© The Author 2006. Published by Oxford University Press.
ARTICLE |
Survival Effects of Postmastectomy Adjuvant Radiation Therapy Using Biologically Equivalent Doses: A Clinical Perspective
Affiliation of authors: National Health and Medical Research Council Clinical Trials Centre, University of Sydney, Australia
Correspondence to: Val Gebski, BA, MStat, National Health and Medical Research Council Clinical Trials Centre, Level 5, Bldg. F, 88 Mallett St., Camperdown NSW 2050, Australia (e-mail: val{at}ctc.usyd.edu.au).
 |
ABSTRACT |
|---|
 TopNotes Abstract IntroductionPatients and methodsResultsDiscussionReferences |
|---|
Background: Postmastectomy radiation therapy reduces locoregional recurrence among women with operable breast cancer, but whether it improves survival has been controversial. We reanalyzed the results from 36 unconfounded trials (i.e., addition of radiation therapy was the sole discriminant between treatments being compared) that were identified in previous meta-analyses, which provided 38 comparisons. Methods: We used three predefined treatment categories for individual patient data: 1) a biologically equivalent dose (BED) of 40–60 Gy in 2-Gy fractions with an appropriate target volume, 2) an inadequate or excessive dose of radiation therapy, and 3) an inappropriate target volume. Effects of radiation therapy on 5-year and 10-year survival in each of the treatment categories were estimated from a cohort of 13 199 patients from the published rates or, if these were unavailable, from the published survival curves. We also used this categorization to reanalyze data from Early Breast Cancer Trialists' Collaborative Group (EBCTCG) postmastectomy studies. At 10 years, 16 (84%) of the 19 comparisons in our study coincided with those reported by the EBCTCG. All statistical tests were two-sided. Results: Twenty-five of the 38 available comparisons had used optimal and complete radiotherapy (i.e., category 1). Of these 25 comparisons, 17 had 5-year data, and these data showed that adjuvant radiation therapy was associated with a 2.9% absolute increase in survival (odds ratio [OR] of death = 0.87, 95% confidence interval [CI] = 0.79 to 0.96; P = .006). Thirteen category 1 trials had data at 10 years, and these data showed that adjuvant radiation therapy was associated with a statistically significant 6.4% absolute increase in survival (OR of death = 0.78, 95% CI = 0.70 to 0.85; P<.001). No statistically significant change in survival was observed among category 2 (OR of death = 0.91, 95% CI = 0.75 to 1.11) or 3 (OR of death = 0.97, 95% CI = 0.61 to 1.55) trials. Among the 33 EBCTCG studies, odds of local recurrence were reduced more among category 1 trials (80% lower) than among category 2 (70% lower) or 3 (64% lower) trials (Pheterogeneity<.001). Odds of all-cause death were also lower among category 1 trials (13% lower) than among category 2 (3% lower) or 3 (26% higher) trials (Pheterogeneity = .01). Conclusions: Adjuvant radiation therapy with an optimal BED and target volume was statistically significantly associated with improved survival for up to 10 years.
|
INTRODUCTION |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Comprehensive meta-analyses of patients in randomized trials of radiation therapy for operable breast cancer have failed to resolve a fundamental question: whether the reduction of locoregional recurrence that is associated with postoperative radiation therapy is reflected in improved overall survival. In general, these meta-analyses have concluded that postoperative radiation therapy was not statistically significantly associated with increased overall survival at 10 years (1–11). Moreover, they have shown that any reduction in breast cancer mortality has been offset by mortality from late side effects of radiation therapy, including heart disease. This conclusion has profoundly affected multidisciplinary management of operable breast cancer. Modern radiation therapy substantially reduces radiation to the heart by avoiding direct irradiation of it and raises the possibility that a more effective dose to the target volume that avoids direct irradiation of the heart could be associated with improved outcomes in patients treated with postoperative adjuvant radiation therapy (12).
Controversy remains, however, and its origins are easy to identify (12–14). The main criticism of the published meta-analyses of adjuvant radiation therapy trials is that only the presence or absence of radiation therapy was evaluated, not the parameters of radiation therapy such as target volume, and dosage. In contrast, meta-analyses of adjuvant systemic therapy have analyzed the outcomes associated with chemotherapy separately, according to single-agent versus multiple-agent regimens, and analyzed the outcomes of adjuvant tamoxifen separately according to the duration of treatment and the total amount of drug given.
In the past two decades, new equipment and techniques for radiation therapy have led to a better understanding of the dose–response relationships in the control of subclinical breast cancer (15,16). This dose–response relationship was primarily developed from the results of studies other than those included in the trial meta-analyses. Of the 32 series considered to develop the dose–response relationship, only three series (9%) overlapped with these meta-analyses. The optimal dose range—i.e., the range that offers the greatest chance of locoregional control of breast cancer at the lowest cost in locoregional morbidity—appears to be 40–60 Gy in 2-Gy fractions. Schedules using different dose and fraction schemes other than 2 Gy can be converted to an equivalent 2-Gy schedule (17). These optimal biologically equivalent doses (BEDs) permit comparison of the total dose delivered by different fractionation schedules. It is now clear that the dose of postoperative radiation therapy in some trials was inadequate or excessive according to the current understanding of the dose–response relationships for minimal residual breast cancer. Consideration of the dose of radiotherapy in the trials has been largely neglected, with the exception of the work by Cuzick et al. (4,5), who described dose in terms of rad-equivalent therapy but, in their final analysis, did not discriminate according to dose.
The pattern and extent of locoregional recurrence after mastectomy are related to the size and characteristics of the primary tumor and to the extent of regional lymph node involvement. Assumptions about the relative risk of recurrence at different sites (chest wall, axilla, supraclavicular fossa, and/or internal mammary lymph nodes) have led to empirical restrictions of the radiation therapy target volume. These restrictions have compromised earlier clinical trials investigating the role of radiation therapy. A good example is the restriction of radiation therapy in what is known as the hockey-stick technique (18). In this technique, radiation therapy after mastectomy is delivered to the internal mammary lymph node chain and supraclavicular fossa only. In contrast, it is now generally accepted that an appropriate target volume should include the chest wall and the areas with the regional lymph nodes.
