1) Create a public database of 50,000 lung cancer screening imaging datasets | |
• Leverage available data and phantom validation tools to ensure this is a high-quality image collection. | |
•Ensure that there is sufficient diversity (e.g., demographic, geographic, disease states, scanner models). | |
•Ensure that major lung cancer sub-populations (e.g., small cell) are represented in such a collection. | |
•Provide reliable sources of “Ground truth” information including lung nodule biopsy results. | |
2) Increase resource to support and curate new and existing databases and related technology infrastructure | |
•Centralized Database Example: The Cancer Imaging Archive (TCIA). | |
•Distributed Database Example: Early Lung Imaging Confederation (ELIC). | |
•Technology Example: Automated insertion of lung nodules in existing CT scans. | |
3) Investigate a New Global Patient Data Electronic Submission Infrastructure | |
•Allow global patients to opt-in to having their data used by researchers with appropriate de-identification. | |
•Hospitals and healthcare facilities will inform patients of this opportunity to advance research. | |
•A global opt-in strategy will need informed consent templates for different regions of the world. | |
4) Improve the QIBA Small Lung Nodule Profile | |
•Add support for volume measurement of part-solid lung nodules. | |
•Add support for the ability to use multiple CT scanners over time to measure lung nodules. |