NEW YORK (GenomeWeb) – Provista Diagnostics presented two studies at last week's San Antonio Breast Cancer Symposium as it continues development efforts for its Videssa Breast proteomic breast cancer test.
The company is also participating in the US government's Cancer Moonshot program, working within the program's Blood Profiling Atlas Project to analyze sources of pre-analytical variation in its test and those of other companies, Judy Wolf, Provista's chief medical officer, told GenomeWeb.
Additionally, the company is slowly expanding its early-access program for Videssa and preparing for health economic and clinical utility studies for the test, Wolf said.
The Videssa test is intended to aid physicians in the interpretation of mammograms, helping them to distinguish between benign and malignant lesions that appear during imaging and to interpret results in patients with dense breasts, which can be difficult to assess via mammography.
Mammogram findings are graded using the Breast Imaging-Reporting and Data System (BI-RADS), a 0 to 6 scale with likelihood of malignancy rising with the number grade. A score of 1, for instance, is considered negative, while a score of 6 is a proven malignancy. Scores in the 3 to 4 range can be challenging to interpret, though.
The Videssa test measures 15 proteins — a mix of serum protein biomarkers and tumor-associated autoantibodies — to rule out cancer in women with either a borderline BIRADS score or where interpretation of a result is difficult due to dense breast tissue.
One of the company's two SABCS presentations focused on this latter group of patients. In that study Provista looked at 454 patients with BI-RADS scores of 3 or 4 and with available information on their breast density. In the non-dense patients, the Videssa test performed with sensitivity of 92.3 percent and specificity of 86.6 percent, with negative predictive value of above 99 percent. In the dense patients, the test had sensitivity of 88.9 percent and specificity of 81.2 percent, with negative predictive value likewise above 99 percent, demonstrating, the authors said, that the test is effective for ruling out breast cancer in both non-dense and dense-breast patients.
Wolf suggested that these results and the test more generally stand to become particularly relevant as awareness grows of the difficulty of using imaging in patients with dense breasts.
A number of states now require that patients with dense breasts be notified that their mammogram results may not be accurate, but, she said, "it's not clear yet what to do about it."
One potential follow up tool is MRI, but these scans are typically quite expensive. The Videssa test could present another option, at least for excluding those patients who test negative from further follow-up.
"Because our test is not looking at the anatomic picture, [but] at the physiologic and biochemical changes and reactions to breast cancer, the density of the breast doesn't really matter," Wolf said.
Better methods for screening women with dense breasts "is something that wasn't even recognized as a need a few years ago, and I think is just going to become a more and more obvious need as women now are going to know they have dense breasts and know what that means," she said. "So, this is a potentially huge area where [the test] can improve outcomes and save money by [distinguishing] women who may not need more imaging from those who do."
The second study Provista presented at SABCS looked at the stability of the Videssa assay over time, comparing test results at baseline and at a six-month follow-up visit. Looking at samples from 235 women with BI-RADS scores of 3 or 4, the researchers found that 198, or 84.3 percent, of them received the same Videssa result at both visits.
Contained within this overall result were larger and smaller variations for the different analytes that make up the assay, with some showing 100 percent concordance between the two measurements and some showing considerably lower concordance.
Wolf said she believed the discordance in Videssa results between the two time points was due both to technical and biological variability.
"I personally believe that, six months apart, there are going to be some biologic differences that are expected," she said. "And if, for instance, half of that sixteen percent is from that and the other half is from the preciseness of the test, then if we can improve preciseness of the test we know what is left must be biologic change."
"I think as we do the assay more and, in particular, as there are more women who have more than one [Videssa] test, we'll be able to zero in on what can change over time just from biologic and physiologic changes versus the accuracy of the test," she said.
Wolf added that Provista is working within the Cancer Moonshot program's Blood Profiling Atlas Project to assess sources of preanalytical variation in its assay.
"What [our] lab has done and what we're going to be sharing and working on within the Moonshot program is all the pre-analytics to look at the preciseness and accuracy of the test," she said. "So [issues] like: how long [a sample] has been frozen, the time of day, the day of the week [it was run], the operator performing the assay, [and] the machine that was used."
The Blood Profiling Atlas Project brings together collaborators from industry, government, and academia to share and make freely available raw data from various molecular testing assays along with relevant clinical information like patient diagnoses and outcomes and sample preparation and handling protocols. Participants include AstraZeneca, Eli Lilly and Company, Epic Sciences, Memorial Sloan Kettering Cancer Center, Foundation Medicine, Genentech, Guardant Health, Novartis, Personal Genome Diagnostics, Pfizer, Thermo Fisher Scientific, the University of Michigan, and the University of Southern California.
"The purpose of the initiative is to have companies that are working on [molecular diagnostics] pool their knowledge together … so we're not replicating the same thing over and over again without sharing that information," Wolf said, noting that Provista was one of the few proteomics-focused firms involved.
The company also continues to lay the groundwork for clinical utility and health economic studies on the Videssa test, she said, noting that it recently met with around 40 breast surgeons from around the country to get input on how best to build a patient registry for conducting such research.
"What we would hope to do is involve those physicians as well as some others that we've worked with in the past to do kind of a clinical registry where we assess the clinical utility based on how they actually use it in practice, and then we can also get health economic data by looking at, for instance, did that change their pattern of advanced imaging? Did it change the pattern of biopsy rates?" she said.
Wolf said the company hoped to finalize the protocol for this work in the first quarter of next year.