A final consideration is to distinguish between studies in which radiation therapy is given as an adjuvant therapy to improve survival and studies in which it is given to compensate for less radical surgery. An example of the former is the comparison between radical mastectomy alone and radical mastectomy plus postoperative radiation therapy. An example of the latter is the comparison between radical mastectomy and local excision of the tumor plus radiation therapy (4), in which any evaluation of the role of radiation therapy is confounded by the difference in surgery. In such studies (19,20), equivalence of survival and quality of life with a reduction in treatment-related morbidity are competing endpoints. Such trials (19,20), confounded by the surgical intervention, were not included in this study or in the previous EBCTG overviews (9,10).
We assessed the association between postmastectomy radiation therapy for early breast cancer and overall survival in a meta-analysis of 36 randomized trials containing 38 comparisons that were unconfounded (addition of radiation therapy was the sole discriminant between treatments being compared). We considered the specific issues of radiation dosage and target volume coverage. We also investigated whether studies using optimal BED and appropriate target volumes observed the greatest benefit of radiation therapy.
|
PATIENTS AND METHODS |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Eligible Trials
We considered any randomized controlled trial that had been included in earlier published meta-analyses and systematic reviews of radiation therapy. A literature search by the Cochrane Breast Cancer Group identified no other studies whose results were published or available after 2002.
Trials were considered eligible for inclusion in our study if the following criteria were met: 1) Studies were of operable breast cancer that was initially treated by mastectomy. (Stage I and stage II disease and selected cases of stage III disease were considered operable.) 2) Studies were randomized controlled clinical trials that compared adjuvant radiation therapy with no such therapy. This treatment was the sole discriminating factor between the two arms of the trial. Other treatments, including extent of surgery, endocrine therapy, and chemotherapy, if given, had to be common to each arm. Three studies that reported the use of randomization but may have instead used date of birth as the allocation method (21,22) were included in the primary analysis but were excluded in a sensitivity analysis. Thirty-eight unconfounded randomized comparisons from 36 trials were identified, with data being available from a total cohort of 13 199 patients. Thirty-three of these comparisons coincide with the studies appearing in the EBCTCG analysis. Because it had access to individual patient data for some studies, the EBCTCG was able to provide comparisons in addition to or different from those in published reports (23), and sometimes a strict relationship between our studies and the EBCTCG is not always possible.
Radiation Therapy
Planned radiation therapy for each trial was classified into the following three major categories: Category 1, optimal radiation therapy, included studies that delivered optimal radiation therapy, which was defined as doses in the range of 40–60 Gy in 2-Gy fractions (where 50 Gy = 5000 rads) (15) or as a BED (17) to the chest wall, axillary lymph nodes, and the supraclavicular fossa with or without the internal mammary lymph nodes. Category 2, inadequate or excessive radiation therapy, included studies that delivered inadequate or excessive radiation therapy, which was defined as either doses of less than 40 Gy in 2-Gy fractions (or, for other fractionation schedules, the calculated BED being less than 40 Gy) or of greater than 60 Gy in 2-Gy fractions (or for other fractionation schedules the calculated BED being more than 60 Gy). The BED was calculated by use of 
/
, a ratio reflecting the weighting of the dose per fraction in the schedule to the total dose delivered, equal to 10, standardized to 2-Gy fractions. Category 3, incomplete tissue coverage, included studies in which radiation therapy provided incomplete tissue coverage by restricting the target volume to areas of less than the area of the chest wall and regional lymph nodes. Techniques for which the target volume was restricted were considered to be inappropriate because an area at risk of recurrence received no radiation therapy.
Studies included in category 2 had treatment that by design delivered an inadequate or excessive dose irrespective of the target volume. Studies included in category 3 had treatment that delivered an inappropriate target volume irrespective of dose. Any study satisfying the criteria for both categories 2 and 3 was included in category 2.
Statistical Analyses
The primary outcomes in this study were 5-year and 10-year overall survival rates that were calculated on an intention-to-treat analysis. All eligible trials with available data on these outcomes were included. When data from published studies were not available, we contacted the authors to obtain the 5-year and 10-year survival rates, as appropriate. If the overall survival rates at 5 and 10 years, measured as odds ratios (ORs), were not provided in the reports or from the authors, we estimated these rates directly from the number of randomly assigned patients and the published survival curves. These estimates did not account for censoring, but censoring would have no substantial effect on the width of the confidence intervals (CIs) because the follow-up for most of the studies was 10 years or more. Odds ratios (and their 95% confidence intervals) were obtained for each comparison, and the pooled estimate was calculated by use of an inverse variance–weighted average of the individual studies (24). Chi-square tests were used to test for statistical heterogeneity. Survival data were analyzed by using software provided by the Cochrane Collaboration (25). Comparisons between the treatments were performed with chi-squared tests at a 5% statistical significance level.
The presence of positive lymph nodes reflects an increased risk of death, and the benefit of adjuvant radiation therapy in this patient group is of particular clinical interest. If studies whose recruitment was restricted to lymph node–positive patients or if comparisons for this patient subgroup were provided in published reports, these comparisons could be combined to provide estimates of the effect of radiation therapy in this patient cohort.
