Diagnostic accuracy of the Videssa protein-based liquid biopsy for breast cancer in suspicious mammography.

Abstract

e13034 Background: The Videssa protein-based blood test (Provista Diagnostics) is a combined proteomic biomarker assay that aims to detect established breast cancer – rendering it useful in patients with abnormal or difficult-to-interpret mammograms. Since the incidence of malignancy for BI-RADS 4 biopsies is approximately 20%, a significant percentage of patients undergo needle biopsy unnecessarily. The Videssa assay provides an alternative diagnostic method by way of a non-invasive liquid biopsy. Methods: Our goal was to assess whether this liquid biopsy combined with 3D tomographraphic mammography could accurately predict disease status to reduce the amount of unnecessary tissue biopsies. An IRB-approved, prospective, single-arm study had patients with BI-RADS4 lesions with calcifications requiring tissue biopsy undergo the blood test prior to stereotactic core biopsy. Subsequent results and biopsy pathology were correlated. Results: 46 patients were initially entered with BI-RADS4 calcifications. 9 patients had DCIS (19.6%) and 37 patients had benign calcifications (80.4%). No patient had invasive cancer. Liquid biopsy results were elevated in 2 patients with DCIS (22% sensitivity) and in 5 patients with benign biopsies (10.8%). At interim analysis, the liquid biopsy demonstrated a 28.5% positive predictive value (PPV) and 82% negative predictive value (NPV). Tumor grade did not affect results. We subsequently discontinued using the blood test for BI-RADS4 calcifications and enrolled 13 patients with non-palpable solid lesions to determine the correlation. The liquid biopsy was elevated in 1 of 4 patients with invasive cancer (25% sensitivity) and was not elevated in all 9 patients with benign biopsies (100% NPV). At one-year follow-up, 3 of the 5 patients with an elevated blood assay and negative biopsy have had no incidence of cancer on subsequent imaging. The remaining 2 patients are scheduled for follow-up. Conclusions: In our study, the NPV of the Videssa liquid biopsy is not robust enough to defer tissue biopsy in any patient with indeterminate calcifications on imaging, nor is it specific enough to help predict results in high risk solid lesions. We do not see this liquid biopsy changing our current practice.