Post-operative radiotherapy for ductal carcinoma in situ of the breast(review)

Transcription

1 Cochrane Database of Systematic Reviews Post-operative radiotherapy for ductal carcinoma in situ of the breast(review) GoodwinA,ParkerS,GhersiD,WilckenN GoodwinA,ParkerS,GhersiD,WilckenN. Post-operative radiotherapy for ductal carcinoma in situ of the breast. Cochrane Database of Systematic Reviews 2013, Issue 11. Art. No.: CD DOI: / CD pub7. Post-operative radiotherapy for ductal carcinoma in situ of the breast(review) Copyright 2013 The Cochrane Collaboration. Published by John Wiley& Sons, Ltd.

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure Figure Figure Figure Figure Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 1 All ipsilateral recurrence.. 27 Analysis 1.2. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 2 Ipsilateral Invasive recurrence. 28 Analysis 1.3. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 3 Ipsilateral DCIS recurrence. 28 Analysis 1.4. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 4 Incidence of Ipsilateral recurrence by surgical excision Analysis 1.5. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 5 Incidence of ipsilateral breast recurrence by age Analysis 1.6. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 6 Incidence of Ipsilateral Recurrence by histology Analysis 1.7. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 7 Incidence of All Ipsilateral Recurrence Analysis 1.8. Comparison 1 Post-operative radiotherapy versus surgery alone, Outcome 8 Size of DCIS lesion ADDITIONAL TABLES APPENDICES WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT DIFFERENCES BETWEEN PROTOCOL AND REVIEW INDEX TERMS i

3 [Intervention Review] Post-operative radiotherapy for ductal carcinoma in situ of the breast Annabel Goodwin 1, Sharon Parker 2, Davina Ghersi 3, Nicholas Wilcken 4 1 Medical Oncology/Cancer Genetics, Concord Hospital, Concord, Australia. 2 Centre for Primary Health Care and Equity, University of NSW, Sydney, Australia. 3 Research Translation Group, National Health and Medical Research Council, Canberra, Australia. 4 Medical Oncology, Westmead and Nepean Hospitals, Westmead, Australia Contact address: Nicholas Wilcken, Medical Oncology, Westmead and Nepean Hospitals, Department of Medical Oncology and Palliative Care, Westmead Hospital, Westmead, NSW, 2145, Australia. nicholas.wilcken@sydney.edu.au. Editorial group: Cochrane Breast Cancer Group. Publication status and date: Stable (no update expected for reasons given in What s new ), published in Issue 11, Citation: Goodwin A, Parker S, Ghersi D, Wilcken N. Post-operative radiotherapy for ductal carcinoma in situ of the breast. Cochrane Database of Systematic Reviews 2013, Issue 11. Art. No.: CD DOI: / CD pub7. Background A B S T R A C T The addition of radiotherapy (RT) following breast conserving surgery (BCS) was first shown to reduce the risk of ipsilateral recurrence in the treatment of invasive breast cancer. Ductal carcinoma in situ (DCIS) is a pre-invasive lesion. Recurrence of ipsilateral disease following BCS can be either DCIS or invasive breast cancer. Randomised controlled trials (RCTs) have shown that RT can reduce the risk of recurrence, but assessment of potential long-term complications from addition of RT following BSC for DCIS has not been reported for women participating in RCTs. Objectives To summarise the data from RCTs testing the addition of RT to BCS for treatment of DCIS to determine the balance between the benefits and harms. Search methods We searched the Cochrane Breast Cancer Group Specialised Register (2 June 2011), Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2008, Issue 1), MEDLINE (2 June 2011), EMBASE (2 June 2011) and the World Health Organization s International Clinical Trials Registry Platform (WHO ICTRP; 2 June 2011). Reference lists of articles and handsearching of ASCO (2007), ESMO (2002 to 2007), and St Gallen (2005 to 2007) conferences were performed. Selection criteria RCTs of breast conserving surgery with and without radiotherapy in women at first diagnosis of pure ductal carcinoma in situ (no invasive disease present). Data collection and analysis Two authors independently assessed each potentially eligible trial for inclusion and its quality. Two authors also independently extracted data from published Kaplan-Meier analysis (survival curves) and reported summary statistics. Data were extracted and pooled for four trials. Data for planned subgroups were extracted and pooled for analysis.there were insufficient data to pool for long-term toxicity from radiotherapy. 1

