at SciVerse ScienceDirect

The Breast 22 (2013) S129eS136

Contents lists available

The Breast

journal homepage: www.elsevier .com/brst

Original article

Hypofractionation in the era of modulated radiotherapy (RT)

Kent W. Mouwa, Jay R. Harris b,*

aHarvard Radiation Oncology Program, Boston, MA, USAbDana-Farber Cancer Institute and Brigham & Women's Hospital, Harvard Medical School, Boston, MA, USA

Keywords:Breast-conserving therapyHypofractionation

* Corresponding author. Dana-Farber Cancer Institut450 Brookline Ave., Boston, MA 02215-5450, USA. Tel.:632 2290.

E-mail address: jharris@lroc.harvard.edu (J.R. Harr

0960-9776/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.breast.2013.07.024

a b s t r a c t

The use of radiation therapy (RT) as a component of breast-conserving therapy (BCT) has been shown toreduce the risk of local-regional recurrence and improve overall survival. As has been the commonpractice in the United States and Continental Europe, the majority of studies that demonstrated thesebenefits utilized daily radiation doses ranging from 1.8 to 2.0 Gray (Gy) per day given for approximately 5weeks. However, due to geographic limitations, patient preferences, and financial considerations, therehave been continued attempts to evaluate the efficacy and safety of abbreviated or hypofractionatedcourses of whole-breast radiation. Two key factors in these attempts have been: 1) advances in radio-biology allowing for a more precise estimation of equivalent dosing, and 2) advances in the delivery of RT(‘modulation’) that have resulted in substantially improved dose homogeneity in the target volume.Hypofractionated schedules have been compared to conventional radiation courses in several random-ized controlled trials, as well as many prospective and retrospective experiences. These studies, nowwithabout 10 years of follow-up, have demonstrated equivalent rates of local-regional recurrence, disease-free survival, and overall survival. The rates of toxicity have generally not been increased with hypo-fractionated regimens; however, certain toxicities may take decades to manifest. The generalizability ofthese results is unclear, as the majority of patients in the trials were elderly with early-stage hormone-receptor positive disease. Nevertheless, there is now sufficient evidence to recommend hypofractionatedwhole breast RT for a substantial percentage of patients.

� 2013 Elsevier Ltd. All rights reserved.

Introduction

Today, many women with early-stage invasive breast cancerare candidates for breast-conserving therapy (BCT), which con-sists of breast-conserving surgery (BCS) followed by radiation toall or a portion of the treated breast. The goal of BCT is to provideoncologic outcomes equivalent to those of mastectomy whilepreserving a cosmetically-acceptable breast. Mature data fromseveral large prospective randomized trials have shown thatoverall survival is equivalent with BCT or mastectomy [1e3].Currently, the majority of women diagnosed with early-stagebreast cancer in North America and Europe are treated usingBCT approaches.

Historically, RT has been delivered using ‘conventional’ frac-tionation, with daily fractions sizes of 1.8e2 Gy per day to a totaldose of 45e50 Gy with or without addition of a boost dose to the

e, Dana Building, Room 1622,þ1 617 632 2291; fax: þ1 617

is).

All rights reserved.

site of the primary tumor. However, interest in hypofractionatedaccelerated courses of whole breast radiation continues to grow,and mature data is now available from several trials comparingconventional and hypofractionated treatment regimens.

Conventional WBI

The importance of RT following BCS has been confirmed inmultiple randomized trials, and decades of follow-up are nowavailable for trials that compared BCS with or withoutconventionally-fractionated WBI. Results from the available ran-domized trials have been analyzed by the Early Breast CancerTrialists’ Collaborative Group (EBCTCG) [4,5]. Whereas the indi-vidual trials did not have the statistical power to detect improvedsurvival with the addition of radiation, the EBCTCG meta-analysisdemonstrated that the improvement in local control provided byradiation translated into an overall survival benefit. The originalanalysis showed that radiation provided similar proportional re-ductions in local recurrence in all subgroups of patients, and thatprevention of local recurrence at 5 years translated into improvedsurvival at 15 years, at an approximate 4:1 ratio [4].

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136S130

In the most recent EBCTCG update, 7 additional trials in low-riskpatients were added, longer follow-upwas obtained in the initial 10trials and, importantly, local and distant recurrences were analyzedtogether to determine the effect of radiation on 10-year “firstrecurrence” and 15-year breast cancer death rates [5]. (Technically,the probabilities of local and distant recurrence are not statisticallyindependent, and therefore valid estimates of the separate effectsof radiation on local and distant recurrence cannot be obtained [6].)Overall, WBI reduced the 10-year risk of any first recurrence with arate ratio of about half (0.52), from 35.0% to 19.3% (2P< .00001),and it reduced the 15-year risk of breast cancer death with a rateratio of about one-sixth (0.82), from 25.2% to 21.4% (2P ¼ .00005).The risk of death from any cause was similarly lower in the radia-tion arm (34.6% vs 37.6%; 2P ¼ .03), indicating that WBI does notsignificantly increase non-breast cancer deaths.

The EBCTCG analysis shows the benefit of radiation in pre-venting recurrence is largest in the first year, but remains sub-stantial throughout the first decade, whereas the benefit in breastcancer deaths becomes apparent only after several years and con-tinues well into the second decade. (Interestingly, 5 years oftamoxifen therapy also reduced the annual rate of any recurrenceover the first 10 years by about one-half [rate ratio, RR ¼ 0.53], butreduced the annual rate of breast cancer death by about one-third[RR ¼ 0.68] [7].) The risk reduction in recurrence and death withradiation was present for women with both negative and positivelymph nodes. Benefit was seen in patients of all ages, and withvarious tumor grades and sizes, although the absolutemagnitude ofthe benefit varied. Based on the data from this update, the new “4:1ratio” was stated to be between the reduction in first recurrence at10 years (as opposed to the reduction in local recurrence at 5 years)and the reduction in mortality at 15 years.

The rates of ipsilateral breast tumor recurrence (IBTR) with BCSandWBI have been reduced considerably over time, and the 5-yearrate is now about 2% [8]. The reasons for this improvement areseveral-fold and include better imaging with mammography, morethorough pathologic evaluation of the resected specimens (espe-cially margin status), and improvements in systemic therapy that,when combined with radiation, substantially reduce IBTR [9].

