ORLANDO - Oncotype DX was prognostic and predicted
estrogen-receptor (ER)-positive, node-negative breast cancer patients' response
to adjuvant tamoxifen in a study presented by Soonmyung Paik, MD, at the
American Society of Clinical Oncology 41st Annual Meeting (abstract 510). "The
advantage of Oncotype DX is that it is quantitative in nature," Dr. Paik said.
Prior studies have shown the commercially available 21-gene
Recurrence Score (RS) assay, also known as Oncotype DX (see Figure), was
independently prognostic of distant metastasis in tamoxifen-treated patients.
It also showed a linear relationship between increased recurrence score and
increased benefit of chemotherapy, thus allowing women with low-risk scores to
"The academic question remained as to whether the Oncotype
DX assay is a purely prognostic factor or predictive of tamoxifen response,"
said Dr. Paik, director of the Division of Pathology at the National Surgical
Adjuvant Breast and Bowel Project (NSABP), Pittsburgh.
Dr. Paik and his colleagues sought to discover if RS could
predict prognosis, response to hormonal therapy, or both in patients
participating in the NSABP B-14 trial, a 1980s' study designed to determine
whether tamoxifen was more effective than placebo in prolonging disease-free
and overall survival.
For about 20% of the B-14 patients, 645 women, the
investigators could evaluate tumor tissues extracted from paraffin blocks: 355
in the placebo arm and 290 in the tamoxifen cohort. The original B-14 trial
measured ER status using ligand binding.
The assay proved prognostic in the placebo arm of B-14, with
low-risk patients (RS < 18) having a 10-year distant-recurrence-free survival
of 86%, intermediate risk (RS between 18 and 30) 62%, and high risk (RS
≥ 31) 69% (P = .0001).
The assay also was shown to be a predictive factor for women
receiving tamoxifen. For low-risk patients, there was a 51% relative benefit
with tamoxifen use and a 7.2% absolute benefit at 10 years (P = .04).
For intermediate-risk patients, researchers found a 46% relative benefit and a
17.3% absolute benefit (P = .02). And in high-risk women, they found a
5% relative benefit and a 1.6% absolute benefit (P = .82).
The researchers also looked at whether individual components
of the RS predicted benefit of tamoxifen. ER level measured by ligand binding
correlated with tamoxifen benefit: the higher the ER, the higher the benefit,
Dr. Paik said. Quantitative measurement of ER gene expression level using the
Oncotype DX assay proved highly predictive of benefit from tamoxifen, with a
hazard ratio of 0.43 at the highest ER level, compared with 4.86 for the lowest
ER level (ER negative). "This ER-tamoxifen interaction was highly significant
with a P value of .0006," Dr. Paik said. He noted that HER2 or
progesterone-receptor status did not significantly predict tamoxifen response.
Dr. Paik pointed out that there was no absolute correlation
between ER gene expression and the RS score itself. "Those patients with a high
ER score can also have a low or high RS," he said, "so you cannot use ER level
to predict RS."
He concluded that "the Oncotype DX assay captured both
prognosis and response to tamoxifen treatment, and could predict the likelihood
of recurrence in node-negative, ER-positive breast cancer patients. With
further validation, Oncotype DX may eventually help us determine which patients
will benefit from hormonal treatment. However, we do not recommend the use of
Oncotype DX to exclude patients from receiving endocrine therapy at this
During the discussion, Joseph A. Sparano, MD, professor of
medicine, Albert Einstein Comprehensive Cancer Center, Bronx, New York, called
Dr. Paik's conclusions conservative and supported by the data presented. He
said the findings raise a question of whether the assay can identify
ER-positive disease that doesn't benefit from hormone therapy. "This could
spare some patients up to 10 years of unnecessary treatment," Dr. Sparano said.
"Although this is an exciting potential application of this technology, it is
premature to withhold therapy from those with a high recurrence score." Dr.
Paik agreed with this comment.
Dr. Sparano indicated that two prospective, randomized clinical trials, now
under development or under way, will attempt to evaluate breast cancer markers
and tests to determine their clinical use. The studies are the Program for the
Assessment of Clinical Cancer Tests (PACCT) and Microarray in Node Negative
Disease May Avoid Chemotherapy (MINDACT). "Completion of these two important
studies will hopefully let us know precisely what type of molecular profiling
may be applied in every day clinical trials," Dr. Sparano said.