Skip to Main Content

An official website of the United States government

Principal Investigator
Name
Jonathan Walsh
Degrees
Ph.D
Institution
Unlearn.AI
Position Title
Chief Scientific Officer and Founder
Email
About this CDAS Project
Study
PLCO (Learn more about this study)
Project ID
PLCO-1957
Initial CDAS Request Approval
Jul 28, 2025
Title
Building a Non-small Cell Lung Cancer and Colorectal Cancer Machine Learning Model
Summary
Since 2017, Unlearn has been at the forefront of applying AI to clinical development through the use of digital twins—AI-generated predictions of patient-level outcomes. Our deep learning models are trained on longitudinal, multivariate data from clinical trials, registries, and observational studies, enabling accurate forecasts of disease progression and patient outcomes. We are now expanding into oncology, with an initial focus on non-small cell lung cancer (NSCLC) and colorectal cancer (CRC). By leveraging patient-level datasets from PCLO alongside other data osurces, our goal is to develop a machine learning model for NSCLC and CRC.
Aims

1. Data Harmonization
Unlearn will aggregate, standardize, and harmonize datasets from PCLO. This critical first step ensures data consistency and quality, forming a strong foundation for reliable model development.
2. Model Development
Using the harmonized dataset, Unlearn will develop a machine learning model tailored to NSCLC and CRC. The model will be validated on a hold-out subset to assess its predictive accuracy and robustness in simulating disease progression.
3. Strategic Collaborations
Unlearn will establish partnerships with biotech and pharmaceutical companies to integrate digital twins—AI-generated predictions of patient-level outcomes—into clinical trials. These digital twins can serve as external controls in single-arm trials or be used to reduce sample size and increase statistical power in randomized controlled trials through covariate adjustment.

Collaborators

None at this time