4 Main results Four RCTs involving 3925 women were identified and included in this review. All were high quality with minimal risk of bias. Three trials compared the addition of RT to BCS. One trial was a two by two factorial design comparing the use of RT and tamoxifen, each separately or together, in which participants were randomised in at least one arm. Analysis confirmed a statistically significant benefit from the addition of radiotherapy on all ipsilateral breast events (hazards ratio (HR) 0.49; 95% CI 0.41 to 0.58, P < ), ipsilateral invasive recurrence (HR 0.50; 95% CI 0.32 to 0.76, p=0.001) and ipsilateral DCIS recurrence (HR 0.61; 95% CI 0.39 to 0.95, P = 0.03). All the subgroups analysed benefited from addition of radiotherapy. No significant long-term toxicity from radiotherapy was found. No information about short-term toxicity from radiotherapy or quality of life data were reported. Authors conclusions This review confirms the benefit of adding radiotherapy to breast conserving surgery for the treatment of all women diagnosed with DCIS. No long-term toxicity from use of radiotherapy was identified. P L A I N L A N G U A G E S U M M A R Y Post-operative radiotherapy for ductal carcinoma in situ Ductal carcinoma in situ (DCIS) is characterised by the development of cancerous cells in the milk ducts of the breast and is commonly diagnosed by mammography screening. Surgical removal of the breast offers a good prognosis, however many women and clinicians prefer breast conserving surgery (BCS), the removal of the DCIS plus a rim of normal breast tissue, as there is no guarantee that DCIS will progress to invasive cancer. This approach means that most of the normal breast is saved. The main risk of inadequately removing all the DCIS is either a recurrence of DCIS or the development of invasive breast cancer at a later time with the risk that this can progress to metastatic disease (cancer that has spread). Radiotherapy (RT) is treatment using ionising radiation. Giving RT after BCS is thought to reduce the risk of developing recurrent disease (either DCIS or invasive breast cancer). This review aimed to assess both the benefit of adding RT to treatment and any potential long or short-term harm it may cause. Shortterm harm includes skin rash and redness, or inflammation of lung tissue. Potential long-term side effects from RT include vascular disease (heart and major blood vessel disease), damage to the lungs, development of lung cancer, or osteoradionecrosis (bone damage resulting in bone death). The review identified four large randomised controlled trials (3925 women) that compared treatment with breast conserving surgery alone and breast conserving surgery with the addition of RT. The addition of RT reduced the risk of a recurrence of either DCIS or invasive cancer in the treated breast by 51%. Older trials of breast conserving surgery followed by RT for invasive breast cancer have shown long-term toxicity from the addition of RT. We found no evidence of increased toxicity from the use of RT although some trials did not report on the causes of non-breast cancer deaths (deaths which potentially could be related to side effects). The number of non-breast cancer deaths reported were similar in both radiotherapy and control groups. Changes in delivery of RT between older and more recent trials and a subsequent decrease in exposure of normal tissue may account for this finding. Longer follow up of trial participants is required before a definite conclusion can be drawn, however radiotherapy techniques are continuing to improve and future patients are likely to experience a further decrease in exposure of nearby normal tissues. Overall survival was high and similar between each group whether radiotherapy was used or not. There were no reports of short-term toxicity from use of radiotherapy, or quality of life data. B A C K G R O U N D Description of the condition Ductal carcinoma in situ (DCIS) of the breast is a pre-malignant condition defined by Harris et al as a group of closely related lesions whose common histological feature is the proliferation of cancer cells within the ducts without invasion of the surround- 2

5 ing stromal tissue (Harris 1992). DCIS is a precursor to invasive breast cancer. There are shared chromosomal changes between adjacent in situ and invasive cancers that demonstrate their clonal, evolutionary relationship (Burstein 2004). It is estimated that up to 35% of individuals with DCIS may develop an invasive carcinoma over 10 years (Levine 2001). DCIS itself presents a very low risk of metastatic disease. However, about half the subsequent local recurrences (recurrence in the same area) will be invasive. For an invasive recurrence there is a 15% to 20% risk of metastatic disease at eight years (Solin 2001) and reduced survival. The 8-year rate of breast cancer specific mortality reported by Silverstein et al (Silverstein 1998) for 35 patients with an invasive local recurrence was 14.4%. DCIS is not normally palpable (NBCC 2003) and prior to the introduction of widespread mammographic screening programmes the proportion of patients diagnosed with breast cancer who had DCIS was relatively small. DCIS is often found before clinical symptoms in women attending screening mammography by detection of microcalcifications or a soft-tissue density, or both. Screening has increased the numbers of women diagnosed with DCIS from around 4% of all breast biopsies to 10% (Millikan 1995) and DCIS now accounts for 20% to 30% of all mammographically detected breast cancers (Mokbel 2002). The overall rate of breast cancer recurrence, including ipsilateral (the same breast) and contralateral (the other breast) disease, is 1% to 2% for all women diagnosed with DCIS. For those women diagnosed with DCIS who have mastectomy the rate is 0.5% to 1% per year in the contralateral breast (either DCIS or invasive disease). After breast conserving surgery the local recurrence rate is approximately 16% at five years, which reduces to approximately 8% with the addition of radiotherapy (RT) (Burstein 2004). Several factors are postulated to be associated with an increased risk of a local recurrence after BCS and RT. The strongest factor is whether DCIS has been completely excised or not. Margins that are clear or are greater than 1 mm from the margin are associated with a lower risk of local recurrence than those with margins that are either involved or less than 1 mm in width. Another factor associated with a higher risk of local recurrence is age (less than 40 years); however surgical margins, higher tumour grade, or necrosis and more advanced clinical disease are all possible confounders. Pathological features that may increase the risk of local recurrence are higher grade DCIS, especially with the presence of comedo necrosis (larger, faster growing, consisting of a higher nuclear grade and more likely to produce microinvasion), and an area of DCIS greater than 4 cm. Presentation of more advanced disease with clinical symptoms such as a palpable mass or nipple discharge, rather than a screen-detected lesion, is also associated with a higher local recurrence risk. Description of the intervention When post-operative RT is used after breast conserving surgery, the standard dose of RT that is administered in tangential fields is 44 to 50 Gray (Gy) delivered to the whole breast in fractions of 180 to 200 cgy on a daily basis. Some patients may receive a boost to the site of DCIS (a higher dose of radiation than the dose given to the surrounding breast tissue). A boost to the tumour site has been shown to reduce the risk of a local recurrence for patients who have an invasive breast cancer, and this technique has been extrapolated to use for DCIS (Burstein 2004). Why it is important to do this review Little is known of the natural history of DCIS. It is not possible to identify which cases of DCIS will progress to invasive cancer but estimates indicate that women who have had DCIS are four to 12 times more likely to develop subsequent invasive breast cancer (NBCC 2003). In terms of management, mastectomy presents a near total avoidance of recurrence (Morrow 2002) without the need for RT but requires more surgery. RT is frequently used because of the potential risk of recurrence following breast conserving surgery. Randomised controlled trials have shown a lower risk of recurrence with this approach. What remains unclear is the balance between reducing recurrence with potential long-term and short-term harms due to RT. The decision to choose mastectomy rather than breast conserving surgery followed by RT involves consideration of the psychological impact of having a mastectomy as well as the reduction in risk of disease recurrence (Reaby 1998). All patients spend a significant amount of time attending hospital for treatment, which impacts on daily life. Short-term complications of RT include the temporary discomfort of a skin reaction (such as a rash or redness) and radiation pneumonitis. The recognised long-term complications from RT following breast conserving surgery for DCIS would be similar to those following surgery for invasive breast cancer since RT for DCIS follows the same treatment protocol as for invasive breast cancer. These include a modest increase in vascular deaths (EBCTCG 2000), pulmonary toxicity including lung cancer and pulmonary fibrosis (a persistent cough or breathlessness), and contralateral breast cancer. It is also recognised that the newer RT techniques in use today, such as conformal fields, result in a lower RT dose to the surrounding normal tissues and may result in fewer long-term complications. Over the last 10 to 15 years, RT technique improvements have reduced the incidence of pneumonitis and pulmonary fibrosis, and in all probability vascular problems. It is, therefore, important to evaluate whether the addition of RT to breast conserving surgery influences both local recurrence and overall survival. It is also important to weigh up benefits in terms of local control and survival with the side effects experienced with this treatment. There has been no direct prospective randomised controlled trial comparing breast conserving surgery with mastectomy for DCIS. 3

