MetaStat, Inc. presented positive results from an analytical validation study demonstrating the analytical accuracy, reproducibility, and precision of its MetaSite Breast test at the Tumor Metastasis meeting of the American Association for Cancer Research (AACR) in Austin, Texas. The MetaSite Breast™ test predicts the likelihood of distant cancer metastasis in patients with early-stage ER-positive invasive breast cancer and allows physicians to both optimize treatment decisions and ultimately improve patient outcomes. The test identifies and measures the number of micro-anatomical structures (MetaSites) consisting of a Mena protein expressing tumor cell, an endothelial cell, and a perivascular macrophage, all in direct contact.

MetaSites are hypothesized to be the portal of entry for cancer cells into the blood stream contributing to the development of metastasis. In this study, MetaStat assessed the analytical precision and accuracy of the fully-automated clinical MetaSite Breast test using formalin-fixed, paraffin-embedded (FFPE) tissue samples from patients with invasive breast cancer. MetaSite Breast™ was shown to have analytical precision of greater than 97% with a mean percent coefficient of variation of 6.6% (n=35).

Pathologist quality review showed a high degree of precision between two pathologists by ROC analysis with AUCs (area-under-the-curve) ranging from 0.97 to 1.0 over three independent sample set reviews (n=105). Additionally, overall assay performance using different instruments, operators and tumor sections performed over three consecutive days was highly reproducible with AUCs for MetaSite score classification ranging from 0.91 to 0.96. Importantly, MetaSite Breast™ showed a high degree of analytical accuracy with the reference standard (semi-manual pathologist MetaSite/TMEM counting) by ROC analysis with AUCs of 0.84 and 0.90 for low and high risk cut-points, respectively.

The reference standard method was originally developed at the Albert Einstein College of Medicine where they have previously demonstrated the number of MetaSites/TMEMs in tumors to be predictive of distant metastatic disease in ER-positive breast cancer patients (Rohan et al. JNCI 2014).