Review

Predictors of recurrence for ductal carcinoma in situ after

breast-conserving surgery
John R Benson, Gordon C Wishart

Ductal carcinoma in situ (DCIS) constitutes a major public health problem, with up to half of screen-detected cancers Lancet Oncol 2013; 14: e348–57
representing pure forms of DCIS without evidence of invasion. A proportion of cases detected with routine screening Cambridge Breast Unit,
would not have progressed to a life-threatening form of breast cancer during the patient’s lifetime, and overdiagnosis Addenbrooke’s Hospital,
Cambridge, UK
of breast cancer is a cause for concern. Once DCIS has been detected, treatment is obligatory and present technologies
(Prof J R Benson DM); and
do not allow accurate risk stratification such that intensity of treatment can be tailored to risk of recurrence and Anglia Ruskin University,
progression to invasive disease. Present management strategies are based on prognostic and predictive information Cambridge, UK (J R Benson,
derived from conventional histopathological and host factors. With increasing molecular characterisation of these Prof G C Wishart MD)
preinvasive lesions, data will be available for how factors such as oestrogen receptor, progesterone receptor, HER2, Correspondence to:
Prof J R Benson, Cambridge
and indicators of proliferative activity can provide additional information about both prognosis and benefit from
Breast Unit, Addenbrooke’s
adjuvant treatments such as radiotherapy and hormonal therapy. Low-risk patients are especially poorly defined in Hospital, Hills Road, Cambridge
terms of need for adjuvant therapies, which can be associated with both short-term adverse sequelae and long-term CB2 0QQ, UK
effects (eg, cardiotoxicity) that can affect all-cause mortality. Optimum risk prediction in the future is likely to be john.benson@addenbrookes.
nhs.uk
achieved by integration of both conventional and molecular factors, which should be incorporated into a validated
predictive model to help with clinical decision making.

Introduction treatments and those at lower risk can avoid adjuvant

Ductal carcinoma in situ (DCIS) is a heterogeneous therapies such as radiotherapy and tamoxifen, which can
disease with increased prominence after the introduction impair quality of life and increase health-care costs.
of breast screening programmes. Before the advent of Clinical trials have shown that both these modalities
screening, DCIS represented only 2–5% of symptomatic reduce invasive recurrence in the ipsilateral breast, but no
breast cancers, compared with almost 20% of newly survival benefit specific to breast cancer has been shown
diagnosed symptomatic cases in the present era and up to from the addition of radiotherapy, tamoxifen, or both to
half of screen-detected breast cancer.1–4 Not all patients with lumpectomy (wide local excision) alone.12–14
DCIS will progress to invasive disease, with proportionate DCIS represents a late-stage disease in terms of
estimates ranging from 25% to 50% depending on grade molecular progression, and many genetic mutations
of the lesion.5 Not all invasive cancers arise from lesions occur before invasion.15,16 No genes are yet recognised as
that are recognised histologically as carcinomas in situ, main drivers of invasion, but quantitative changes in
although a phase involving increased epithelial expression of genes associated with angiogenesis, cell
proliferation is likely to precede an invasive cancer. The adhesion and motility, and formation of extracellular
frequency of DCIS in routine autopsy studies (ranging matrix are probably important.17,18 There are few
from <1% to 14·3%)6,7 suggests that some non-invasive differentially expressed genes between hyperplasia and
lesions detected mammographically, and subsequently DCIS, and normal breast epithelium is committed at a
excised, would have been of no clinical importance. This very early stage to low-grade versus high-grade lesions.
occurrence raises the issue of non-obligate progression, Moreover, high-grade DCIS tends to recur as
with resultant uncertainties of biological behaviour and higher grade invasive cancer, and there is generally
dilemmas in clinical management.8 Although the risk of retention of molecular subtypes from in-situ to invasive
death after any treatment for DCIS is less than 2% after disease. Low-risk patients are particularly poorly defined
10 years, recurrence rates vary for different treatments, and in terms of need for adjuvant therapies; a molecular
half the cases of ipsilateral breast tumour recurrence after profile allows for improved risk assessment. Many trials
breast-conserving surgery alone or combined with are not powered to detect differences in outcome
radiotherapy are invasive disease.9,10 The main goal of attributable to patient and tumour characteristics, and
treatment is to prevent recurrence of invasive disease, subset analyses use small numbers of patients and
minimise treatment-related morbidity, and optimise short follow-up duration. Furthermore, many studies
cosmesis. Furthermore, within older age groups life originate from single institutions with idiosyncratic
expectancy is likely to be determined not by a diagnosis of practices in adequacy of surgical excision and
DCIS, but by competing causes of death.11 pathological assessment.
The identification of molecular and biological markers In this Review we discuss conventional risk factors
that provide prognostic and predictive information to add for ipsilateral breast tumour recurrence after
to standard clinical and pathological factors is an area of breast-conserving surgery for DCIS, and consider how
current research. A molecular risk profile will be clinically molecular risk factors might be integrated into clinical
useful if higher-risk patients can be selected for specific algorithms for more accurate prediction of invasive

