NEW YORK (GenomeWeb) – Microarray analysis could provide an efficient, genome-wide approach to diagnosing hematological cancers, according to a new study published last month in Oncotarget.
Led by Jess Peterson and Svetlana Yatsenko, scientists from the University of Pittsburgh performed G-banded chromosome analysis, fluorescent in situ hybridization (FISH), and microarray studies on 27 samples from patients with hematological malignancies. They found that 90 percent of chromosomal abnormalities identified by karyotyping or FISH were also detectable by microarray.
"The array is detecting a loss or deletion of genomic material or gain or amplification of genetic material," Yatsenko told GenomeWeb. "The microarray is a tool that is able to discover these copy number changes to the precision of the base pair. It can exactly tell you which genes are deleted in a particular cancer sample, and that is directly related to the method of therapy, or prognosis, or even classification of the disease, something clinicians can use to directly monitor the patient."
Using this increased genomic resolution, the scientists also found that about 50 percent of their samples had additional abnormalities that were not seen with either karyotyping or FISH analysis.
For the study, the researchers used custom-designed arrays manufactured by Agilent. "In the Cancer Genome Atlas there are about 500 genes," associated with cancer, Yatsenko said. Many off-the-shelf arrays cover only these genes, but she and her colleagues found another 400 genes to put on the array after an extensive search of recent literature. They included genes related to those associated with blood cancers, as well as genomic regions that are generally associated with translocations.
Physicians looking to classify hematological malignancies have historically used genomic alterations detected by karyotype analysis, Yatsenko said. Over the years, other genes have been identified as the primary triggers of cancer progression. She noted that as many as 50 genes may play an orchestrated role in a particular patient's disease, she said.
"Even if it has the same name for different patients, the pathology behind that could be different and response [to treatment] could be different," for each patient, Yatsenko said.
Karyotyping is well established and good at whole genome analysis, Yatsenko said, but it has a low resolution, about 10 million base pairs, because it requires looking through a microscope.
"The karyotype analysis is not that useful when talking about molecular diagnostics," Yatsenko said. "We may not identify abnormalities even if they are present in the cancer genomes." Karyotyping also requires cells that are actively dividing. That's not always possible with cancer patients, who may be under treatment or whose cells have accumulated mutations that change the normal course of cell division.
FISH was developed in the 1990s and was aimed to find smaller chromosomal abnormalities not picked up by karyotyping, but it has its own limitations. Like karyotyping, it resolution is still lower than resolution of microarray analysis. "The main problem is that you have to know what you're looking for, it's a targeted technique," Yatsenko explained, "so you have to suspect some abnormality and use probes from a targeted region to prove it. If lucky, you may find this but there is a limitation in the available probes." Not so with microarrays. "You do not have to know prior to testing which region of the genome is rearranged. You just test the entire genome."
The limitations of the technologies have clinical consequences. In acute lymphoblastic leukemia, a disease commonly found in children, certain deletions in genes like PAX5 and IKZF1 are considered pathogenic, Yatsenko said. "When IKZF1 is deleted that significantly affects the patient prognosis. In that case the clinician may select more aggressive treatment rather than standard therapy that could be associated with survival." The deletions could be small, variable in size and undetectable even with FISH technology. "In this situation there is no way other than microarrays to detect these deletions," she said. "As many as 40 percent [of deletions] are simply missed with FISH, because they are smaller than the probes used to find them," she said.
Microarrays provide an efficient, higher-throughput technology for cancer analysis, Yatsenko said. "It's more innovative and takes into account not only the known and specific abnormalities, but also individual changes in the genome that affect the treatment and prognosis of the patient."
Microarrays aren't perfect, which is why Yatsenko said she and her colleagues tried to compare microarrays with existing technologies. They cannot detect transpositions, nor can they detect clonality. They look at DNA from all the cells of a sample so it's impossible to tell whether some cells have only accumulated a few or all of the detected mutations.
But the ability to inform personalized medicine makes it a potentially valuable tool for oncologists. Moreover, microarrays can eliminate inefficiencies inherent in FISH testing. "Sometimes you end up with 20 FISH probes and you still can't get a reliable diagnosis," Yatsenko said, where a single microarray experiment is more likely to yield a definite answer.
Arrays are available to clinicians but are not widely used. Not all clinicians and pathologists are aware of their availability and not all of them may have the knowledge of how arrays can be help inform the treatment of their patients, Yatsenko said. "This is why we structured our paper to show differences between the techs and show the advantages and limitations of each in the clinical setting," she said.
Reimbursement is another hurdle for microarrays to clear before they reach widespread use in the clinic. Many insurance providers consider the technology experimental and will not reimburse for testing, limiting even clinicians who are aware of the technology's benefits.
"If there is no reimbursement, clinicians are just stuck to the old methods," Yatsenko said.