OncoForge Conceptual Simulations

A conceptual simulator study for cancer-biology literacy, signal interpretation, tradeoff comparison, synthetic profiles, and non-medical report generation.

Research brief

Synthetic data, signal heatmaps, tumor/healthy curves, treatment-cocktail abstractions, and report comparison.

Problem: Cancer biology concepts are hard to compare without transparent modeling language.
Hypothesis: A conceptual simulator can clarify signals, tradeoffs, and uncertainty without offering medical direction.
Current boundary: Not medical advice, not clinical prediction, not treatment recommendation.

Conceptual biology simulator

Study Overview

OncoForge is a study in model literacy: how to represent cancer-related signals, tradeoffs, and uncertainty in a way that can be inspected without becoming medical advice.

The simulator uses synthetic profiles to model tumor burden, healthy-cell preservation, signal heatmaps, and report comparisons without implying patient prediction.

Its scientific value comes from transparency: assumptions, result drivers, and conceptual outputs stay visible together.

Primary audience: learners, researchers, and developers exploring model structure.
Primary data: synthetic profiles, not patient records.
Primary output: literacy, comparison, and report generation.

Variables

State variable

Tumor burden

A conceptual pressure signal that can rise, fall, or plateau across a synthetic run.

Model Lens

Choose the output that matters most for the first demo.

Mechanism clarity

Signal movement clarifies what changed in the synthetic model.

Study record

Concept

Model-literacy simulator

The study begins as a way to make abstract cancer-biology tradeoffs visible.

Build

Synthetic profile set

Synthetic profiles can show signal movement without using real patient information.

Review

Report discipline

Exports and model cards keep each run inspectable and bounded.

Next experiments

01
Create synthetic profiles for high inflammation, low immune visibility, and resistant-growth examples.
02
Add model cards explaining every variable in plain language.
03
Export run history as HTML, JSON, and CSV.
04
Test local AI summaries against strict non-medical language boundaries.

Related research area All studies