(Learn more about this study)
Initial CDAS Request Approval
Apr 8, 2020
Automated assessment of non-cancer diagnoses and findings in NLST
The focus of our work is in developing machine learning and automated algorithmic evaluations of multi-modal data, including CT imaging, in evaluating non-oncologic diseases. While the focus of the NLST was in detecting and treating early stage lung cancer, the large population of screened, high-risk individuals also presents opportunities for assessment of non-oncologic diseases. Prior NLST-approved projects include those in emphysema evaluation, assessment of interstitial lung changes, and changes in lung findings over time. We seek to build upon these efforts by assessing the variety of non-oncologic diseases labeled in the dataset and enhancing our existing algorithms to detect these diseases, correctly classify them, monitor changes over time, and assess whether quantification assists in predictions.
1. Develop an algorithm to detect and classify non-oncologic diseases.
2. Evaluate whether time-sequence data, including imaging, is helpful in outcome predictions.
3. Evaluate quantified measures and whether these aid classifications and predictions.
Michael Muelly, imvaria inc.