We also undertook a secondary analysis of the published report by the EBCTCG (10) by applying the three-category classification system, described above, for radiation therapy. Individual patient data from the EBCTCG report also largely coincided with those that we considered and had 10-year follow-up data; thus, their survival rates would approximately reflect the 10-year results in our approach. The EBCTCG report contains data on substantially more comparisons than can be identified in the literature, which was the source for our primary analyses. The studies that we considered in our primary analysis substantially overlapped (33 of 38 comparisons) those reported by the EBCTCG. One comparison included in the EBCTCG report assessed preoperative radiation therapy and was not part of the comparisons that we considered. Results from our reanalyses of these studies were used to explore the associations between radiation therapy and local recurrence, breast cancer mortality, or non–breast cancer mortality and were used to assess the consistency of our primary analyses by examining the associations between radiation therapy and all-cause mortality. The EBCTCG report allowed comparison of the odds ratios of the treatment arms from the follow-up of individual patients and contained more information than the 5- and 10-year proportions, provided in the original publications or extracted from the survival curves for the studies in our analysis. An analysis of the effect of radiation therapy using the published observed death rates minus the expected death rates and the corresponding variances is also possible. Such an analysis would produce comparisons of hazard ratios and would have confidence intervals that would be wider than those that we have reported. Such an analysis would produce results consistent with the approach that we have adopted. Apart from the EBCTCG report's including trials for which published data were not available, because the follow-up in most of their studies exceeded 10 years, the overall survival results between the two analyses should be comparable. These analyses were also performed with the Cochrane Library software by use of chi-squared tests with a statistical significance level of 5%.
Finally, we developed a strategy for estimating the absolute survival rate associated with optimal postoperative radiation therapy. This strategy considered the relationship between optimal postoperative radiation therapy and cause-specific mortality and investigated the relationship between this association and that between deaths from any cause and the risk of breast cancer-specific death. All statistical tests were two-sided.
|
RESULTS |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Primary Analysis
Forty randomized comparisons comparing postoperative radiation therapy with no radiation therapy were identified from previously published meta-analyses (4,5,10,11) are presented in Table 1. Thirty-eight of the 40 trials provided comparisons of the effect of radiation therapy that were not confounded. Of these 38 comparisons, 25 used optimal radiation therapy with appropriate target volume (category 1), seven used inadequate or excessive doses of radiation therapy (category 2), and six used inappropriate target volumes (category 3) (Fig. 1). Among these 38 comparisons, 26 (68%) had data at a follow-up of 5 years, and 19 (50%) had data at a follow-up of 10 years. Among the 25 trials in category 1, 17 (68%) had data at a follow-up of 5 years that could be extracted, and 13 (52%) had data at a follow-up of 10 years.
|
|
In the primary analysis, we calculated the BED for comparisons in all three categories (Table 2). We used the calculated BED to classify these comparisons into the respective categories and to investigate the effect of radiation therapy for those studies in category 1. The 26 studies with follow-up data at 5 years included a total of 13 199 patients with such a follow-up, and the 19 studies with follow-up data at 10 years included a total of 8921 patients with such a follow-up. All of these follow-up data were available for this primary analysis. Thirteen studies compared megavoltage, one compared orthovoltage, two compared both orthovoltage and megavoltage, and one compared radiation energy not stated. For eight of the 38 comparisons, no systemic chemotherapy was given and for the comparisons in the primary analysis, no systemic chemotherapy was given in four of the early studies—the Stockholm study (26), the Edinburgh study (27), NSABP-04 (28,29), and the Wessex study (30); the rest of the studies included chemotherapy and/or hormone therapy.
|
To investigate the association between radiation therapy and survival after follow-ups of 5 and 10 years, we calculated the pooled weighted estimate of this association for studies in each of the three categories. At a follow-up of 5 years among studies in category 1, we found a statistically significant 13% relative survival advantage associated with radiation therapy (OR of death from any cause = 0.87, 95% CI = 0.79 to 0.96; P = .006), compared with no radiation therapy (Fig. 2). This advantage translated into an absolute 2.9% increase in survival (or 29 lives per 1000 patients treated) and a number needed to treat of 34 (i.e., on average, for every 34 patients treated, one life would be saved over 5 years). At a follow-up of 10 years among studies in category 1, we found a statistically significant 22% increase in relative survival associated with radiation therapy (OR = 0.78, 95% CI = 0.70 to 0.85; P<.001), compared with no radiation therapy (Fig. 3). This increase corresponded with an absolute 6.4% increase in survival (64 per 1000) and a number needed to treat of 16. When we considered trials of high-risk patients (i.e., patients with lymph node–positive disease) separately, we found that an absolute 5.2% increase in survival (52 per 1000) at a 10-year follow-up was associated with adjuvant radiation therapy (OR = 0.80, 95% CI = 0.64 to 1.0; P = .05, with no evidence of heterogeneity), compared with no radiation therapy.
|
|
Among studies with inadequate or excessive radiation therapy (category 2) or with radiation therapy to an inappropriate target volume (category 3), we found no statistically significant association between radiation therapy and survival after 5 or 10 years of follow-up. For example, among category 2 studies with 10 years of follow-up, we found a stronger association between survival and radiation therapy with an inadequate or excessive dose (OR of death = 0.91, 95% CI = 0.75 to 1.11) than between survival and no radiation therapy (Fig. 3). Among category 3 studies with 10 years of follow-up, we found a stronger association between survival and radiation therapy with an inappropriate target volume (OR of death = 0.97, 95% CI = 0.61 to 1.55) than between survival and no radiation therapy (Fig. 3). However, because of a smaller number of events in category 2 and 3 studies than in category 1 studies, these results were not statistically significantly different from each other (Pheterogeneity = .14). Some of the protocols included systemic chemotherapy, but the associations between radiation therapy and survival did not differ statistically significantly between regimens without (26–30) and with (19,22,23,32–40) chemotherapy (Pheterogeneity = .10, at a follow-up of 5 years and Pheterogeneity = .96 at a follow-up of 10 years).
Differential Follow-Up
To examine the potential bias associated with differential follow-up times, we determined whether those category 1 studies with better outcomes at a follow-up of 5 years were more likely to have 10-year survival data than studies with poorer outcomes at a follow-up of 5 years. When we combined the 5-year and the 10-year survival data for the 13 studies that delivered optimal radiation therapy and had both 5- and 10-year follow-up data, we found a 14% increase in survival was associated with radiation therapy (OR of death = 0.86, 95% CI = 0.78 to 0.95; P = .004), compared with no radiation therapy. However, for the four studies in category 1 with no available follow-up data at 10 years, we found a 3% decrease in survival associated with radiation therapy (OR of death = 1.03, 95% CI = 0.65 to 1.62; P = .9). However, results from a test for heterogeneity (Pheterogeneity = .38) indicated that the differences between these associations can be attributed to chance.