Conventional WBI is well tolerated. Long-term cosmetic out-comes are quite good, and significant treatment-associatedmorbidity is rare. In a series of more than 400 patients with stageI/II breast cancer treated with WBI, the rate of unacceptablecosmetic results was 6.7% at 11 years, and was primarily limited topatients who received doses higher than 60 Gy [10]. Rates of othercomplicationsdincluding rib fracture, pericarditis, and tissuenecrosisdhave also been shown to be quite low, and these risks arelikely further decreasedwith the use of current techniques [11]. Thetime courses for both local recurrence and treatment complicationsare long, so long-term follow-up is needed to assess efficacy andsafety. There is now extensive long-term experience worldwideusing conventional WBI.

Rationale for hypofractionation

The basis for the ‘conventional’ scheme of 1.8e2 Gy daily frac-tions extends from an early radiobiologic observation that normaltissues are generally more sensitive to fraction size than tumorcells. Delivering a small fraction of radiation daily for several weekswas developed as a technique to exploit this putative difference infraction size sensitivity, and to spare normal tissue relative to tu-mor. However, much of this early experimental work was carriedout using squamous carcinomas, and the applicability to other tu-mor types has only recently been re-examined [12e14].

The relationship between radiation dose (D) and survivingfraction of cells (S) can be modeled using a linear-quadratic (LQ)

formula with two constants e one (a) that is proportional to doseand another (b) that is proportional to the square of the dose:

S ¼ e�ðaDþbD2Þ

The LQ formula can be adapted to account for multiple fractiontreatments:

S ¼�e�ðadþbd2Þ�n

where n is the number of fractions and d is the fractional dose. Thefraction size sensitivity of both normal tissues and tumors can berepresented by the alpha/beta (a/b) ratio, which is the dose atwhich the linear and quadratic contributions to cell kill are equal.Early-reacting normal tissues, such as skin andmucosa, have highera/b ratios of 8e12 Gy and are relatively less sensitive to fraction sizethan are late-reacting normal tissues, such as muscle, bone, andnervous tissue, which have much lower a/b ratios of approximately2e5 Gy. For years, it was felt that the vast majority of tumors had a/b ratios that were similar to early-reacting normal tissue, andtherefore, that the maximum therapeutic benefit of radiation couldbe achieved by increasing fractionation.

More recently, it has become apparent that different tumortypes respond very differently to radiation, and that certain tumortypes, such as breast and prostate carcinoma, actually exhibitfraction size-dependent survival characteristics (and therefore, a/bratios) that more closely resemble late-reacting normal tissues[13,14]. For these tumors, using small fraction sizes to improve thetherapeutic ratio would be of minimal benefit, and instead, largerfraction sizes may be equally effective in controlling tumor whilestill minimizing late toxicities.

Attempts to use larger fraction sizes (i.e., hypofractionation) inthe treatment of breast cancer are dependent upon accurate esti-mates of equivalent dose. Decades of experience withconventionally-fractionated breast radiation have led to use of adose-fractionation schedule (classically, 50 Gy in 25 2-Gy fractions)that has a well-characterized tumor control and toxicity profile.However, as fraction size increases, the total dose required tomaintain similar effects on tumor and normal tissue decreases.Therefore, any comparison between hypofractionated and con-ventional regimens is dependent upon an accurate estimate ofequivalent dose. An under-estimation of equivalent dose in ahypofractionated course could lead to inferior tumor control,whereas over-estimation could lead to unacceptable toxicity.Because late-reacting tissue toxicity can take years to manifest,accurate estimates of dose equivalence have evolved slowly.

With a refined understanding of the radiobiologic consequencesof fraction size and better estimates of equivalent dose, interest hasbeen renewed in the use hypofractionated courses of radiation forbreast cancer. Given the potential for similar tumor control andcosmetic outcomes with hypofractionated treatment, many logisticand economic factors favor such hypofractionated courses. Fewertreatment visits may improve patient satisfaction and quality of lifeby minimizing the psychological and physical strain associatedwith treatment [15,16]. Shorter courses of therapy could alsoimprove compliance with radiation in elderly and geographicallyisolated patients, populations shown to have lower compliancewith radiation following BCS [17e19]. Finally, some forms ofhypofractionated radiation could improve efficiency and decreasethe cost of treatment [20,21].

WBI techniques

In addition to a better understanding of the biologic equivalenceof accelerated and conventional dose schedules, technical

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136 S131

improvements in the delivery of breast irradiation have alsocontributed to increased use of acceleratedwhole breast irradiation(AWBI). Many aspects of AWBI radiation planning and delivery aresimilar to conventional WBI, which may allow it to be more easilyincorporated into existing practice.

All contemporary WBI begins with CT-based planning. Patientsare typically treated supine with arms above the head. Other po-sitions, such as prone or lateral decubitus, can be useful for patientswith large or pendulous breasts, but there is less experience usingthese positions in the hypofractionated setting. Left-sided tumorscan be treated using a heart block or breath-holding techniques toavoid direct heart irradiation, but significant patient cooperationand special in-room patient-position monitoring are required forthe breath-hold technique.

WBI is usually delivered via tangent fields using high-energy X-rays. Forward planning techniques allow addition of sub-fields tomodulate the beam to optimize dose homogeneity, which isparticularly important in thehypofractionated setting. Becausedailyfraction sizes are larger, doses to hotspots (areas of increased dosewithin the treatment volume) are amplified in the hypofractionatedsetting, which could result in greater toxicity compared to a similarplan delivered using conventional fractionation. With the use ofsubfields and higher beam energies, such as 10 MV or a mixture of6 MV and 15 MV, adequate dose homogeneity can be obtained in alarge percentage of patients. IntensityModulated RadiationTherapy(IMRT) using fields from multiple angles is not required and canresult in increased low-dose irradiation of adjacent structures.

Table 1Published AWBI trials.

