Post‐mastectomy radiotherapy for women with early breast cancer and one to three positive lymph nodes
Abstract
Objectives
This is a protocol for a Cochrane Review (intervention). The objectives are as follows:
To assess the effects of post‐mastectomy radiotherapy (PMRT) in women diagnosed with early breast cancer and found to have one to three positive axillary lymph nodes.
Background
Description of the condition
Breast cancer is the most common malignancy to affect women; in 2018, more than 2 million breast cancer cases were diagnosed globally, which accounted for 24.5% of all cancers in women (Bray 2018). The American Cancer Society has estimated that in 2020 around 276,480 new cases of invasive breast cancer will be diagnosed in the USA, with an estimated 42,690 deaths (Siegel 2020). Early diagnosis through screening, and advancements in supportive (adjuvant) treatments over the years, have resulted in improved outcomes for breast cancer survivors. This is reflected in the 1.5% average reduction in the age‐adjusted death rate per year between 2008 and 2017 in the USA compared to a 0.3% average rise of age‐adjusted new cases in the same period (SEER 2020).
The development of local and regional recurrence (LRR) of breast cancer is a dreaded outcome that affects 5% to 15% of women diagnosed with breast cancer after mastectomy (removal of all breast tissue) and radiotherapy (EBCTCG 2014). LRR after primary breast cancer treatment is typically associated with an increased risk of concurrent and future spread of cancer elsewhere in the body (Van Tienhoven 1999). The 10‐year relative survival (cancer survival in the absence of other causes of death) after LRR has been shown to be in the region of 25% to 50% even after attempts to remove the cancer recurrence (Van Tienhoven 1999; Chagpar 2003). The commonest area of LRR after mastectomy is the chest wall (53%) followed by lymph nodes above and below the collar bone (26%) and in the armpit (13%) (Katz 2000; Wallgren 2003; Nielsen 2006). A palpable lump in these sites, painful enlarged lymph nodes and in some cases elevated red patches on the chest wall are common symptoms. The risk of LRR after mastectomy substantially increases with the number of axillary lymph nodes containing breast cancer. The Early Breast Cancer Trialists' Collaborative Group (EBCTCG) individual participant data meta‐analysis demonstrated that the risk of LRR is more than doubled in women who had a mastectomy with one to three lymph nodes affected with cancer compared to node‐negative women with breast cancer (EBCTCG 2005).
Description of the intervention
Breast cancer is a systemic disease and radiotherapy is a well‐established additional treatment (adjuvant) that aims to reduce LRR (Fisher 1998; Fisher 2002). Radiotherapy involves radiation treatment using X‐rays or other forms of ionising radiation and doses are measured in Gray (Gy). Radiotherapy is administered in fractions to achieve maximum cancer control while attempting to reduce local complications. The standard practice in the UK for post‐mastectomy chest wall radiotherapy is 40 Gy in 15 fractions over a period of three weeks based on the evidence from the UK START A & B trials (START TRIAL A; START TRIAL B). Amongst women with node‐positive or high‐risk node‐negative breast cancer, the addition of regional nodal radiotherapy is dependent on risk stratification based on the characteristics of their cancer. The role of regional nodal irradiation has been evaluated in two large randomised controlled trials (RCTs) and both failed to identify any overall survival benefit, but did show a significant reduction in breast cancer recurrence (Poortmans 2015; Whelan 2015). In contrast, a Danish population‐based cohort study showed an overall survival benefit of 4% for those women who received radiotherapy to lymph nodes near their breast bone (internal mammary) (Thorsen 2016). However, due to the lack of conclusive evidence of overall survival benefit seen among internal mammary‐ and node‐irradiated women with breast cancer in RCTs (Hennequin 2013; Poortmans 2015), this radiotherapy is recommended only in people with a high index of suspicion of nodal involvement or confirmed significant axillary nodal metastatic cancer (Poortmans 2015; NICE 2018).
How the intervention might work
Radiotherapy has been shown to reduce the 10‐year risk of LRR by two‐thirds in women affected with breast cancer (EBCTCG 1995). However, the improvement in LRR has not been shown to translate into a consistent survival advantage. The EBCTCG overview showed that for every 1.5 LRR prevented in the first 10 years after radiotherapy, one breast cancer‐specific death could be prevented in 20 years (EBCTCG 2014). The absolute benefit in survival advantage might be much higher if we consider the advances in radiotherapy techniques over the last 20 years (EBCTCG 1995; Truong 2005). These include methods used to protect the heart, lungs and major blood vessels inside the chest wall from radiotherapy‐induced morbidity and mortality in women with breast cancer. The findings of the EBCTCG overview (EBCTCG 2000), and long‐term outcomes of some seminal RCTs (Ragaz 2005; Overgaard 2007), were considered by the St Gallen Consensus Guidelines (2009) which recommended post‐mastectomy radiotherapy for women affected with breast cancer who had a 20% or greater 10‐year risk of LRR (Goldhirsch 2009).
