A simple blood draw might one day catch cancer before you ever feel a symptom. That’s no longer just a hope — it’s increasingly looking like a near-term reality, and new data set to be presented at the American Association for Cancer Research’s 2026 Annual Meeting is making the case more forcefully than ever.
Exact Sciences is bringing a wave of findings to AACR 2026, centered on Cancerguard, its multi-cancer early detection test. The headline: a refined combination of methylation and protein biomarkers is catching more early-stage cancers with fewer false alarms — a balance that has eluded researchers in this space for years. For the estimated 70 percent of cancers that arise in types with no currently recommended screening protocol, the implications are hard to overstate.
Two Signals, One Test
Here’s what makes the data interesting. Of all the true positive cancer signals Cancerguard identified in the new analysis, 47.1% came from methylation biomarkers alone, while 7.4% were caught by protein biomarkers alone, and 45.5% required both signals together to surface. Take away either layer, and nearly half the detections — or more — vanish. “We designed Cancerguard as the first-of-its-kind multi-biomarker test because no one signal tells the whole story,” said Tom Beer, M.D., chief medical officer for multi-cancer early detection at Abbott’s cancer diagnostics business. “By combining biomarkers, we can detect cancer earlier, when it matters most.”
Perhaps equally telling: of the 2.6% false positives in the dataset, not a single one was simultaneously positive for both biomarker types. That’s not a coincidence. It’s the architecture working as intended — each layer serving as a check on the other, quietly doing the math that a single-signal test simply can’t.
The DETECT-A Legacy
Cancerguard didn’t emerge from a vacuum. Its intellectual ancestor is the CancerSEEK test, which powered the DETECT-A study — the first large-scale, prospective, interventional trial of a blood-based multi-cancer early detection approach. That study enrolled more than 10,000 women and was designed to detect nine distinct cancer types. The data it generated helped lay the scientific foundation for everything that followed.
Now, multiyear outcomes from DETECT-A are drawing formal recognition. The AACR Cancer Prevention Research Award is going to the team behind a 2024 publication covering the study’s long-term results — and the findings are striking. Every patient who was treated for a stage I or II cancer detected through the study remained alive and cancer-free after a median four-year follow-up. Every single one. “Long-term follow-up provides critical insight into how multi-cancer early detection can shape the future of cancer screening,” Beer noted. “With nearly 70 percent of cancers occurring in types without recommended screening, these findings highlight the potential for MCED to increase early detection and improve outcomes.”
What Cancerguard Actually Does
The current version of the test is a lab-developed product that screens for more than 50 cancer types — including the six deadliest, among them pancreatic and ovarian cancers, which are notoriously difficult to catch early and carry dismal survival rates when found late. Cancerguard hits 68% sensitivity for those deadliest cancers and detects more than one in three early-stage cancers overall, according to data released ahead of the conference. It covers cancer types responsible for at least 80% of U.S. diagnoses annually.
Still, sensitivity in the 60s isn’t perfection. Catching one in three early-stage cancers means two in three aren’t caught. That’s the honest math, and it’s worth sitting with. But compare that to the alternative — no screening at all for pancreatic cancer, no standard test for ovarian cancer — and the calculus shifts considerably. The optimized methylation-protein classifier being unveiled at AACR is designed to push that sensitivity higher while holding the line on specificity, which is the technical way of saying: find more cancers, not more false alarms.
A Broader Moment for Early Detection
AACR 2026, running April 17 through 22, is shaping up as something of a landmark gathering for the early detection field. Beyond the Cancerguard data, the conference is featuring work from AACR grantees on molecular alterations in the pancreas that could enable earlier interception of one of oncology’s most feared diagnoses. The association also announced its 2026 Scientific Achievement Award recipients, with the broader theme of the meeting leaning heavily into prevention and detection rather than treatment alone — a quiet but meaningful shift in emphasis.
A new AACR partnership aimed specifically at accelerating research on early detection in high-risk populations was also announced earlier this year, signaling that the institutional momentum behind this science isn’t slowing down.
The question that lingers over all of it — the one that will determine whether any of this changes medicine at scale — is whether payers, regulators, and primary care physicians can move fast enough to match the science. The test exists. The data is accumulating. What happens next is largely a story about systems, not biology. And systems, as any reporter covering health care knows, have a way of taking their time.