Trial Canadian RMH/GOC

No. patients 1234 1410

Med. follow-up 12 years 9.7 years

Patient/tumorcharacteristics

pT1-2 N0, breast width�25 cm

T1-3 N0-1, age <75 years

Surgery BCS þ ALND BCS�ALND

Margin status No tumor on ink Macroscopically clear margins

Systemic therapy HT 42%, CT 11% HT 65%, CT 3%, HTþ CT 11%

Comments No boost or RNI 14 Gy boost 75%, RNI 21%

Total dose/no.fractions/weeks

50/25/5 42.5/16/3.2 50/25/5 42.9/13/5 39/13/5

IBTR 6.7% 6.2% 12.1% 9.6%^ 14.8%^

OSa 84.4% 84.6% 68% (pooled)Cosmesis

Goodeexcellent 71.3% 69.8%

Grade 2e3 skin/subcuttissue toxicity

7.7%/10.4% 8.9%/11.9%

Any photographicchange

35% 42% 27%p< 0.001 (comparing all 3)

Overall fair/pooroutcome

61% 66% 51%p< 0.001 (comparing all 3)

Any mod/markedchange (per physician)

IBTR ¼ ipsilateral breast tumor recurrence; HT ¼ hormonal therapy; CT ¼ chemotherapregional nodal irradiation.p¼ 0.027 comparing 39 Gy and 42.9 G arms in RMH/GOC.

a 10 year overall survival (OS) data for Canadian and RMH/GOC; 5 year OS data for ST

Conventionally-fractionated treatments are delivered daily tothe entire breast in 1.8e2 Gy fractions, with total whole-breastdoses of 45e50 Gy. Two large randomized trials have investigatedthe impact of a 10 or 16 Gy boost dose to the primary site followingBCS and conventionally-fractionated WBI, and both showed thataddition of a boost significantly decreased local recurrence rates(from 10.2% to 6.2% at 10 years in the EORTC [European Organisa-tion for Research and Treatment of Cancer] trial and from 4.5% to3.6% at 5 years in the Lyon trial) [22,23]. Provider-assessed rates offibrosis and telangiectasias were increased with addition of a boost,but patient-reported cosmetic outcomes did not differ. In the U.S., aboost is typically given using an en face electron beam.

Standard tangent fields cover a substantial percentage of level Iand II axillary nodes. “High tangent” techniques can be used to treata greater percentage of the axilla. For appropriate patients, theaddition of a third or fourth field to treat supraclavicular and in-ternal mammary nodal regions may reduce the risk of recurrenceand improve survival [24].

AWBI trials

Design

Several large randomized trials have investigated AWBI in thetreatment of early breast cancer (Table 1) [25e28]. In each of thesetrials, hypofractionated schedules were compared against con-ventional WBI of 50 Gy in 2 Gy daily fractions.

START A START B

2236 2215

9.3 years 9.9 years

pT1-3a N0-1, age >18 years pT1-3a N0-1, age >18 years

BCS�ALND (85%), mastectomy (15%) BCS�ALND (92%), mastectomy (8%)

�1 mm �1 mm

HT 54%, CT 11%, HTþ CT 25% HT 72%, CT 7%, HTþ CT 15%

10 Gy boost 61%, RNI 14% 10 Gy boost 43%, RNI 7%

50/25/5 41.6/13/5 39/13/5 50/25/5 40/15/3

7.4% 6.3% 8.8% 5.5% 4.3%

88.9% 88.7% 89.3% 89% 92%

50.4% 49.5% 43.9% 45.3% 37.9%

y; BCS ¼ breast-conserving surgery; ALND ¼ axillary lymph node dissection; RNI ¼

ART A and B.

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136S132

The hypofractionated trial with the longest published follow-upwas a non-inferiority study conducted in Canada using a hypo-fractionated regimen of 42.5 Gy in 2.66 Gy daily fractions deliveredin a little over 3 weeks [28]. A total of 1234 women with T1-2,pathologically node-negative invasive breast cancer were enrolledfollowing BCS. Negative surgical margins and a negative axillarydissection were required, and no boost dose or regional nodalirradiation was allowed. To minimize dose heterogeneity, womenwith a breast width >25 cm were excluded. The primary outcomewas local recurrence of invasive cancer in the treated breast, andsecondary outcomes included rates of regional or distant disease,and provider-assessed breast cosmetic results and late toxic effects.

Over a similar time period, a randomized trial was also con-ducted at the Royal Marsden Hospital/Gloucestershire OncologyCentre (RMH/GOC) in which 1410 women younger than 75 years ofage with T1-3, N0-1 operable breast cancer were randomized toconventional fractionation or to one of two accelerated courses:39 Gy in 13 fractions or 42.9 Gy in 13 fractions [14,27]. All courseswere delivered over 5 weeks. The primary endpoint of the studywas late change in breast appearance, and the fractionationschedules in the experimental arms were chosen based on theextremes of the accepted estimates of the fraction sizes consideredto be equivalent to 50 Gy in 2 Gy fractions in terms of late normaltissues responses. Assuming linearity in the response between thetwo test doses, the hypofractionated dose iso-effective with con-ventional fractionation could be determined by interpolation, andan a/b ratio for late change in breast appearance could be calcu-lated. Microscopic margin status was not required, but all womenhad macroscopically clear margins. A 14 Gy tumor bed boost in 7fractions was used in 75% of women, and 21% underwent regionalnodal irradiation to the supraclavicular and axillary lymph nodes.

Initial data from the RMH/GOC trial was used to inform thedesign of two additional hypofractionation trials in the UK: STARTAand START B. Similar to the RMH/GOC trial, START A enrolledwomen with pT1-3a, N0-1 invasive breast cancer and randomizedto treatment with conventional or accelerated WBI over 5 weeks[25]. Microscopically-clear surgical margins of �1 mm wererequired, but unlike in the RMH/GOC trial, patients were allowed toenroll in START A after mastectomy (although this populationcomprised only 15% of patients in the trial). Approximately 30% ofpatients were node-positive, the majority of whom were managedwith surgery alone (83.5%) rather than surgery and regional nodalirradiation (14%). Nearly all patients (>90%) were treated withhormonal therapy and/or chemotherapy. In START A, fractionationin the experimental armswas 39 or 41.6 Gy in 13 fractions deliveredover 5 weeks. These doses were altered slightly from the RMH/GOCtrial based on initial RMH/GOC results suggesting higher rates oflate toxicity in the 42.9 Gy arm. Like the RMH/GOC trial, START Awas designed to allow interpolation between the two 13-fractionregimens in order to identify the schedule equivalent to 50 Gy in25 fractions in terms of late normal tissue effects. Unlike RMH/GOC,START A was powered to compare local tumor control at this level.Use of a boost was employed in 60% of the patients, and consisted of10 Gy in 5 fractions.

