Researchers at the University of Michigan along with a crew of Boston scientists have discovered a 22-protein panel of biomarkers that was able in preliminary studies to identify prostate cancer with 82 percent sensitivity and 88 percent specificity.
The news is noteworthy because it could lead to the development of an alternative diagnostic to standard prostate specific antigen tests, which return false positives around 80 percent of the time and therefore trigger unnecessary prostate biopsies.
Also noteworthy is that, unlike other prostate cancer research groups that use mass spec-based methods for biomarker discovery, this group used a protein array-based method to discover their panel of biomarkers.
"Initially, we envision this new test could be used as a supplement to PSA. A physician might suggest a patient with elevated PSA have this test before a biopsy to better determine whether it's a cancerous or benign condition," said study co-author Arul Chinnaiyan, an associate professor of pathology and urology at the University of Michigan. "In the future, I think this could replace PSA."
According to a paper published last week in the New England Journal of Medicine, the researchers, led by Chinnaiyan, Xiaoju Wang, and Jianjun Yu, first discovered the panel of biomarkers using phage display protein microarrays. They then validated the protein biomarker panel on 128 blood serum samples.
"Initially, we envision this new test could be used as a supplement to PSA. A physician might suggest a patient with elevated PSA have this test before a biopsy to better determine whether it's a cancerous or benign condition. In the future, I think this could replace PSA."
Of the 128 samples, they found that eight of 68 controls and 11 of 60 prostate cancers were misclassified.
"These numbers appear to be a great improvement over what traditional PSA screening gives," said Mark Scholz, the medical director of the Prostate Oncology Specialists group in Marina del Rey, Calif.
The researchers first produced a phage library that expressed proteins expressed by cancerous cells. They then incubated that library with control serum to remove non-cancer-specific phage peptides. The library was then incubated with prostate cancer serum to identify autoantibodies against prostate cancer.
The researchers scanned 2,300 autoantibodies and initially narrowed that group down to 186 that reacted with blood serum from men with prostate cancer. Those 186 autoantibodies were then tested against 59 prostate cancer samples and 70 control samples. Finally, a panel of 22 compounds that best distinguished prostate cancer blood samples from control samples was selected out.
Using these 22 markers, only two of 70 controls incorrectly tested positive for prostate cancer, and seven of 59 prostate cancer samples were falsely negative.
"We sort of systematically tried different numbers of biomarkers," said Chinnaiyan. "Twenty-two is not necessarily the magic number, but it appears to be an optimal one. If we used too many or too few markers, the accuracy went down a bit."
Next, the researchers validated their findings using 128 serum samples. They found that eight of 68 controls incorrectly tested positive, and 11 out of 60 prostate cancer samples incorrectly tested negative. This means that 88 percent of the time, samples that were not cancerous were correctly identified, and 82 percent of the time, samples that were cancerous tested positive.
Since the test worked well for identifying prostate cancer in intermediate PSA ranges that do not always suggest cancer, the researchers suggested that the biomarker panel could be used to help determine whether a patient should undergo a biopsy.
"These numbers are definitely in the acceptable range, and they're certainly better than the PSA," said Chinnaiyan. "The next step is to validate on a larger, multi-institutional cohort."
Chinnaiyan said the cohort could involve thousands of patients, including those from the Harvard Cancer Center and a number from Sweden.
He also said that his group's test, which is based on protein microarrays, may have advantages over mass spec-based prostate cancer tests because chips could be more easily adopted than mass specs by clinics.
"I think the goals [of our test and mass spec-based tests] are generally the same, just the technology is different," said Chinnaiyan. "Definitely, a protein microarray based test could be easier to do than a mass spec test."
Scholz said Chinnaiyan's technique is a "very exciting" way to detect cancer at a very early stage.
"It would seem to me that this discovery technique should work for all cancers," he said.
He added that in order for the new test to be adopted into the clinic, it must be evaluated in a genuine screening situation with live humans and compared head-to-head with the best PSA technology.
"The best PSA technology incorporates PSA density, which is the ratio of PSA to prostate size, as well as PSA velocity, which is the rate of change in PSA over time," Scholz noted.
He pointed out that the over-diagnosis of innocuous, benign variants of prostate cancer that are not life threatening, and the resulting unnecessary treatment, may be becoming an even bigger problem than the failure to diagnose prostate cancer.
The new technology should help combat over-diagnosis, rather than exacerbating the problem, Scholz said.
Chinnaiyan said that the University of Michigan is evaluating potential commercialization partners for the test.
— Tien-Shun Lee ([email protected])