and positive nodes, requiring chemotherapy and/or RNI. Notably, these patients comprised 59% of our study cohort, for whom the 2011 ASTRO consensus guidelines would not have endorsed HF-WBI²⁷; however, the guidelines were updated in 2018 to recommend HF-WBI as the preferred standard in most women with early-stage breast cancer who do not require RNI.⁶ Additionally, the 2018 guidelines recommend a boost in patients ≤50 years old or those between 51 and 70 years old with high-grade disease, who together comprise 60% of our cohort, and suggest omitting a boost in patients >70 years old with hormone-receptor positive and low-to-intermediate grade tumors, who comprise only 6% of our cohort. The low relapse rates observed in our cohort lend further support to the use of HF-WBI and tumor bed boost in these patients as promulgated by the 2018 guidelines.

The major HF-WBI trials did not include DCIS, although they included microinvasive disease, which has a natural history similar to DCIS.²⁸ In contrast, 20% of the patients in our cohort had pure DCIS, and only 2 recurrences were observed. The updated ASTRO guidelines state that HF-WBI may be used as an alternative in the treatment of DCIS, but they stop short of making a definitive recommendation in the absence of randomized data⁶; however, the recently closed TROG 07.01 trial, which randomized HF-WBI versus CF-WBI in women with DCIS, will inform future recommendations. Retrospective data suggest HF-WBI is equally effective in treating DCIS,^29–31 which the results of our trial also support.

It is not yet standard to hypofractionate RNI in the United States. With a maximum follow-up of 10 years for patients treated with RNI on our trial, it is encouraging that we have not observed any brachial plexopathies or other late arm or shoulder toxicities; however, the small sample size limits any conclusions we can make regarding hypofractionated RNI. In a companion phase 2 postmastectomy RT (PMRT) trial at our institution that tested the same fractionation scheme, all patients were treated with RNI and there have similarly been no reports of brachial plexus toxicity.³² The favorable results reported from this trial have led to the opening of Alliance A221505, a randomized phase 3 trial of HF-PMRT versus CF-PMRT that will help to inform definitive recommendations regarding the use of hypofractionation for RNI in all settings.

Current HF-WBI regimens take 4 weeks or longer to complete when followed by a boost, reduced from 6 to 7 weeks with CF-WBI and boost. Shortening the course of adjuvant breast RT increases patient convenience, satisfaction, and ability to return to work sooner, enabling greater access to breast RT and therefore to breast conservation, while decreasing health care costs.^4,5,33–35

In addition to our trial, there have been other efforts to further shorten the course of WBI. One approach uses established 3-week HF-WBI regimens but delivers a concurrent rather than sequential boost so that total treatment time is not lengthened. Radiation Therapy Oncology Group 1005 is a randomized phase 3 trial testing 40 Gy in 15 fractions with a concurrent boost to 48 Gy versus standard HF-WBI or CF-WBI with a sequential boost; the trial has completed accrual and is currently maturing. IMPORT HIGH is another phase 3 trial testing a concurrent boost in addition to a risk-adapted dose design; 3-way randomization is among HF-WBI with a sequential boost, reduced dose (36 Gy in 15 fractions) WBI with 40 Gy in 15 fractions to a partial breast volume and a concurrent tumor bed boost to 48 Gy, and an identical third arm with a concurrent boost to 53 Gy. Preliminary results were recently reported and showed low rates of moderate to marked adverse effects in all 3 arms; follow-up for the primary endpoint of local relapse continues.³⁶

Another approach is testing extremely hypofractionated regimens. The randomized phase 3 UK FAST trial tested 30 Gy or 28.5 Gy in 5 once-weekly fractions against 50 Gy in 25 fractions in postmenopausal women with early-stage, node-negative tumors; 10-year results were recently reported, demonstrating low rates of local recurrence in all arms and similar late toxicity between the 28.5 Gy and 50 Gy arms but increased toxicity in the 30 Gy arm.³⁷ Results of this trial led to the randomized UK FAST-Forward trial, which is testing 27 Gy or 26 Gy in 5 daily fractions against 40 Gy in 15 fractions in a higher-risk population including younger, postmastectomy, and node-positive women; both RNI and boost are allowed. Preliminary results showed mild acute skin reactions in all arms; follow-up continues for endpoints of tumor control and late toxicity.³⁸

Our trial has several limitations, including a non-randomized design, which limits comparative analyses with other HF-WBI regimens. Furthermore, although a strength of the trial is its inclusion of high-risk women with encouraging results, the heterogeneity of the study cohort combined with the low number of local relapses limits subgroup analysis. Finally, we will continue to monitor disease control, cosmesis, and toxicity because these are expected to evolve well beyond 5 years.

Conclusions

We conducted a phase 2 trial of one of the shortest courses of HF-WBI, delivered in 3 weeks inclusive of a standardized sequential boost. Our study cohort was heterogeneous and included a high percentage of young women and women requiring chemotherapy and/or nodal radiation, which lends further support to the broad adoption of hypofractionation as endorsed by the recent ASTRO guidelines. With 5 years of follow-up, there continues to be excellent locoregional control and low toxicity with favorable cosmetic outcomes.

Summary.

This phase 2 trial delivered 36.63 Gy in 11 fractions to the whole breast followed by 13.32 Gy in 4 fractions to the tumor bed over three weeks in a heterogeneous cohort of patients. Five-year locoregional control was 97.7% with low rates of acute and late toxicity and favorable breast cosmesis. These results support the continued study of shorter whole breast radiation treatment schedules that maintain optimal treatment outcomes while increasing patient convenience.