www.thelancet.com/oncology Vol 14 August 2013 e348

 Review

recurrence risk and provision of guidance to tailor invasive malignancy was unclear. Nonetheless, an
intensity of treatments for DCIS. expert panel concluded that an adequate negative
margin exists when the tumour is not touching ink, and
Conventional predictors of local recurrence recommended this margin as the standard definition
Much of the evidence for recurrence risk and conventional for invasive cancer.27
histopathological factors comes from non-randomised The issue of margin width might differ for DCIS
studies with small numbers of events. Nonetheless, lesions, which often have a discontinuous growth
evidence suggests that determinants of ipsilateral breast pattern, and demands special consideration for surgical
tumour recurrence after conservation surgery for DCIS clearance.27 Up to 40% of DCIS lesions are estimated to
are multifactorial and involve factors related to the grow discontinuously, and skip lesions can be separated
patient, tumour, and treatment. The relative importance by a distance exceeding a designated negative margin.
of these factors remains controversial, and specific Although DCIS lesions usually grow continuously along
indices (eg, the Van Nuys Prognostic Index) and any single ductal tree, different duct systems are
nomograms have been devised that attempt to intertwined and not spatially exclusive. Complex,
incorporate independent predictors of local recurrence in three-dimensional DCIS lesions with arborisation can
a clinically useful instrument. appear multifocal when viewed in two dimensions on a
pathological slide with sectioning of more than one duct
Margin status system. Stereoscopic interrogation of mastectomy
The intuitive association between higher rates of local specimens shows that only 8% of DCIS cases have gaps
recurrence and margin positivity for invasive disease has of greater than 10 mm between lesions and that 85% of
been confirmed for DCIS in several studies,19,20 using the gaps are smaller than 5 mm.28 Therefore, some have
definition of tumour at the inked resection margin.21 In advocated a wider margin of clearance for DCIS lesions
the National Surgical Adjuvant Breast and Bowel Project of up to 10 mm.29 Dunne and colleagues30 examined the
(NSABP) B-24 trial,22 women undergoing risk of ipsilateral breast tumour recurrence in more than
breast-conserving surgery with radiotherapy for DCIS 4000 patients undergoing breast-conserving surgery
were randomly assigned to tamoxifen or placebo. with radiotherapy for DCIS, in both randomised and
Both groups included patients with negative prospective or retrospective non-randomised trials.
(ie, no tumour at ink; n=1341) and positive (n=457) When specific margin thresholds were examined,
surgical margins. Invasive and non-invasive local a margin of 2 mm was associated with significantly
recurrence rates were significantly higher for lower risk of ipsilateral breast tumour recurrence than
margin-positive patients than for margin-negative was a margin of smaller than 2 mm (1 mm or no tumour
patients in both the placebo group (recurrence rate at ink; odds ratio [OR] 0·53, 95% CI 0·26–0·96).
30·89 vs 16·05 per 1000 patients per year) and the By contrast, a margin of more than 5 mm did not lead to
tamoxifen group (17·37 vs 12·4 per 1000 patients lower rates of ipsilateral breast tumour recurrence
per year). Moreover, tamoxifen conferred a (OR 1·51, 95% CI 0·51–5·0, p>0·5). The researchers
proportionately greater reduction in local recurrence for concluded that 2 mm is an adequate margin when
patients with positive or unknown margins than for patients receive radiotherapy, but the absolute number
those with negative margins (44% vs 22%), although the of events in each group was relatively low. This
benefits of tamoxifen for DCIS seem to be less than minimum margin recommendation of 2 mm is
those for invasive cancer and are restricted to oestrogen endorsed by the aforementioned expert panel27 and
receptor (ER) positive disease.22,23 accords with present guidance from the National
Many studies of local recurrence are retrospective Institute of Health and Care Excellence31 and other
with no standardised methods for pathological consensus opinions.32
assessment of margins in terms of orientation, inking, The frequency of multifocality warrants careful
specimen sectioning (perpendicular or tangential assessment of margins and justifies a wider margin
shaved), and reporting practices. Although rates of mandate for DCIS than for invasive disease. A
recurrence are minimised by negative margin status,24 meta-analysis by Wang and colleagues,33 based on
no direct relation exists between margin width and randomised controlled trials and observational studies,
ipsilateral breast tumour recurrence for invasive examined the relation between margin thresholds and
disease.25 The definition of a negative margin used in ipsilateral breast tumour recurrence for DCIS treated
prospective randomised controlled trials is tumour not with breast-conserving surgery with or without
touching ink, and there is no conclusive evidence that radiotherapy (randomly allocated or otherwise).
wider margins are necessary for satisfactory rates of 1066 events of ipsilateral breast tumour recurrence were
local control in patients receiving systemic therapies.26 assessed in 7564 patients undergoing either breast-
Until recently, there was widespread agreement that a conserving surgery alone (565 events) or combined
margin mandate of 2 mm was appropriate, although with radiotherapy (501 events). Binomial likelihood
whether this margin pertained equally to DCIS and techniques were used to model the effect of margin

e349 www.thelancet.com/oncology Vol 14 August 2013

 Review

width on ipsilateral breast tumour recurrence by Tumour characteristics

estimation of treatment-specific effects for each margin Despite being a heterogeneous group of lesions, three
threshold, and with the assumption of a linear risk of basic categories of DCIS are recognised. High-grade DCIS
recurrence with time. Risk of ipsilateral breast tumour lesions (figure 1A) are characterised by very irregular and
recurrence (irrespective of radiotherapy) gradually pleomorphic nuclei, coarse and clumped chromatin,
reduced with progressive increments in margin width prominent nucleoli, and frequent mitoses. There is
of 2 mm, 5 mm, and 10 mm. When analysis was negligible architectural differentiation between cells, a
confined to the subgroup receiving both breast- solid growth pattern, and necrosis is usually present. By
conserving surgery and radiotherapy, the probability of contrast, low-grade DCIS lesions (figure 1B) consist
either a 5 mm or 10 mm margin threshold being the predominantly of cells with monomorphic, regular nuclei
best option (lowest rate of ipsilateral breast tumour with fine chromatin, small nucleoli, and few mitoses.
recurrence) was 0·97 compared with 0·02 for a margin Intermediate-grade DCIS lesions show much less nuclear
greater than 2 mm. A 10 mm margin was optimum, pleomorphism than do high-grade lesions, with evidence
with an OR of 0·17 (95% credible interval 0·12–0·24) of cellular polarisation. At present there is a focus on
compared with an OR of 0·38 (95% credible interval nuclear grading, which is a better biological predictor of
0·28–0·51) for a margin of greater than 2 mm, and an recurrence than comedo necrosis (which can be present
overall probability of it being the best option of 0·96. in both low-grade and intermediate-grade lesions) and
Further studies are needed to assess which groups of architectural differentiation.39 Lagios and colleagues40
patients will benefit from these wider margins of reported rates of recurrence with longer-term follow-up in
excision, which could yield rates of local control a group of 79 patients treated with excision only. 19% of
equivalent to breast-conserving surgery combined with high-grade lesions recurred within 26 months, compared
radiotherapy.29 The effect of margin width on ipsilateral with 10% of intermediate-grade lesions after 87 months
breast tumour recurrence might be modulated by and no low-grade lesions after 124 months of follow-up.
volume of the resected specimen. In a study of Other studies12,41 have reported nuclear grade to be the
146 patients with a median follow-up of 7·2 years, Vicini
and colleagues34 noted that adequacy of excision was A
determined by both margin width and excision volume.
Thus for specimens of less than 60 cm³, 5-year actuarial
rates of ipsilateral breast tumour recurrence did not
differ between patients with margins of 2 mm or less
and patients with margins of 5 mm or more. By contrast,
for specimens exceeding 60 cm³, rates of ipsilateral
breast tumour recurrence were 11% and 2%, respectively,
for these two margin categories (p=0·04). These
findings could indicate the complex interaction of ductal
systems in three dimensions and more complete
excision of the ductal system with greater volumes of
tissue resected. Oncoplastic techniques could allow
wider margins of excision (eg, 10 mm) while achieving
satisfactory cosmetic outcomes and perhaps avoiding
the need for adjuvant treatments such as radiotherapy.35 B
Low-grade and intermediate-grade DCIS lesions are
more likely to be associated with a discontinuous growth
pattern, and patients with these lesions could derive
lower absolute benefit from radiotherapy. Therefore
wider margins might be appropriate for these types of
DCIS. The Early Breast Cancer Trialists’ Collaborative
Group (EBCTCG)36 noted that longer-term mortality for
invasive breast cancer was affected by reduction in
locoregional disease attributable to more extensive
surgery and radiotherapy. With improvements in
surgical margins of clearance, the absolute benefits
from radiotherapy for local control will be reduced in
the future. However, at present the identification of any
group of patients with DCIS for whom there is no
Figure 1: Ductal carcinoma in situ lesions in breast tissue
statistically significant benefit from radiotherapy is not (A) High-grade lesion (reduced by 36% from ×100). (B) Low-grade lesion
possible.37,38 (reduced by 36% from ×200).

