LangChain IL-6 Pathway Clinical Trial Matching Workflow
Evaluating anti-inflammatory drugs like IL-6 inhibitors across cardiovascular clinical trials requires stateful patient matching. Build a LangGraph and LangChain agent to query patient EHR records against trial inclusion and exclusion criteria for cardiac biomarker profiles. Reach greater than 90 percent precision in cohort qualification.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop a stateful LangGraph agent using LangChain to match patient electronic health records against complex IL-6 cardiovascular clinical trial eligibility rules.
How work is evaluated
Evaluates precision, recall, and state transition logic of the trial matching workflow.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
correct_eligibility_decision
Verifies that eligible status matches ground truth annotation.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
matching_precision
Precision of trial participant qualification decisions. • target: 0.91 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build a stateful patient trial matching workflow using LangGraph state graphs
Extract structured lab markers (hsCRP, IL-6, troponin) from unstructured EHR clinical notes
Implement inclusion/exclusion logic verification using LangChain tools
Generate patient qualification audit logs for trial recruitment coordinators
Reference links and supporting material
Synthetic electronic health record summaries annotated with cardiac lab values and clinical trial inclusion outcomes.
Challenge setup
Evaluates precision, recall, and state transition logic of the trial matching workflow.
Challenge input
JSON containing patient record text and lab values
LangChain
Required framework for building LLM tools and integrations.
LangGraph
Enables state graph routing for deterministic eligibility checks.
Evaluated output
JSON with eligible status, satisfied criteria array, and failed criteria array
- Verifies that eligible status matches ground truth annotation.
- Precision of trial participant qualification decisions. • target: 0.91 • range: 0-1
- Benchmark: TrialMatch-Bench
- Matching Precision target: 0.91
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- LangChain · Required
- Langchain · Optional
- Llama 3.2 · Optional
- LangChain · Required
- Langchain · Optional
Evaluation contract
- correct_eligibility_decision · Weight 1
- matching_precision · Weight 1
Metrics
- matching_precision · Target 0.91
Benchmarks
- TrialMatch-Bench
- biotech
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing patient record text and lab values
Output: JSON with eligible status, satisfied criteria array, and failed criteria array
Execution harness: Python
Evaluation tools
- LangChain · Agent core framework
- LangGraph · Stateful flow orchestrator
Recipe freeze marker
No recipe freeze time is recorded.