How Accurate Are Blood Tests for Early Breast Cancer Detection?

For healthcare providers, the more useful question may be: accurate in which patient populations, at which stage, and in which clinical context?

Blood-based cancer detection is moving rapidly from research toward clinical practice. But the technologies now available or under investigation are not interchangeable. Some are designed to identify signals from dozens of cancers. Others focus specifically on one cancer, such as breast cancer. And the biomarkers they analyze, from circulating tumor DNA to RNA-based gene expression to proteins, reflect fundamentally different approaches to detecting disease.

For physicians considering these technologies, understanding those distinctions is essential.

Blood-based detection is not one technology

A useful way to evaluate emerging blood tests is to ask four questions:

1. What biological signal does the test measure?
2. How well does it detect early-stage disease?
3. How often does it correctly identify patients without cancer?
4. And what clinical question was the test designed to answer?

Those questions reveal meaningful differences among four tests receiving attention in this space. The data and tests below will be discussed in the context of breast cancer.

Galleri®: detects DNA-based methylation patterns shed by cancer cells across multiple cancers

GRAIL's Galleri is a multi-cancer early detection (MCED) test that evaluates cell-free DNA methylation patterns in blood and predicts a cancer signal of origin.

Galleri demonstrates high specificity across cancers, but its sensitivity varies considerably by cancer type and stage. This distinction is particularly relevant when considering its performance in breast cancer.

In clinical validation data, overall sensitivity for breast cancer was 30.5%.¹ Performance increased markedly as disease advanced, which is typical for a test measuring DNA:

• Stage 0: Not reported
• Stage I: 2.6%
• Stage II: 47.5%
• Stage III: 85.5%
• Stage IV: 90.9%

The stage gradient is clinically important. Galleri detected approximately 1 in 38 Stage I breast cancers in these validation data, compared with more than 9 in 10 Stage IV cancers. The limitation is consistent with a broader challenge for circulating tumor DNA approaches: early, localized tumors may release relatively small quantities of tumor-derived DNA into circulation, which is especially true in breast cancer.

More recent literature reviewing Galleri's performance continues to cite the 30.5% overall breast-cancer sensitivity and 2.6% Stage I sensitivity, showing the distinction between detecting advanced disease and detecting breast cancer at its earliest stages.²

For breast-care clinicians, that context is essential. Galleri was developed as an MCED test capable of identifying signals associated with many cancer types. It was not designed as a breast-specific supplemental screening test, and it should not replace recommended breast cancer screening.

Cancerguard®: detects DNA-based methylation patterns shed by cancer cells combined with a few proteins across multiple cancers

Exact Sciences' Cancerguard is another MCED approach, but its mechanism differs slightly. The test evaluates DNA methylation patterns shed by cancer cells but also includes a small number of tumor-associated proteins.

Its development data illustrates why cancer type and stage matter when discussing test performance.³ In current development data for the optimized methylation-protein classifier, excluding breast and prostate cancers, reported sensitivity increased substantially with advancing stage:

• Stage I: 24.8%
• Stage II: 57.8%
• Stage III: 81.5%
• Stage IV: 90.3%

Overall specificity was 97.4% in these test-development data.³

Breast cancer presents an important distinction. Breast-specific performance data provided for comparison indicate substantially lower sensitivity in earlier-stage disease:

• Stage 0: Not reported
• Stage I: 8%
• Stage II: 40%
• Stage III: 61%
• Stage IV: 100%

These stage-specific results should be interpreted cautiously because they are not the same population used to calculate the breast-and-prostate-excluded performance figures above, and the source and study design should be considered when making cross-test comparisons.⁴

Importantly, Cancerguard is not indicated for breast cancer screening. Exact Sciences' current intended-use information states that the test is not indicated for screening of breast cancer, prostate cancer, or precancerous lesions. Although development studies have demonstrated detection of breast cancer signals, Exact Sciences directs patients and clinicians to continue currently recommended cancer screening and states that Cancerguard is not a replacement for established screening or diagnostic modalities.

This distinction is particularly relevant for clinicians evaluating blood-based tests for early breast cancer detection. A test's ability to detect some breast cancers in development studies does not necessarily mean it has been clinically indicated for breast cancer screening. Intended use, cancer-specific performance, stage-specific sensitivity, and the population in which a test was validated all matter when determining its potential role in care.

Syantra: detects a breast-cancer-associated RNA-based gene-expression signature

Syantra takes a breast-specific approach to blood-based cancer detection. Its Onco-ID™ Breast platform analyzes RNA-based gene-expression biomarkers from whole blood and uses machine-learning-based software to identify a molecular signature associated with breast cancer. The technology measures RNA, not DNA-based signals in the bloodstream, which correlates closer to the body's response to cancer, rather than relying solely on tumor-derived DNA circulating in the bloodstream.

Clinical performance has been evaluated through the International Identify Breast Cancer (IDBC) study. In a blinded independent test set of 695 evaluable women, including 96 participants with breast cancer and 599 without cancer, the Syantra DX Breast Cancer test demonstrated:⁵

• Sensitivity: 79.2%
• Specificity: 94.3%
• Overall accuracy: 92.2%

Importantly for understanding its potential role in early detection, 59% of the breast cancers in the blinded test set were Stage I and 25% were Stage II. The published interim analysis did not report sensitivity separately for each cancer stage.

Performance also varied among clinically relevant subgroups. Among women younger than 50, the test demonstrated 91.7% sensitivity and 99.0% specificity, although this subgroup included only 12 cancer cases. Among women with extremely dense breasts (BI-RADS density category D, n=52), reported sensitivity was 88.9% and specificity 95.3%. Confidence intervals were wide, however, reflecting the smaller sample size.