6 There have been four large randomised controlled trials comparing breast conserving surgery alone with the addition of radiotherapy. These are the NSABP B-17, EORTC10853, UKCCCR, and SweDCIS trials. All concur that the addition of radiotherapy reduces the risk of local recurrence in all subgroups. Other recent changes to radiotherapy include the addition of a boost to the site of DCIS, which is especially used for patients with involved margins. Types of interventions Any trial in which RT (of any kind) was the primary treatment comparison after BCS. This included: BCS (either lumpectomy, quadrantectomy, or segmental mastectomy) with or without RT (any standard schedule of treatment). Trials where patients received tamoxifen were included provided this was given in both study arms and patients differed only in respect to receiving radiotherapy or not. O B J E C T I V E S To summarise the data from randomised controlled trials testing the addition of radiotherapy (RT) to breast conserving surgery (BCS) for treatment of DCIS to determine the balance between benefits and harms. M E T H O D S Criteria for considering studies for this review Types of studies 1. Randomised controlled clinical trials (RCTs). 2. Any trial including only patients with pure DCIS (no invasive cancer), or any trial including patients with pure DCIS which stratified by absence or presence of DCIS and where patients with pure DCIS could be separated out. 3. Any trial comparing BCS (lumpectomy, quadrantectomy, segmental mastectomy) with or without RT. 4. Trials where patients received tamoxifen were included provided this was given in both study arms and patients differed only in respect to receiving radiotherapy or not. Types of participants Women with a histological diagnosis of DCIS for the first time (not recurrent or metastatic disease) with no prior history of malignant disease (other than in situ carcinoma of the cervix, or Basal Cell Carcinoma (BCC) or Squamous Cell Carcinoma (SCC) of the skin). No invasive breast cancer. No age limit. Types of outcome measures Primary outcomes Efficacy: ipsilateral local recurrence (both DCIS and invasive cancer); contralateral breast cancer (both DCIS and invasive cancer); metastatic disease (if reported); breast cancer mortality; all-cause mortality. Toxicity: long-term complications from radiotherapy; vascular mortality (including acute myocardial infarction, sudden cardiac death, congestive cardiac failure, and cerebrovascular disease); pulmonary toxicity; lung cancer; osteoradionecrosis. Secondary outcomes Quality of life assessment (individual trial instruments), if available. Search methods for identification of studies Electronic searches See: Cochrane Breast Cancer Group methods used in reviews. There were no language or publication restrictions. (a) Cochrane Breast Cancer Group Specialised Register. The Specialised Register was searched (2 June 2011). Details of the search strategy used by the Group for the identification of studies for the Register, and the procedure used to code references, are outlined in the Group s module ( clabout/articles/breastca/frame.html). Studies with the keywords radiotherapy, dcis, or ductal carcinoma or text words dcis, ductal carcinoma in situ, or intraductal carcinoma in situ were extracted for consideration. 4