www.thelancet.com/oncology Vol 14 August 2013 e350

 Review

most significant predictor of recurrence on both univariate analysis. This finding contrasted with those from other
and multivariate analysis. Differences in rates of studies41 suggesting that older age and postmenopausal
recurrence between low-grade and high-grade lesions can status were associated with lower rates of ipsilateral
disappear with longer-term follow-up; the crucial variable breast tumour recurrence, which were almost twice as
is time since diagnosis rather than potential to progress.42 high for premenopausal women. Silverstein’s group
In a systematic review that used pooled, random-effects subsequently showed age to be an independent
risk estimates from randomised and observational data, prognostic factor for local recurrence, which led to
Wang and colleagues43 noted that tumour grade predicted incorporation of age into the Van Nuys Prognostic Index
ipsilateral breast tumour recurrence after breast- by the division of patients into three groups (≥61 years,
conservation therapy for DCIS (risk estimate 1·81, 95% CI 40–60 years, and ≤39 years).48 These findings have
1·53–2·13). In the updated analysis of the UK, Australia, encouraged aggressive approaches to treatment of DCIS
and New Zealand (UK/ANZ) DCIS trial,14 rates of in younger patients with either mastectomy or
ipsilateral breast tumour recurrence did not differ between radiotherapy after breast-conserving surgery.
low-grade and intermediate-grade DCIS, which suggests
that distinction between these categories might be Family history
unnecessary and that grades should be reclassified as low Genetic factors could affect not only the incidence of
or intermediate, high, and very high. Cytonuclear grade DCIS but also the risk of ipsilateral breast tumour
should be the basic method for the assessment of intrinsic recurrence. McCormick and colleagues49 recorded a
biological aggressiveness. significantly higher chance of a positive family history of
Although rates of ipsilateral breast tumour recurrence breast cancer in patients who recurred after
are generally higher for tumours with a component of breast-conserving surgery and radiotherapy for DCIS,
comedo necrosis (irrespective of adjuvant therapies) than than in those who maintained local control (40% vs 18%,
for those without,44 the presence of necrosis might be a p=0·03). Therefore, mastectomy should be considered in
weaker predictor of ipsilateral breast tumour recurrence patients with a strong family history of DCIS, even if they
than cellular architecture and nuclear grade. Division have localised lesions that would otherwise be suitable
into comedo versus non-comedo necrosis cannot for breast-conserving surgery alone.50
clinically stratify patients for recurrence risk. High-grade
lesions without comedo necrosis can have similar Molecular predictors of local recurrence
biological behaviour and marker profiles to high-grade The risk of local recurrence after breast-conservation
lesions with necrosis.45 Furthermore, comedo necrosis is therapy for invasive disease can be predicted from gene
associated with a lower risk estimate for ipsilateral breast microarray data, which classifies breast tumours into
tumour recurrence than is tumour grade (risk estimate distinct biological subtypes. The basal subtype seems to
1·71, 95% CI 1·36–2·16).43 Estimates of the extent of be associated with a higher risk of local recurrence than
DCIS are often inaccurate, whether based on radiological do luminal subtypes, and could have stronger and more
or pathological criteria, which restricts the usefulness of consistent associations than some of the conventional
interpretation of tumour size as a predictive factor for histopathological factors such as grade, subtype, and
local recurrence. Some elements of a DCIS lesion can be nodal status.51 Studies52–55 examining expression of steroid
non-calcified, which can lead to underestimation of size, receptors (ER, progesterone receptor [PR]), HER2, Ki-67,
whereas benign calcification can be misinterpreted as p53, and EGFR in pure DCIS, compared with the in-situ
malignant when core biopsy samples show calcification component of mixed lesions in invasive ductal carcinoma
within DCIS lesions. Moreover, accurate and reproducible (IDC), have identified distinct molecular characteristics.
pathological measurement of size according to the Van Gene expression analyses, comparing in-situ with
Nuys Prognostic Index needs a serial sequential tissue invasive components of DCIS-IDC lesions, have shown
protocol with size computed from serial sections and not very similar molecular profiles.16,56 A matched-pair
from the largest diameter of a single section.46 analysis of DCIS and IDC was done to identify potential
These differences in pathological processing could markers of the transition from in-situ to invasive
account for variation in distribution of tumour sizes malignancy.56 With the use of laser-capture
between studies and hinder predictive accuracy of this microdissection and microarray analysis to establish
variable. Nonetheless, there is a positive correlation epithelial-specific gene expression profiles, researchers
between the maximum size measurement of DCIS identified several progression-specific candidate genes,
lesions and recurrence risk,47 with a recurrence risk including GREM1, involved in epithelial–mesenchymal
estimate for tumour size of 1·63 (95% CI 1·30–2·06).43 signalling, and DSE, involved in induction of tumour-
reactive cytotoxic T lymphocytes. A similar study57
Age at diagnosis and menopausal status compared gene-expression profiles in 41 breast tissue
Initial studies by Silverstein and colleagues47 showed that samples, comprising non-neoplastic (n=4), pure DCIS
age was not a predictor of local recurrence after (n=5), in-situ component of DCIS-IDC (n=22), and IDC
breast-conserving surgery for DCIS on univariate (n=10) with identification of 147 genes differentially

