SAN ANTONIOA retrospective analysis of data from the
National Surgical Adjuvant Breast and Bowel Project (NSABP) Protocol B-24
suggests that estrogen-receptor (ER) expression predicts response to tamoxifen
(Nol-vadex) among patients with ductal carcinoma in situ (DCIS).
Speaking at the 25th Annual San Antonio Breast Cancer Symposium
(abstract 30), D. Craig Allred, MD, professor of pathology, Baylor College of
Medicine, reported that tamoxifen treatment was associated with an approximate
50% reduction in the risk of subsequent breast cancer among ER-positive
patients, compared with placebo, but had little benefit among those who were ER
The study group was drawn from the original B-24 trial, which
randomized 1,804 patients with DCIS to lumpectomy and local radiation plus
either tamoxifen or placebo. In the original trial, the use of tamoxifen was
associated with a 37% reduction in all breast cancer events at 5 years.
Estrogen-receptor status was not a prerequisite for enrollment in the B-24
In the current study, researchers from Baylor and the NSABP
evaluated tamoxi-fen response among all women with known receptor status in the
B-24 trial. ER results were available on a subset of 676 patients344 in the
placebo group and 332 in the tamoxifen group.
Estrogen-receptor results for 450 of the patients had been
determined in a central reference laboratory using a standardized
immunohistochemistry method. The remaining 226 results were obtained from
outside laboratories and were determined primarily by immunohistochemistry
procedures. However, differences were apparent in the sensitivity of the
various assays used by the outside laboratories (see discussion below).
In the study population overall, 23% of the tumors were ER
negative, and 77% were ER positive. In the tamoxifen group, 20% were ER
negative, and 80% were ER positive. Among patients in the placebo group, 25%
had ER-negative tumors, and 75% had ER-positive tumors. These differences were
not statistically significant, Dr. Allred said.
Similarly, no significant differences were noted between the
two arms in the distribution of several other variables that might influence
outcome, including tumor size, surgical margins, comedo necrosis, and patient
When the investigators analyzed patient outcome according to
receptor status and tamoxifen treatment, they found little advantage to
tamoxifen in the patients with ER-negative tumors. At a median follow-up of 8.7
years, the percentage of ER-negative women who had experienced subsequent
breast cancer was 23% in the tamoxifen-treated group and 26% in the placebo
In contrast, among ER-positive women, the percentage of breast
cancer events was 10% in the tamoxifen group, compared with 23% in the placebo
groupabout a 50% reduction for ER-positive women treated with tamoxifen.
"The Kaplan-Meier curves for time to first breast cancer
event show that 90% of tamoxifen-treated women with ER-positive DCIS were
disease free at 9 years," Dr. Allred said, "compared with 70% to 80%
of those in the other groups. This again emphasizes the benefit of tamoxifen in
In formal statistical analyses of results, ER-positive patients
treated with tamoxi-fen had a relative risk of subsequent breast cancer of
0.41, a highly significant 59% reduction (P = .0002). Among ER-negative
women, the relative risk was 0.80a 20% risk reduction that was not
significant (P = .51).
Effect of Different Assays
"The trend for an apparent benefit from tamoxifen in
ER-negative DCIS is puzzling," Dr. Allred said. "A comparison of ER
results between the central and outside laboratories may shed some light on
The central reference laboratory used in the study relied on a
standardized assay that had been validated as useful in predicting response to
hormonal therapy in several published studies of invasive breast cancer.
Results from the
central laboratory showed that 20% of the tumors were ER negative. However, the
negative rate from the outside labs was 30%, or 50% higher.
"Studies assessing ER status by immunohistochemistry in
invasive breast cancer have shown that the false-negative rate can be fairly
high in laboratories using diverse assays that have not been standardized or
validated, and the same may be true for DCIS," Dr. Allred said. "This
idea is supported by comparing the clinical results from outside laboratories
in this study to results from the central laboratory."
When the researchers restricted their analysis to the
ER-negative results determined only in the outside laboratories, tamoxifen
appeared to reduce subsequent breast cancer by 57% in "ER-negative"
DCIS. But an analysis restricted to ER-negative results obtained from the
central reference laboratory found "absolutely no benefit" to
tamoxifen in ER-negative DCIS.
"It is likely that the ER-negative group defined by
outside laboratories contained many false-negative resultsup to 50%giving
a false impression of response to tamoxifen," Dr. Allred said. "This
type of error results from tests with insufficient sensitivity. This is a
danger because most physicians probably would not give tamoxifen to patients
with ER-negative DCIS, denying those with false-negative results a chance of
benefiting from the drug."