Why it is important to do this review
Post‐mastectomy radiotherapy (PMRT) is currently recommended for women with breast cancer who have four or more lymph nodes involved with metastatic cancer. There is still no international consensus on whether to offer PMRT in women with breast cancer and one to three axillary lymph nodes affected with cancer (low volume axillary metastatic disease). The National Comprehensive Cancer Network (NCCN) guideline recommends giving strong consideration to providing PMRT to women with one to three metastatic axillary lymph nodes (Salerno 2017). However, there is currently no risk stratification model based on demographic or cancer characteristics to enable clinicians to make these difficult choices and decisions. Age, location of the tumour in the inner aspect of the breast, nodal ratio (i.e. the number of lymph nodes with cancer versus the total number of lymph nodes removed during axillary surgery), lymphovascular invasion (cancer invading the small blood vessels and lymphatics in the breast tissue) and oestrogen receptor (ER) negativity (i.e. an absence of female hormone molecule binding sites on the surface of cancer cells) (Truong 2005; Garg 2007), are some of the variables that have been shown to be predictors of LRR in PMRT‐women with low volume axillary nodal disease. The predictive role of age was considered by both NCCN and the European Society of Breast Cancer Specialists and they recommend PMRT in young women affected with breast cancer and metastasis in one to three axillary lymph nodes (Cardoso 2012; Recht 2016; Salerno 2017). Similarly, the St Gallen breast cancer meeting in 2019 recommended PMRT in triple‐negative breast cancer (i.e. where there is a lack of expression of oestrogen, progesterone and human epidermal growth factor receptor‐2 protein molecule binding sites on the surface of cancer cells) with one to three positive axillary lymph nodes (Burstein 2019).
The SUPREMO trial is the only ongoing RCT that might shed light on the role of modern adjuvant systemic treatment (anti‐hormonal tablets or chemotherapy) in PMRT for women diagnosed with breast cancer and found to have metastasis in one to three lymph nodes. Since the result of this trial is not expected to be published before 2024, this review will try to bridge the gap in the literature by evaluating the long‐term outcomes of published RCTs to address the uncertainty surrounding the use of PMRT in women with breast cancer who have a low volume of axillary lymph node disease.
Objectives
To assess the effects of post‐mastectomy radiotherapy (PMRT) in women diagnosed with early breast cancer and found to have one to three positive axillary lymph nodes.
Methods
Criteria for considering studies for this review
Types of studies
All randomised controlled trials (RCTs) evaluating post‐mastectomy radiotherapy (PMRT) in women diagnosed with early breast cancer and low‐volume axillary metastatic disease ‐ defined as the involvement of cancer cells in one to three axillary lymph nodes after sentinel lymph node biopsy (SLNB) or axillary lymph node dissection (ALND).
Types of participants
We will include women diagnosed with early breast cancer and found to have one to three positive axillary lymph nodes. This includes women who have a macroscopic (≧ 2 mm in size) deposit of cancer in the axillary lymph nodes (macrometastases). We will endeavour to contact the corresponding authors of trial reports to obtain data on macrometastasis if data are missing, or described only as part of a subgroup analysis within manuscripts.
We will include women diagnosed with breast cancer treated with mastectomy and SLNB without any further axillary surgery as well as those undergoing ALND with or without initial SLNB. We will include women undergoing either simple or modified radical mastectomy, while those undergoing radical mastectomy will not be included in the review. To improve the generalisability of the results, we will include women of all ages and ethnicities, who have been diagnosed with breast cancer, with any tumour size(s) and histological types.
Types of interventions
We will include radiotherapy using X‐rays (electron and photon radiation). The total radiation dose administered for treatment should be consistent with the current recommendations, i.e. 40 Gy to 50 Gy in 15 to 25 fractions over three to five weeks. We will include post‐mastectomy radiotherapy given to the ipsilateral chest wall, axilla, supraclavicular fossa, and internal mammary nodes.
We will exclude studies in which women diagnosed with breast cancer received only intra‐operative radiation, brachytherapy and radiotherapy given using gamma rays.