The START B trial was a non-inferiority trial that randomizedpatients to conventional radiation over 5 weeks versus an accel-erated course of 40 Gy in 15 fractions over 3 weeks, a commonly-used (but empirically-derived) hypofractionated schedule at thetime of trial conception [26]. Unlike the RMH/GOC and START Atrials, total treatment time differed between the control andexperimental arms in START B, and therefore, estimates of a/b ra-tios for tumor and/or normal tissues could not be performed. Pa-tient and tumor characteristics in START B were similar to START A,although tumors were slightly less advanced (10% of tumors �3 cmcompared to 22% in START A, and 23% node-positive patients

compared to 29% in START A). Likely for these reasons, the numberof mastectomies, as well as the percentage of patients treated witha boost dose or RNI, were lower than in START A.

Tumor control

Ipsilateral tumor control was measured in each of these hypo-fractionation trials, and ten-year follow-up is now available from allfour trials. Ipsilateral tumor recurrence was a primary endpoint inthe Canadian, START A, and START B trials. The RMH/GOC trial wasdesigned to test the fractionation sensitivity and dose response oflate effects, but ipsilateral tumor control data was also reported.

Given its large size and long follow-up, the Canadian trial inparticular has been widely considered to be practice-changing. Thecumulative incidence of local recurrence at 10 years was 6.7% inwomen treated with 50 Gy in 25 fractions over 5 weeks comparedto 6.2% in women treated with 42.5 Gy in 16 fractions over 3.5weeks. When combined with non-invasive recurrences, the overallrisk of relapse at 10 years was 7.5% in the conventional fractionationarm versus 7.4% in the hypofractionation arm. Neither of thesedifferences was statistically significant. Subgroup analyses showedthat treatment effect was similar regardless of age, tumor size, ERstatus, or use of systemic therapy. The hypofractionated regimendid appear to be less effective in patients with Grade 3 tumors,where the 10-year risk of local recurrence was 15.6% compared to4.7% in the conventional fractionation arm. At 10 years, 84.4% and84.6% of patients were alive in the conventional and hypofractio-nation arms, respectively, and causes of death were similar be-tween arms.

The RMH/GOC trial was designed as a START pilot study togenerate an estimate of the a/b ratio for late changes in breastappearance. However, ipsilateral tumor recurrence data was alsoreported. The risk of ipsilateral tumor recurrence at 10 years was12.1% in the 50 Gy arm, 9.6% in the 42.9 Gy arm, and 14.8% in the39 Gy arm. There was no statistical difference between the 50 Gyarm and either of the hypofractionated arms; however, the differ-ence between the 42.9 and 39 Gy arm was statistically significant(p¼ 0.027). There was no difference in survival among the groups.

Five-year results from the START A and START B trials werepublished in 2008, and 10-year updates were recently presented inabstract form [29]. At 5 years, patients treated with conventionalfractionation in START A had 3.6% rate of local recurrence,compared to 3.5% in the 41.6 Gy arm and 5.2% in the 39 Gy arm. Thispattern persisted over time, and at 10 years, the cumulative inci-dence of local recurrence had increased to 7.4% in the 50 Gy armversus 6.3% in the 41.6 Gy arm and 8.8% in the 39 Gy arm. Theserates were not statistically different, although there was a trendtoward inferior control in the 39 Gy arm.

In START B comparing 50 Gy in 25 fractions over 5 weeks to40 Gy in 15 fractions over 3 weeks, the published 5-year rate oflocal-regional tumor relapse was 3.3% in the conventional armversus 2.2% in the hypofractionated arm, a non-significant differ-ence. Interestingly, the rate of distant relapse was significantlylower in the 40 Gy arm (HR 0.69, p¼ 0.01), which contributed tosignificantly lower rates of breast cancer mortality (HR 0.75,p¼ 0.02) and overall mortality (HR 0.76, p¼ 0.03) in this arm. At 10years, there continued to be no significant difference in local-regional relapse between arms, with rates of 5.5% in the 50 Gyarm versus 4.3% in the 40 Gy arm. Ten-year distant relapse andsurvival data has not yet been presented.

Toxicity

The toxicity and cosmetic implications of AWBI were assessed ineach of the randomized trials and were in general quite favorable.

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136 S133

In the Canadian trial, nurse-assessed skin and soft tissue toxicitieswere not different between standard and hypofractionated arms ateither 5 or 10 years, with the majority of women having no or mildtoxic effects. Global cosmetic outcome as assessed using the EORTCscale also did not vary between arms, with approximately 70% ofwomen in both arms having excellent or good cosmetic outcomesat 10 years.

In the RMH/GOC trial, appearance was scored based on photo-graphic assessment by observers blinded to treatment arm. Changein breast appearance (compared to the contralateral breast)was graded using a 3-point scale. Outcomes were also judgedclinically using a 4-point scale to assess the individual endpointsof shrinkage, distortion, edema, induration, telangiectasias, armswelling and shoulder stiffness. In addition, an overall cosmeticoutcome (excellent, good, fair, or poor) was assigned. The risk ofdeveloping any photographically-assessed radiation effect wassignificantly lower in the 39 Gy arm than in the 42.9 Gy or the 50 Gyarm (27% in the 39 Gy arm versus 42.3% in the 42.9 Gy arm and35.4% in the 50 Gy arm, p< 0.001 comparing all 3 schedules). Byclinical assessment, there were significant differences among thethree arms in all endpoints except arm edema. In each case, therewere fewer side effects in the 39 Gy arm than in the 42.9 Gy arm.The use of a boost significantly increased rates of induration andtelangiectasias, but did not result in a change in photographicbreast appearance.