Sensitivity Analysis
In a sensitivity analysis, we excluded the three category 1 studies [Manchester P and Manchester Q (21) and Finnish (22)] for which the randomized treatment allocation may have used date of birth. This randomization method is undesirable because allocation concealment cannot be guaranteed and the distribution of odd and even numbers is not evenly distributed throughout a year (41). Although we have included these comparisons in our primary analysis, the sensitivity analysis will provide an indication of the impact on effect of radiation therapy when these comparisons are included. Exclusion of these trials resulted in 15 of 24 category 1 trials with data at 5 years of follow-up and a 3.4% absolute increase in survival (34 per 1000) associated with adjuvant radiation therapy (OR of death = 0.85, 95% CI = 0.77 to 0.95; P = .003). Also, exclusion of these trials resulted in 12 of 24 category 1 trials with data at 10 years of follow-up and a 7.1% absolute increase in survival (71 per 1000) associated with adjuvant radiation therapy (OR = 0.75, 95% CI = 0.67 to 0.82; P<.001). Thus, the results of these analyses support the conclusions of the primary analysis.
Applying the Radiation Therapy Classification to the EBCTCG Overview
To investigate the association between radiation therapy and all-cause mortality, we applied our classification system to all the ECBTCG trials of postmastectomy radiation therapy (Fig. 4), whose follow-up periods averaged 10 years (10). The EBCTCG studies had reported a 1.3% absolute increase in all-cause mortality associated with postoperative radiation therapy (with either mastectomy or breast-conserving surgery) and a 1.26% absolute increase in all-cause mortality when the analysis was restricted to the postmastectomy radiation trials, with neither increase being statistically significant. However, in the 23 studies from their report that were classified as category 1, statistically significant improved overall survival was observed, the absolute survival benefit associated with radiation therapy was 3.9% (39 per 1000), and the relative benefit was 13% (OR of death = 0.87, 95% CI = 0.80 to 0.94, P<.001). This association was stronger than that among studies in category 2 (OR = 0.97, 95% CI = 0.87 to 1.09) or category 3 (OR = 1.26, 95% CI = 1.03 to 1.53), which showed substantial harm. The heterogeneity of the associations over the three classification groups was statistically significant (Pheterogeneity = .006).
|
When we applied our classification system to the EBCTCG studies to investigate the association between radiation therapy and isolated local recurrence, breast cancer mortality, or non–breast cancer mortality, data were available for only nine studies in category 1, four in category 2, and one in category 3 (Fig. 5). We found that radiation therapy was associated with an 80% reduction in local recurrence in category 1 studies, a 70% reduction in local recurrence in category 2 studies, and a 64% reduction in category 3 studies. There was statistically significant heterogeneity among these categories (Pheterogeneity<.001). In contrast, we detected no evidence of statistically significant heterogeneity among types of radiation therapy associated with breast cancer–specific mortality, but we found some (although non–statistically significant) evidence for a greater difference in non–breast cancer mortality among studies in categories 2 and 3 (Pheterogeneity = .07). Among the nine category 1 studies, we found a 20% reduction in breast cancer death associated with radiation therapy (OR = 0.8, 95% CI = 0.73 to 0.88) and a 15% increase in non–breast cancer deaths associated with radiation therapy (OR = 1.15, 95% = CI 0.97 to 1.36).
|
Estimating Survival Benefits of Optimal Radiation Therapy
We estimated the association between survival and optimal radiation therapy that used BED to appropriate target volumes after a follow-up of 10 years (Table 3) among patients in various risk groups. We used the associations between optimal radiation therapy (category 1) and the risk of breast cancer death (20% reduction) or the risk of non–breast cancer death (15% increase) directly from the EBCTCG data (Fig. 5), and we applied these results to a theoretical patient whose 10-year survival was estimated as 50% and whose 10-year risk of death from breast cancer was 80%. If this patient were to receive optimal radiation therapy, she would experience a 13% relative reduction in the risk of death from breast cancer (which was similar to our results observed when all 23 studies in Fig. 4 were analyzed) and a 6.5% absolute increase in 10-year survival. When we applied these same assumptions to a patient at high risk, i.e., with a 33% chance of survival for 10 years (and a 90% 10-year risk of death from breast cancer), we found an 11% absolute reduction in her risk of death from any cause. When we applied these same assumptions to a patient at low risk, with a 66% chance of survival for 10 years (and 65% 10-year risk of death from breast cancer), the absolute reduction in risk of death was 2.6%.
|
To determine the degree of benefit of radiation therapy, we reanalyzed data from the trials investigating radiation therapy and breast cancer. We found that increased survival was associated with radiation therapy among patients at moderate to high risk of breast cancer death provided that they received the optimal dose of radiation therapy to the complete target volume. We believe that these findings would probably be strengthened by additional analyses that are more extensive and that adopt clinically relevant classification criteria similar to ours and use both published and unpublished studies and individual patient data.
|
DISCUSSION |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
In this analysis, we restricted our meta-analyses to trials that used the optimal radiation therapy dose delivered to appropriate target volumes and found that radiation therapy was associated with improved overall survival. We also found that, over all such trials with available data and among those eligible trials included in the EBCTCG overview (23 of 26 have common comparisons), improved 5-year survival and 10-year survival was associated with optimal radiation therapy. When we compared trials that used BED to the appropriate volumes with trials using inadequate or excessive doses or inappropriate target volumes, we found statistically significant heterogeneity in survival results, which indicated that these groups should be considered separately. Although our categorization of trials was developed with knowledge of the trial results, the categorization and subsequent analyses were also supported in the following ways. First, we based our choice of a BED of 40–60 Gy in 2-Gy fractions on studies that were largely independent of these trials and on the ability of the radiation dose to control microscopic residual disease at acceptable toxicity (15). Second, evidence from randomized trials of other outcomes (isolated local recurrence and cause-specific deaths) (10) appears to consistently support our findings: in particular, among the EBCTCG trials with available data, trials that used optimal radiation therapy found a statistically significantly association of radiation therapy with a lower risk of local recurrence than trials that used another radiation therapy regimen, as expected. Furthermore, studies that used optimal radiation therapy might show a larger reduction of breast cancer mortality and a smaller reduction of non–breast cancer mortality (because of less radiation to inappropriate target volumes). Among EBCTCG trials that reported cause-specific mortality, the results were consistent with this expectation but not definitive, possibly because of the small number of trials with available data.
