Cambridge, UK

Ending trial-and-error cycling through advanced therapies.

Escora CareTech builds molecular and clinical decision tools that predict biologic response and stratify patients into clinically actionable subgroups, starting with moderate-to-severe atopic dermatitis.

Inputs Blood RNA · serum biomarkers Tape-strip skin-derived transcriptomics Clinical EASI · DLQI treatment hx Lifestyle behaviour environment Foundation model platform Multi-modal omics + lifestyle data Treatment-annotated clinical outcomes Benchmarked vs. TARC, eosinophils, IgE Treatment-escalation score Low Intermediate High risk INADEQUATE BENEFIT Clinically actionable subgroups Trial enrichment Treatment sequencing Switch pathway
Founded at Cambridge

Escora is co-founded by Dr Miranda Robbins (MRC Laboratory of Molecular Biology, University of Cambridge) and Dr Gehad Youssef (Milner Therapeutics Institute, University of Cambridge), bringing together translational neuroscience, machine learning for high-dimensional biomedical data, biomarker robustness, and clinical-trial biomarker analysis.

The problem

Advanced therapy selection is still trial-and-error.

Patients cycle through biologic and targeted therapies for months before non-response is clear, and sponsors run heterogeneous trial populations that dilute signal.

In atopic dermatitis, this is especially visible. Anti-Type-2 pathway agents, JAK inhibitors, and OX40, OX40L, and IL-31 therapies now compete across overlapping patient groups, with no robust molecular way to match patients to the right pathway before treatment begins.

NICE guidance for advanced atopic dermatitis uses 16-week response assessment points, with adequate response commonly defined using EASI improvement and DLQI improvement. Patients who do not meet these criteria have spent four months on an advanced therapy without sufficient benefit. The clinical and economic cost of that delay is avoidable.

What Escora does

Foundation models for treatment-escalation decisions.

Escora's platform applies foundation models to multi-modal omics and lifestyle data, combined with treatment-annotated clinical outcomes, to predict biologic response and stratify patients into clinically actionable subgroups.

Each treatment-escalation score targets a specific clinical decision point and a validated clinical endpoint, with rigorous benchmarking against established clinical and serum biomarkers.

Enrich trials for likely responders

Molecularly defined cohorts for stronger Phase 2 and Phase 3 signal in advanced therapy trials, reducing trial failure driven by heterogeneous populations.

Guide treatment sequencing

Earlier, better-informed decisions at defined clinical checkpoints, before and during treatment, rather than waiting for visible non-response.

Interpretable responder signatures

High-dimensional multi-modal data translated into interpretable signatures, with external validation across independent cohorts as a core design requirement.

How it works

From decision point to actionable score.

A structured pipeline that turns a specific clinical question into a validated, interpretable treatment-escalation score.

01

Define the decision point

For example: inadequate benefit from anti-Type-2 therapy at 16 weeks. Every score is anchored to a specific, clinically meaningful threshold rather than a general response prediction.

02

Integrate multi-modal data

Foundation models trained on multi-modal omics, lifestyle data, and clinical features, linked to treatment-annotated outcomes across individual patient trajectories.

03

Benchmark rigorously

Confound-aware, leakage-controlled evaluation, benchmarked head-to-head against established clinical and serum biomarkers including TARC/CCL17, eosinophils, and IgE.

04

Return an actionable score

Low, intermediate, or high risk of inadequate benefit for a specific therapy pathway. Features link back to pathways and disease biology so the score is explainable, not a black box.

Lead programme

Moderate-to-severe atopic dermatitis.

Escora's lead programme predicts inadequate benefit from anti-Type-2 pathway therapy in moderate-to-severe atopic dermatitis, evaluating blood and skin-derived molecular signals against clinically meaningful 16-week outcomes.

Sample report format
Anti-Type-2 pathway score
Risk of inadequate benefit at 16 weeks
Low Intermediate High
High risk of inadequate benefit
Endotype
Type-2 low
Decision point
EASI 50 + DLQI
Data sources
Blood + tape-strip
Benchmark
TARC · Eos · IgE

The clinical context

Moderate-to-severe atopic dermatitis has more advanced therapy options than ever before. Anti-Type-2 pathway biologics, JAK inhibitors, and emerging OX40, OX40L, and IL-31 agents sit alongside one another with no robust molecular way to select between them for an individual patient.

A substantial fraction of treated patients do not meet 16-week response criteria such as EASI 50 plus DLQI improvement after four months on an advanced therapy.

The platform

The score integrates blood, skin-derived tape-strip transcriptomics, clinical features, and lifestyle data into a foundation-model evaluation linked to clinically meaningful 16-week outcomes, with pathway-level framing (Type-2-high, Type-2-low, mixed endotypes) for utility across the anti-Type-2 class.

Why pathway-level matters

Pathway-level scores work across drugs of the same mechanism rather than against a single product. That makes them directly useful both for incumbents and for sponsors developing alternative-mechanism therapies who need to show differentiated effect in molecularly defined subgroups.

For pharma

Trial enrichment for alternative-mechanism therapies.

Escora supports sponsors developing advanced therapies by identifying molecularly defined patient subgroups most likely to show benefit, strengthening trial design, biomarker strategy, and evidence generation for differentiated positioning.