Because treatment times were the same across arms of theRMH/GOC trial, the a/b ratio for late effects could be calculated. Thea/b ratio for any photographic change in breast appearance was3.6 Gy and did not vary if scores from the first 1e2 years werecensored (to eliminate potential contribution from transient sur-gery- or radiation-induced changes). That there were significantdifferences in many of the cosmetic outcomes between the 39 and42.9 Gy arms highlights the sensitivity of late tissues to total doseand fraction size. An estimation of the relationship between doseand change in breast appearance suggests that in the range of dosestested, a 5% increase in total dose corresponds to a 9% increase inthe number of patients experiencing marked change in breastappearance.

Five-year toxicity data has been reported for START A and STARTB, and 10-year updates were also recently presented. In START A,there were no differences in physician- or patient-reportedcosmetic outcomes between 50 Gy and 41.6 Gy, but outcomeswere significantly better in the 39 Gy arm. At 10 years, the rate ofany physician-assessed moderate/marked effect in the conservedbreast was 50.4% in the conventional arm compared to 49.5% in the41.6 Gy arm and 43.9% in the 39 Gy arm. The difference between the39 Gy and 50 Gy arms was statistically significant.

Because treatment occurred over 5 weeks in all treatment arms,data from the RMH/GOC and START A trials were analyzed togetherto determine the a/b ratio for late-reacting normal tissue and local-regional relapse.When10-year data is included for both trials, the a/b ratio for adverse events assessedbyphotographwas3.1Gy (95%CI:2.0e4.2) and the a/b ratio for local-regional relapsewas 3.5 Gy (1.2e5.7 Gy). The authors concluded that the similarity in these valuesdemonstrated that both breast cancer and normal tissue are simi-larly sensitive to fraction size, and thus smaller fraction sizes are asgentle on breast cancer as they are on normal tissue.

In START B, there was a trend toward improved patient- andphysician-reported outcomes in the 40 Gy arm compared to the50 Gy arm at 5 years. By 10 years, these differences had increased,and the rates of any marked/moderate toxicities were lower in the40 Gy arm compared to the 50 Gy arm (37.9% vs 45.3% at 10 years).Importantly, rates of severe late toxicities such as symptomatic ribfracture, lung fibrosis, ischemic heart disease, or cardiac-relateddeath were similarly low across treatment arms in both trials.

Late adverse events may continue to accrue for many years, andit is important that patients continue to be monitored for bothdisease recurrence and late toxicity. However, the balance ofavailable data suggests no trend toward increased toxicity withhypofractionation and if anything, a suggestion of improvedcosmetic outcomes with hypofractionation.

AWBI patient selection

The majority of patients enrolled in the first-generation AWBItrials was older womenwith early-stage hormone receptor-positivecancers. For example, in the Canadian trial, 80% of primary tumorswere less than 2 cm in size, 70% were estrogen receptor positive,and all women had node-negative disease. Women enrolled in theSTART trials were allowed to have larger primary tumors, andapproximately 25% were lymph-node positive.

In each trial, subgroup analyses were performed in an attemptto identify patients for whom hypofractionated treatment may beinferior to conventional treatment. In the Canadian trial, no dif-ference in local control was noted between hypofractionated andconventional treatment arms when patients were analyzed by age(<50 vs �50), tumor size (<2 cm vs �2 cm), estrogen receptorstatus, or use of systemic therapy. However, in an unplanned subsetanalysis, women with grade 3 tumors had a 15.6% rate of localrecurrence in the hypofractionated arm compared to 4.7% in theconventional arm (p¼ 0.01 for interaction among grades 1e3).

In a post-hoc meta-analysis of ten-year data from the RMH/GOCand START A and B trials, no differences in local control were notedbetween fraction sizes �2 Gy versus >2 Gy when patients wereanalyzed by age (<50 vs�50), type of surgery (BCS vsmastectomy),nodal status (positive vs negative), use of a boost, or use of systemictherapy. Unlike the Canadian trial, patients with grade 3 tumors didnot have higher rates of local-regional relapse when treated using ahypofractionated regimen; instead, therewas a trend toward bettercontrol with fraction sizes larger than 2 Gy. In a retrospectiveanalysis of 1335 early-stage patients with Grade 3 tumors treatedbetween 1990 and 2000 in British Columbia, local control rateswere similar in women treated with hypofractionated andconventionally-fractionated regimens [30]. The balance of evidencesuggests that hypofractionated regimens are equivalent to con-ventional fractionation for all tumor grades.

All women in the Canadian trial were pathologically node-negative following axillary dissection, and no regional nodal irra-diation was allowed. Nearly all women in the START trials alsounderwent surgical axillary lymph node staging, and women with1e3 positive lymph nodes (N1 disease) were allowed to enroll.Regional nodal irradiation (RNI) was allowed, but was used in onlyabout 10% of women. When included, RNI was delivered to thesupraclavicular fossa (with or without inclusion of the axilla) usingthe same dose fractionation as used for the breast. In none of theSTART hypofractionated arms has there been an indication ofincreased bracial plexopathy, arm edema, shoulder stiffness, orother toxicities potentially related to treatment of the lymphatics.However, the total number of patients treated with RNI was rela-tively small, and toxicity may take longer than 10 years to manifest.Therefore, providers may be somewhat more hesitant to usehypofractionation when planning to treat the regional lymphatics.

A boost dose to the surgical cavity was not allowed in the Ca-nadian trial. In START A and B, participating centers were requiredto treat patients according to their institutional boost policy.Approximately 60% of patients in START A and 40% of patients inSTART B received a boost consisting of 10 Gy in 5 fractions deliveredvia an en face electron beam field. In the RMH/GOC trial, a subset ofpatients was sub-randomized to receive a boost, while other pa-tients were offered a boost electively (in total, 75% received a

Table 2On-going AWBI trials.