Our analysis has several limitations. Survival data at 5 and 10 years of follow-up were lacking in several trials, which may have led to a biased estimate in favor of postoperative radiation therapy because trials without such data tended to have less favorable estimates of risk reduction in the EBCTG overview. Further, our analyses of survival at 5 years and 10 years of follow-up also ignore late effects beyond 10 years. Despite these limitations, when we examined trials from the EBCTCG overview, we found that improved overall survival was associated with the use of an effective BED to an appropriate target volume, compared with other regimens. These supplementary analyses are, however, limited by the small number of trials with available data. Analyses that are more extensive and that adopt clinically relevant classification criteria similar to those that we used with both published and unpublished studies and individual patient data could strengthen these findings. Finally, as noted above, the categories, although designed before our analyses, were designed with knowledge of trial results, which could lead to biased results. This is a legitimate criticism, but it is not stronger than criticisms of other systematic reviews in breast cancer (such as the subclassification of chemotherapy regimens into single-agent or anthracycline-containing regimens). We have been conservative in including three studies whose treatment allocation may have been by date of birth, which could have led to questionable randomization or allocation concealment. When we excluded these studies in a sensitivity analysis, the association between radiation therapy and improved outcome was strengthened.
We have calculated the most plausible or realistic estimates of improved outcome associated with modern radiation therapy to assist doctors and patients in clinical decision making. We further used these results to calculate the risk of death for individual patients. We have also suggested that a 20% relative risk reduction in breast cancer death and a 15% relative increase in non–breast cancer deaths were plausible estimates of the risks associated with radiation therapy in the range of 40–60 Gy equivalents on basis of the data from the EBCTCG. Thus, among patients with a higher risk of death from breast cancer, a greater absolute benefit in increased breast cancer–specific survival, compared with survival from other causes, is associated with radiation therapy (42). Among patients with an expected 10-year survival of less than 50%, with 80% of deaths from breast cancer, we might expect a more than 6% absolute reduction in all-cause mortality, so that only about 16 patients need be treated with radiation therapy in the equivalent range of 40–60 Gy to prevent one death. About half of the studies included in our analysis were of lymph node–positive disease—that is, the patients were at high risk. According to the Nottingham Group, the expected 10-year overall survival for high-risk patients (lymph node–positive patients or patients with high-grade tumors) varies between 13% and 53% (43). Our results indicate that appropriate adjuvant radiation therapy is associated with a 16% reduction in the risk of death in such women. However, this estimate may be conservative because contemporary population studies indicate that modern radiation techniques are associated with a decreased risk of excess non–breast cancer deaths, compared with earlier radiation techniques (44–47).
A patient with breast cancer is entitled to make an informed choice about whether to receive radiation therapy. Consequently, the clinician should provide information about the evidence-based risks associated with various options to assist her in her choice (48). The evidence suggests that the survival of these patients with breast cancer would be increased with radiation therapy, although the amount of the increase would depend on the risk profile of the patient. Determining whether the side effects and the inconvenience of radiation therapy offset the long-term survival benefits would be more of an issue for patients at lower risk, for whom the smaller absolute increases in survival may be insufficient to justify the toxicity of treatment, than for patients at higher risk. The results of both our primary analysis and our reanalysis of the EBCTCG indicate that the balance between breast cancer and non–breast cancer deaths should favor radiation therapy for women with a high risk of death from breast cancer, particularly when the proportion of non–breast cancer deaths is low (i.e., there is a low risk of non–breast cancer deaths among all women at high risk of breast cancer death). The average age at presentation with breast cancer is approximately 60 years. In our opinion, many patients with breast cancer would choose to receive adjuvant radiotherapy and would accept a possible relatively small increase in the risk of death from other causes more than 10 years later.
Our conclusions are further supported by the results of the meta-analysis examining the benefit of radiation therapy after breast-conserving surgery, in which a statistically significant 8% relative reduction in all-cause death was associated with radiation therapy (hazard ratio = 0.92, 95% CI = 0.85 to 0.98) (49). Our meta-analysis of randomized trials of BED radiation therapy after mastectomy found that improved survival was associated with optimal radiation therapy. Consequently, we recommend that postmastectomy radiation therapy be considered as part of standard care for all women at high risk.
|
NOTES |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Supported by a grant from the National Health and Medical Research Council, Australia.
|
REFERENCES |
|---|
|
TopNotesAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
(1) Stjernsward J. Decreased survival related to irradiation postoperatively in early operable breast cancer. Lancet 1974;2(7892):1285–6.[Medline]
(2) Levitt SH, McHugh RB, Song CW. Radiation therapy in the postoperative treatment of operable cancer of the breast. Cancer 1976;39:933–40.[CrossRef][Web of Science]
(3) Levitt SH, McHugh RB. Radiation therapy in the postoperative treatment of operable cancer of the breast. Cancer 1977;39:924–32.[CrossRef][Web of Science][Medline]
(4) Cuzick J, Stewart H, Peto R, Baum M, Fisher B, Host H, et al. Overview of randomized trials of postoperative adjuvant radiation therapy in breast cancer. Cancer Treat Rep 1987;71:15–29.[Web of Science][Medline]
(5) Cuzick J, Stewart H, Peto R, Fisher B, Kaae S, Johansen H, et al. Overview of randomized trials comparing radical mastectomy without radiation therapy against simple mastectomy with radiation therapy in breast cancer. Cancer Treat Rep 1987;71:7–14.[Web of Science][Medline]
(6) Cuzick J. Overview of adjuvant radiation therapy for breast cancer. Recent Results Cancer Res 1989;115:220–5.[Medline]
(7) Cuzick J, Stewart H, Rutqvist L, Houghton J, Edwards R, Redmond C, et al. Cause-specific mortality in long-term survivors of breast cancer who participated in trials of radiation therapy. J Clin Oncol 1994;12:447–53.[Abstract]
(8) Auquier A, Rutqvist LE, Host H, Rotstein S, Arriagada R. Post-mastectomy megavoltage radiation therapy: the Oslo and Stockholm trials. Eur J Cancer 1992;28:433–7.[CrossRef][Medline]
(9) EBCTCG Early Breast Cancer Trialists' Collaborative Group. Effects of radiation therapy and surgery in early breast cancer. An overview of randomized trials. N Engl J Med 1995;333:1444–55.