We engage before, during, or after a trial: before to define enrichment hypotheses, during to guide biomarker collection, and after to interpret responder and non-responder subgroups.

The advanced atopic dermatitis space is fragmented enough that the right molecular subgroup framing carries both clinical and commercial weight. A therapy that underperforms in an unselected population can show compelling signal in a molecularly enriched one.

What we look for in collaborations

Treatment-linked clinical outcome data (responders and non-responders)
Baseline molecular or biomarker samples at or before treatment initiation
Access to responder and non-responder cohorts across therapy classes
Interest in enrichment, treatment sequencing, or retrospective signal analysis

Use cases

Phase 2 enrichment for likely responders in advanced atopic dermatitis
Retrospective responder analysis to recover signal from completed trials
Biomarker strategy for alternative-mechanism therapies (OX40, OX40L, IL-31)
Patient segmentation for trial design and cohort definition
Evidence generation for differentiated market positioning
Scientific approach

Designed for signatures that survive real-world deployment.

Score development is built around robustness under the conditions that matter: cohort shift, platform differences, ancestry variation, treatment-line heterogeneity, and the unforgiving test of an independent dataset.

VALIDATION FRAMEWORK Discovery Held-out External A External B vs. TARC vs. Eos+IgE Anti-Type-2 score evaluated evaluated evaluated evaluated compared compared Serum TARC/CCL17 baseline baseline baseline baseline comparator n/a Eosinophils + IgE baseline baseline baseline baseline n/a comparator External cohorts span European and East Asian ancestry. All evaluation points pre-specified. Escora score Established biomarker comparator Benchmark comparison point

Why the scores are trustworthy

Tested across independent cohorts

Held-out independent academic and clinical-trial datasets built into the evaluation plan. Cross-cohort evaluation across European and East Asian ancestry is a default, not an afterthought.

Cohort, batch, ancestry, and treatment-line aware

Explicit handling of technical batch, cohort, platform, ancestry, treatment-line, and concomitant-therapy confounding. Leakage controls applied throughout pipeline design.

Benchmarked against established biomarkers

Every signature is benchmarked head-to-head against TARC/CCL17, eosinophils, and IgE. We report where the score adds value and where it does not.

Features link to pathways and biology

Attribution methods, pathway enrichment, and attribution-stability analysis ensure model-prioritised features correspond to known disease biology, not statistical artefact.

Multi-modal transcriptomics

Bulk and targeted transcriptomic discovery in blood and lesional and non-lesional tape-strip samples, with platform-agnostic feature engineering across assay types.

Endpoints anchored to clinical standards

Outcomes aligned to NICE response criteria (EASI 50 plus DLQI improvement) and regulatory standards (EASI 75, IGA 0/1), with time-to-event modelling where appropriate.

Pipeline

A treatment-sequencing pipeline for advanced disease.

The lead programme addresses the largest single decision in advanced atopic dermatitis therapy today. The pipeline extends across the major treatment-decision points in advanced disease care.

Programme
Decision
Stage
Anti-Type-2 score
Inadequate benefit from anti-Type-2 pathway therapy at the 16-week clinical decision point.
Lead programme
JAK score
Response stratification for the oral JAK inhibitor class (upadacitinib, abrocitinib, baricitinib).
In design
OX40 / OX40L score
Response stratification for OX40 and OX40L agents as they move through late-phase trials.
Roadmap
IL-31 score
Itch-pathway response stratification for IL-31 receptor blockade.
Roadmap
Commercial model

Pharma partnerships first. Clinical decision support next.

Near term

Pharma partnerships. Trial-enrichment scoring and biomarker stratification for Phase 2 and Phase 3 programmes in advanced atopic dermatitis, supporting sponsors developing alternative-mechanism therapies to demonstrate effect in molecularly defined subgroups.

Mid term

Central-lab assay. A targeted RNA panel ordered before advanced therapy initiation, with results returned as a low, intermediate, or high risk score for the relevant pathway.

Long term

Reimbursed clinical decision support. Prospective evidence and a health-economic case demonstrating that the test improves treatment selection and reduces cost per patient achieving disease control.

Why now

Multi-modal atopic dermatitis cohorts with linked treatment response have reached the scale where signature translation is realistic. Methods for evaluating signature stability across cohorts and platforms have matured. The advanced therapy landscape is fragmented enough that sequencing decisions carry both clinical and commercial weight.

Team

Founders

MR

Dr Miranda Robbins

Co-founder

MRC Investigator Scientist at the MRC Laboratory of Molecular Biology, University of Cambridge, supported by an AstraZeneca BlueSky Fellowship. Research spanning synaptic biology, tau pathology, neural circuit function, and imaging technology for neurodegeneration. PhD in Molecular Neuroscience at Cambridge. MSc Integrative Neuroscience, University of Edinburgh. BSc Molecular Biology, UCL.

GY

Dr Gehad Youssef

Co-founder

Senior postdoctoral researcher at the Milner Therapeutics Institute, University of Cambridge. Translational machine learning research on high-dimensional biomedical data, signature robustness under deployment shift, interpretable biomarker discovery, leakage and confound-aware evaluation, and clinical-trial biomarker analysis. PhD UCL.

Contact

Get in touch.

Pharma partnerships, cohort collaborations, or investor enquiries

Send us a message and we will get back to you. You can also reach us directly at escoracaretech@gmail.com.