e351 www.thelancet.com/oncology Vol 14 August 2013

 Review

expressed between pure DCIS and the in-situ component Two presentations at the 2012 San Antonio Breast
of DCIS-IDC. Two genes, LOX and SULF1, might Cancer Symposium addressed molecular prediction of
participate in progression of DCIS to invasive disease, recurrence after conservative treatment of DCIS. The first
with evidence that the in-situ component of DCIS-IDC examined gene expression profiles in 108 cases of
exhibits molecular changes facilitating invasion of recurrent DCIS (66 cases of DCIS only, 42 cases of IDC)
surrounding tissue before any detectable morphological with a 14-gene panel, and showed that molecular changes
changes occur. associated with invasive recurrence differed from those
Expression of biomarkers indicative of an abrogated in non-invasive recurrence.62 In the second, 314 cases of
response to cellular stress might predict a subset of DCIS underwent immunohistochemical analysis of
patients with DCIS with worse outcomes. Findings from molecular phenotype to investigate correlation with
a study58 of 70 cases of DCIS with (n=32) or without recurrence risk.63 Cases were assigned to one of four
(n=38) recurrence suggested that a small panel of phenotypes according to ER, PR, and HER2 expression
biological markers could predict the recurrence of DCIS, (luminal A: ER/PR positive and HER2 negative;
with high expression of p16 or COX-2 and Ki-67 luminal B: ER/PR positive and HER2 positive; HER2
conferring a significant risk of both in-situ and invasive type: ER/PR negative and HER2 positive; triple negative:
recurrence. The small sample size of the study prevented ER/PR negative and HER2 negative). The predictors of
the researchers from determining whether the marker overall recurrence were tumour grade, margin status,
panel was an independent predictive variable on and molecular phenotype, with patients in the luminal B
multivariate analysis when combined with conventional and HER2 groups at significantly heightened risk. The
recurrence factors such as tumour grade, lesion size, and strongest predictors of invasive recurrence were
margin clearance. A nested case-control study59 of more ER negativity, HER2 positivity, and molecular phenotype,
than 1000 patients treated with breast-conserving surgery with all subtypes except luminal A associated with
recorded a significantly higher risk of invasive recurrence significantly increased risk.
(19·6%) for tumours triple-positive for p16, COX-2, and This study has pragmatic appeal for the management
Ki-67 compared with triple-negative tumours (4·1%). of DCIS. It used immunohistochemistry to measure
A competing risk model was used to identify high-risk expression of ER, PR, and HER2. These markers are
groups on the basis of a combination of clinicopathological often used to select patients for appropriate adjuvant
and molecular predictors; a surgical margin clearance of therapy after surgery for early invasive breast cancer.
1 mm or greater, palpability, and either (ER–, HER2+, Molecular phenotyping on the basis of
Ki-67+) status or (p16+, COX-2+, Ki-67+) status was immunohistochemistry of three receptors could be
associated with increased 8-year risk of ipsilateral breast introduced with little additional cost and few changes to
tumour recurrence. The idea of an immunohistochemical routine clinical practice. Patients with a high-risk
panel for DCIS is appealing and analogous to how molecular phenotype could be selected for adjuvant
receptor expression is used in invasive carcinoma to therapies (radiotherapy with or without tamoxifen) after
predict response to therapy. Moreover, it is likely to be breast-conserving surgery. Another benefit is
adopted by clinicians if validated and shown to be identification of a low-risk group (luminal A) for whom
straightforward to use and interpret. the avoidance of adjuvant therapy after surgery might be
The 21-gene recurrence score (Oncotype DX) is used to safe. This study has provided a straightforward risk
stratify recurrence risk in patients with ER-positive stratification based on an immunohistochemical
invasive breast cancer and to help to select patients who molecular classification; further investigation and
might benefit from chemotherapy.60 A 12-gene subset of validation in additional datasets is needed to establish
the original 21-gene array has been used to develop and whether this panel of markers could replace traditional
validate a DCIS score that divides patients into low-risk, risk factors, or be amalgamated into a mathematical
intermediate-risk, and high-risk for 10-year local model or nomogram. The finding that both tumour
recurrence.61 The DCIS score was based on results from grade and margin status retained significance on
327 of 670 patients recruited to the Eastern Cooperative multivariate tests suggests that a combined approach (as
Oncology Group (ECOG) E5194 trial37 with small, is used for invasive breast cancer) might be best.
localised DCIS lesions and excision margin widths of
3 mm or greater. Patients did not receive radiotherapy Radiotherapy after breast conservation therapy
after breast-conserving surgery and tamoxifen was for DCIS
optional. A continuous DCIS score model was used to The EBCTCG overview64 included data for 3729 women
predict the 10-year risk of ipsilateral breast tumour undergoing breast-conserving surgery for localised DCIS
recurrence, and disease-free survival correlated accurately with or without radiotherapy. Treatment with
with DCIS scores. However, this method was developed radiotherapy roughly halved the rate of ipsilateral breast
in radiotherapy-naive patients, and further validation is tumour recurrence, but did not affect contralateral events
needed in datasets containing patients treated by or distant disease. For patients treated with radiotherapy
radiotherapy. the absolute 10-year risk of relapse in the ipsilateral

