Open-source infrastructure for oncology clinical decision-making
Get clear, evidence-based treatment strategies in minutes. Upload a patient profile to receive standard and aggressive options, built on global clinical guidelines and references. Transparent and based on internationally recognized standards.
Why it matters and how it works
To prescribe a treatment, an oncologist or clinical pharmacologist spends 2–4 hours of manual work: opening the NCCN PDF, cross-checking the ESMO guideline, re-reading the MoH protocol, verifying the national formulary for drug availability, looking up renal/hepatic dose adjustments, adding supportive care, remembering vaccinations and opportunistic-infection prophylaxis. Every time, for every patient, from scratch. Any missed contraindication can cost a life.
OpenOnco automates the grunt work: the logic is the same as a classic multidisciplinary team (MDT), augmented by an analytical layer. Several specialists around the patient, case discussion, an agreed plan, return to the case when new data arrives. We simply formalize this as a structured engine — each «virtual specialist» is a versioned module with its own rules and source list. The clinician receives a ready draft plan with all citations and only verifies and tunes it for the specific patient.
- RF 310 red flags across 65 diseases
- DOSE renal / hepatic / age / weight adjustments
- LINK biomarker ↔ regimen ↔ monitoring
Why start using it today
Less time on manual NCCN/ESMO/MoH cross-checking, more patients seen. Fewer missed contraindications, less harm.
An LLM is not the decision-maker — a declarative rule engine is, with public code and a public KB. Plans are built without external LLM calls; only scientific databases (PubMed, ClinicalTrials.gov, DailyMed, openFDA) are queried. The clinician sees every «why» alongside every «what».
TP53, CD30, MYD88, eGFR, hepatic function — every flag automatically rewrites the plan: contraindications, dose adjustments, supportive care, monitoring.
Fresh labs or clinician decisions update both plans automatically — no need to re-sweep all the sources by hand.
Each drug in the plan is tagged: whether it is registered in Ukraine (MoH) and whether it is reimbursed by the state medical-guarantees programme (NHSU). The clinician immediately sees what is free, what is by prescription, and what has to be sourced separately. Access is metadata shown next to the recommendation, not a filter — regimen choice is driven by evidence, not by registration status.
A separate mode generates a plain-language version of the plan for the patient: no Latin, no acronyms, with explanations of why each step was chosen and what to watch for between visits. Same plan, two voices — clinical for the oncologist, human for the patient.
MIT-style. No paywall, no restrictions for public hospitals. Open-source means it can’t quietly disappear or be locked behind investors tomorrow.
65 diseases, 310 red flags, 250 indications, 186 regimens, 95 treatment algorithms — clinical sign-off received. The first real-patient plans were rated strong by practising oncologists. The KB grows weekly, but it is already the densest open evidence-to-plan layer for Ukrainian oncology that exists. No reason to wait.
Every missed biomarker can cost a life. Every hour of manual cross-checking is an hour the patient waits for a decision. The tool is ready today — start with a virtual patient and see your typical case through a layer of open standards.