Skip to main content
Premium Trial:

Request an Annual Quote

Will Microarray Profiles Ever Become Regular Features of Clinical Oncology?


ANAHEIM, Calif. — Microarray profiling has yet to become a reliable tool in molecular oncology, but, then again, the technology's strengths have yet to be fully exploited in that area. In fact, they may first prove to be more important in other areas, such as drug development, according to two leading lights of targeted oncology.

"I'm not against microarrays," Richard Simon, chief of the US National Cancer Institute's Biometric Research Branch, said at the start of his talk during the American Association for Cancer Research meeting, here this week.

Simon and Todd Golub, the Charles A. Dana Investigator in Human Cancer Genetics at the Dana-Farber Cancer Institute, spoke on the future of microarray profiling in clinical oncology in a forum entitled Expression Profiling of Tumors: Transforming Clinical Oncology or Stuck on the Research Bench?

While both speakers agreed that expression profiles have not yet made much of an impact on clinical decisions, Simon said profiling was more well-suited — initially, anyway — to a supporting role in drug development. He added that array-based diagnostics are probably not favored for cancers that have only one treatment, and for drugs that are already sold without a companion test.

Golub said microarrays can enter the clinic earlier as analysis and standardization evolves, and the introduction of microRNA tumor profiling.

Before expression profiles can gain broad acceptance, the methods researchers use need to be standardized and cleaned up, said Simon. In many researchers' hands, expression profiles reveal results that, while real, may not be broadly applicable because study populations and conditions are so tightly controlled, he said. Also, much research is conducted without first specifying a hypothesis, and models built using part of a particular dataset are sometimes then used to evaluate other parts of that dataset, he added.

Microarray profiles are powerful tools to identify biomarkers that can be developed for use in diagnostics — probably using other, more widely familiar technologies, such as RT-PCR, said Simon. Two main factors prevent all drug-targeting diagnostics from entering the clinical oncology arena, he told Pharmacogenomics Reporter at the AACR meeting this week. "One is the financial incentive, the other is, 'Is the classifier really going to be used?' he said.

To overcome the first of Simon's stumbling blocks — the financial incentive — he said companies will likely need to drive the validation and reproducibility studies necessary to establish a diagnostic in a clinical setting, as opposed to a tightly controlled study setting. In particular, developing a drug and a diagnostic together probably provides extra incentive for drug makers and diagnostic companies, he said. "In fact, with a new treatment, you have a company that wants to get their drug approved [and a diagnostic is] potentially targeting that drug to the right patients and helping [the drug maker]. In terms of getting the drug approved they'll need an assay approved to do it," he added.

Pharmacogenomic data "will be looked for in all treatments [AstraZeneca] is currently developing," said Mary Stuart, a research physician at AstraZeneca. "There is a big incentive for companies to" find biomarkers that can help stratify patients in clinical trials, and for the possible development of diagnostics, she said. "It's much more important that drugs be used in patients that respond."

In situations where only one treatment exists for a life-threatening illness, Simon sees little point in developing a diagnostic, whether alongside a drug or following one. "They're going to say, 'Well, your classifier may not be perfect, so I'm going to take my chance with the treatment anyway,'" he said.

In clinical oncology, Oncotype Dx, a test developed for breast cancer by Genomic Health using the NCI's tissue bank, is Simon's model of the likely future of microarrays. To identify genes capable of estimating recurrence risk in women on tamoxifen with node-negative, estrogen-receptor-positive breast cancer, researchers used microarray analysis. For the actual test, which is now run only in CLIA-approved labs, interrogation of the 21-gene set is performed by RT-PCR from paraffin-embedded tissue samples, a medium for which microarrays are not suited, Simon said.

To Todd Golub, microarray expression profiling has already had a major impact. Tumor banking has improved to keep up with the quality demands of the technology, clinical trials routinely feature use of the technology, and many drugs under development anticipate the use of pharmacogenomic tools as relevant diagnostics, he said.

But Golub is optimistic about expression arrays' diagnostic abilities, once the kinks are ironed out. In particular, his group's research with microRNA may lead to better tumor-typing methods than are possible using mRNA, he said. Using poorly differentiated gut tumor, the microRNA profiles could distinguish 11 of 17 tumor types, while mRNA profiles identified only one.

The evolution of technical knowledge should help usher expression profiling into the clinic, said Golub. Among the achievements he sees as instrumental is the Gene-Set Enrichment Analysis pioneered by the Dana-Farber Cancer Institute to link changes in expression with phenotypes.

In some ways Golub and Simon agree about the barriers facing expression arrays in the cancer clinic. The variability in sample collection methods, RNA isolation, and the technical performance of arrays is still too high to allow direct data comparison, necessitating the creation of a "classifier" that would let researchers at different sites, with different materials, draw the same conclusions, Golub said.

— CW

Filed under

The Scan

Study Tracks Off-Target Gene Edits Linked to Epigenetic Features

Using machine learning, researchers characterize in BMC Genomics the potential off-target effects of 19 computed or experimentally determined epigenetic features during CRISPR-Cas9 editing.

Coronary Artery Disease Risk Loci, Candidate Genes Identified in GWAS Meta-Analysis

A GWAS in Nature Genetics of nearly 1.4 million coronary artery disease cases and controls focused in on more than 200 candidate causal genes, including the cell motility-related myosin gene MYO9B.

Multiple Sclerosis Contributors Found in Proteome-Wide Association Study

With a combination of genome-wide association and brain proteome data, researchers in the Annals of Clinical and Translational Neurology tracked down dozens of potential multiple sclerosis risk proteins.

Quality Improvement Study Compares Molecular Tumor Boards, Central Consensus Recommendations

With 50 simulated cancer cases, researchers in JAMA Network Open compared molecular tumor board recommendations with central consensus plans at a dozen centers in Japan.