www.thelancet.com/oncology Vol 14 August 2013 e352

 Review

breast, from either DCIS or invasive disease, was 15·2% p=0·45) and suggested possible detrimental effects of
lower than for patients who did not receive radiotherapy radiotherapy on all-cause mortality. Nonetheless, a
(12·9% vs 28·1% respectively; two-sided p<0·0001). Cochrane review66 concluded that there were no significant
A larger proportion of recurrence events were prevented long-term toxic effects from radiotherapy. There is a
in older women (≥50 years) treated with radiotherapy balance between the adverse cardiac effects of radiotherapy
than in younger women (17·0% vs 10·6% respectively; and its effect on local relapse, with a theoretical mortality
two-sided p<0·004), but the treatment was effective in rate of 2–3% for excision only compared with 1–2% for
both age groups and effectiveness was not influenced by excision plus radiotherapy.46 Omission of radiotherapy,
other factors such as extent of surgical resection with its financial and physical costs, might be appropriate
(including margin status), mode of detection, grade or for some smaller, non-high-grade DCIS lesions without
size of tumour, and administration of tamoxifen. necrosis, and this policy has been endorsed by the National
Moreover, there persisted an absolute benefit of 18% at Comprehensive Cancer Network in the USA where up to
10 years for women with small, low-grade tumours a third of DCIS cases are estimated to be treated with wide
excised with clear resection margins (no tumour at ink), excision only.67 Nonetheless, updated results of the ECOG
who might have been expected to have little benefit from trial at 10 years37 showed an equivalent rate of ipsilateral
radiotherapy. The benefits of radiotherapy in terms of breast tumour recurrence (15%) in patients deemed to be
local control were greater for patients with high-grade either at lower or higher risk, which suggested that
lesions and involved margins (tumour touching ink). lower-grade DCIS lesions have similar potential for
In the NSABP B-17 trial, the rate of local recurrence at recurrence but could take longer to relapse.
15 years was 35% in the excision-only group compared
with 19·8% in the group that received excision combined Tamoxifen
with radiotherapy (relative reduction of 43%); distant Encouraging results from the NSABP B-17 trial prompted
disease-free survival and overall survival was identical the NSABP B-24 trial,22 which randomly assigned
for both groups.17 When invasive recurrences were 1804 women undergoing breast-conserving surgery
analysed separately, the relative reduction from combined with radiotherapy for DCIS (including patients
radiotherapy was 52% (hazard ratio [HR] 0·48, 95% CI with positive surgical margins) to placebo or 5 years of
0·33–0·69, p<0·001). Invasive ipsilateral breast tumour tamoxifen. After 15 years’ follow-up, 16·6% of patients in
recurrence increased the risk of death related to breast the placebo group had relapsed locally versus 13·2% in
cancer (HR 1·75, 95% CI 1·45–2·96, p<0·001). the tamoxifen group, with a significant difference for
Despite no overall survival benefit, these results led to invasive recurrence that seemed only confined to ER
reassertion of recommendations for breast irradiation positive cases.12 Disease-free survival or overall survival
in most patients with DCIS managed with did not differ between placebo and tamoxifen groups.
breast-conserving surgery. Nonetheless, the trial Allred and colleagues68 retrospectively investigated the
protocol precluded any detailed analysis of relative ER and PR status of 41% of the study population of the
benefit of radiotherapy in relation to tumour type, grade, NSABP B-24 trial, and noted a significant decrease in any
and extent. 75–90% of cases had DCIS lesions of 10 mm breast cancer event (HR 0·58, 95% CI
or smaller, which contrasts with other studies 0·415–0·81, p=0·001) or invasive recurrence (HR 0·53,
documenting less than half of tumours in this size 95% CI 0·34–0·82, p=0·005) at 10 years for ER positive
category.46 The European Organisation for Research and patients (76%) given tamoxifen. Investigators recorded
Treatment of Cancer (EORTC) trial13 was of similar no significant reductions in recurrence for separate
design to the B17 trial and yielded similar results at analyses of ipsilaterally invasive (p=0·10) and non-
15 year follow-up; the risk of ipsilateral breast tumour invasive (p=0·39) cancer, and contralaterally invasive
recurrence was halved, but distant disease-free survival (p=0·06) and non-invasive (p=0·14) cancer, which
and overall survival did not differ between radiotherapy suggests that further stratification is essential to identify
groups. The UK/ANZ trial14 showed reductions in
ipsilateral breast tumour recurrence similar to the 1 2 3
NSABP B17 and EORTC trials at a median follow-up of Tumour size ≤15 16–40 ≥41
12·7 years. A decrease in breast cancer mortality of (mm)
1–2% at years 15–20 might have been expected in the Surgical margins ≥10 1–9 <1
radiotherapy groups of these trials, in which half of local (mm)
recurrences were invasive disease.64 However, this effect Pathological Non-high grade, Non-high grade, High grade, with or
classification without necrosis with necrosis without necrosis
was not noted, but could yet occur at a later timepoint.
A meta-analysis of four randomised trials of Age (years) ≥61 40–60 ≤39

conservatively treated DCIS undertaken by Viani and The scores for each predictor are added to give a total index score.
colleagues65 before the EBCTCG overview64 showed a
Table: Scores for calculation of the University of Southern California/Van
non-significant reduction in overall survival in the
Nuys Prognostic Index
radiotherapy group (relative risk 1·08, 95% CI 0·65–1·78,

e353 www.thelancet.com/oncology Vol 14 August 2013

 Review

patients for whom tamoxifen might be beneficial after undergoing breast-conserving surgery should receive it.
breast-conserving surgery and radiotherapy. Some reports suggest a higher proportion of invasive
In the UK/ANZ trial,14 patients in both the radiotherapy recurrences in irradiated patients, which could relate to
and non-radiotherapy groups were randomly assigned to delays in diagnosis attributable to breast fibrosis.70
receive tamoxifen or no systemic therapy. Tamoxifen Patients with screen-detected DCIS might have higher
yielded a 29% reduction in risk of recurrence, but for all-cause mortality if toxic effects of radiotherapy exceed
patients receiving radiotherapy much of the benefit of any oncological gains in terms of breast-cancer-specific
tamoxifen disappeared. There was no additional benefit mortality. At present, molecular testing is not routinely
from concomitant tamoxifen for patients receiving offered to patients with DCIS. Characterisation of
radiotherapy, although tamoxifen had a longer-term molecular events associated with progression of DCIS is
preventive effect on the contralateral breast. These a major challenge, with the goal of identification of
findings could limit detection of any advantage from additional predictive information complementary to that
anastrozole over tamoxifen for postmenopausal women derived from routine clinical and pathological factors.
with hormone-sensitive DCIS treated with breast- Ultimately, this information could help to select tumours
conserving surgery (not mastectomy) in the International most likely to recur or progress to invasive carcinoma
Breast Cancer Intervention Study II.69 after breast-conserving surgery. Present treatment
options for DCIS are acknowledged as excessive for
Management of patients with DCIS many patients, and an aim of molecular profiling is to
Although mastectomy for DCIS results in very low rates ensure that adjuvant treatments are restricted to patients
of chest wall and distant recurrence, it confers no survival at highest risk of invasive recurrence.
advantage and represents overtreatment for most Van Zee’s group at Memorial Sloan-Kettering Cancer
patients with DCIS. Conversely, breast-conserving Center have developed a DCIS nomogram, and used
surgery alone is associated with higher rates of ipsilateral continuous risk estimation models to calculate 5 and
breast tumour recurrence for most subgroups of women, 10 year probability of ipsilateral breast tumour recurrence
and half of local recurrences are invasive disease. in a retrospective study of 1681 patients treated with
Uncertainties relating to innate tumour biology, nuances breast-conserving surgery.71 The model incorporates ten
of pathological processing, and potential harms of conventional clinicopathological and treatment-related
adjuvant therapies have generated difficulties for variables, with 200 bootstrap samples used for validation.
clinicians managing localised forms of DCIS amenable Values for concordance index and estimates of
to breast-conserving surgery. The most contentious issue concordance probability were similar to other clinical
is adjuvant radiotherapy and whether all patients decision-making tools. However, this nomogram does

DCIS diagnosis

Very low-risk Low-risk Intermediate-risk High-risk

(LNG, ≤10 mm) (USC/VNPI=4,5,6) (USC/VNPI=7,8,9) (USC/VNPI=10–12)

USC/VNPI=7 USC/VNPI=8,9

Trial WLE WLE WLE WLE WLE WLE Mastectomy

(surgery
vs no
surgery) Margins >2 mm Margins ≥3 mm Margins <3 mm Margins 5–10 mm Margins <5 mm

MP good MP good MP poor MP good MP poor MP good MP poor MP good MP poor MP good MP poor

WLE alone WLE alone WLE+RT WLE alone? WLE+RT WLE+RT/TAM Mastectomy WLE+RT/TAM? Mastectomy RT/TAM?