7 (b) MEDLINE (2007 to February 2008). See Appendix 1. For the 2011 review update, MEDLINE was searched until 2 June (c) EMBASE via OVID (2002 to February 2008). See Appendix 2. For the 2011 review update, EMBASE via OVID SP was searched from February 2008 until 2 June See Appendix 3 for the updated search strategy. (d) Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2008, Issue 1). See Appendix 4. (e) The WHO International Clinical Trials Registry Platform (IC- TRP) search portal ( for all prospectively registered and ongoing trials on 2 June See Appendix 5. Searching other resources (a) References from published studies. These were checked for further trials. (b) Conference proceedings. Relevant breast cancer conference proceedings are regularly handsearched and identified studies are included in the Cochrane Breast Cancer Group Specialised Register. The conference proceedings not yet included in the Register and so handsearched were St Gallen (2005 to 2007), American Society of Clinical Oncology (ASCO) (2007), and European Society of Medical Oncology (ESMO) (2002 to 2007). (c) Experts in the field. If necessary, experts in the treatment of DCIS were to be contacted in an attempt to identify other potential data sources. Data collection and analysis Selection of studies In the original and 2011 review update, two authors (AG and NW) independently assessed each potentially eligible trial for: (a) inclusion in the review (according to the eligibility criteria) and (b) quality. A third author was to resolve any discrepancies regarding eligibility or quality (there were none). Where necessary, additional information was sought from the principal investigator of the trial concerned. Data extraction and management Two authors (AG and SP) independently extracted data from the included studies using standard extraction forms. If required, a third author (NW) resolved any discrepancies regarding the extraction of quantitative data. For those studies with more than one publication, we extracted data from all the publications but considered the final or updated version of each study as the primary reference. Assessment of risk of bias in included studies Studies fulfilling the eligibility criteria were independently evaluated for quality using the Risk of bias tool described in the Cochrane Handbook for Systematic Reviews of Interventions (Cochrane Collaboration 2008). Given the nature of the intervention, it was not possible nor practical to expect blinding of the intervention or outcome assessment. We assessed the following four domains of the risk of bias tool: sequence generation, allocation concealment, blinding, and incomplete outcome data. Measures of treatment effect All time-to-event outcome data were extracted as follows. Two authors (AG and SP) extracted the hazard ratio (HR) and associated variances for all ipsilateral breast events indirectly using the methods described by Parmar For three trials we used data extracted from published Kaplan-Meier curves (NSABP 2001; SweDCIS 2008; UKCCCR 2003). For the EORTC trial (EORTC 2006) we used available summary statistics to determine the HR and its variance as there was no survival curve for all ipsilateral breast events. Time to event analyses were conducted, if possible, for: (i) time to local recurrence (both DCIS and invasive breast cancer); (ii) time to breast cancer recurrence including contralateral breast cancer and metastatic disease (counting death as failure); (iii) time to breast cancer death; and (iv) time to death (any cause). All extracted data were entered into Review Manager software (RevMan 2008). Dealing with missing data Where necessary, additional information was sought from the principal investigator of the trial concerned. Data synthesis All data were analysed by intention to treat. A fixed-effect model was used for the primary analysis. Subgroup analysis and investigation of heterogeneity We conducted the following subgroup analyses of risk of local recurrence: (i) surgical margins clear or not; (ii) age under 50 years compared to over 50 years; and (iii) documented presence of marked comedo necrosis or not. R E S U L T S 5

8 Description of studies Six trials fulfilled the inclusion criteria. Two of the trials identified have closed to accrual but follow-up is continuing and no data are available from these trials (NCT ; RTOG 98). Four trials were included in our review (EORTC 2006; NSABP 2001; SweDCIS 2008; UKCCCR 2003). Results of the search Trials were selected for inclusion by two authors (AG and NW) and there were no disagreements regarding trial inclusion. The Cochrane Breast Cancer Group Specialised Register search (17th January 2008) identified 35 potentially eligible trials. The search of CENTRAL identified 150 potentially eligible trials, the MEDLINE search identified 124 potentially eligible trials, and the EMBASE search identified 307 potentially eligible trials. The review of conference abstracts and reference lists did not identify any additional trials. Two authors (AG and NW) read the Journal of Clinical Oncology and identified a more recent update of the results of one trial (SweDCIS 2008) that was not identified as part of the search strategy. In total, 617 references were screened for inclusion in our meta-analysis (Figure 1). 6

9 Figure 1. 7

10 According to our selection criteria, 45 references to potentially eligible trials were identified and reviewed for inclusion. Seven trials were excluded: four trials included patients with invasive breast cancer only (Assertohn 1999; Clark 1992; Goel V 1997; Lee 1994), two were references to trial registers for which we were unable to locate any results, one trial was not a RCT (Silverstein 1991). Two trials identified from a trial register are ongoing and the investigators were contacted, Mr Ronald Kaggwa (NCT ) and Dr Beryl McCormick (RTOG 98); no results are available (see ongoing trials). Sixteen were duplicate reports of the same trial reference. Twenty references to eligible RCTs were identified. In total these 20 references reported on four trials which were subsequently included in our meta-analysis. In the 2011 update, a repeat search was carried out and a recent meta-analysis by the Early Breast Cancer Trialists Collaborative Group was found. We have decided not to replicate these efforts in the 2011 review update. Included studies Four studies were included in the review (See the Characteristics of included studies table). The EORTC trial (EORTC 2006): RCT investigating the role of RT after local excision of ductal carcinoma in situ (DCIS) of the breast. Participants in this multicentre trial were 1010 women recruited in Europe between March 1986 and July The trial has published data with a median follow up of 10.5 years. The NSABP trial (NSABP 2001): RCT investigating the role of RT after lumpectomy for DCIS of the breast. Participants in this multicentre trial were 818 women recruited in the United States of America and Canada between October 1985 and December The trial has a mean follow up of 10.7 years. The SweDCIS trial (SweDCIS 2008): RCT investigating the role of RT after sector resection for DCIS of the breast. Participants in this multicentre trial were 1067 women recruited in Sweden between September 1987 and December The trial has a mean follow-up data for a mean of 8.4 years. The UKCCCR trial (UKCCCR 2003): a multicentre RCT investigating the effectiveness of adjuvant RT and tamoxifen for DCIS of the breast. This trial has a 2 by 2 factorial design. Patients were either randomised to both treatments, or randomised to either one with an elective decision regarding the other treatment. Recruitment from breast screening programmes commenced in May 1990 in the United Kingdom and September 1991 in Australia and New Zealand and concluded in August In total, 1030 women were randomised to radiotherapy or no radiotherapy in this trial that included 1701 women. The trial has published follow-up data for a median follow up of 4.4 years. Excluded studies One trial was not a randomised trial (Silverstein 1991). Two references retrieved were registered trial reports and the authors were unable to locate any further details about these trials. Four trials included patients with invasive breast cancer only (Assertohn 1999; Clark 1992; Goel V 1997; Lee 1994). In the 2011 review update, one trial by Polgar 2008 was excluded. Ongoing studies Two trials are ongoing (see the table of ongoing studies). One is a randomised multicentre phase II trial (NCT ) to compare ipsilateral breast cancer relapse and breast cancer metastasis. The trial includes women with completely excised low-risk oestrogen or progesterone receptor positive DCIS who are receiving adjuvant tamoxifen or anastrozole and randomised to adjuvant radiotherapy or observation. Contact was made with a trial investigator (Mr Ronald Kaggwa). This trial has closed to accrual and patient numbers are small (63). There are no results currently available. We contacted a trial investigator (Dr Beryl McCormick) for the second trial (RTOG 98). This trial has closed to recruitment with just over 600 patients. The recruitment target was 1800 and, due to limited follow up, no results are currently available. Risk of bias in included studies The risk of bias was low as the quality of all four included studies was high (see Risk of bias table). Incomplete outcome data The SweDCIS trial (SweDCIS 2008) excluded 21 randomised participants from the analysis due to protocol violations, however this is unlikely to influence the results. The NSABP trial (NSABP 2001) excluded five randomised participants as they were lost to follow up (two in the control and three in the RT arm). The UKCCCR trial (UKCCCR 2003) included all randomised participants. The EORTC trial (EORTC 2006) excluded eight participants who were lost to follow up but included ineligible participants who were randomised (analysis with and without these participants did not change the results). Effects of interventions Benefit of radiotherapy (RT) 8

