NEW YORK (360Dx) – Provista Diagnostics has presented data from a multicenter prospective study of its proteomic Videssa Breast diagnostic test for breast cancer that indicates the test could significantly reduce unnecessary biopsies in women with borderline imaging results.
Published today in Clinical Breast Cancer, the study consisted of two prospective clinical trials looking at a total of 561 women under the age of 50 who received biopsies after inconclusive imaging results, and found that use of the test could have reduced unneeded biopsies in these groups by up to 67 percent.
The study represents the first peer-reviewed prospective data Provista has published on the test, said Judy Wolf, the company's chief medical officer.
The Videssa test is intended to help doctors make decisions when confronted with difficult-to-interpret imaging results, specifically in the case of patients presenting with Breast Imaging-Reporting and Data System (BI-RADS) scores of 3 or 4.
The BI-RADS is a 0 to 6 scale for grading mammogram findings where the likelihood of malignancy rises 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. Typically, doctors will schedule patients with a score of 3 for follow-up imaging, while they generally biopsy patients with a score of 4. Imaging false positives are very high for patients in this score range, however, with biopsies returning a negative result in between 75 percent and 95 percent of patients, Wolf said.
In the patient population used for the Clinical Breast Cancer study, 91 percent of the biopsies performed based on the patients' imaging results were negative, suggesting that a large percentage of them could be avoided if a test like the Videssa assay were used to rule out those patients whose imaging findings were most likely benign.
"As a physician, I would say as a kind of caveat, that not every biopsy that isn't cancer is unnecessary," Wolf said. "But many potentially are."
The Videssa test is based on a mix of serum protein biomarkers and tumor-associated autoantibodies, some of which were initially discovered by the Biodesign Institute of Arizona State University's Joshua LaBaer and Karen Anderson, who also are members of Provista's advisory board.
Provista signed a deal with ASU in 2013 to evaluate the applicability of the researchers' autoantibody technologies and markers for diagnostic purposes and then expanded that deal a year later.
In the study published this week, the researchers first measured 11 serum protein markers and 33 tumor-associated autoantibodies and combined this with demographic and clinical data from the 351 patients in the first trial cohort to develop its classifier for identifying likely benign patients. They then validated this classifier, which used data from 18 total markers — a mix of serum proteins and autoantibodies — in a second cohort of 210 women.
In the first cohort, the test distinguished between malignant and benign lesions with a sensitivity of 92.3 percent and a specificity of 85.3 percent and a negative predictive value of 99.3 percent. In the validation cohort, sensitivity was significantly lower, at 66.7 percent, while specificity was 81.5 percent and negative predictive value was 98.8 percent.
Looking across both cohorts, imaging led to 341 patients receiving biopsies, which detected 30 cancers, for a false-positive rate of 90.6 percent, while Videssa Breast would have recommended 111 patients receive a follow-up biopsy, reducing false positives by 67 percent. Of these 111 patients, 83 would have been false positives, yielding a 75 percent false positive rate for the test.
The study authors noted that a possible factor in the drop in sensitivity between the training and validation trials was the drop in breast cancer prevalence between the two. In the training set, prevalence was 7.7 percent, while it was 2.9 percent in the validation cohort. In both cases, this was well below the prevalence of around 20 percent that the authors said would be expected based on the literature.
This low prevalence, they noted, could be due to the study's short follow-up time of six months. "A six-month follow-up period may not be sufficient to identify all cancers in BI-RADS 3 and 4 patients," they wrote. "An additional 12-month follow-up may have yielded an increased cancer incidence."
With its first prospective data published, Provista is now planning a clinical utility and health economics study for the test.
That study will "follow doctors who are using this test in clinical practice [to see] what are they actually doing with these women." Wolf said. "Is this actually changing clinical practice as we suspect is might based on the result of our study? What is the health economic impact of that?"
She said the company has developed the study protocol and plans soon to submit it to an independent review board. In addition to consulting with physicians on the study design, the company has been working with insurers to make sure the questions they might have around utility and health economics are addressed, she said.
In addition to the version of Videssa for use in women under 50, Provista has a test intended for use in the 50-and-up population that uses a combination of 11 serum proteins and autoantibodies to help interpret imaging findings.
The company is also developing the test as a tool for cancer detection in women with dense breasts, which can be difficult to assess using mammography. In data presented last year at the San Antonio Breast Cancer Symposium, the test identified cancers in patients with dense breasts with sensitivity of 88.9 percent and specificity of 81.2 percent, and with negative predictive value, likewise, above 99 percent.
Provista began a limited commercial release of Videssa last year, offering the test mainly to its clinical trial collaborators, and has been slowly expanding its availability to a larger group of early-access users, Wolf said.