Early Breast Cancer Trialists' Collaborative Group, "Effects of
Radiotherapy and Surgery in Early Breast Cancer - An Overview of the
Randomized Trials", The New England Journal of Medicine, November 30,
1995, Vol. 333, No. 22, pp. 1444-1456

Background. Randomized trials of radiotherapy and surgery for early
breast cancer may have been too small to detect differences in
long-term survival and recurrence reliably. We therefore performed a
systematic overview (meta-analysis) of the results of such trials.

Methods. Information was sought on each subject from investigators who
conducted trials that began before 1985 and that compared local
therapies for early breast cancer. Data on mortality were available
from 36 trials comparing radiotherapy plus surgery with the same type
of surgery alone, 10 comparing more-extensive surgery with
less-extensive surgery, and 18 comparing more-extensive surgery with
less-extensive surgery plus radiotherapy. Information on mortality was
available for 28,405 women (97.4 percent of the 29,175 women in the
trials).

Results. The addition of radiotherapy to surgery resulted in a rate of
local recurrence that was three times lower than the rate with surgery
alone, but there was no significant difference in 10-year survival;
among a total of 17,273 women enrolled in such trials, mortality was
40.3 percent with radiotherapy and 41.4 percent without radiotherapy
(P = 0.3). Radiotherapy was associated with a reduced risk of death
due to breast cancer (odds ratio, 0.94; 95 percent confidence
interval, 0.88 to 1.00; P = 0.03), which indicates that, after 10
years, there would be about 0 to 5 fewer deaths due to breast cancer
per 100 women. However, there was an increased risk of death from
other causes (odds ratio, 1.24; 95 percent confidence interval, 1.09
to 1.42; P = 0.002). This, together with the age-specific death rates,
implies, after 10 years, a few extra deaths not due to breast cancer
per 100 older women or per 1000 younger women. During the first decade
or two after diagnosis, the excess in the rate of such deaths that was
associated with radiotherapy was much greater among women who were
over 60 years of age at randomization (15.3 percent vs. 11.1 percent
[339 vs. 249 deaths]) than among those under 50 (2.5 percent vs. 2.0
percent [62 vs. 49 deaths]). Breast-conserving surgery involved some
risk of recurrence in the remaining tissue, but no significant
differences in overall survival at 10 years were found in the studies
of mastectomy versus breast-conserving surgery plus radiotherapy (4891
women), more-extensive surgery versus less-extensive surgery (4818
women), or axillary clearance versus radiotherapy as adjuncts to
mastectomy (4370 women).

Conclusions. Some of the local therapies for breast cancer had
substantially different effects on the rates of local recurrence -
such as the reduced recurrence with the addition of radiotherapy to
surgery - but there were no definite differences in overall survival
at 10 years.

Trials of Radiotherapy

"Figure 2 shows survival among the approximately 16,000 women in the
35 trials of radiotherapy from whom individual data on survival were
collected, categorized according to nodal status.  There was no
statistically significant effect of radiotherapy in women with
node-positive or node-negative cancer."

"Overall, about one third more women in the radiotherapy groups than
in the non-radiotherapy groups died of "non-breast-cancer"causes (7.7
percent vs. 5.7 percent [527 vs. 391]), but this difference occurred
partly because those assigned to radiotherapy had slightly longer
recurrence-free survival and were therefore at risk for death without
recurrence for slightly longer.  After we allowed for this, there was
an increase of only about one quarter in such deaths (odds ratio, 1.24
+/- 0.08; 95 percent confidence interval, 1.09 to 1.42; P=0.002).
This increase of about one quarter was found among women in all age
groups: under 50, 50 through 59, and 60 or older, at randomization.
But, at least during the first decade or two after diagnosis, that
absolute excess was much greater among those who were 60 or older at
randomization (15.3 percent vs. 11.1 percent [339 vs. 249 deaths] that
among those under 50 (2.5 percent vs. 2.0 percent [62 vs 49])."

Trials Comparing More Extensive with Less Extensive Surgery

"Overall, 48.0 percent of the women assigned to more extensive surgery
and 50.1 percent of those assigned to less extensive surgery died;
this corresponds to a nonsignificant reduction of 3 percent in the
odds of death. ... Data on causes of death were available for only 53
percent of the women who died without a recurrence of breast cancer;
these data also showed no significant differences."

"Figure 4 shows survival according to nodal status for approximately
3400 women in trials comparing more extensive with less extensive
surgery.  The less extensive surgery was total or radical mastectomy
in some of these trials and simple mastectomy in all the others, since
data on individual patients were not available from the trial of
breast-conserving surgery.  No difference in survival was apparent
among either women with node-positive cancer or those with
node-negative disease."

"Among the women whose outcomes are summarized in Figure 4,
more-extensive surgery involved a nonsignificant reduction in the rate
of recurrence; 48.8 percent of those treated with more-extensive
surgery and 50.3 percent of those with less-extensive surgery had a
reported recurrence (odds ratio, 0.98 +/- 0.05 with no significant
heterogeneity among different trials or among different types of
surgery)."

"Some of the local therapies for breast cancer had substantially
different effects on the rates of local recurrence, but there were no
definite differences in overall 10-year survival.  It has long been
accepted that radiotherapy can delay or prevent local or regional
recurrence in women with early breast cancer, as may more extensive
surgery.  More recently, it has appeared that radiotherapy can also
produce a small increase in the rate of death from causes other than
breast cancer.  In this extensive overview, we confirmed these
findings, but we could not assess separately the effects of treatment
on deaths from cardiovascular or other specific causes or the
relevance of particular details of radiologic or surgical technique.
Our findings indicate, however, that the absolute excess rate of
non-breast-cancer mortality during the first decade or so after
radiotherapy is strongly related to age.  Among women who were under
50 when they underwent irradiation, the apparent excess is just a few
deaths not due to breast cancer per 1000 women, whereas among women
who we 60 or older at the time of radiotherapy, it is a few per 100.
As Table 2 suggests, the excess may persist for more than 10 years.
If such a proportional excess persists indefinitely, the absolute
excess might become appreciable even among women who were under 50
when they received radiotherapy.  Although the radiotherapy techniques
differed substantially among the studies, the overall result still
provides a valid measure of the value of such treatment."