11 The presence of Kaplan-Meier survival curves for ipsilateral breast recurrence for three trials (EORTC 2006; NSABP 2001; UKCCCR 2003) and summary statistics (HR, the 95% confidence interval (CI) and observed events in each group) for the EORTC trial (EORTC 2006) made pooled analysis of the HR possible using the methods published by Parmar (Parmar 1998). We calculated the HR and its associated variance for each of the four trials for all ipsilateral breast events and pooled this result using RevMan 5. The pooled HR for all ipsilateral breast events (Analysis 1.1; Figure 2) was statistically significant in favour of RT (HR 0.49; 95% CI 0.41 to 0.58, P < ). Figure 2. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.1 All ipsilateral recurrence. Individual trial results were all consistent with the pooled HR for all ipsilateral breast events. The EORTC trial (EORTC 2006) found the 10-year local relapse-free rate was 85% with adjuvant radiotherapy compared to 74% without (HR 0.53, log rank P < 0.001). The UKCCCR trial (UKCCCR 2003) found the absolute risk of all ipsilateral events was reduced by 8.9% (from 13.7% in the control group to 4.8% in the RT group). The SweDCIS trial (SweDCIS 2008) found the absolute risk reduction was 16% at 10 years, corresponding to a relative risk (RR) of 0.40 (95% CI 0.30 to 0.54), for all ipsilateral breast events. The NSABP trial (NSABP 2001) found the cumulative incidence of all ispilateral events at 12 years was 31.7% in the control group compared to 15.7% in the RT group (RR 0.43; 95% CI 0.32 to 0.58, P < ). There were Kaplan-Meier survival curves for ipsilateral invasive recurrence for two of the four trials (NSABP 2001; UKCCCR 2003). The pooled HR for ipsilateral invasive recurrence (Analysis 1.2; Figure 3) was statistically significant (HR 0.50; 95% CI 0.32 to 0.76, P = 0.001). The pooled results of the two trials showed a trend towards benefit which was consistent with each individual trial s result. The EORTC trial (EORTC 2006) found the risk of invasive local recurrence was reduced by 42% (P = ). The NSABP trial (NSABP 2001) found that invasive breast tumour recurrence was reduced from 16.8% to 7.7% (P < ) (RR 0.38; 95% CI 0.25 to 0.59, P < ). The UKCCCR trial (UKCCCR 2003) found that the HR for use of RT and incidence of ipsilateral invasive disease was 0.45 (95% CI 0.24 to 0.85, P = 0.01); in absolute terms reduced from 5.3% to 2.5%. The SweDCIS (SweDCIS 2008) found ipsilateral invasive disease was reduced from 12.3% to 7.2%. Figure 3. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.2 Ipsilateral Invasive recurrence. 9