Trial FAST (UK)

No. patients 915

Accrual dates 2004e2007

Entry criteria �50 years old, tumor <3 cm, N0, microscopically clear margins

Arm 1a 50/25/5 (whole breast)

Arm 2 28.5/5/5 (whole breast)

Arm 3 30/5/5 (whole breast)

Trial IMPORT HIGH (UK)

No. patients 840 (goal)

Accrual dates 2009ecurrent

Entry criteria Any one of: 18e49 years old, �pT2, grade 3,margin <2 mm, LVI, involved ALNs

Arm 1 40/15/3 (whole breast) followed by tumor bed boostof 16/8/1.5

Arm 2 36/15/3 (whole breast) with integrated indexquadrant boost to 40/15/3 and tumor bed boostto 48/15/3

Arm 3 36/15/3 (whole breast) with integrated index quadrantboost to 40/15/3 and tumor bed boost to 53/15/3

Trial IMPORT LOW (UK)

No. patients 2018

Accrual dates 2006e2010

Entry criteria �50 years old, unifocal pT1N0, grade 1e2, no LVI,margins �2 mm

Arm 1 40/15/3 (whole breast)

Arm 2 36/15/3 (whole breast) with integrated index quadrantboost to 40/15/3

Arm 3 40/15/3 (partial breast)

Trial SHARE (France)

No. patients 2800 (goal)

Accrual dates 2010ecurrent

Entry criteria �50 years old, menopausal, unifocal T1N0 tumor,margins �2 mm

Arm 1 50/25/5 (whole breast) followed by tumor bed boostof 10e16/5e8/1.5

Arm 2 40/15/3 or 42.5/16/3.2 (whole breast)

Arm 3 40/10/1 (partial breast)

Trial RTOG 1005 (USA)

No. patients 2150 (goal)

Accrual dates 2011ecurrent

Entry criteria Stage 1e2 and 1 of: <50 years old, involved ALNs,LVI, �2 close margins, 1 close margin þ EIC,1 focally-positive margin, ER-/PR-, grade 3,Oncotype �20 OR Stage 0 and <50 years old andgrade 3 OR Stage 0e2 after neoadjuvant chemotherapy

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136S134

boost). Combined analysis of the local-regional relapses from allthree trials did not reveal an interaction between boost and frac-tionation schedule. Cosmetic outcomes have been reported for theRMH/GOC trial, and rates of telangiectasias and breast indurationwere significantly higher in the boost group [14]. Currently, there isno consensus among experts on the use and fractionation of a boostin hypofractionated regimens, and the issue is being investigated inseveral on-going hypofractionation trials.

Systemic therapy (chemotherapy and/or hormonal therapy) wasused in a majority of patients in the published hypofractionatedtrials. In each of the trials, use of systemic therapy was not speci-fied, but was left to the policy of participating institutions. Use ofsystemic therapy did not appear to influence recurrence rates inhypofractionated versus conventionally-fractionated arms of theCanadian trial, or across fractionation arms in compiled data fromthe RMH/GOC, STARTA and START B trials (Table 1). In the Canadiantrial, use of systemic therapy was not a predictor of a worsecosmetic outcome. Likewise, in a non-randomized cohort ofwomen treated using similar fractionation, neither rates of acuteradiation dermatitis nor late skin toxicity were higher in womenreceiving chemotherapy [31]. However, there continues to beconcern that the additive effects of cytotoxic chemotherapy andhypofractionated radiation may increase the risk of poor cosmeticoutcome, and for this reason, use of hypofractionated regimens hasbeen cautiously implemented in patients also receivingchemotherapy.

Women with pure ductal carcinoma in situ (DCIS) wereexcluded in the published trials of hypofractionation, althoughthey represent a significant percentage of patients eligible for BCT.There is not strong radiobiologic evidence to suggest DCIS re-sponds differently to hypofractionation than invasive cancer, andretrospective experiences suggest that hypofractionation is aseffective as conventional fractionation in providing durable localcontrol and preventing invasive recurrences for patients with DCIS[32,33]. If adequate surgical margins can be obtained, hypo-fractionation may be a reasonable management strategy for pa-tients with pure DCIS.

Approximately one quarter of women enrolled in the publishedhypofractionated trials were younger than 50. In neither the Ca-nadian trial nor in combined analysis of the START trials was therean interaction between age (�50 years versus <50 years) andfractionation regimen. However, because the number of youngpatients was small and because younger age has been associatedwith more aggressive biology and higher rates of local recurrence,there has been reluctance by some radiation oncologists to usehypofractionation in young breast cancer patients. In addition,these younger patients havemany years to exhibit very late toxicity.

With the recent update of the START trials, long-term oncologicand cosmetic outcomes are now available for several thousandwomen treated with AWBI using fraction sizes of 2.5e3 Gy. Thecumulative evidence suggests that this approach is safe and effec-tive for many women with early-stage invasive disease. At ourinstitution (Dana-Farber/Brigham & Women’s Cancer Center), weare expanding the use of hypofractionated treatment based on themost recent data. Currently, we favor a hypofractionated course of42.5 Gy in 16 fractions (‘Canadian’) with the optional inclusion of a5e10 Gy boost delivered in 2e4 fractions. We are offering thisapproach towomen�50 years oldwith invasive cancer andwomen�60 years old with DCIS. Women receiving systemic chemotherapyare no longer excluded. Patients are not eligible if the radiation plancontains hotspots >107% despite optimization, or if a supra-clavicular field is planned. Indications for a 5e10 Gy boost includeyounger age (50e59 years for invasive cancer), grade 3 or triple-negative disease, pure DCIS, or other high risk features at thediscretion of the radiation oncologist.

Table 2 (continued )

Arm 1 50/25/5 or 42.5/16/3.2 followed by partial breast boostof 12e14/6e7/1.5

Arm 2 40/15/3 with integrated partial breast boost to 48/15/3

LVI ¼ lymphovascular invasion; ALN ¼ axillary lymph nodes; EIC ¼ extensiveintraductal component.

a Total dose/number of fractions/weeks of treatment.

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136 S135

Future directions

Additional follow-up data are being obtained on the first-generation AWBI trials, and multiple second-generation AWBI tri-als have already been launched (Table 2). Trial designs vary, andvarious schedules of AWBI with or without a sequential or con-current boost are being compared against conventional WBI oraccelerated partial breast irradiation (APBI).