Figure 2: Algorithm for treatment of DCIS

MP includes ER, PR, HER2, and Ki-67 status, and Oncotype DX score. DCIS=ductal carcinoma in situ. LNG=low nuclear grade. USC/VNPI=University of Southern California/Van Nuys Prognostic Index.
WLE=wide local excision. MP=molecular profile. RT=radiotherapy. TAM=tamoxifen.

www.thelancet.com/oncology Vol 14 August 2013 e354

 Review

not include molecular predictors such as ER, HER2, and predict recurrence or progression to invasion, would help
Ki-67 status and probably underestimates the true with decision making after breast-conserving surgery for
heterogeneity of DCIS lesions and the effect of innate DCIS (figure 2). Some of the key molecular events that
biological proclivity on specific risk of invasive recurrence. characterise progression of DCIS have been identified,
The University of Southern California/Van Nuys and the next logical step is to develop and validate
Prognostic Index was initially developed retrospectively predictive models. These models should be user friendly
from single-institution data in 1996, with an updated and allow individualisation of treatments, and could lead
version available in 2003.48 The index attempts to to withholding of not only radiotherapy and hormonal
incorporate independent predictors of local recurrence therapies, but even surgical excision for selected patients.
with appropriate weighting into a clinical decision-making A comprehensive management strategy should integrate
instrument that can offer both patients and clinicians clinicopathological features and molecular profiling, and
guidance when they choose treatments for DCIS. The take into account patient-related factors such as age,
index assigns a score (1–3) for each of four predictors: comorbidities, breast size, and personal preference.
pathological classification (grade, presence or absence of Contributors
necrosis), margin width, tumour size, and age, with Both authors contributed equally to the idea, design, planning, and
cutoffs established from statistical analysis of data from a writing of this Review. JRB was responsible for editing and ensuring
consistency of style throughout.
large series of patients with non-randomised treatments
(table). Updated analysis of local recurrence-free survival Conflicts of interest
We declare that we have no conflicts of interest.
in the index dataset of more than 1500 patients with pure
DCIS has enabled recommendations for treatment References
1 American Cancer Society. American Cancer Society breast cancer
options based on a combination of index scores and facts and figures 2011. http://www.cancer.org/Research/
margin widths. According to these data, rates of CancerFactsFugures/BreastCancerFactsFugures/breast-cancer-facts-
and-figures-2011-2012 (accessed Feb 10, 2013).
ipsilateral breast tumour recurrence of less than 20% at
2 Tabar L, Smith RA, Vitak B, et al. Mammographic screening: a key
12 years can be achieved with wide excision only for factor in the control of breast cancer. Cancer J 2003; 9: 15–27.
patients with scores of 4, 5, or 6 irrespective of margin 3 Silverstein MJ, Lagios MD, Recht A, et al. Image-detected breast
width, and for patients with a score of 7 if the margin cancer: state of the art diagnosis and treatment. J Am Coll Surg
2005; 201: 586–97.
width is 3 mm or greater. Radiotherapy should be 4 Independent UK Panel on Breast Cancer Screening. The benefits
recommended for patients with a score of 7 when margin and harms of breast cancer screening: an independent review.
width is less than 3 mm, a score of 8 when margin width Lancet 2012; 380: 1778–86.
is 3 mm or greater, and a score of 9 when margin width is 5 Rosen PP, Braun DW, Kinne DE. The clinical significance of
pre-invasive breast cancer. Cancer 1980; 46: 919–25.
5 mm or greater. All other permutations warrant 6 Nielsen M, Jensen J, Andersen J. Precancerous and cancerous
mastectomy; chest wall radiotherapy can be indicated. breast lesions during lifetime and at autopsy: a study of 83 women.
Whether a margin of 10 mm or greater alone can act as Cancer 1984; 54: 612–15.
7 Bartow SA, Pathak DR, Black WC, et al. Prevalence of benign,
a surrogate predictor, and avoid the need for radiotherapy atypical and malignant breast lesions in populations at different
is controversial. In a group of 366 patients who had risk for breast cancer: a forensic autopsy study. Cancer 1987;
margin widths of 10 mm or greater, 30 patients (8%) 60: 2751–60.
8 Anderson TJ. Genesis and source of breast cancer. Br Med Bull 1991;
developed local recurrence of whom 26 had received 47: 305–18.
excision only.46 Despite the significant benefits derived 9 Silverstein MJ, Lagios MD, Martino S, et al. Outcome after local
from radiotherapy, local control at 12 years for excision recurrence in patients with ductal carcinoma in situ of the breast.
J Clin Oncol 1998; 16: 1367–73.
alone was similar to results from the NSABP B-17 trial
10 Romero L, Klein L, Ye W, et al. Outcome after invasive recurrence in
for patients treated with excision plus radiotherapy. patients with ductal carcinoma in situ of the breast. Am J Surg 2004;
The index dataset uses serial sequential tissue processing, 188: 371–76.
which has restricted its generalisability and impeded 11 Gotzsche PC, Olsen O. Is screening for breast cancer with
mammography justifiable? Lancet 2000; 355: 129–34.
external validation. Nonetheless, the index is
12 Wapnir I, Dignam JJ, Fisher B, et al. Long term outcomes of
multifactorial, and incorporation of molecular markers invasive ipsilateral breast tumour recurrences after lumpectomy in
of risk together with standardised pathological processing NSABP B-17 and B-24 randomised clinical trials for DCIS.
J Natl Cancer Inst 2011; 103: 478–88.
might promote greater concordance across studies and
13 Bijker N, Meijnen P, Peterse JL, et al. Breast conserving treatment
increase predictive accuracy. with or without radiotherapy in ductal carcinoma in situ: ten-year
results of European Organisation for Research and Treatment of
Cancer Randomised Phase III Trial 10853—a study by the EORTC
Conclusions Breast Cancer Cooperative Group and EORTC Radiotherapy Group.
In the past 30 years, clinicians have benefited from the J Clin Oncol 2006; 24: 1–8.
use of prognostic or predictive models in early invasive 14 Cuzik J, Sestak I, Pinder S, et al. Effect of tamoxifen and
radiotherapy in women with locally excised ductal carcinoma in
breast cancer, such as Adjuvant!72 and Predict,73 which situ: long term results from the UK/ANZ DCIS trial. Lancet Oncol
incorporate not only tumour size, grade, and lymph node 2011; 12: 21–29.
status, but variables such as age and ER, HER2, and Ki-67 15 Seth A, Kitching R, Landberg G, et al. Gene expression profiling of
status. A similar approach, which combines conventional ductal carcinomas in situ and invasive breast tumours.
Anticancer Res 2003; 23: 2043–51.
risk factors for recurrence with biological markers that