(10) EBCTCG Early Breast Cancer Trialists' Collaborative Group. Favourable and unfavourable effects on long-term survival of radiation therapy for early breast cancer: an overview of the randomized trials. Lancet 2000;355:1757–70.[CrossRef][Web of Science][Medline]
(11) Whelan TJ, Julian J, Wright J, Jadad AR, Levine ML. Does locoregional radiation therapy improve survival in breast cancer? A meta-analysis. J Clin Oncol 2000;18:1220–9.
(12) Dodwell DJ, Langlands AO. Cardiac morbidity of post-operative adjuvant radiation therapy for breast cancer. A review. Australas Radiol 1994;38:154–6.[CrossRef][Medline]
(13) Edland RW. Presidential address: does adjuvant radiation therapy have a role in the postmastectomy management of patients with operable breast cancer—revisited. Int J Radiat Oncol Biol Phys 1988;15:519–35.[Web of Science][Medline]
(14) Van de Steene J, Soete G, Storme G. Adjuvant radiation therapy for breast cancer significantly improves overall survival: the missing link. Radiother Oncol 2000;55:263–72.[CrossRef][Web of Science][Medline]
(15) Denham JW. The radiation dose-response relationship for control of primary breast cancer. Radiother Oncol 1986;7:107–23.[CrossRef][Web of Science][Medline]
(16) Fletcher GH. Local results of irradiation in the management of localized breast cancer. Cancer 1971;29:545–51.[Web of Science]
(17) Barton M. Tables of equivalent dose in 2 Gy fractions—a simple application of the linear quadratic formula. Int J Radiat Oncol Biol Phys 1995;31:371–8.[CrossRef][Web of Science][Medline]
(18) Harris JR, Hellman S. Put the ‘hockey-stick’ on ice. Int J Radiat Oncol Biol Phys 1988;15:497–9.[CrossRef][Web of Science][Medline]
(19) Kaae S, Johansen H. Five year results: two random series of simple mastectomy with postoperative irradiation versus extended radical mastectomy. Breast Cancer 1962;87:82–8.
(20) Langlands AO, Prescott RJ, Hamilton T. A clinical trial in the management of operable cancer of the breast. Br J Surg 1980;67:170–4.[Web of Science][Medline]
(21) Easson E. Postoperative radiotherapy in breast cancer. In: Forrest A, Kunkler P, editors. Prognostic factors in breast cancer. Edinburgh (Scotland): Livingston; 1968. p. 118–27.
(22) Klefstrom P, Grohn P, Heinonen E, Holsti L, Holsti P. Adjuvant postoperative radiation therapy, chemotherapy, and immunotherapy in stage III breast cancer. II. 5-year results and influence of levamisole. Cancer 1987;60:936–42.[CrossRef][Web of Science][Medline]
(23) Griem KL, Henderson C, Gelman R, Ascoli D, Silver B, Recht A, et al. The 5-year results of a randomized trial of adjuvant radiation therapy after chemotherapy in breast cancer patients treated with mastectomy. J Clin Oncol 1987;5:1546–55.
(24) Egger M, Davey Smith G, Altman DG. Systematic reviews in health care: meta-analysis in context. London (UK): BMJ Books; 2001.
(25) Cochrane Library. Issue 4. Oxford (UK): Update software; 2002.
(26) Rutqvist LE, Pettersson D, Johansson H. Adjuvant radiation therapy versus surgery alone in operable breast cancer: long-term follow-up of a randomized clinical trial. Radiother Oncol 1993;26:104–10.[CrossRef][Web of Science][Medline]
(27) Stewart HJ, Jack WJ, Everington D, Forrest AP, Rodger A, McDonald CC, et al. South-east Scottish trial of local therapy in node negative breast cancer. Breast 1994;3:31–9.[CrossRef]
(28) Fisher B, Montague E, Redmond C, Deutsch M, Brown GR, Zauber A, et al. Findings from NSABP protocol no. B-04—comparison of radical mastectomy with alternative treatments for primary breast cancer. Cancer 1980;46:1–13.[CrossRef][Web of Science][Medline]
(29) Fisher B, Redmond C, Fisher ER, Bauer M, Wolmark N, Wickerham L, et al. Ten-year results of a randomized clinical trial comparing radical mastectomy and total mastectomy with or without radiation therapy. N Engl J Med 1985;312:674–81.[Abstract]
(30) Turnbull AR, Turner DT, Chant AD, Shepherd JM, Buchanan RB, Fraser JD. Treatment of early breast cancer. Lancet 1978;2(8079):7–9.[CrossRef][Web of Science][Medline]
(31) Velez-Garcia E, Carpenter JT, Moore M, Vogel CL, Marcial V, Ketcham A, et al. Post surgical adjuvant chemotherapy with or without radiation therapy in women with breast cancer and positive axillary nodes: a South-Eastern Cancer Study Group (SEG) trial. Eur J Cancer 1992:28A:1833–7.
(32) Ragaz J, Jackson SM, Le N, Plenderleith IH, Basco JJ, Wilson VE, et al. Adjuvant radiation therapy and chemotherapy in node-positive premenopausal women with breast cancer. N Engl J Med 1997;337:956–62.