12 There were Kaplan-Meier survival curves for ipsilateral DCIS recurrence for two trials (NSABP 2001; UKCCCR 2003). The pooled HR for ipsilateral DCIS recurrence (Analysis 1.3; Figure 4) was just stastically significant in favour of RT (HR 0.61; 95% CI 0.39 to 0.95, p=0.03). This was consistent with each individual trial result. The NSABP trial (NSABP 2001) found the incidence of non-invasive ipsilateral breast tumour recurrence was reduced from 14.6% to 8.0% (RR 0.49; 95% CI 0.32 to 0.76, P = 0.001). The EORTC trial (EORTC 2006) found the risk of non-invasive recurrence was reduced by 48% (P = ); in absolute terms from 14% to 7%. The UKCCCR trial (UKCCCR 2003) found the HR for ipsilateral DCIS was 0.36 (95% CI 0.19 to 0.66, P = ). The SweDCIS (SweDCIS 2008) found ipsilateral DCIS was reduced from 14.8% to 4.9%. Figure 4. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.3 Ipsilateral DCIS recurrence. All trials reported excellent survival in both the RT and control groups. In the NSABP trial (NSABP 2001), overall survival was 86% (control) compared to 87% (RT) at 12 year follow up. In the EORTC trial (EORTC 2006), 10-year survival was 95% in both groups, and 10-year metastasis free survival 96% in both groups. SweDCIS trial (SweDCIS 2008) did not report survival but the number of women with either breast cancer death or other death was similar in each group (8.4% in the RT group compared to 9.6% in the control group, corresponding to a survival of 92% and 90% respectively). The UKCCCR trial (UKCCCR 2003) did not report survival according to randomisation to RT or not; including all trial participants (not all randomised to RT) survival was 97.9% in the control group and 96.2% in the RT group. Contralateral breast events were similar in both RT and control groups for all trials. In the NSABP trial (NSABP 1998) rates were 3.3% in the control group and 5.7% in the RT group (overall 3% invasive and 1.5% non-invasive). Both groups were combined for the 12 year follow-up (NSABP 2001). In the EORTC trial (EORTC 2006) rates were 5.6% in the control group and 7.7% in the RT group (95% CI 0.87 to 2.30, P = 0.16). The UKCCCR trial (UKCCCR 2003) found similar contralateral breast event rates (2%) in each group (HR 0.82, P = 0.65). SweDCIS trial (SweDCIS 2008) reported 6.5% with contralateral breast carcinoma in the RT arm and 5.9% in the control arm. Subgroup analysis of benefit The EORTC trial (EORTC 2006) and SweDCIS trial (SweDCIS 2008) both conducted subgroup analyses of the benefit from RT. The EORTC trial found using multivariate analysis that young age (40 years or less), intermediately or poorly differentiated DCIS, cribiform or solid growth pattern, or doubtful margins were risk factors for recurrence; all subgroups benefited from addition of RT. The SweDCIS trial found (on subgroup analysis) minimal benefit for RT in women under the age of 50 years (who had a 6% risk reduction) but marked benefit for RT in women over age 60 years (with an 18% risk reduction) that was not confounded by focality, lesion size, completeness of excision, or detection mode. A separate report from the NSABP trial (1999) found a benefit for RT in all groups according to the pathological feature studied, but the presence or absence of comedo necrosis was an independent predictor of recurrence risk. When moderate or marked comedo 10

13 necrosis was present the RR was 0.30, compared to a RR of 0.52 in its absence. We planned to analyse the benefit of RT in the following subgroups: whether DCIS was completely excised or not, young age or older age, and the presence or absence of comedo necrosis. We performed an unplanned analysis based upon lesion less than 10mm compared to greater or equal to 10mm of DCIS. No subgroup information was provided for the UKCCCR trial (UKCCCR 2003). Information regarding completeness of excision was available for 2272 women: 73.5% of participants in the EORTC trial (EORTC 2006), 70.8% of participants in the NSABP trial (NSABP 2001), and 89.2% of participants in the SweDCIS trial (SweDCIS 2008). Data were extracted and pooled for these partipants (Analysis 1.4; Figure 5). In the group with incomplete excision the calculated RR was 0.55 (95% CI 0.38 to 0.78, P = 0.001), and with complete excision the RR was 0.48 (95% CI 0.39 to 0.59, P < ). Figure 5. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.4 Incidence of Ipsilateral recurrence by surgical excision. The trials defined young age differently. The EORTC trial (EORTC 2006) reported data for women under or over age 40 years. In the small group of 41 women under 40 years, the HR calculated from reported summary statistics was 0.33 (95% CI 0.13 to 0.88, P = 0.03) and for women over age 40 years the HR was 0.59 (95% CI 0.43 to 0.81, P = 0.001). The NSABP trial (NSABP 2001) and SweDCIS trial (SweDCIS 2008) reported data on women under or over age 50 years. Data were pooled for these two trials (Analysis 1.5; Figure 6). For women under age 50 years the RR was 0.67 (95% CI 0.48 to 0.93, P = 0.02) and for women over age 50 years the RR was 0.35 (95% CI 0.27 to 0.46, P < ). 11

14 Figure 6. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.5 Incidence of ipsilateral breast recurrence by age. Ipsilateral breast event rates according to histological subtype were available for two trials: the NSABP trial (NSABP 2001), available for 83.1% of trial participants; and the EORTC trial (EORTC 2006), available for 76.4% of trial participants. Data were pooled (Analysis 1.6; Figure 7) and a benefit from RT was found both in the presence and absence of comedo necrosis. In the presence of comedo necrosis the overall RR was 0.44 (95% CI 0.33 to 0.59, P < ). In the absence of comdeo necrosis the RR was 0.60 (95% CI 0.45 to 0.81, P = ). The SweDCIS trial (SweDCIS 2008) reported on a cohort of 271 women (25.4% of trial participants): if comedo necrosis was present the HR was 0.40 (95% CI 0.23 to 0.67), and if comedo necrosis was absent the HR was 0.07 (95% CI 0.01 to 0.42) with P = for the test for interaction. 12

15 Figure 7. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.6 Incidence of Ipsilateral Recurrence by histology. Ipsilateral breast event rates were calculated based upon the size of the original tumour as an unplanned analysis in response to requests for this information. Three trials (NSABP 2001, EORTC 2006 and SweDCIS 2008) had reports of the trials which included event rates for ipsilateral breast events based on DCIS size and data were pooled (Analysis 1.8; Figure 8). This information was available for 1742 of the 3925 participants (44.4%). The size of DCIS was less than 10 mm for 903 participants and greater or equal to 10 mm for 839 participants. For DCIS less than 10mm the HR was 0.43 (0.30 to 0.61). For DCIS greater than or equal to 10 mm the HR was 0.32 (0.22 to 0.46). 13