Given the promising results of the published trials utilizingmoderate hypofractionation (2.5e3 Gy per fraction), more extremehypofractionated regimens are now being tested. The UK FAST Trialaccrued 915 early-stage patients treated with BCS and randomizedto 50 Gy in 25 fractions, 30 Gy in 5 fractions, or 28.5 Gy in 5 frac-tions, all delivered over 5 weeks [34]. The doses in the 30 Gy and28.5 Gy experimental arms were chosen because they are predictedto be equivalent to the 50 Gy arm for an a/b ratio of 4.0 or 3.0 Gy,respectively. The primary endpoint was 2-year change in photo-graphic breast appearance, and initial results were recently re-ported. Compared to the control arm, women treated with 28.5 Gyin 5 fractions had a similar risk of any change in photographicbreast appearance at 2 years (risk ratio 1.15, p¼ 0.49) while womentreated with 30 Gy in 5 fractions were significantly more likely tohave a change in appearance (risk ratio 1.70, p< 0.001). Similarly, 3-year rates of clinically-assessedmoderate ormarked adverse effectswere higher in the 30 Gy arm (17.3%) than in the 28.5 Gy (11.1%) or50 Gy (9.5%) arms. Although these initial results suggest that28.5 Gy in 5 fractions has a similar normal tissue toxicity profile asconventionally-fractionated radiation, it is uncertain how thelinear-quadratic model used to estimate dose equivalence trans-lates to increased fraction sizes beyond the a/b ratio. Therefore,longer follow-up will be required to determine the effect of largefraction sizes on both normal tissue toxicity and tumor control.

The Radiation Therapy Oncology Group is conducting a largemulti-center randomized trial (RTOG 1005) investigating thetiming of boost irradiation. Beginning in 2011, womenwith DCIS orearly-stage invasive cancer and at least one risk factor are beingrandomized to whole breast radiation (using conventional or ‘Ca-nadian’ fractionation) followed by sequential boost versus ahypofractionated course of 40 Gy in 15 fractions over 3 weeks to thewhole breast with a simultaneous integrated cavity boost to a totaldose of 48 Gy. Other open randomized trials in both the US andEurope are using similar schema to investigate the role of frac-tionation, dose, and treatment time on both disease control andnormal tissue toxicity (Table 2).

Conclusions

Interest in the use of AWBI continues to grow. The results of theCanadian hypofractionation trial have beenwidely considered to bepractice-changing, and recently-updated results from the STARTtrials also indicate that hypofractionated regimens represent a goodtreatment choice for many appropriately-selected women. Howev-er, the applicability of the results of these trials to broader patientpopulations, including younger woman with more advanced

or hormone receptor-negative disease, remains uncertain. Forth-coming results from second-generation AWBI trials are addressingseveral important issues, andwill likely guide the expanded used ofAWBI.

Conflict of Interest Statement

None declared.

Acknowledgment

Portions of this article were published in: Irradiation in early-stage breast cancer: conventional whole-breast, accelerated partial-breast, and accelerated whole-breast strategies compared. MouwKW, Harris JR. Oncology (Williston Park). 2012. 26(9):820e30 and areused with permission of the publisher.

References

[1] Fisher B, Anderson S, Bryant J, Margolese RG, Deutsch M, Fisher ER, et al.Twenty-year follow-up of a randomized trial comparing total mastectomy,lumpectomy, and lumpectomy plus irradiation for the treatment of invasivebreast cancer. N Engl J Med 2002;347(16):1233e41.

[2] van Dongen JA, Voogd AC, Fentiman IS, Legrand C, Sylvester RJ, Tong D, et al.Long-term results of a randomized trial comparing breast-conserving therapywith mastectomy: European Organization for Research and Treatment ofCancer 10801 trial. J Natl Cancer Inst 2000;92(14):1143e50.

[3] Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgerywith radical mastectomy for early breast cancer. N Engl J Med 2002;347(16):1227e32.

[4] Clarke M, Collins R, Darby S, Davies C, Elphinstone P, Evans E, et al. Effects ofradiotherapy and of differences in the extent of surgery for early breast canceron local recurrence and 15-year survival: an overview of the randomisedtrials. Lancet 2005;366(9503):2087e106.

[5] Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, et al. Effectof radiotherapy after breast-conserving surgery on 10-year recurrenceand 15-year breast cancer death: meta-analysis of individual patientdata for 10,801 women in 17 randomised trials. Lancet 2011;378(9804):1707e16.

[6] Gelman R, Gelber R, Henderson IC, Coleman CN, Harris JR. Improved meth-odology for analyzing local and distant recurrence. J Clin Oncol 1990;8(3):548e55.

[7] Early Breast Cancer Trialists’ Collaborative GroupDavies C, Godwin J, Gray R,Clarke M, Cutter D, et al. Relevance of breast cancer hormone receptors andother factors to the efficacy of adjuvant tamoxifen: patient-level meta-anal-ysis of randomised trials. Lancet 2011;378(9793):771e84.

[8] Arvold ND, Taghian AG, Niemierko A, Abi Raad RF, Sreedhara M, Nguyen PL,et al. Age, breast cancer subtype approximation, and local recurrence afterbreast-conserving therapy. J Clin Oncol 2011;29(29):3885e91.

[9] Morrow M, Harris JR, Schnitt SJ. Surgical margins in lumpectomy for breastcancer e bigger is not better. N Engl J Med 2012;367:79e82.

[10] Delouche G, Bachelot F, Premont M, Kurtz JM. Conservation treatment of earlybreast cancer: long term results and complications. Int J Radiat Oncol BiolPhys 1987;13(1):29e34.

[11] Pierce SM, Recht A, Lingos TI, Abner A, Vicini F, Silver B, et al. Long-term ra-diation complications following conservative surgery (CS) and radiationtherapy (RT) in patients with early stage breast cancer. Int J Radiat Oncol BiolPhys 1992;23(5):915e23.

[12] Bentzen SM, Ruifrok AC, Thames HD. Repair capacity and kinetics for humanmucosa and epithelial tumors in the head and neck: clinical data on the effectof changing the time interval between multiple fractions per day in radio-therapy. Radiother Oncol 1996;38(2):89e101.

