Available
Immunotherapy
Immune checkpoint inhibitors
PD-1, PD-L1 and CTLA-4 inhibitors, alone or in combination with chemotherapy.
- Objective response
- Grade ≥3 irAE risk
- Organ-specific irAEs
POST /v1/models/ici/predict
New ADC and bispecific models now in early access
See who immunotherapy is likely to help, before the first infusion.
Cure-IO turns routine clinical information into an explainable estimate of treatment response and immune-related side-effect risk. Each prediction shows which factors raised or lowered it, so clinicians and patients can talk it through together.
Prediction · ICI-v2.3Live
Objective response
0%
95% CI —
Grade ≥3 irAE risk
0%
95% CI —
Top drivers · responseΔ pts
Example patient, cycled every few seconds. Supports a conversation; it does not make the decision.
Model suite
One platform for every major class of cancer immunotherapy.
Each model is trained for its own therapy class and its own toxicity profile. License one, or connect all four through a single API key.
Early access
ADC
Antibody–drug conjugates
Targeted payload therapies such as HER2 and TROP2-directed conjugates.
- Response probability
- ILD / pneumonitis risk
- Ocular toxicity
POST /v1/models/adc/predict
Early access
Bispecifics
Bispecific antibodies
T-cell engagers and dual-target antibodies across solid and blood cancers.
- Response probability
- CRS grade ≥2
- ICANS risk
POST /v1/models/bispecific/predict
In development
CAR-T
CAR-T cell therapy
Autologous CAR-T in lymphoma, myeloma and leukaemia.
- Complete response
- CRS and ICANS
- Prolonged cytopenia
POST /v1/models/cart/predict
How it works
From routine data to a clear, explainable estimate.
STEP 1
Pull routine data
Age, tumour type and stage, ECOG, regimen, line of therapy and standard bloods such as NLR, albumin and LDH. No new tests.
STEP 2
Run the model
The therapy-class model returns a calibrated probability with a confidence interval, in under a second.
STEP 3
Explain every number
Per-patient attributions show which factors raised or lowered the estimate and by how much.
STEP 4
Talk it through
A plain-language summary supports the consultation and the MDT. The clinician stays in charge of the decision.
Interactive demo
Change a patient’s details and watch the drivers move.
This demo uses illustrative weights so you can see how Cure-IO presents a prediction. It is not the validated model and must not be used for real patients.
Estimate
model: ici-demo · illustrative–
Likely response
Probability of objective response (RECIST) on this regimen.
––
Side-effect risk
Probability of a grade 3 or higher immune-related adverse event.
–Raises the estimateLowers the estimateΔ = percentage points vs. cohort baseline
Conversation summary
–
Illustrative demo only. Cure-IO supports shared decision-making between patients and clinicians. It does not diagnose, and it does not replace clinical judgement.
Built for two worlds
The same science, shaped for the people who use it.
For pharma and biotech
Enrich trials, read real-world response and de-risk your IO pipeline.
- Responder enrichment and stratification for Phase II/III
- Toxicity forecasting for safety monitoring plans
- Per-study licences, private VPC or on-prem deployment
Bring explainable predictions into clinic, the MDT and the consent conversation.
- Embedded in the EHR through SMART on FHIR
- Patient-friendly summaries for shared decisions
- Full audit trail, SSO and local data residency
Safety and governance
Designed to be questioned.
Cure-IO is decision support. Every output carries its confidence interval, its drivers and its model version, so the people relying on it can check how it got there. Prospective validation and regulatory work are being run with clinical partners.
Explainable by defaultPer-patient attributions on every prediction
Calibrated outputsProbabilities with 95% confidence intervals
Bias monitoringPerformance reported by age, sex and ethnicity
Privacy firstDe-identified inputs, UK GDPR and HIPAA-ready controls
Versioned modelsEvery result tied to an immutable model card
Human in the loopNever a replacement for clinical judgement
Get started
Bring explainable immunotherapy predictions to your team.
Book a 30-minute walkthrough for your clinical, data or pharma team, or start free in the Sandbox.