Gene-expression profiling of tumors may soon become a routine part of how an oncologist decides on the best way to treat an expanding number of malignancies. While many technologies are in early stages of clinical validation, few have made it to the market. As in the early days of 2-D electrophoresis, though, finding meaningful predictive differences in a sea of data remains elusive.
Increasing progress in validating multiple-gene signatures has reduced the firehose effect. In February 2007, the FDA okayed Agendia’s MammaPrint for the prediction of breast cancer returning within 5 to 10 years from initial diagnosis. It became the first multivariate in vitro diagnostic assay to be approved by the agency.
Fifteen percent of stage I breast cancers recur, and among those patients with recurrences, chemotherapy will benefit a minority. Since oncologists can’t predict who will respond the most to chemo, standard practice is to prescribe it to all patients. Treatment regimens cost between $15,000 and $30,000 per patient and have serious side effects.
Companies are thus developing gene-signature tests that aid in identifying patients most likely to benefit from treatment for various types of cancers. Thus far insurance providers have been willing to pay for tests because they potentially save on chemotherapy costs.
Leveraging Gene Signatures
Agendia, for example, designed its RNA-based test to help clinicians decide on the risks of a particular tumor recurring. MammaPrint is based on a microarray analysis of a panel of 70 genes in a tumor sample. It is indicated for use by physicians as a prognostic tool along with other clinicopathological factors including tumor size and lymph-node involvement.
FDA sanction was based on a study in 302 women up to age 60 at five European centers. The trial confirmed the test was useful in predicting time to distant metastasis in stage 1 and stage 2 disease where tumor size is equal to or less than 50 mm and when the cancer is lymph-node negative.
The firm is currently conducting another study to compare the relative accuracy of MammaPrint to traditional prognostic assessment methods. It intends to enroll 6,000 patients by late 2011, focusing on patients where there is discordant assessment between the two methods. To date over 3,100 patients have been enlisted.
Genomic Health’s Oncotype DX Breast Cancer Assay is a 16-gene expression test approved to predict the likelihood of chemotherapy benefit and recurrence risk for patients with early-stage, ER+ breast cancer. It is performed on a tumor sample to get a Recurrence Score, or a number between zero and 100 that shows the chance of the breast cancer returning within 10 years of the original diagnosis. The 16 genes used in the test were selected from three clinical trials that demonstrated their link to breast cancer recurrence.
In 2006, the NCI launched the TAILORx trial to further validate the Oncotype DX Breast Cancer Assay. In October, over 10,000 women had reportedly been enrolled in the study, and final results are expected in 2015. NCI notes that it is the first study to integrate a gene-expression assay into the decision-making process.
The Oncotype DX Colon Cancer Assay is a 12-gene assay approved for the prediction of colon cancer recurrence for patients with stage 2 disease. Genomic Health is also developing tests for ductal carcinoma in situ, colon, lung, and renal cancers, as well as melanoma.
ChipDX offers a colon cancer prognostic tool for research use only. Earlier in December the firm said that the 163-gene signature used in this test stratifies colon cancer patients into high- and low-risk groups for recurrence with greater accuracy than current methods.
Reporting in the December issue of the British Journal of Cancer, the company identified the predictive patterns from 232 stage 1–4 colon cancer patients. Multivariate analysis showed the classifier to be associated with approximately three- to fourfold increased risk of recurrence when tested in 60 stage 2 and 3 patients.
ChipDX has an online platform that performs multigene diagnostic analysis on whole-genome Affymetrix GeneChip data. The firm is reviewing regulatory requirements and potential partners to take it to the market as a diagnostic tool.
In an approach that can potentially obviate the need for tissue tumor samples, Source MDx says it has developed a test for prostate cancer using whole blood to measure RNA transcripts for six cancer- and immune-related genes. The diagnostic combined with PSA testing can significantly improve accuracy compared to PSA testing alone, the company says.
In a two-year prospective study of approximately 500 men conducted by Dana Farber Cancer Research Institute and Source MDx, the six-gene test demonstrated sensitivity and specificity greater than 90% in discriminating prostate cancer patients from healthy normal subjects.
But are physicians likely to use these prognostic tools in making treatment decisions? Earlier this month Genomic Health reported results from a meta-analysis of 912 breast cancer patients. In the study physicians who used Oncotype DX changed their treatment decisions in over one-third of their patients, leading to an overall reduction in chemotherapy use of approximately 28%. The analysis also found that the use of Oncotype DX led to the addition of chemotherapy in approximately 4% of patients who, prior to testing, were considered low risk.
Emily Faucet, senior director of corporate communications at Genomic Health, commented on getting physicians to use the test, “It’s challenging, because we are changing a treatment paradigm. But we have been able to generate the level of clinical evidence that’s necessary, and that gave physicians the confidence to use the test and payers to reimburse.
“As a result, the test has been in the market for seven years, and more than 10,000 physicians have used it on more than 175,000 early-stage breast cancer patients,” Faucet continued. “Right now about 50% of women who qualify for the test get it.”
As validation of these genomic analyses tools accrue and insurers remain willing to pay for them, molecular diagnostics and prognostics may soon become a part of standard clinical practice.