e355 www.thelancet.com/oncology Vol 14 August 2013

 Review

16 Ma XJ, Salunga R, Tuggle JT, et al. Gene expression profiles of 39 Holland R, Peterse JL, Millis R, et al. Ductal carcinoma in situ:
human breast cancer progression. Proc Natl Acad Sci USA 2003; a proposal for a new classification. Semin Diagn Pathol 1994;
100: 5974–79. 11: 167–80.
17 Burstein HJ, Polyak K, Wong JS, et al. Ductal carcinoma in situ of 40 Lagios MD, Margolin FR, Westdahl PR, et al. Mammographically
the breast. N Engl J Med 2004; 350: 1430–41. detected ductal carcinoma in situ: frequency of local recurrence
18 Baxter N, Virnig BA, Durham SB, et al. Trends in the treatment of following tylectomy and prognostic effect of nuclear grade on local
ductal carcinoma in situ of the breast. J Natl Cancer Inst 2004; recurrence. Cancer 1989; 63: 618–24.
96: 443–48. 41 Silverstein MJ, Barth A, Poller DN, et al. Ten-year results
19 Schnitt SJ. Risk factors for local recurrence in patients with invasive comparing mastectomy to excision and radiation therapy for
breast cancer and negative surgical margins of excision. ductal carcinoma in situ of the breast. Eur J Cancer
Am J Clin Pathol 2003; 120: 485–88. 1995; 9: 1425–27.
20 Schnitt SJ, Abner A, Gelman R, et al. The relationship between 42 Solin LJ, Kurtz J, Fourquet A, et al. Fifteen-year results of breast
microscopic margins of resection and the risk of local recurrence in conserving surgery and definitive breast irradiation for the
patients with breast cancer treated with breast-conserving surgery treatment of ductal carcinoma in situ of the breast. J Clin Oncol
and radiation therapy. Cancer 1994; 74: 1746–51. 1996; 14: 754–63.
21 Kell MR, Morrow M. An adequate margin of excision in ductal 43 Wang S-Y, Shamliyan T, Virnig BA, Kane R. Tumour
carcinoma in situ. BMJ 2005; 331: 789–90. characteristics as predictors of local recurrence after treatment of
22 Fisher B, Dignam J, Wolmark N, et al. Tamoxifen in treatment of ductal carcinoma in situ: a meta-analysis. Breast Cancer Res Treat
intraductal breast cancer: National Surgical Adjuvant Breast and 2011; 127: 1–14.
Bowel Project B-24 randomised controlled trial. Lancet 1999; 44 Habel LA, Daling JR, Newcomb PA, et al. Risk of recurrence after
353: 1993–2000. ductal carcinoma in situ of the breast.
23 Allred DC, Bryant L, Land S, et al. Estrogen receptor expression as a Cancer Epidemiol Biomarkers Prev 1998; 7: 689–96.
predictive marker of effectiveness of tamoxifen in the treatment of 45 Solin L, Yeh IT, Kurtz J, et al. Ductal carcinoma in situ (intraductal
DCIS: findings from NSABP Protocol B-24. Breast Cancer Res Treat carcinoma) of the breast treated with breast conserving surgery and
2003; 76: 36. definitive irradiation: correlation of pathologic parameters with
24 Smitt MC, Nowels K, Carlson RW, et al. Predictors of re-excision outcome of treatment. Cancer 1993; 71: 2532–42.
findings and recurrence after breast conservation. 46 Silverstein MJ, Lagios MD. Ductal carcinoma in situ of the breast.
Int J Radiat Oncol Biol Phys 2003; 57: 979–85. In: Benson JR, Tuttle T, Gui G, eds. Early breast cancer: from
25 Singletary SE. Surgical margins in patients with early-stage breast screening to multidisciplinary management. Boca Raton, FL: CRC
cancer treated with breast conservation therapy. Am J Surg 2002; Press Taylor and Francis Group, 2013: 341–60.
184: 383–93. 47 Silverstein MJ, Poller D, Craig P, et al. Predicting local recurrence
26 Wapnir I, Anderson S, Mamounas E, et al. Survival after IBTR in in patients with intraductal breast carcinoma (DCIS).
NSABP node negative protocols B-13, B-14, B-19, B-20 and B-23. Proc Am Soc Clin Oncol 1995; 14: 117.
J Clin Oncol 2005; 23: 517. 48 Silverstein MJ. The University of Southern California/Van Nuys
27 Kaufmann M, Morrow M, von Minckwitz G, et al. Locoregional Prognostic Index for ductal carcinoma in situ of the breast.
treatment of primary breast cancer: consensus recommendations Am J Surg 2003; 186: 337–43.
from an International Expert Panel. Cancer 2010; 116: 1184–91. 49 McCormick B, Rosen PP, Kinne D, et al. Ductal carcinoma in situ
28 Faverley DR, Burgers L, Bult P, Holland R. Three-dimensional of the breast: an analysis of local control after conservation surgery
imaging of mammary ductal carcinoma in situ: clinical and radiotherapy. Int J Radiat Oncol Biol Phys 1991; 21: 289–92.
implications. Semin Diagn Pathol 1994; 11: 193–98. 50 Hiramatsu H, Bornstein BA, Recht A, et al. Local recurrence after
29 Silverstein MJ, Lagios MD, Groshen S, et al. The influence of conservative surgery and radiation therapy for ductal carcinoma in
margin width on local control of ductal carcinoma in situ of the situ: possible importance of family history. Cancer J Sci Am 1995;
breast. N Engl J Med 1999; 340: 1455–61. 1: 55–61.
30 Dunne C, Burke JP, Morrow M, Kell MR. Effect of margin status on 51 Nguyen PL, Taghian AG, Katz MS, et al. Breast cancer subtype
local recurrence after breast conservation and radiation therapy for approximated by estrogen receptor, progesterone receptor and
ductal carcinoma in situ. J Clin Oncol 2009; 27: 1615–20. HER2 is associated with local and distant recurrence after breast
conserving therapy. J Clin Oncol 2008; 26: 2373–78.
31 NICE. Breast cancer (early & locally advanced): diagnosis and
treatment. NICE Clinical Guideline 2009; 80. 52 Farabegoli F, Champeme MH, Bieche I, et al. Genetic pathways in
the evolution of breast ductal carcinoma in situ. J Pathol 2002;
32 Toi M, Winer EP, Inamoto T, et al. Identifying gaps in the
196: 280–86.
locoregional management of early breast cancer: highlights from
the Kyoto Consensus Conference. Ann Surg Oncol 2011; 53 Mylonas I, Makovitzky J, Jeschke U, Briese V, Friese K, Gerber B.
18: 2885–92. Expression of Her2/neu, steroid receptors (ER and PR), Ki67 and
p53 in invasive mammary ductal carcinoma associated with ductal
33 Wang S-Y, Chu H, Shamliyan T, et al. Network meta-analysis of
carcinoma in situ (DCIS) versus invasive breast cancer alone.
margin threshold for women with ductal carcinoma in situ.
Anticancer Res 2005; 25: 1719–23.
J Natl Cancer Inst 2012; 104: 507–16.
54 Park K, Han S, Kim HJ, Kim J, Shin E. HER2 status in pure ductal
34 Vicini F, Kestin LL, Goldstein NS, et al. Relationship between
carcinoma in situ and in the intraductal and invasive components
excision volume, margin status and tumor size with the
of invasive ductal carcinoma determined by fluorescence in situ
development of local recurrence in patients with carcinoma in situ
hybridization and immunohistochemistry. Histopathology 2006;
treated with breast conserving surgery. J Surg Oncol 2001;
48: 702–07.
76: 245–54.
55 Steinman S, Wang J, Bourne P, Yang Q, Tang P. Expression of
35 Down SK, Burger A, Hussein M. Oncological advantages of
cytokeratin markers, ER-alpha, PR, HER-2/neu, and EGFR in pure
oncoplastic breast conserving surgery in treatment of early breast
ductal carcinoma in situ (DCIS) and DCIS with co-existing invasive
cancer. Breast J 2013; 19: 5–63.
ductal carcinoma (IDC) of the breast. Ann Clin Lab Sci 2007;
36 Early Breast Cancer Trialists’ Collaborative Group. Effects of 37: 127–34.
radiotherapy and of differences in the extent of surgery for early
56 Schuetz CS, Bonin M, Clare SE, et al. Progression-specific genes
breast cancer on local recurrence and 15 year survival: an overview
identified by expression profiling of matched ductal carcinomas
of the randomised trials. Lancet 2005; 366: 2087–106.
in situ and invasive breast tumors, combining laser capture
37 Hughes L, Wang M, Page D, et al. Local excision alone without microdissection and oligonucleotide microarray analysis. Cancer Res
radiation for ductal carcinoma in situ of the breast: a trial of the 2006; 66: 5278–86.
Eastern Cooperative Oncology Group. J Clin Oncol 2009;
57 Castro NP, Osorio CABT, Torres C, et al. Evidence that molecular
27: 5319–24.
changes in cells occur before morphological alterations during the
38 Wong JS, Kaelin CM, Troyan SL, et al. Prospective study of wide progression of breast ductal carcinoma. Breast Cancer Res 2008.
excision alone for ductal carcinoma in situ of the breast. 10: R87.
J Clin Oncol 2006; 24: 1031–36.

