Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials

Authors: Early Breast Cancer Trialists' Collaborative Group (EBCTCG)
Source: Lancet. 2005 366:2087-2106
Affiliation: University of Oxford, Oxford, UK

Background

Numerous randomized, controlled trials have studied therapies directed at improving local control in early breast cancer, especially the role of adjuvant radiotherapy. Every 5 years since 1985, the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) has centrally reviewed these trials and performed meta-analyses based on individual patient data. The last EBCTCG meta-analysis ( Lancet 2000 ) demonstrated that the addition of adjuvant radiation therapy (RT) after breast conserving surgery improved local control (by 68%) and breast cancer-specific survival (by 14%), but the improvement in overall survival was marginal (5%) and did not reach statistical significance. In contrast, post-mastectomy radiation in patients with positive lymph nodes improved local control, breast cancer-specific survival, and overall survival. This is an update of the EBCTCG meta-analysis that evaluates the impact of local control on 15-year (yr) survival.

Methods

• Design: Meta-analysis of randomized, controlled trials in which individual patient data has been obtained and updated
• Patients: 42,000 patients (pts) from 78 unconfounded randomized controlled trials with early stage breast cancer that began recruitment by 1995
• Treatments compared:
  • RT vs. No RT (23,500 pts)
  • More vs. Less Surgery (9,300 pts)
  • More Surgery vs. RT (9,300 pts)
• Endpoints: Rates of breast cancer recurrence (local vs. regional/distant), breast cancer-specific survival, overall survival, and secondary malignancy.
  • Contralateral breast cancer was not counted as a local recurrence
  • Underlying causes of death were investigated further

Results

• RT after Breast Conserving Surgery
  • RT decreases 5-yr local recurrence (LR) from 26% to 7% (70% relative decrease)
  • RT decreases 15-year breast cancer mortality by 5.4% (HR 0.83)
  • RT decreases 15-year overall mortality by 5.3% (p=0.005)
• RT after Mastectomy/Axillary Lymph Node Dissection in Node-Positive Women
  • RT decreases 5-yr LR from 23% to 6% (similar relative decrease)
  • RT decreases 15-yr breast cancer mortality by 5.4%
  • RT decreases 15-yr overall mortality by 4.4% (p=0.0009)
    • In node-negative women, RT did NOT decrease breast cancer mortality, and the absolute local recurrence improvement was small
• In all trials comparing RT vs. No RT, complications were assessed:
  • RT increased risk of contralateral breast cancer (RR 1.18)
  • RT increased risk of non-breast cancer mortality (RR 1.12)
    • Heart disease mortality (RR 1.27)
    • Lung cancer mortality (RR 1.78)
• In order to assess the relationship between Local Control and Breast Cancer Mortality in general, trials were separated into 3 groups based on absolute reductions in LC: (<10%, 10-20%, >20%).
  • A 20% reduction in local recurrence translates into a 15-yr breast cancer mortality reduction of 5.2%
  • Authors propose a "4-to-1" relationship between LR at 5 yrs and breast cancer mortality at 15 yrs

Discussion

• There is a constant ratio between local control and reduction of breast cancer specific mortality ("4-to-1" rule)
• RT results in a relatively consistent, proportional reduction in local recurrence (~70%), and therefore the absolute magnitude of effect on breast cancer-specific survival depends on the magnitude of the risk of local recurrence.
• Competing excess mortality from RT (heart disease and lung cancer) results in a reduction of overall survival that is significant, but is still less than the improvement it confers in breast cancer-specific survival.
  • The death rate after 15 years from causes other than breast cancer is higher than before 15 years. The long follow-up is generally from the much older trials (1960s and 70s).
  • Although the major impact on breast cancer mortality is realized in the first 15 years, late toxicity from radiation continues indefinitely.
• Modern radiotherapy techniques and equipment are considerably safer than those employed in many of these earlier trials. It is now less common to treat internal mammary lymph node chains that overlie the heart and lungs. It is also standard practice at present to observe radiation dose limits to heart and lung tissue.

Comments

• Compared to the last EBCTCG meta-analysis that addressed radiotherapy in early breast cancer, longer follow-up has yielded statistically significant relationships:
  • RT after breast conserving surgery decreases both breast cancer mortality AND overall mortality
• By considering other local treatments that impact local recurrence, the authors have modeled the impact of 5-yr local control on 15-yr breast cancer mortality (the "4-to-1" rule). These relationships will be helpful in designing future trials to assess the impact on survival of newer, emerging local therapies, such as partial breast irradiation.
• An interesting omission from this overview was the relationship between TREATMENT ERA and OVERALL SURVIVAL, although the authors hint that modern radiotherapy may lead to decreasing rates of non-breast cancer and contralateral breast cancer events.