The nascent field of clinical proteomics got a big boost this week as several companies announced they have either begun marketing their proteomic-based ovarian cancer tests, or have moved closer to that goal.
On Monday, Laboratory Corporation of America launched its OvaSure screening test, a six-biomarker assay developed in collaboration with Yale University School of Medicine to detect the presence of early-stage ovarian cancer in high-risk patients.
Two days later, Vermillion announced that it had filed its OVA1 ovarian tumor triage test with the US Food and Drug Administration for 510(k) pre-market approval.
And a spokeswoman from proteomic diagnostic developer Correlogic told ProteoMonitor this week that the company plans to “shortly” file for regulatory approval of its OvaCheck test for early detection of epithelial ovarian cancer.
Ovarian cancer has long been considered a promising application area for clinical proteomics. Despite being one-tenth as common as breast cancer — around 22,000 women were diagnosed with ovarian cancer in 2007, according to the American Cancer Society — its mortality rate is about three times higher because it is difficult to detect in its earliest stages: nearly 80 percent of patients are not diagnosed until they are in the advanced stages of the disease.
Currently, screening strategies for ovarian cancer rely on pelvic and rectal exams or ultrasound because there are no reliable biomarkers for detecting the disease. The most commonly used biomarker in the field, CA-125, is useful for detecting late-stage ovarian cancer, but is a very poor predictor of early-stage disease. As a result, a number of groups have turned to proteomics as a potential alternative.
The road to the clinic has been a bumpy one, however. Correlogic was actually one of the earliest players in the field, and began developing its test in 2002 based on a study led by Lance Liotta and Emanual Petricoin that was published that year in the Lancet. The study was one of the first to show that a proteomic pattern could discriminate ovarian cancer from normal samples, but eventually drew criticism from some in the research community who questioned the reproducibility of the study.
Commercializing the test was further delayed in 2004 when the FDA warned Correlogic that it would need to file for premarket approval of the test [see PM 02-27-04]. The company hit another regulatory snag last year when the FDA released its draft guidance on in vitro diagnostic multivariate index assays, signaling its goal to begin regulating tests that use algorithms to interpret gene and protein data, a category that includes OvaCheck [see PM 02-15-07].
Now, however, the firm is on track to file with the agency, Annette Fribourg, vice president of government and public relations at Correlogic, told ProteoMonitor. “We have been conducting trials on the OvaCheck test in anticipation of a regulatory submission,” she said.
“We have several blood tests in the pipeline, and the OvaCheck test will be the first of our tests, and we’re expected to make our regulatory submission for Ovacheck shortly,” she added.
Fribourg declined to provide details on the company’s filing status, however, noting that “there are still some issues that we are discussing with the FDA.”
Vermillion, meantime, has also encountered a few hurdles on its path towards commercialization. The company, formerly Ciphergen, originally began developing its ovarian cancer test in 2004 after licensing a set of protein biomarkers from Johns Hopkins University [see PM 08-20-04]. After selling its SELDI business to Bio-Rad in 2006 and focusing its efforts exclusively on diagnostic development, the company said in late 2007 that it expected to file for FDA approval for the test in early 2008 [see PM 11-08-07].
In March, Vermillion announced that the test met its primary endpoints in a clinical study of 550 patients conducted at 27 clinical sites within the US. The test is designed to stratify women with pelvic masses into high-risk and low-risk categories to help determine if they should be referred to a specialist prior to surgery.
Vermillion said that its biomarker panel ruled out malignancy with approximately 95-percent certainty and exhibited around 90-percent sensitivity in detecting malignant ovarian tumors.
This week, the company submitted its 510(k) filing, representing “an important milestone for Vermillion and a significant step toward the commercialization of OVA1,” Gail Page, president and CEO of Vermillion, said in a statement.
Vermillion officials could not be reached for further comment, but the company said it plans to host a “roundtable teleconference” to discuss the OVA1 test with investors on July 15.
On the Fast Track?
LabCorp’s OvaSure test, meanwhile, has had the quickest route to market, though it still took several years to commercialize. In addition, the company has not yet filed for FDA approval for OvaSure, but is instead marketing it as a laboratory-developed test under the Clinical Laboratories Improvement Act.
The OvaSure test grew out of a study that Yale researchers, led by Gil Mor, associate professor of reproductive sciences, published in PNAS in 2005. In that paper, Mor and colleagues described a panel of four biomarkers that could discriminate between disease-free patients and those with early-stage ovarian cancer, but the sensitivity and specificity of the test were only around 95 percent — not predictive enough for a commercial test.
“Although the data from the Phase III trial has not been published, our internal data that we’ve generated to date at LabCorp on OvaSure has been validated with 99 percent correlation to the performance characteristics in the Phase II trial.”
LabCorp signed a license agreement with Yale to commercialize the test in 2006, and earlier this year, Mor’s team published a Phase II study in Clinical Cancer Research that included two additional markers, one of which was CA-125. These markers increased the sensitivity to 95.3 percent and the specificity to 99.4 percent. By comparison, CA-125 alone has sensitivity of 72 percent and sensitivity of 95 percent.
