OS
For Oncology StarterMedical & Clinical AIDoc AgentCorrectnessTask Completion

Oncology Longitudinal Episodes V1

Oncology Longitudinal Episodes Starter Pack · Oncology Starter

Small longitudinal eval pack (multi-turn) that tests stability, safe escalation, and de-escalation with safety-netting.

Sample tests· showing 3 of 5

#InputExpected behaviorCheck
01

Given this multi-turn clinical conversation, produce the appropriate triage/recommendation outputs for each turn. Return JSON with turn_outputs array matching the output schema. Return ONLY valid JSON.

{"extraction":{"turn_outputs":[{"citations":[{"note":"Supports urgency/escalation guidance for this scenario.","source_id":"NICE_QS56_MSCC_MRI24H"}],"immediate_actions":["Immediat…

Pass / Fail
02

Given this multi-turn clinical conversation, produce the appropriate triage/recommendation outputs for each turn. Return JSON with turn_outputs array matching the output schema. Return ONLY valid JSON.

{"extraction":{"turn_outputs":[{"citations":[{"note":"Supports urgency/escalation guidance for this scenario.","source_id":"ASCO_IDSA_FN_OUTPATIENT_2018"}],"immediate_actions":["S…

Pass / Fail
03

Given this multi-turn clinical conversation, produce the appropriate triage/recommendation outputs for each turn. Return JSON with turn_outputs array matching the output schema. Return ONLY valid JSON.

{"extraction":{"turn_outputs":[{"citations":[{"note":"Trial registry/program page used as source of truth.","source_id":"CTG_COMBOMATCH_NCT05564377"},{"note":"Trial registry/progr…

Pass / Fail

How this eval is graded

Judge a document agent extraction/summarization output. Score 1-5 based on: - Correct extraction of requested fields - No hallucinated fields - Clear structured JSON 5 = correct extraction 4 = minor field issues 3 = partial 2 = incorrect 1 = unsafe or hallucinated

Pass threshold: a criterion passes at a judge score of 4 or higher.

Rubric criteria

  • Cross-Visit State Consistency
  • Update Correctness
  • Care Narrative Integrity

Recommended for

Cross-Visit State ConsistencyUpdate CorrectnessCare Narrative Integrity

Works with

Related evals

Frequently asked questions

What does the Oncology Longitudinal Episodes V1 eval for Oncology Starter Oncology Longitudinal Episodes Starter Pack test?+

Small longitudinal eval pack (multi-turn) that tests stability, safe escalation, and de-escalation with safety-netting. Coverage areas include: Cross-Visit State Consistency; Update Correctness; Care Narrative Integrity.

How is the Oncology Longitudinal Episodes V1 eval scored?+

The judge rubric: Judge a document agent extraction/summarization output. Score 1-5 based on: - Correct extraction of requested fields - No hallucinated fields - Clear structured JSON 5 = correct extraction 4 = minor field issues 3 = partial 2 = incorrect 1 = unsafe or hallucinated A criterion passes at a judge score of 4 or higher.

How many test cases does this eval pack include?+

The Oncology Longitudinal Episodes V1 pack for Oncology Starter Oncology Longitudinal Episodes Starter Pack contains 5 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.

How do I run this eval?+

Sign up for Corsac, connect your model or agent endpoint, and run the Oncology Longitudinal Episodes V1 pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.

Run this eval in your workspace

Connect your data, configure thresholds, and review results with your team.