AI-Powered Deep Proteomics Enables Early, Non-Imaging Detection of Breast Cancer With MRI-Comparable Accuracy

At the San Antonio Breast Cancer Symposium (SABCS) in San Antonio, Texas, Justin Drake, Ph.D., discusses emerging clinical findings from Astrin Bioscience’s Certitude Breast test, a breakthrough, AI-powered proteomics blood test designed to detect breast cancer as early as stage 0. Unlike traditional imaging—particularly for women with dense breasts who often lack reliable screening options—Certitude Breast offers MRI-comparable accuracy through a simple at-home collection or in-office blood draw. By leveraging proteomics rather than genomics, the test identifies cancer-related protein signals earlier and more reliably, positioning it as a significant supplemental tool to guide women toward appropriate next steps in their breast cancer screening and care.

Pharmacy Times: Certitude Breast showed MRI-comparable accuracy using less than 1 mL of plasma. From a clinical implementation standpoint, what key steps need to happen for pharmacists and other frontline providers to confidently use this test in routine screening—particularly for women with dense breasts who currently lack reliable imaging options?

Justin Drake, Ph.D: They don't have reliable imaging options. So, you know, a mammogram for women with dense breasts is about 50% to 60% sensitive, so you leave a large percentage of women that really have negative mammograms or inconclusive mammograms. And right now, the FDA guidance is to talk to your physicians. They really don't have another option that gives them clarity on next steps.

So what we feel is our test, Certitude, allows us to fit into that gap as a supplemental test post-mammogram to help guide women toward an MRI if it's positive or toward another round of screening potentially a year from now if the test remains negative. Right now, our test has a negative predictive value of 99.9%. So if a test is negative, we're fairly confident that that test, you know, you are truly negative for breast cancer. So again, we're not trying to replace all the imaging modalities. We're trying to provide a supplemental option to guide women on the next steps in their potentially their cancer care journey, if needed.

Pharmacy Times: Your classifier achieved high specificity and sensitivity across stages 0–2, which is a major limitation of existing liquid biopsy technologies. Can you walk us through which proteomic signatures were most informative in distinguishing early-stage disease and how AI strengthened that signal detection compared with traditional genomic approaches?

Justin Drake, Ph.D: Yeah, so as you mentioned, a lot of the current NSAID tests that are out there, the multi-cancer early detection tests, they work very well for late-stage disease. And for some other cancer types, they work really well for even stage 0 to 2. But for breast cancer in particular, those tests are not very sensitive. So there's a few reasons for that. One is that cells don't really shed early in disease, so getting the DNA, which is what they're typically using to detect, is not as high abundance.

So they have a sensitivity problem because they cannot find the analytes. So we thought of a different approach. So we moved into proteomics because what's happening in the protein space is, you know, cells are constantly talking to each other, right? So whether it's a cancer cell to a normal cell, a cancer cell to another cancer cell, an immune cell to a cancer cell, etcetera, those cells are in constant communication to each other via proteins. So we're capturing those proteins in real time, and that's giving us the sensitivity that we need to provide early-stage detection from 0 to 2 in this particular asset.