Personalized Health Care (PHC) in Cancer - Genomic Biomarkers for Pharmaceutical Development - page 36

shows the genes included in this panel; the algorithm used to calculate a recurrence score from
these genes is described in detail by Paik et al. [98] .
A recurrence score of <18 is considered low risk, whereas RS ≥18 and <31 is considered inter-
mediate risk, and RS ≥31 is considered high risk for distance recurrence. A randomized trial of
2617 node-negative, hormone receptor positive women taking tamoxifen versus placebo was
used to evaluate the 21-gene assay [99] . Of the 668 patients taking tamoxifen, 93% of patients
in the low risk group were free from distance recurrence compared to only 69% of the high
risk group [98,99] . The probability of developing a distant recurrence was 31% in the high risk
group, 14.4% in the intermediate group, and 6.9% in the low risk group. Not only does the recur-
rence score predict distant recurrence for all age categories and across tumor sizes in multiple
trials, but it also predicts overall survival [98] . The 21-gene assay has been endorsed by ASCO as
a tumor marker [90] and the National Comprehensive Cancer Network (NCCN) Breast Cancer
Panel considers it an option for patients with ER+ node-negative breast cancer for guiding
decision making regarding adjuvant chemotherapy. To further increase the ability to interpret
Oncotype DX data for therapeutic decision making, the TAILORx study is testing BC patients
with the Oncotype DX and then assigning patients with intermediate risk of recurrence to either
hormonal therapy alone or chemotherapy and hormonal therapy [100] . This trial will provide
key information on this prognostic test regarding the intermediate risk group, where it is cur-
rently unclear whether these patients benefit from the addition of chemotherapy [100] .
In contrast to Oncotype DX, MammaPrint (Agendia) was developed as a general prog-
nostic test for premenopausal untreated patients, and its role as a predictive marker for a
particular treatment response was not examined when initially developed. This test can
also be used to identify patients who may benefit from adjuvant chemotherapy + hormone
therapy, and who are at greatest risk of relapse following treatment [101,102] . Development
of this prognostic test originated with a microarray analysis of the relative expression lev-
els of 25,000 genes in tumors from a cohort of 78 premenopausal patients younger than 55
years without nodal metastasis who received no systemic therapy [103] . By associating the
expression of each gene with clinical outcome, a prognostic algorithm based on the expres-
sion levels of 70 genes was established. A threshold to separate patients with good and poor
prognoses was determined in this discovery cohort of 78 patients, and the average gene
expression level of the 70 genes in patients with a good prognosis was calculated [103] . For
any new case, a correlation coefficient for the 70 gene expression classifier is developed by
comparing the expression levels to the average value for the good prognosis group from
the first study cohort. Tumors are classified as having a good prognosis if the correlation
coefficient is above 0.4. Turning the 70-gene assay into the MammaPrint assay was achieved
by designing a custom microarray with fewer probes than the original array of 25,000
genes, which allows eight patients to be tested on a single microarray slide [104] . As the
MammaPrint assay is based on microarray technology, it requires RNA to be obtained from
frozen tissue. The inability of this test to be applied to FFPE samples limits its clinical utility.
In a cohort of 80 postmenopausal women with ER+ and node-negative breast cancer,
the MammaPrint assay classified 27 patients as low risk and these patients had excellent
clinical outcomes [105] . The clinical utility of the MammaPrint assay is currently being
tested in a randomized clinical trial called MINDACT (Microarray In Node-Negative and
1 to 3 positive lymph node Disease may Avoid Chemo Therapy), a prospective randomized
study comparing the 70-gene signature with common pathological criteria in selecting BC