While CA-125 is not the most useful biomarker on its own, “in combination with others it becomes very good,” Mor told ProteoMonitor.
The Yale researchers are currently conducting a Phase III trial for the test and expect to publish their results by the end of the year.
Mor said that the study is using samples from National Institutes of Health’s Prostate, Lung, Colon, and Ovarian cancer repository as well as from the non-profit Gynecological Oncology Group. “The results will be evaluated some time in July and we will have a final report maybe in August or September,” he said.
In the meantime, LabCorp determined that the test is accurate enough to begin offering it to physicians. “Although the data from the Phase III trial has not been published, our internal data that we’ve generated to date at LabCorp on OvaSure has been validated with 99-percent correlation to the performance characteristics in the Phase II trial,” Myla Lai-Goldman, executive vice president and chief medical officer of LabCorp, told ProteoMonitor.
Mor declined to provide details of the study’s findings prior to publication, but said that the results so far have been “excellent.”
Lai-Goldman said that demand is high for an effective early screening test, and described OvaSure as “a novel assay” in the area of ovarian cancer.
“There has not been any consistently reliable screening test to detect ovarian cancer,” she said. While CA-125 is effective in some cases, “the challenge with CA-125 when used alone is that it’s frequently normal in individuals with early-stage ovarian cancer and only becomes elevated in more advanced disease.
“In fact, it’s only shown to be elevated in about 50 percent of patients with Stage I disease, and it can be positive in patients who don’t have cancer, so we really have not had a test with appropriate sensitivity and specificity,” Lai-Goldman said.
The company currently offers OvaSure as a laboratory-developed test, but Lai-Goldman noted that LabCorp plans to “continue to seek guidance on regulatory status of our offering from the FDA.”
LabCorp is also keeping an eye on Correlogic’s test. The reference lab signed a licensing agreement for the firm’s ovarian cancer test in 2002, as did Quest Diagnostics, but it appears that it has not played an active role in developing it.
“I believe that [Correlogic] has been doing further studies, and certainly we will continue to track what they are doing as well in the development of their test,” said Lai-Goldman, without elaborating.
A Brighter Future Ahead?
Lai-Goldman said that LabCorp sees promise in proteomic-based tests for certain applications. As an example, she noted a collaboration that the company announced in January with Duke University Medical Center to commercialize a four-protein assay for the early detection of lung cancer [see PM 01-17-08].
The test is being developed as a follow-on for patients who have already had a spiral CT scan, “which has a very high degree of sensitivity, but a lower specificity,” she said. “That’s another area in which patients have not had good access to screening tests, and where we think proteomic testing has the potential for significant contribution.”
The primary advantage of proteomic-based tests, Lai-Goldman said, is that “one isn’t putting the whole dependence of sensitivity and specificity on a single biomarker, but [is] able to use combinations of biomarkers, which, with carcinogenesis, is a better way to achieve the specificity and sensitivity that we need for this type of assay.”
Mor agreed, noting that while proteomics diagnostic development has had some “setbacks, “LabCorp’s launch of the test indicates that “that there is hope when using these technologies.”
Specifically, he said, “the main message is that the combination of multiple proteins may represent the right approach in order to develop a test for the detection of a disease.”
In addition, he said, the test’s accuracy is evidence that researchers in the field need to work with “real proteins” as opposed to protein fragments from mass spectrometry, which is “very non-specific.”
Mor noted that there are still a few difficulties ahead, however. One challenge for multiple-marker test development, for example, is identifying the optimal number of proteins to use in a panel.
“The risk is that by putting in too many [proteins], you may introduce a lot of variability into the test. So the challenge is to find the appropriate number in the combination,” he said. “I don’t want to say that six is the magic number – absolutely not. But the more markers you introduce, the potential of having variability in your test also increases, and the lower the number, the more restrained you are, so that’s a problem.”
In addition, he said that the ovarian cancer test still needs to be validated in longitudinal studies. “The sensitivity and specificity may change in the longitudinal studies, but we think it will stand much, much better than CA-125 alone,” he said.
Lai-Goldman said that LabCorp plans to look out for additional tests to add to its proteomics portfolio. “We will continue to look for tests from either universities or other sources where the data is promising, and where we feel that by collaborating we can produce a test that patients need, where there is an unmet medical need.”
Other clinical proteomics firms are also optimistic about the future of the field. “We see dramatic breakthroughs occurring in the next two- to four-year year time span, as proteomic and metabolomic tests begin to enter the mainstream of medical practice,” Correlogic’s Fribourg said.
She noted that the company is also developing a serum-based triage test for colorectal cancer as well as a serum-based assay for breast cancer. “These will be minimally invasive, and less costly than many existing technologies; both factors will save medical and financial resources and enhance patient compliance,” she said.