www.thelancet.com/oncology Vol 14 August 2013 e356

 Review

58 Gauthier ML, Berman HK, Miller C, et al. Abrogated response to 66 Goodwin A, Parker S, Ghersi D, et al. Post-operative radiotherapy
cellular stress identifies DCIS associated with subsequent tumor for ductal carcinoma in situ of the breast.
events and defines basal-like breast tumours. Cancer Cell 2007; Cochrane Database Syst Rev 2009; 4: CD000563.
12: 479–91. 67 Carlson RW, Berry DA, Burstein HJ, et al. NCCN clinical practice
59 Kerlikowske K, Molinaro AM, Gauthier ML, et al. Biomarker guidelines in oncology: breast cancer, 2008. http://www.nccn.org/
expression and risk of subsequent tumors after initial ductal professionals/physician_gls/pdf/breast.pdf (accessed June 19, 2013).
carcinoma in situ diagnosis. J Natl Cancer Inst 2010; 102: 627–37. 68 Allred DC, Anderson SJ, Paik S, et al. Adjuvant tamoxifen reduces
60 Joh JE, Esposito NN, Kiluk JV, et al. The effect of Oncotype DX subsequent breast cancer in women with estrogen receptor-positive
recurrence score on treatment recommendations for patients with ductal carcinoma in situ: a study based on NSABP Protocol B-24.
estrogen receptor-positive early stage breast cancer and correlation J Clin Oncol 2012; 30: 1268–73.
with estimation of recurrence risk by breast cancer specialists. 69 Cuzick J. IBIS II: a breast cancer prevention trial in post-
Oncologist 2011; 16: 1520–26. menopausal women using the aromatase inhibitoranastrozole.
61 Solin L, Gray R, Baehner F, et al. A quantitative multigene RT-PCR Expert Rev Anticancer Ther 2008; 8: 1377–85.
assay for predicting recurrence risk after surgical excision alone 70 Schwartz G. Treatment of subclinical ductal carcinoma in situ of
without irradiation for ductal carcinoma in situ (DCIS): a the breast by local excision and surveillance: an updated personal
prospective validation study of the DCIS Score from ECOG E5194. experience. In: Silverstein MJ, Recht A, Lagios M, eds. Ductal
34th Annual San Antonio Breast Cancer Symposium; San Antonio, carcinoma in situ of the breast. Philadelphia, PA: Lippincott
TX; Dec 6–10, 2011. S4–6 (abstr). Williams & Wilkins, 2002: 308–21.
62 Sakr RA, Chandarlapaty S, Giovanni C et al. Molecular predictors 71 Rudloff U, Jacks LM, Goldberg JI, et al. Nomogram for predicting
for type of recurrence following conservative treatment for DCIS. the risk of local recurrence after breast conserving surgery for
35th Annual San Antonio Breast Cancer Symposium; San Antonio, ductal carcinoma in situ. J Clin Oncol 2010; 30: 5319–24.
TX; Dec 4–8, 2012. PD04–05 (abstr). 72 Olivoto IA, Bajdik CD, Ravdin PM, et al. Population-based
63 Williams KE, Barnes NLP, Cheema K, Dimipoulos N, Landberg G, validation of the prognostic model ADJUVANT! for early breast
Bundred NJ. Molecular phenotypes of DCIS predict invasive and cancer. J Clin Oncol 2005; 12: 2716–25.
DCIS recurrence. 35th Annual San Antonio Breast Cancer 73 Wishart GC, Bajdik CD, Dicks E, et al. PREDICT Plus: development
Symposium; San Antonio, TX; Dec 4–8, 2012. PD04–06 (abstr). and validation of a prognostic model for early breast cancer that
64 Early Breast Cancer Trialists Collaborative Group. Overview of the includes HER2. Br J Cancer 2012; 107: 800–07.
randomized trials of radiotherapy in ductal carcinoma in situ of the
breast. J Natl Cancer Inst Monogr 2010; 41: 162–77.
65 Viani GA, Stefano EJ, Afonso SL, et al. Breast-conserving surgery
with or without radiotherapy in women with ductal carcinoma in
situ: a meta-analysis of randomized trials. Radiat Oncol 2007;
2: 28–39.

e357 www.thelancet.com/oncology Vol 14 August 2013