16 Figure 8. Forest plot of comparison: 1 Post-operative radiotherapy versus surgery alone, outcome: 1.8 Size of DCIS lesion. Number required to treat We pooled results from the four trials (EORTC 2006; NSABP 2001; SweDCIS 2008; UKCCCR 2003) to calculate the incidence of all ipsilateral breast recurrence (Analysis 1.7). Ipsilateral recurrence was 11.6% for women who received RT compared to 23.9% for women treated with surgery alone. Nine women require treatment with RT to prevent one ipsilateral breast recurrence. Potential harm from radiotherapy (RT) Toxicity was an outcome of interest and all four trials were assessed for this information. For the UKCCCR trial (UKCCCR 2003), toxicity data were not provided for the intervention and control groups but only for the whole group of 1694 participants. Attempts to contact trial investigators were not successful. Information about cause of death was not reported for non-breast cancer death for the SweDCIS trial (SweDCIS 2008). Contact was made with a trial investigator and this information is not yet available. Overall, there was a low number of deaths in each of the four trials. It is recognised that longer follow up will increase the incidence of any long-term toxicity. Median follow up for ranged from 4.4 years to 10.5 years for the four trials. Three trials had information about the cause of non-breast cancer deaths. There were seven deaths due to vascular causes in the NSABP trial (NSABP 2001), six in the control arm and one in the RT arm; the first report of the EORTC trial (EORTC 2000) reported two vascular deaths (one in each arm) and subsequently (EORTC 2006) reported seven vascular deaths (arm not specified). The UKCCCR trial (UKCCCR 2003) reported six vascular deaths. There were no reports of pulmonary toxicity in any of the trials. Reports of a second cancer did not always specify the type of malignancy. In the NSABP trial (NSABP 2001), one woman died of either lung cancer or metastatic breast cancer (could not be determined); 18 second cancers occurred in the control group, including two lung cancers and three endometrial cancers; and 20 second cancers occurred in the RT group, including one lung cancer, one oesophageal cancer, one bone cancer, one soft tissue cancer, two endometrial cancers and one cancer in the lymph nodes. In the EORTC trial (EORTC 2006), there was one mediastinal lymphoma (RT arm) in the first EORTC trial report (EORTC 2000) and in the subsequent report there were 13 malignancies other than breast cancer (type and arm of study not specified). The UKCCCR trial (UKCCCR 2003) reported nine deaths due to another malignancy (type and arm of study not specified). There were no reports of osteoradionecrosis in any trial. There was no information about any short-term toxicity associated with RT in any of the four trials. There were no quality of life data reported in any of the trials. D I S C U S S I O N 14

17 Summary of main results This systematic review has confirmed the individual trial results for a benefit of adding RT to breast conserving surgery for DCIS. Results confirm a statistically significant reduction in ipsilateral breast events using data extracted and pooled from four trials that were identified by a comprehensive search strategy (HR 0.49; 95% CI 0.41 to 0.58, P < ). All four trials were large multicentre trials with excellent randomisation procedures, use of allocation concealment, and intention-to-treat analysis. Blinding of treatment was not done and is not appropriate for an intervention trial using RT. There was no heterogeneity present between trials. The characteristics of women included in these four trials reflect the population of women commonly diagnosed with DCIS (Table 1). In these studies, the majority of women had DCIS diagnosed by screening mammography and which was clinically impalpable. The reported median ages of women were just over 50 years. This further strengthens the applicability of the results of this metaanalysis to most women seen in clinics for treatment of DCIS. Overall completeness and applicability of evidence The validity of our results is high due to the quality of the included studies and our comprehensive search strategy. Two smaller trials were identified whose results are not yet available but, when available, these results are unlikely to influence the overall result. It is possible that unpublished eligible trials are reported in conference abstracts which were not included in the Cochrane Breast Cancer Group Specialised Register or handsearched and potentially could be missed by our search strategy. We welcome contact by any trial investigators if this is the case. We feel it is unlikely there are any other trials with sufficient patient numbers to influence our results. A small number of trial participants were excluded from analysis in some of the included trials but all participants were accounted for. This is, therefore, unlikely to influence any of the trial results. The benefit from the addition of RT to breast conserving surgery was homogeneous between each individual trial and was confirmed on extraction and pooling of data. The following subgroups were planned for analysis prior to our review; young age compared to older age, presence of marked or moderate comedo necrosis compared to its absence, and complete surgical excision of DCIS compared to incomplete excision. Pooling of results for all of these subgroups showed benefit from the use of RT. Data were available for extraction from two trials regarding recurrent ipsilateral DCIS and recurrent ipsilateral invasive disease (a total of 1848 participants). Even with reduced patient numbers and a lower event rate, there was a benefit from RT for ipsilateral invasive disease and for recurrent ipsilateral DCIS, the benefit for RT was just statistically significant. A benefit for the addition of RT was seen whether the size of the DCIS lesion was small (less than 10 mm) or large (greater than or equal to 10 mm). Quality of the evidence To assess the balance between the benefit from RT and possible toxicity from treatment, the cause of death was assessed for all studies. Overall, long-term harm was not well reported in these trials. With the available information, there was no evidence of excess deaths attributable to the addition of RT, either due to vascular disease, pulmonary toxicity, or second malignancies. Rate of death due to any cause was low in both arms of all trials and was similar between trials. However, if long-term toxicity due to RT does occur, a longer follow-up period may be required to show such an effect. The actual development of disease (such as vascular disease or malignancy) was not reported, only cause of death. As RT techniques continue to improve, exposure of nearby normal tissues is reduced and any long-term harm associated with the use of RT may be reduced even further for future patients. A reassuring finding was no reporting of excess deaths in the RT arm compared to the group treated by surgery alone. Short-term toxicity and quality of life was not reported in any of the trials. This review was not able to assess therefore the impact of any potential short-term harm from treatment with radiotherapy. Short term side effects of radiation therapy have however been well reported. In future it would be beneficial to women undergoing this treatment to have access to comprehensive information concerning these, and the outcomes of different treatment options which impact on quality of life and influence complex psychosocial treatment decisions. This may be studied in non-randomised controlled trials. A U T H O R S C O N C L U S I O N S Implications for practice This result confirms the benefit of radiotherapy following breast conserving surgery for DCIS and supports its use for all women as the overall benefit was large and all subgroups analysed showed benefit for the use of radiotherapy. There was no reported longterm toxicity from the use of radiotherapy and no excess deaths from any cause were reported in the groups treated with radiotherapy. Short term toxicity and quality of life were not reported. Clinicians therefore need to ensure that comprehensive information relating to potential side effects is made available to women undergoing this treatment. Implications for research Ideally, long-term follow up of patients treated in clinical trials should occur to allow identification of any long-term toxicity; including assessment of general health, such as development of vascular disease, in addition to recording cause of death. Assessment of short-term toxicity would be of benefit in assessment of the overall effect of treatment for a group with an excellent prognosis. 15