(33) Ragaz J, Jackson S, Le N, Manji M, Wilson K, Olivotto I, et al. Postmastectomy radiation outcome in node positive breast cancer patients among N1-3 versus N4+ subset: impact of extracapsular spread. Update of the British Columbia randomized trial. Proc ASCO 1999; abstract 274. Available at: http://www.asco.org/ac/1,1003,_12-002636-00_18-0017-00_19-0014529,00.asp [last assessed: September 14, 2004].
(34) Blomqvist C, Tiusanen K, Elomaa I, Rissanen P, Hietanen T, Heinonen E, et al. The combination of radiation therapy, adjuvant chemotherapy and tamoxifen in stage II breast cancer. Long-term follow-up results of a randomised trial. Br J Cancer 1992;66:1171–6.[Web of Science][Medline]
(35) Overgaard M, Hansen PS, Overgaard J, Rose C, Andersson M, Bach F, et al. Postoperative radiation therapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 1997;337:949–55.
(36) Overgaard M, Jensen M, Overgaard J, Hansen PS, RoseC, Andersson M, et al. Postoperative radiation therapy in high-risk postmenopausal breast-cancer patients given adjuvant tamoxifen: Danish Breast Cancer Cooperative Group DBCG 82c randomised trial. Lancet 1999;353:1641–8.[CrossRef][Web of Science][Medline]
(37) Olson JE, Neuberg D, Pandya KJ, Richter MP, Solin LJ, Gilchrist KW, et al. The role of radiation therapy in the management of operable locally advanced breast carcinoma. Cancer 1997;79:1138–49.[CrossRef][Web of Science][Medline]
(38) Schmoor C, Bastert G, Dunst J, Bojar H, Christmann D, Unbehaun V, et al. Randomized trial on the effect of radiation therapy in addition to 6 cycles CMF in node-positive breast-cancer patients. Int J Cancer 2000;86:408–15.[CrossRef][Web of Science][Medline]
(39) Hayat H, Brufman G, Borovik R, Chaichick S, Rath P, Robinson E, et al. Adjuvant chemotherapy and radiation therapy vs. chemotherapy alone for stage II breast cancer patients. Ann Oncol 1990;1(suppl):21.
(40) Morgan DA, Galea MH, Berridge J, Blamey RW, Ellis IO. After mastectomy, high-grade, node positive breast cancer patients benefit from radio-therapy. Clin Oncol 1992;4:335.[CrossRef]
(41) Levitt SH, McHugh R. Radiotherapy in the postoperative treatment of operable cancer of the breast. Cancer 1977;39:924–32.[CrossRef][Web of Science][Medline]
(42) Overgaard M. Overview of randomised trials in high risk breast cancer patients with adjuvant systemic therapy with or without postmastectomy irradiation. Semin Radiat Oncol 1999;9:292–9.[CrossRef][Web of Science][Medline]
(43) Galea MH, Blamey RW, Elston CE, Ellis IO. The Nottingham Prognostic Index in primary breast cancer. Breast Cancer Res Treat 1992;22:207–19.[CrossRef][Web of Science][Medline]
(44) Giordano G, Kuo Y, Freeman J, Buchholz T, Hortobagyi G, Goodwin J. Risk of cardiac death after adjuvant radiotherapy for breast cancer. J Natl Cancer Inst 2005;97:419–24.
(45) Cuzick J. Radiotherapy for breast cancer. J Natl Cancer Inst 2005;97:406–7.
(46) Taghian A, Jeong J, Mamounas E, Anderson S, Bryant J, Deutsch M, et al. Patterns of locoregional failure in patients with operable breast cancer treated by mastectomy and adjuvant chemotherapy with or without tamoxifen and without radiotherapy: results from five national surgical adjuvant breast and bowel project randomized clinical trials. J Clin Oncol 2004;22:4247–54.
(47) Darby S, McGale P, Taylor C, Peto R. Long-term mortality from heart disease and lung cancer after radiotherapy for early breast cancer: prospective cohort study of about 300,000 women in US SEER cancer registries. Lancet Oncol 2005:6;557–65.[CrossRef][Web of Science][Medline]
(48) McDonald B, Swanton J. Medical negligence: the doctor's duty to warn of risks and the role of expert evidence. Aust Law J 1993;67:145–8.
(49) Vinh-Hung V, Verschraegen C. Breast-conserving surgery with or without radiation therapy: pooled-analysis for risks of ipsilateral breast tumor recurrence and mortality. J Natl Cancer Inst 2004;96:115–21.
(50) Lythgoe JP, Palmer MK. Manchester regional breast study—5 and 10 year results. Br J Surg 1982;69:693–6.[Web of Science][Medline]
(51) Lythgoe JP, Leck I, Swindell R. Manchester regional breast study. Lancet 1978;1(8067):744–7.[Web of Science][Medline]
(52) Host H, Brennhovd IO. The effect of post-operative radiation therapy in breast cancer. Int J Radiat Oncol Biol Phys 1977;2:1061–7.[Web of Science][Medline]
(53) Management of early cancer of the breast. Report on an international multicentre trial supported by the Cancer Research Campaign. Br Med J 1976;1:1035–8.
(54) Cancer Research Campaign working party Cancer Research Campaign (King's/Cambridge) trial for early breast cancer A detailed update at the tenth year. Lancet 1980;2(8184):55–60.[CrossRef][Medline]
(55) McArdle CS, Crawford D, Dykes EH, Calman KC, Hole D, Russell AR, et al. Adjuvant radiation therapy and chemotherapy in breast cancer. Br J Surg 1986;73:264–6.[Web of Science][Medline]
(56) Ryden S, Ferno M, Moller T, Aspegren K, Bergljung L, Killander D, et al. Long-term effects of adjuvant tamoxifen and/or radiation therapy. Acta Oncol 1992;31:271–4.[Web of Science][Medline]
(57) Tenvall-Nittby L, Tengrup I, Landberg T. The total incidence of locoregional recurrence in a randomized trial of breast cancer TNM stage II. Acta Oncol 1993;32:641–6.[Web of Science][Medline]
(58) Gervasio H, Alves H, Rito A, Albano J, Gordilho M, Pimentel D, et al. Phase III study: adjuvant chemotherapy versus adjuvant radiation therapy plus chemotherapy in women with node-positive breast cancer. A retrospective analysis of 18 years. Breast J 1998;4(suppl):S88.
