NEW YORK – UK-based Angle, developer of the Parsortix cell isolation platform, has collected what it believes is enough data to pursue its first clinical diagnostic application, an ovarian cancer test designed to noninvasively assess whether a pelvic mass is benign or cancerous.
According to the company, recent results from its clinical verification study suggest the test could reduce both false positives and false negatives by 50 percent or more compared to the imaging and blood protein tests currently used.
Angle CEO Andrew Newland said that the new results demonstrate two things: First, it reconfirms that the Parsortix system can isolate intact, living cancer cells. Second, it illustrates why those cells are an ideal sample for this diagnostic application.
"The reason we have such high specificity [compared to other biomarkers] is because we have this great sample," he said. "With tests for ovarian cancer that are based on measuring serum protein, those protein markers can be upregulated for reasons which aren't cancer. Even if you have an absolutely perfect test to measure them, the thing that you're measuring is not actually specific.
"That's why other approaches have had such high false positives. It's not because they're not good tests; it's just that what they're measuring is not a good indicator," he added.
Angle's new study, led by investigators at the University of Rochester Medical Center has not yet been published in the peer-reviewed literature, but the firm made public its headline results this month, reporting that a gene expression assay analyzing cells isolated from patient's blood samples could discriminate individuals with malignant ovarian cancer from those with benign pelvic masses with high accuracy.
The researchers recruited 144 patients being clinically assessed after presenting with a pelvic mass, and of these, 48 were ultimately found to have a malignant tumor following surgery.
Angle analyzed blood samples, first capturing circulating cells with its Parsortix system, a microfluidic platform that is able to isolate intact, living cells, differentiating likely tumor cells and cell clusters from other circulating cells based on morphological features like size and deformability.
In contrast to other platforms that target cells with certain surface proteins, Parsortix is epitope independent and can capture a variety of CTCs regardless of their class.
Investigators then used an assay Angle developed called Landscape+, which measures the expression levels of 164 genes. Comparing data for the cancer cases and non-cancer cases in the cohort, the team ultimately derived a predictive algorithm that included RNA expression levels of 23 different genes added to physician's clinical risk assessment and patient age.
According to Angle, the area under the receiver operating characteristic curve for this ovarian cancer algorithm was 95.4 percent. Using a cutoff of 40 percent risk to define positive versus negative results, this would correspond to a sensitivity of 90 percent and a specificity of 93 percent. Compared to physician's initial risk assessment for patients in this study cohort, use of the algorithm with this cutoff would have reduced both the rate of false positives and false negatives by at least 50 percent, the company said.
Pre-surgical ovarian cancer risk tools serve mainly to help triage patients, ideally sending women who are at a higher risk to have their surgery performed by a gynecologic oncologist as opposed to a gynecologist. Patients may also be managed conservatively, with close monitoring instead of having surgery, but risk assessment algorithms have not been embraced by professional guidelines for informing that decision.
"In terms of commercialization, we've been developing this initially as a lab-developed test," Newland said. "We have two clinical labs, one which is in the United States in Philadelphia, which is the larger, lead effort. And then the secondary one is in the United Kingdom … and those two facilities are going to be accredited soon."
According to Newland, the company is also still exploring other commercial strategies including pursuing testing deals with large clinical centers or partnering with a larger-scale clinical lab company that could adopt and perform the test more broadly.
"We're also currently exploring whether the [gene expression aspect of the] test can be undertaken on standardized sequencing platforms. If so, then it could be partnered up with some of the big sequencing companies, and they could offer that potentially as a product rather than us [doing it] as a service," Newland added.
Angle would compete in the clinic with a variety of clinical and proteomic algorithms, including the US Food and Drug Administration-cleared OVA1 assay marketed by Aspira Women's Health.
In its pivotal validation data, Aspira reported a sensitivity of about 96 percent for OVA1 when combined with clinical impression but only a 35 percent specificity. The firm has since coupled the original assay with a second test, Overa, which bumps specificity up to 69 percent. The company sells OVA1 on its own or the two tests as a combined reflex under the name OVA1plus.
Other competitors include the ROMA algorithm marketed by Fujirebio, which weighs two protein biomarkers, CA-125 and HE4, alongside a variety of clinical factors. Potential new entrants are also exploring exosome-based technologies to extract more predictive proteomic signals and new targets like protein glycosylation.
Several professional guidelines groups endorse the use of biomarker assays to aid in the direction of high-risk women to specialist surgeons, but absent utility evidence that proves these tools improve outcomes, their ultimate value is still treated skeptically.
Aspira has collected some clinical utility data for OVA1, publishing an analysis of patient charts in 2016 that showed that among 122 patients with OVA1 results indicating elevated ovarian cancer risk, 89 percent were either referred to a gynecologic oncologist for surgery or had a gynecologic oncologist available during surgery. But the firm hasn't shared any prospective analyses indicating that this triage actually improves health outcomes or quality of life.
This gap has stymied insurance coverage for ovarian cancer risk tests across the board. CMS contractor National Government Services, for example, currently maintains a non-coverage policy for all multi-marker serum tests related to ovarian cancer testing.
"To date, none of the multi-marker serum tests addressed in this policy have been shown to reliably screen, improve quality of life, or decrease mortality in women with ovarian cancer, [and] coverage of these tests must await larger, prospective, non-industry funded data on long-term outcomes including quality of life, improvement in survival, and impact on mortality in women with ovarian cancer," the local coverage determination reads.
Although its new data is unpublished, Angle's report is in line with a previous study, published in October in Obstetrics & Gynecology, which analyzed results from 183 patients, concluding that a smaller panel of eight genes and four serum biomarkers could discriminate malignant tumors with an AUC of 95 percent.
Newland highlighted the fact that Angle's newest study was performed in a real-world setting, as opposed to the retrospective case-control trials that usually feature in early assay development and validation.
"These are women who all had a condition. They're actually the intended use population, so we think that that brings a lot of credibility to what we're doing," he said.
In terms of future applications, Newland said that prostate cancer offers a similar use case, where Parsortix and downstream gene expression analysis would aid physician decision-making. The company has partnered with Solaris Health, a network of US urology clinics that serve over 700,000 patients a year, with plans to begin studies in the next few months looking at the potential to guide biopsy decisions in men suspected of having a prostate tumor. Angle expects to have results in hand by the end of next year.
The company is also interested in exploring whether its platform could be used to serially monitor the thousands of women with pelvic masses who are not recommended to undergo surgery.
"Obviously, it's expected that they don't have cancer, that they've got a benign condition. But obviously, there's a risk that some of them are missed. And if you did have ovarian cancer, then you really don't want to be waiting, so we perceive that there's an opportunity for that population to have a simple blood test on a repeat basis, just to maintain a higher degree of certainty that they haven't actually got cancer," Newland said.