Skip to main content
Back to challenges
Challenge

Build a Google ADK & OpenAI o3 Multi-Cancer Diagnostic Utility Agent

Multi-Cancer Early Detection (MCED) blood tests generate complex real-world data requiring clear diagnostic utility mapping and patient counseling. Build a Google Agent Development Kit (ADK) agent powered by OpenAI o3 reasoning to evaluate patient MCED test results and generate clinical action plans.

Data ScienceHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Construct an agent using Google ADK and OpenAI o3 reasoning to interpret MCED test outcomes and generate evidence-based downstream workups.

Delivery guide

How work is evaluated

Evaluation

Evaluates agent ability to generate appropriate follow-up diagnostic pathways from MCED signals.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

correct_imaging_modality

Checks if contrast-enhanced CT or EUS is included for pancreatic tissue of origin.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

guideline_conformance_score

Conformity score against standard NCCN workup protocols for suspected malignancies. • target: 0.95 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Orchestrate clinical workflow agents using Google Agent Development Kit (ADK)

  • Leverage OpenAI o3 complex reasoning capabilities to weigh tissue of origin (TOO) probabilities

  • Formulate standardized diagnostic workup recommendations based on NCCN guidelines

  • Generate patient-facing plain language explanations that prevent over-diagnosis confusion

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic dataset of 30 MCED blood test patient results, covering positive and negative signals across multiple organ systems.

Challenge setup

Evaluates agent ability to generate appropriate follow-up diagnostic pathways from MCED signals.

Challenge input

JSON containing signal_detected (boolean), predicted_too (string), and patient_age (int).

Google ADK

Agent development kit framework for modular agent architecture.

OpenAI o3

High-capacity reasoning engine for clinical logic.

Evaluated output

JSON with recommended_workup (list), recommended_imaging (list), and clinical_summary (string).

Evaluation checks
  • Checks if contrast-enhanced CT or EUS is included for pancreatic tissue of origin.
  • Conformity score against standard NCCN workup protocols for suspected malignancies. • target: 0.95 • range:...
Configured evaluation evidence
  • Benchmark: NCCN Diagnostic Pathway Benchmark
  • Guideline Conformance Score target: 0.95
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • Google · Required
  • Google ADK · Optional
Policy Serving
  • o3 · Optional

Evaluation contract

  • correct_imaging_modality · Weight 1
  • guideline_conformance_score · Weight 1

Metrics

  • guideline_conformance_score · Target 0.95

Benchmarks

  • NCCN Diagnostic Pathway Benchmark
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing signal_detected (boolean), predicted_too (string), and patient_age (int).

    Output: JSON with recommended_workup (list), recommended_imaging (list), and clinical_summary (string).

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent Framework
  • OpenAI o3 · Reasoning LLM

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Google ADK & OpenAI o3 Multi-Cancer Diagnostic Utility Agent