The investigators also evaluated 19 tumors smaller than 10 mm, reporting sensitivity of 68.4% for these small cancers.

These results are promising, but their context matters. They represent an interim analysis rather than definitive evidence from a population-level breast cancer screening trial. Syantra continues to evaluate Onco-ID Breast in clinical studies. Its ongoing international IDBC study analyzes whole-blood samples collected around the time of mammography and before biopsy or surgery.

For clinicians comparing blood-based breast cancer technologies, Syantra illustrates another fundamentally different detection strategy: whole-blood gene expression combined with machine learning, rather than the circulating tumor DNA approaches used by many multi-cancer early detection tests.

Certitude: deep proteomics focused on breast cancer

Certitude, developed by Astrin Biosciences, approaches detection through a different biological layer: the proteins.

Following a blood draw, Astrin's platform profiles more than 9,000 proteins in a single run and uses AI-based analysis to identify changes associated with breast cancer. Unlike broad MCED tests, Certitude is specifically designed around breast cancer detection, including the clinical challenge presented by dense breasts.

Astrin currently reports approximately 93% sensitivity and 92% specificity for detecting breast cancer, including patients with high breast density.⁶ Stage-specific clinical data demonstrates sensitivity across the breast cancer continuum:

• Stage 0: 84.6% (11/13)
• Stage I: 88.4% (61/69)
• Stage II: 96.9% (95/98)
• Stage III: 87.5% (14/16)
• Stage IV: 100% (2/2)

These results are particularly notable in the context of early-stage disease. In the study population represented here, Certitude identified 88.4% of Stage I breast cancers and 96.9% of Stage II breast cancers. Sensitivity was 84.6% among the 13 Stage 0 cases evaluated. Certitude also reports a >99.9% negative predictive value in women with dense breasts.⁶

The data should be interpreted in the context of the number of cancers represented at each stage. In particular, the Stage 0 estimate is based on 13 cases and the Stage IV estimate on only two cases, resulting in wider confidence intervals. Stage-specific estimates should therefore be considered alongside the study population, confidence intervals, specificity, predictive values, and intended use when assessing clinical performance.

This breast-specific approach creates an important distinction when evaluating emerging blood tests. Multi-cancer tests are designed to detect signals across many cancer types, whereas Certitude is designed to provide another layer of insight within the breast-care pathway. Its stage-specific data provide clinicians with information about performance where early detection matters most, including Stage 0 and Stage I disease.

Designed to support, not replace, standard breast screening, Certitude offers a blood-based option to help inform care decisions when additional clarity may be valuable.

Why breast density makes the question especially relevant

Blood-based detection should be considered in the context of what imaging already does well and where limitations remain.

Mammography remains central to breast cancer screening. The American Cancer Society continues to recommend mammography for breast cancer screening, and blood-based tests should not be interpreted as substitutes for established screening.⁷

Dense breast tissue, however, can make mammographic interpretation more challenging. The National Cancer Institute notes that false-negative mammograms are more common among women with dense breasts.⁸

That creates a clinically meaningful opportunity for complementary technologies.

A blood test does not encounter challenges with breast density in the same way an image does. Rather than attempting to visualize a lesion through overlapping fibroglandular tissue, molecular assays interrogate biological signals circulating in blood.

The result is not a replacement for imaging, but potentially another source of information when imaging alone does not fully resolve the clinical question.

Comparing performance requires more than one number

For HCPs evaluating these tests, headline sensitivity should never be considered in isolation.

A clinically meaningful comparison should consider cancer-specific sensitivity, stage-specific sensitivity, specificity, predictive values (both positive and negative), intended-use population, prevalence, study design, and the downstream pathway following a positive or negative result.

This is particularly important when comparing MCED technologies with a breast-specific test. Performance across 20, 50, or more cancers answers a different question from performance specifically in patients being evaluated for breast cancer.

Predictive values also depend on disease prevalence. A high NPV can be particularly valuable when the intended clinical objective is helping clinicians determine how confidently to move forward after a negative result, but it must always be interpreted in the population in which it was established.

Where could blood tests fit tomorrow?

The most responsible near-term role for blood-based testing is complementary.

Current breast screening recommendations continue to center on mammography, with additional imaging and diagnostic evaluation determined by risk and clinical findings. A suspicious finding may ultimately require biopsy. The American Cancer Society notes that biopsy remains the definitive method for determining whether a breast abnormality is cancer.⁹

Blood-based technologies may nevertheless help address a different need: providing additional molecular information within increasingly personalized screening and diagnostic pathways.

For the busy primary care or OB/GYN managing dense-breast follow-up, that could mean another evidence-based tool to help navigate uncertainty. For the concierge or longevity physician building highly individualized prevention strategies, it could mean incorporating molecular information alongside imaging, family history, breast density, genetics, and other risk factors.

The future is therefore unlikely to be imaging versus blood.

It is more likely to be thoughtful integration of multiple sources of information each selected for the clinical question it is best equipped to answer.

For breast cancer specifically, that distinction may become increasingly important. As blood-based technologies mature, clinicians will need to look beyond the category of "liquid biopsy" and evaluate the biology, validation population, stage-specific performance, and intended use behind each test.

Certitude is designed to support that next layer of insight, not replace, standard screening, but help clinicians and patients make more informed decisions when additional clarity matters.

Learn More

Interested in how deep proteomic science may complement your breast health program? Connect with our team to review the clinical evidence, explore the science behind Certitude, and learn how it may fit into your practice.

Less uncertainty. More confidence.