We will include women diagnosed with breast cancer who also received adjuvant treatments (endocrine, chemotherapy and/or biological agents) that are given to both the intervention and comparison groups. While we will include women who underwent chemotherapy after mastectomy (adjuvant chemotherapy), we will exclude studies that used neoadjuvant chemotherapy (NACT), in which chemotherapy is given before surgery. NACT is usually given to women diagnosed with large breast cancers or cancer that involves multiple axillary lymph nodes, or both. It is given with the intention of reducing the size of cancer in order to facilitate breast‐conserving surgery and to help in the planning of post‐surgery adjuvant treatment. In such cases, progression of the disease or poor response to NACT is the most common reason for offering mastectomy before or after completion of NACT. The decision to give adjuvant radiotherapy to these women with breast cancer is also guided by the pre‐NACT cancer characteristics. These factors detract from the main focus of this review, so we will include only those studies that involve adjuvant chemotherapy.
We will compare PMRT in women with early breast cancer with low volume axillary disease to those who did not receive any radiotherapy.
Types of outcome measures
Primary outcomes
-
Local and regional recurrence (LRR). LRR is defined as the time between the treatment of breast cancer and recurrence of breast cancer in the ipsilateral chest wall, axilla, supraclavicular fossa, infraclavicular fossa or internal mammary nodes. We will consider LRR as the first event after treatment of breast cancer, i.e. prior to the development of any systemic recurrence (i.e. recurrence of breast cancer anywhere else in the body other than those sites involved by LRR). In this review, if it is not possible to extract results for the time to LRR then we will attempt to extract data at 5, 10 and 15 years.
Secondary outcomes
-
Overall survival (OS): defined as the time between the diagnosis of breast cancer or time of surgery to the date of death from any cause.
-
Disease‐free survival (DFS): defined as the duration between the diagnosis of breast cancer or time of surgery to the date of loco‐regional or systemic recurrence or death, whichever occurs first.
-
Time to progression (TTP): defined as the duration between the diagnosis of breast cancer or time of surgery to the date of loco‐regional or systemic recurrence, whichever occurs first.
-
Adverse events, including short‐ and long‐term events. Short‐term adverse events will include erythema, hyperpigmentation, and breast oedema. Long‐term adverse events will include lymphoedema, cardiac toxicity, pulmonary toxicity, bone necrosis, and development of secondary radiation‐induced cancers.
-
Quality of life (QOL): measured using any validated tool(s).
Search methods for identification of studies
Electronic searches
We will search the following databases:
-
The Cochrane Breast Cancer Group's (CBCG's) Specialised Register. Details of the search strategies used by the Group for the identification of studies, and the procedure used to code references, are outlined in the Group's module (https://breastcancer.cochrane.org/specialised-register). Trials with the keywords "breast cancer”, “mastectomy”, “radiotherapy”, “radiation therapy”, “post‐operative radiotherapy” and “post‐mastectomy radiotherapy” will be extracted and considered for inclusion in the review.
-
CENTRAL (The Cochrane Library, latest issue). See Appendix 1.
-
MEDLINE (via OvidSP) from inception to present. See Appendix 2.
-
EMBASE (via OvidSP) from inception to present. See Appendix 3.
-
The WHO International Clinical Trials Registry Platform (ICTRP) search portal (http://apps.who.int/trialsearch/Default.aspx) for all prospectively registered and ongoing trials. See Appendix 4.
-
Clinicaltrials.gov (http://clinicaltrials.gov/). See Appendix 5.
Searching other resources
Bibliographic searching
We will try to identify further studies from the reference lists of relevant trials or reviews. We will obtain a copy of the full article for each reference that reports a potentially eligible trial. Where this is not possible, we will attempt to contact the trial authors to obtain additional information.
Data collection and analysis
Selection of studies
Two review authors (RV and MC) will review the studies identified from the search strategy independently. Each author will apply the selection criteria to identify relevant studies for inclusion. If there is disagreement, a consensus will be reached through deliberation with the help of a third reviewer (SSR). We will use the PRISMA flow diagram to describe the selection process and record all excluded studies in the Characteristics of excluded studies table. We will not apply any restrictions regarding language or publication date of the studies. If we identify relevant studies published in a language other than English, we will endeavour to obtain English translations of those articles.
Data extraction and management
Two review authors (RV and MC) will extract the data and any disagreements will be resolved through discussion with JB and SSR. We will collect data that are available on demographics (age), tumour characteristics (tumour size, grade, surgical margin and receptor status), adjuvant treatments, and outcome measures. We will attempt to perform pooled statistical analysis on outcome measures if there are sufficient data available from the included studies.
We will enter data into RevMan Web for analysis. If needed, we will request further information from the corresponding authors of the studies about the statistical methods, analysis, and results. In circumstances where corresponding authors do not provide the data we request, we will try to extract the necessary information from published results using well‐established statistical methods (Tierney 2007).