18 A C K N O W L E D G E M E N T S The authors would like to thank Nicole Holcroft for her assistance with the development of the search strategies. We would also like to thank Patrick Fitzgerald, Associate Professor Dianne O Connell, and Professor Val Gebski for advice in developing the analysis section of this protocol. R E F E R E N C E S References to studies included in this review EORTC 2006 {published data only} Bijker N, Meijnen P, Peterse J, Bogaerts J, Van Hoorebeeck I, Julien JP, et al. Breast-conserving treatment with or without radiotherapy in ductal carcinoma-in-situ: Tenyear results of European Organisation for Research and Treatment of Cancer Randomized Phase III Trial 10853, a study by the EORTC Breast Cancer Cooperative Group and EORTC Radiotherapy Group. Journal of Clinical Oncology 2006;24(21): Bijker N, Peterse JL, Duchateau L, Julien JP, Fentiman IS, Duval C, et al. Risk factors for recurrence and metastasis after breast-conserving therapy for ductal carcinoma-insitu: Analysis of European Organization for Research and Treatment of Cancer Trial Journal of Clinical Oncology 2001;19(8): Bijker N, Peterse JL, Fentiman IS, Julien JP, Hart AAM, Avril A, et al. Effects of patient selection on the applicability of results from a randomised clinical trial (EORTC 10853) investigating breast-conserving therapy for DCIS. British Journal of Cancer 2002;87: Bijker N, Rutgers EJT, Peterse JL, Fentiman IS, Julien JP, Duchateau L, et al. Variations in diagnostic and therapeutic procedures in a multicentre, randomized clinical trial (EORTC 10853) investigating breast-conserving treatment for DCIS. European Journal of Surgical Oncology 2001;27: Fentiman IS, Julien JP, Mignolet F. EORTC trial 10853: treatment options for completely excised ductal carcinoma of the breast. European Journal of Surgical Oncology 1993; 19:499. Fentiman IS, Julien JP, van Dongen JA, van Geel B, Chetty U, Coibion M. Reasons for non-entry of patients with DCIS of the breast into a randomised trial (EORTC 10853). European Journal of Cancer 1991;27(4): Julien JP, Biijker N, Fentiman IS, Peterse JL, Delledonne V, Rouanet P, et al. Radiotherapy in breast-conserving treatment for ductal carcinoma in situ: first results of the EORTC randomised phase III trial Lancet 2000; 355: NSABP 2001 {published data only} Fisher B, Dignam J, Wolmark N, Mamounas E, Constantino J, Poller W, et al. Lumpectomy and radiation therapy for the treatment of intraductal breast cancer: Findings form the National Surgical Adjuvant Breast and Bowel Project B- 17. Journal of Clinical Oncology 1993;16(2): Fisher B, Dignam J, Wolmark N, Mamounas E, Costantino J, Poller W, et al. Lumpectomy and radiation therapy for the treatment of intraduct breast cancer: Findings from National Surgical Adjuvant Breast and Bowel Project B17. Journal of Clinical Oncology 1998;16(2): Fisher B, Land S, Mamounas E, Dignam J, Fisher ER, Wolmark N. Prevention of invasive breast cancer in women with ductal carcinoma in situ: an update of the national surgical adjuvant breast and bowel project experience. Seminars in Oncology 2001;28(4): Fisher ER, Costantino J, Fisher B, Palekar A S, Redmond C, Mamounas E. Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) Protocol B17. Cancer 1995;75(6): Fisher ER, Dignam J, Tan-Chiu E, Costantino J, Fisher B, Paik S, et al. Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) eight year update of Protocol B17. Cancer 1999;86(3): Nol G, Proudhom MA, Mazeron JJ. [Lumpectomy and radiation therapy for the treatment of intraductal breast cancer: findings from National Surgical Adjuvant Breast and Bowel Project B-17]. Cancer Radiothèrapie: Journal de la Sociètè Française de Radiothèrapie Oncologique 1999;3(4): SweDCIS 2008 {published data only} Emdin SO, Granstrand B, Ringberg AR, Sandelin K, Arnesson L-G, Nordgren H, et al. SweDCIS: Radiotherapy after sector resection for ductal carcinoma in situ of the breast. Results of a randomised trial in a population offered mammography screening. Acta Oncologica 2006;45: Holmberg L, Garmo H, Granstrand B, Ringberg A, Arnesson L-G, Sandelin K, et al. Absolute risk reductions for local recurrence after post operative radiotherapy after sector resection for ductal carcinoma in situ. Journal of Clinical Oncology 2008;26(8): Ringberg A, Nordgren H, Thorstensson S, Idvall I, Garmo H, Granstrand B, et al. Histopathological risk factors for ipsilateral breast events after breast conserving treatment for ductal carcinoma in situ of the breast - Results from the Swedish randomised trial. European Journal of Cancer 2007; 43: