Mastering Disease Heterogeneity in Neuroscience Trials
A longitudinal patient platform that identifies individuals at risk of neurodegenerative disease, continuously assesses their disease state, and models individual progression trajectories from early detection through clinical endpoint analysis.
Global Pharma Sponsors utilizing our trajectory models
Screened & Qualified Trial Candidates
Sample Size Reduction via Regulatory Covariate Adjustment
Foundational R&D from Paris Brain Institute (ICM)
Delivery of Digital Twin Covariates Post Data-Lock
Trial Failure Is Often a Symptom of Disease Heterogeneity, Not Drug Inefficacy.
Promising compounds routinely miss their primary endpoints not because the mechanism of action is flawed, but because trials cannot account for the clinicopathological mismatch. Two patients can enter a study with identical amyloid, tau, or genetic profiles: one deteriorates rapidly, while the other remains functionally stable for 18 months due to cognitive reserve and individual compensation.
Standard trial protocols capture only static, six-month snapshots in an artificial clinic environment. This unmeasured baseline variance drowns the drug’s true treatment effect in statistical noise, forcing sponsors to artificially oversize cohorts or risk trial failure on inconclusive endpoints.
The Static Snapshot Blindness
Standard on-site screening captures a single point in time, offering insufficient resolution to reliably separate rapid clinical progressors from compensating non-progressors.
The Variance Penalty
High unexplained outcome variance erodes statistical power, forcing sponsors to add 20–30% more patients just to overpower the baseline noise.
The Biomarker Screening Bottleneck
Restricting trial entry to amyloid- and tau-positive populations drives screen failure rates above 90%, without guaranteeing a homogeneous clinical progression trajectory.
One Longitudinal Platform from Early Real-World Detection to Primary Endpoint Analysis
We combine longitudinal ecological phenotyping in the community with patient digital twins to capture and neutralize disease progression variance across the full trial lifecycle.
DocMemo Engine — Ecological Phenotyping & Real-World Staging
Focus: Community capture before and alongside site visits.
Identifies individuals with subjective memory complaints in the community before formal clinical screening and longitudinally assesses their cognitive state—capturing subtle cognitive decline and real-world progression dynamics.
Triage biologically aligned candidates, eliminate site screening friction, and capture baseline cognitive risk profiles as explanatory covariates.
CourseMap AI Engine — Spatiotemporal Digital Twins
Focus: Prognostic Covariate Adjustment & Variance Neutralization.
A longitudinal generative modeling architecture that maps non-linear disease progression trajectories across clinical scores, MRI volumetrics, PET imaging, and fluid biomarkers.
Generates patient digital twins at baseline to adjust for individual progression variance in the primary analysis, reducing required sample size by 15–25% without regulatory friction.
De-Risking Primary Endpoints Within Standard Regulatory Frameworks
Qairnel’s prognostic digital twin technology is fully aligned with FDA and EMA guidances on prognostic covariate adjustment in randomized clinical trials.
- FDA Final Guidance: Adjusting for Covariates in Randomized Clinical Trials (May 2023)
- EMA Guideline on Adjustment for Baseline Covariates (EMA/CHMP/295050/2013)
Zero Type I Error Inflation
Digital twins are generated strictly using pre-randomization baseline data, maintaining blinding and fully preserving valid statistical inference under the primary Statistical Analysis Plan (SAP).
Pragmatic Regulatory Adoption & Hybrid Control Horizons
Deployed today as a non-disruptive prognostic covariate within standard randomized arms—delivering immediate power gains with full FDA and EMA endorsement—while progressively building the empirical and regulatory evidence required for future hybrid and external control applications in rare neurological indications.
Data Privacy & Integrity
Fully compliant with HIPAA (US), GDPR (EU), and 21 CFR Part 11 standards.
Validated in Phase II Parkinson’s Disease: The MOVES-PD Study
Joint publication with Sanofi demonstrating significant variance reduction and statistical power gain.
High endpoint variance in early Parkinson’s progression obscuring efficacy signal in this Phase 2 study.
CourseMap AI generated individual prognostic progression twins using baseline clinical and biomarker variables, integrated directly into the standard MMRM statistical model.
Narrowed confidence intervals of the treatment effect, achieving formal statistical significance on the primary endpoint.
Demonstrated identical statistical power for subsequent Phase 3 design with significantly fewer required subjects.
Ready to Master Disease Heterogeneity in Your Next Protocol?
Schedule a 30-minute scientific consultation with our biostatistics and neurology modeling leads.