[13] Vogelius IR, Bentzen SM. Meta-analysis of the alpha/beta ratio for prostatecancer in the presence of an overall time factor: bad news, good news, or nonews? Int J Radiat Oncol Biol Phys 2013;85(1):89e94.

[14] Yarnold J, Ashton A, Bliss J, Homewood J, Harper C, Hanson J, et al. Fraction-ation sensitivity and doseeresponse of late adverse effects in the breast afterradiotherapy for early breast cancer: long-term results of a randomised trial.Radiother Oncol 2005;75(1):9e17.

[15] Kawase E, Karasawa K, Shimotsu S, Izawa H, Hirowatari H, Saito AI, et al. Esti-mationof anxietyanddepression inpatientswithearly stagebreast cancerbeforeand after radiation therapy. Breast Cancer (Tokyo, Japan) 2012;19(2):147e52.

[16] Whelan TJ, Levine M, Julian J, Kirkbride P, Skingley P. The effects of radiationtherapy on quality of life of women with breast carcinoma: results of a ran-domized trial. Ontario Clinical Oncology Group. Cancer 2000;88(10):2260e6.

[17] Athas WF, Adams-Cameron M, Hunt WC, Amir-Fazli A, Key CR. Travel distanceto radiation therapy and receipt of radiotherapy following breast-conservingsurgery. J Natl Cancer Inst 2000;92(3):269e71.

K.W. Mouw, J.R. Harris / The Breast 22 (2013) S129eS136S136

[18] Farrow DC, Hunt WC, Samet JM. Geographic variation in the treatment oflocalized breast cancer. N Engl J Med 1992;326(17):1097e101.

[19] Lazovich DA, White E, Thomas DB, Moe RE. Underutilization of breast-conserving surgery and radiation therapy among women with stage I or IIbreast cancer. JAMA 1991;266(24):3433e8.

[20] Lievens Y. Hypofractionated breast radiotherapy: financial and economicconsequences. Breast 2010;19(3):192e7.

[21] Suh WW, Pierce LJ, Vicini FA, Hayman JA. A cost comparison analysis of partialversus whole-breast irradiation after breast-conserving surgery for early-stage breast cancer. Int J Radiat Oncol Biol Phys 2005;62(3):790e6.

[22] Bartelink H, Horiot JC, Poortmans PM, Struikmans H, Van den Bogaert W,Fourquet A, et al. Impact of a higher radiation dose on local control andsurvival in breast-conserving therapy of early breast cancer: 10-year results ofthe randomized boost versus no boost EORTC 22881-10882 trial. J Clin Oncol2007;25(22):3259e65.

[23] Romestaing P, Lehingue Y, Carrie C, Coquard R, Montbarbon X, Ardiet JM, et al.Role of a 10-Gy boost in the conservative treatment of early breast cancer: resultsof a randomized clinical trial in Lyon, France. J Clin Oncol 1997;15(3):963e8.

[24] Whelan TJ, Olivotto I, Ackerman I, Chapman JW, Chua B, Nabid A, et al. NCIC-CTG MA.20: an intergroup trial of regional nodal irradiation in early breastcancer. J Clin Oncol 2011;29(Suppl.) [abstr LBA1003].

[25] Bentzen SM, Agrawal RK, Aird EG, Barrett JM, Barrett-Lee PJ, Bliss JM, et al. TheUK Standardisation of Breast Radiotherapy (START) Trial A of radiotherapyhypofractionation for treatment of early breast cancer: a randomised trial.Lancet Oncol 2008;9(4):331e41.

[26] Bentzen SM, Agrawal RK, Aird EG, Barrett JM, Barrett-Lee PJ, Bliss JM, et al. TheUK Standardisation of Breast Radiotherapy (START) Trial B of radiotherapyhypofractionation for treatment of early breast cancer: a randomised trial.Lancet 2008;371(9618):1098e107.

[27] Owen JR, Ashton A, Bliss JM, Homewood J, Harper C, Hanson J, et al. Effect ofradiotherapy fraction size on tumour control in patients with early-stagebreast cancer after local tumour excision: long-term results of a randomisedtrial. Lancet Oncol 2006;7(6):467e71.

[28] Whelan TJ, Pignol JP, Levine MN, Julian JA, MacKenzie R, Parpia S, et al. Long-term results of hypofractionated radiation therapy for breast cancer. N Engl JMed 2010;362(6):513e20.

[29] TheUKSTART (Standardisation of Breast Radiotherapy) Trials: 10-year follow-upresults. In: Haviland JS, Agrawal R, Aird E, Barrett J, Barrett-Lee P, Brown J, et al.,editors. CTRC-AACRSanAntonio BreastCancer SymposiumDecember4e8, 2012.

[30] Herbert C, Nichol A, Olivotto I, Weir L, Woods R, Speers C, et al. The impact ofhypofractionated whole breast radiotherapy on local relapse in patients withgrade 3 early breast cancer: a population-based cohort study. Int J RadiatOncol Biol Phys 2012;82(5):2086e92.

[31] Hijal T, Al Hamad AA, Niazi T, Sultanem K, Bahoric B, Vuong T, et al. Hypo-fractionated radiotherapy and adjuvant chemotherapy do not increaseradiation-induced dermatitis in breast cancer patients. Curr Oncol2010;17(5):22e7.

[32] Wai ES, Lesperance ML, Alexander CS, Truong PT, Culp M, Moccia P, et al.Effect of radiotherapy boost and hypofractionation on outcomes in ductalcarcinoma in situ. Cancer 2011;117(1):54e62.

[33] Williamson D, Dinniwell R, Fung S, Pintilie M, Done SJ, Fyles AW. Local controlwith conventional and hypofractionated adjuvant radiotherapy after breast-conserving surgery for ductal carcinoma in-situ. Radiother Oncol2010;95(3):317e20.

[34] Agrawal RK, Alhasso A, Barrett-Lee PJ, Bliss JM, Bliss P, Bloomfield D, et al. Firstresults of the randomised UK FAST Trial of radiotherapy hypofractionation fortreatment of early breast cancer (CRUKE/04/015). Radiother Oncol2011;100(1):93e100.