Assessment of risk of bias in included studies
Two review authors (RV and MC) will assess the risk of bias independently for each included study and one author (SSR) will review the judgements to resolve any disagreement. We will use the risk of bias assessment tool recommended by Cochrane (Higgins 2011). This tool involves seven domains to address the quality of randomisation and the degree of bias arising in an RCT. Each domain is divided into three categories ‐ 'low', 'unclear' or 'high' risk ‐ on the basis of specific criteria described in the tool. These judgements will enable us to categorise studies on the basis of their risk of bias. We will perform sensitivity analysis both with or without trials of low quality to assess the effect of risk of bias on the results.
Measures of treatment effect
We will report time‐to‐event outcome measures (i.e. LRR, OS, DFS and TTP) as hazard ratios (HR) with 95% confidence intervals (CIs). If the HR and associated variance cannot be obtained directly from publications, we will extract the data indirectly from reported results (e.g. log‐rank P‐values) or Kaplan‐Meier survival curves (Parmar 1998; Tierney 2007).
We will report dichotomous outcomes as risk ratios (RR) with 95% CIs. RR values less than one will indicate that post‐mastectomy radiotherapy is the better treatment option, while RR values greater than one will indicate that no radiotherapy after mastectomy is better for women with breast cancer and low volume axillary disease.
We will report continuous outcome measures (i.e. QOL) using the standardised mean difference and 95% CI where different scales are used to measure QOL across studies. If similar scales are used to measure QOL, we will report the mean difference.
Unit of analysis issues
The unit of analysis will be each individual woman diagnosed with breast cancer. Since we are comparing the effect of PMRT against no radiotherapy in low volume axillary disease, we will not consider cross‐over trials and multiple intervention groups. Multiple events per participant will be considered only when the second event is a different outcome measure, that is, if a person develops LRR as the first event and then dies due to breast cancer or any other cause (DFS or OS).
Dealing with missing data
We will contact corresponding authors of studies by email to gather information on missing data. However, in trials where we are unable to retrieve the missing data, we will perform sensitivity analyses by excluding studies assessed as having a high risk of attrition bias.
Assessment of heterogeneity
We will use Chi2 test and I2 statistics for assessing heterogeneity (Cochran 1954; Higgins 2003). For the I2 statistic, a value of 25% to 50% may represent mild heterogeneity, 50% to 75% moderate heterogeneity, and > 75% considerable heterogeneity (Higgins 2011). We will use a random‐effects model for analysis if there is moderate or considerable heterogeneity in the included studies.
Assessment of reporting biases
We will assess publication bias by using a funnel plot. A funnel plot can be used to determine publication bias or systematic heterogeneity; a symmetrically inverted funnel is indicative of a data set without systematic heterogeneity. However, if a funnel plot is asymmetric, this by itself does not confirm publication bias, but could have resulted from other small‐study effects, for example, differences in precision of the included data sets.
Data synthesis
We will use the fixed‐effect model with an inverse variance method for combining RRs or HRs on the log scale (Deeks 2011). If there is moderate to substantial heterogeneity indicated by the I2 statistic, then we will use random‐effects models that employ the DerSimonian and Laird method (DerSimonian 1986).
We will perform data synthesis and statistical analysis in RevMan Web software. If the data are too heterogeneous for a meta‐analysis to provide a meaningful result, we will present the result in a forest plot for visual inspection and provide a narrative synthesis.
Subgroup analysis and investigation of heterogeneity
In this review, if sufficient studies or subgroup analyses within studies are identified, we will investigate the role of age, type of axillary surgery (SLNB alone versus ALND), oestrogen receptor (ER) and human epidermal growth factor‐2 receptor (Her2) status on LRR in women with early breast cancer and low volume axillary nodal disease treated with or without PMRT.
When individual studies indicate that subgroups have been identified, but do not report results, we will attempt to contact the corresponding authors to obtain these results. We will use a random‐effects model for analysis if there is moderate or considerable residual heterogeneity identified amongst the studies included in each subgroup. We will use subgroup analyses to explore the reasons for heterogeneity and to assess the effect of intervention within subgroups on primary and secondary outcome measures.
Sensitivity analysis
We will perform sensitivity analysis if there are sufficient data available to assess bias originating from variations in study quality (for example, by excluding studies considered to be at high risk of bias).
We will also compare subgroup results across trials to those from subgroups within trials if a sufficient number of effect estimates are available.
Summary of findings and assessment of the certainty of the evidence
Two authors (RV and MC) will assess the overall certainty of the evidence by using the GRADE approach (Schünemann 2020). This involves assessing the evidence for each outcome using five domains. These domains relate to:
-
risk of bias of the included studies;
-
inconsistency (i.e. heterogeneity);
-
indirectness (relevance to the review question);
-
imprecision (i.e. confidence intervals); and
-
publication bias.
The GRADE approach will be used to assess the certainty of the evidence for the primary and secondary outcome measures. A Summary of Findings table will be created for the review using GRADEpro GDT software for primary and secondary outcome measures.