(59) Fisher B, Slack NH, Cavanaugh PJ, Gardner B, Ravdin RG. Postoperative radiotherapy in the treatment of breast cancer: results of the NSABP clinical trial. Ann Surg 1970;711–32.
(60) Friedl W, Scheurlen HR, Amberger H, Henningsen B. Radiation therapy in operable breast cancer—10 year results of a prospective randomized trial. J Exp Clin Cancer Res 1984;3:71–7.
(61) Muss HB, Cooper MR, Brockschmidt JK, Ferree C, Richards F, White DR, et al. A randomized trial of chemotherapy (L-PAM vs CMF) and irradiation for node positive breast cancer. Breast Cancer Res Treat 1991;19:77–84.[CrossRef][Web of Science][Medline]
(62) Ahmann DL, Scanlon PW, Bisel HF, Edmonson JH, Frytak S, Payne WS, et al. Repeated adjuvant chemotherapy with phenylalanine mustard or 5-fluorouracil cyclophosphamide and prednisone with or without radiation, after mastectomy for breast cancer. Lancet 1978;1(8070):893–6.[Web of Science][Medline]
(63) Papaioannou A, Lissaios B, Vasilaros S, Miligos S, Papadimitrou G, Kondilis D, et al. Pre- and postoperative chemoendocrine treatment with or without postoperative radiation therapy for locally advanced breast cancer. Cancer 1983;51:1284–90.[CrossRef][Web of Science][Medline]
(64) Faber P, Jedinsky H. Adjuvant chemotherapy in breast cancer—a multicenter trial. Cancer Treat Rev 1979;6(Supp):75–8.
Manuscript received April 29, 2005; revised October 20, 2005; accepted November 21, 2005.
Correspondence about this Article
Editorial about this Article
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2006 98: 1020-1021.
J Natl Cancer Inst 2006 98: 1021.
J Natl Cancer Inst 2006 98: 3-4.
This article has been cited by other articles:
|
|
 |
|
  F.-C. Bidard, Y. M. Kirova, A. Vincent-Salomon, S. Alran, Y. de Rycke, B. Sigal-Zafrani, X. Sastre-Garau, L. Mignot, A. Fourquet, and J.-Y. Pierga Disseminated tumor cells and the risk of locoregional recurrence in nonmetastatic breast cancer Ann. Onc., November 1, 2009; 20(11): 1836 - 1841. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Spillane, F. Noushi, R. A. Cooper, V. Gebski, and R. F. Uren High-resolution lymphoscintigraphy is essential for recognition of the significance of internal mammary nodes in breast cancer Ann. Onc., June 1, 2009; 20(6): 977 - 984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hickey, M. Peate, C.M. Saunders, and M. Friedlander Breast cancer in young women and its impact on reproductive function Hum. Reprod. Update, May 1, 2009; 15(3): 323 - 339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. S. Russell, I. H. Kunkler, G. van Tienhoven, P. A. Canney, J. Thomas, J. Bartlett, M. J. van de Vijver, Y. Belkacemi, J. R. Yarnold, and P. J. Barrett-Lee Postmastectomy Radiotherapy: Will the Selective Use of Postmastectomy Radiotherapy Study End the Debate? J. Clin. Oncol., February 20, 2009; 27(6): 996 - 997. [Full Text] [PDF]
|
|
|
|
 |
|
 N. C Turner and A. L Jones Management of breast cancer--Part I BMJ, July 4, 2008; 337(jul04_1): a421 - a421. [Full Text]
|
|
|
|
|
|
S. H. Yang, K. H. Yang, Y. P. Li, Y. C. Zhang, X. D. He, A. L. Song, J. H. Tian, L. Jiang, Z. G. Bai, L. F. He, et al. Breast conservation therapy for stage I or stage II breast cancer: a meta-analysis of randomized controlled trials Ann. Onc., June 1, 2008; 19(6): 1039 - 1044. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. B. Marks, J. Zeng, and L. R. Prosnitz One to Three Versus Four or More Positive Nodes and Postmastectomy Radiotherapy: Time to End the Debate J. Clin. Oncol., May 1, 2008; 26(13): 2075 - 2077. [Full Text] [PDF]
|
|
|
|
|
|
B. D. Smith, B. G. Haffty, A. Hurria, D. H. Galusha, and C. P. Gross Postmastectomy Radiation and Survival in Older Women With Breast Cancer J. Clin. Oncol., October 20, 2006; 24(30): 4901 - 4907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Shikama, M. Mitsumori, C. Yamauchi, H. Takekawa, K. Arakawa, S. Sasaki, and T. Teshima Patterns of Care Study for Postmastectomy Radiotherapy in Japan: Its Role in Monitoring the Patterns of Changes in Practice Jpn. J. Clin. Oncol., August 1, 2006; 36(8): 499 - 503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Kunkler, N. Russell, P. Canney, K. Venables, and J. Bartlett Re: Survival effects of postmastectomy adjuvant radiation therapy using biologically equivalent doses: a clinical perspective. J Natl Cancer Inst, July 19, 2006; 98(14): 1020 - 1021. [Full Text] [PDF]
|
|
|
|
|
|
C. D. Atkins Re: Survival effects of postmastectomy adjuvant radiation therapy using biologically equivalent doses: a clinical perspective. J Natl Cancer Inst, July 19, 2006; 98(14): 1021 - 1021. [Full Text] [PDF]
|
|
|
|
|
|
L. R. Prosnitz and L. B. Marks Postmastectomy Radiotherapy: Quality Counts! J Natl Cancer Inst, January 4, 2006; 98(1): 3 - 4. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||