Onco
Multidisciplinary tumor board
Simulation
Evidence-based, structured treatment recommendations for oncologists: an agentic architecture that also covers Chain-of-Verification validation, SUT/SGK reimbursement checks and clinical trial matching.
See the workflowOncoHub Council
Evidence-based recommendations
Drawn only from NCCN, ASCO, ESMO and validated institutional guidelines; no unsourced commentary.
Multi-specialist simulation
Core participants are fixed for every case; dynamic specialists are added as the case requires; 6+ disciplines.
SUT / SGK check
Reimbursement status is verified automatically for every recommended drug.
Chain-of-Verification
Staging, biomarker matching and treatment are each verified separately.
Sample case · Metastatic NSCLC
Metastatic NSCLC: Adenocarcinoma, cT(x), N0, M1c, Stage IVB. EGFR exon 19 deletion positive.
| Parameter | Value | Rationale |
|---|---|---|
| cT(x) | Primary mass, size not specified | Primary mass in the right lower lobe bronchus + subpleural lesion in the lower lobe sup. segment (PET-CT FDG+) |
| cN0 | No regional lymph node metastasis | Chest CT: no mediastinal or hilar lymph nodes of pathological size |
| M1c | Multi-organ metastasis | Liver: multiple metastases (largest S5-6, 15×15 mm); Skeleton: multiple bone metastases |
| Stage | Stage IVB | cT(x)N0M1c; Stage IVB because of M1c |
Board recommendation
Osimertinib 80 mg/day NCCN Category 1 · ESMO I-A level of evidence. Recommended as first-line systemic therapy; SGK reimbursement status checked automatically, covered by SUT.
Board preparation takes half the time.
Every case should ideally go through a tumor board; today the infrastructure cannot keep up.
The current challenge
Producing a multidisciplinary decision is one of the most valuable steps in oncology practice, but at most centres the groundwork runs on hand-written notes and spreadsheets. That is slow, inconsistent, and it takes the clinician's attention away from the patient.
- Long preparation: the pre-board process is not standardised; every case is prepared by hand, with notes and spreadsheets
- Access to expertise: assembling a multidisciplinary team for rare tumors is difficult
- Manual SUT checks: reimbursement review is slow and error-prone; guideline fit and coverage sit at separate tables
- Missing-data tracking: biomarker and test follow-up is not systematic; gaps are noticed late
- Guideline lag: NCCN/ESMO updates take time to reach daily practice
- No standardisation: board decisions come in a different format at every centre; they cannot be compared
A board in five steps.
Every case goes through the same structured process.
Knowledge base preparation
The tumor type is identified, and the guidelines, biomarker matrix and SUT reference specific to that type are loaded.
Case summary
Patient data is structured. Missing fields are reported back to the clinician; nothing is assumed.
Board members
Core participants are fixed for every case; dynamic specialists are added according to what the case requires.
Independent assessment
Each specialist produces an independent, evidence-referenced report. Opinions form without consensus pressure.
Consensus and self-check
Reports are synthesised, contradictions are detected, and Chain-of-Verification is run.
Fifty tumor types, seven groups.
A broad range, from the most common to the rarest tumors.
Thoracic & GI
SCLC, colorectal, gastric/GEJ, pancreas, hepatocellular carcinoma, esophagus.
6 typesUrogenital & gynecological
Prostate, bladder, renal cell carcinoma, ovary, endometrium, cervix.
6 typesDermatological & neuroendocrine
Melanoma, head and neck, thyroid, sarcoma, glioma/GBM, meningioma, NET/NEC.
7 typesHematological
DLBCL, Hodgkin, multiple myeloma, AML, ALL, CML, CLL.
7 typesRare & special
Testis, cholangiocarcinoma, mesothelioma, vulva, penis and other rare tumors.
20+ types“Guideline-appropriate” and “reimbursed”, at the same table.
A recommendation can fit the guidelines and still not be reimbursed in Türkiye.
Reimbursement integration
- 01
Automatic check
For every drug the SUT article, indication and conditions are matched. The Council verifies the SGK reimbursement status of each recommended drug before the report is issued.
- 02
Current SUT updates
Changes published in the Official Gazette are added to the knowledge base without disturbing the modular architecture.
NCCN / ASCO / ESMO say: appropriate
General-purpose AI stops here; it does not know national reimbursement.
SUT / SGK: is it covered?
Article, indication and conditions are matched automatically.
Both appropriate and reimbursed
The oncologist leaves the table with a decision that can actually be applied.
Six engineering decisions.
Each one answers a concrete need we met in clinical practice.
Chain-of-Verification
Automatic checkpoints at every processing step. Staging inconsistencies and faulty matches are caught before the report is issued.
Conditional branching
For missing test results it builds a dual scenario: “If positive, A; if negative, B.” The oncologist has a roadmap even while waiting for the result.
Iterative updating
When a new test result arrives the whole report is not regenerated; only the affected sections are updated.
Confidence scoring
Staging confidence and treatment confidence are graded separately: high, medium or low.
SUT / SGK integration
The SGK reimbursement status of every recommended drug is checked automatically. “Guideline-appropriate” and “reimbursed in Türkiye” now sit at the same table.
Modular architecture
New tumor types, guideline updates and knowledge modules are added without disturbing the existing structure.
The difference between general-purpose AI and OncoHub
| Feature | OncoHub | General AI | Portals |
|---|---|---|---|
| SUT / SGK integration | Yes | No | No |
| 50+ tumor type coverage | Yes | Limited | Limited |
| Multi-specialist simulation | Yes | No | No |
| Chain-of-Verification | Yes | No | No |
| Turkish clinical language | Yes | Translated | Limited |
| Missing-data detection | Yes | No | No |
| NCCN / ESMO referencing | Yes | Unclear | Yes |
For a more decisive and efficient board, OncoHub
Write to us about enterprise subscriptions, pilot deployments and demo requests.