In a study involving nearly 2,400 participants, a 34-circular RNA signature outperformed plasma pTau217 in predicting progression to symptomatic Alzheimer’s disease.
Circular Genomics reported that a 34-circular RNA (circRNA) blood signature predicted symptom onset with a hazard ratio of 2.92, 61% stronger than pTau217 alone. The company presented the data in a featured biotech showcase and five posters at the 2026 Clinical Trials on Alzheimer’s Disease conference.
The results come from an NIH-funded study across three independent cohorts totaling nearly 2,400 individuals. When the circRNA signature was combined with pTau217, the hazard ratio for progression rose to 4.83, and the area under the curve (AUC) for Alzheimer’s diagnosis reached 0.967.
Circular Genomics says Alzheimer’s blood diagnostics have long relied on single proteins such as amyloid and tau. Circular RNAs are stable, closed-loop RNA molecules that are enriched in the brain and cross the blood-brain barrier into the bloodstream. The company says they reflect active gene regulation across several brain pathways at once.
“While amyloid and tau tell us the underlying structural pathology is present, they don’t capture the real-time biological response of the brain,” says Nikolaos Mellios, MD, PhD, cofounder and chief scientific officer of Circular Genomics, in a release. “By measuring dynamic synaptic, inflammatory, and metabolic changes before symptoms appear, our 34-circRNA signature captures the active and dynamic disease state.”
Mellios presented the company’s CircPath platform, which is currently available for research use only. The company says the platform offers whole-transcriptome profiling and targeted prognostic applications. Drug developers could use it to stratify patient cohorts, enrich prevention trials with people who are close to progressing, and track pharmacodynamic responses from a routine blood draw.
Timing and Replication Data
Poster P274 examined when circRNA levels shift relative to clinical onset in a longitudinal cohort of 688 cognitively unimpaired individuals from the Knight Alzheimer Disease Research Center. According to the company, the signature detected molecular divergence 4 to 8 years before symptoms emerged, with the most significant shifts 2 to 4 years before cognitive decline.
For 5-year progression prediction, the circRNA model achieved an AUC of 0.870, compared with 0.676 for plasma pTau217 and 0.643 for amyloid PET.
Poster P314 tested the fixed 34-circRNA model in two replication cohorts: a second Knight cohort (n=551) and the A4/LEARN study (n=318). The model produced hazard ratios of 3.83 and 2.93, respectively, for progression from cognitively unimpaired to symptomatic Alzheimer’s disease. Combined with pTau217, the hazard ratios rose to 5.58 in the Knight cohort and 4.58 in A4/LEARN.
Specificity, Diversity, and Laboratory Measurement
Poster P313 evaluated performance across populations. The company reports that the signature performed consistently across African American, Hispanic, and Asian populations, as well as across sexes and APOE4 status, with an AUC above 0.90.
The posters also showed that the signature maintains high specificity for Alzheimer’s disease compared with non-Alzheimer’s dementias, including Parkinson’s disease, frontotemporal dementia, and dementia with Lewy bodies.
For laboratories, the company reports that orthogonal PCR evaluations showed the signal can be measured reliably and consistently using quantitative and digital droplet PCR. Circular Genomics says this supports scalability for future clinical laboratory deployment.
CircPath is not yet offered as a clinical diagnostic. Circular Genomics says the data lay the groundwork for future clinical diagnostic deployment. The company also plans to expand into Parkinson’s disease and ALS.
ID 187465996 © Pop Nukoonrat | Dreamstime.com