2022 World Conference on Lung Cancer (ePosters)-EP01.01-002. Clinical and Molecular Features of Chinese Early-stage Multiple Primary Lung Cancer Patients

Pdf Summary

The National Lung Screening Trial (NLST) has a public database of CT images that is being used for developing and testing algorithms related to lung nodule measurement, detection, and diagnosis. However, these scans are 15-20 years old, and for accurate algorithm development, modern quantitative image standards are necessary. The study analyzed the availability of scans that meet the protocol recommendations of the Quantitative Imaging Biomarker Alliance (QIBA) for scan acquisition and reconstruction.<br /><br />The analysis found that only 2.9% of the NLST images meet the QIBA eligibility criteria. The major reason for exclusion was the slice thickness of the scans. Other organizations also recommend a slice thickness of 1.25 mm or less for lung cancer screening CT protocols. Additionally, the number of images meeting the quality standards and belonging to participants with lung cancer was very small.<br /><br />The study highlights the need for high-quality image databases for software development in screening CT image analysis that represent the CT images currently being obtained for lung cancer screening. Existing public databases such as the NLST may have limitations, and new databases and data distribution approaches are necessary.<br /><br />The results emphasize the importance of carefully utilizing existing databases and the need for the development of newer databases with improved quality. Organizations like the IASLC and Medical Imaging and Data Resource Center are working on developing new databases for better algorithm development.<br /><br />In conclusion, the NLST database has limited availability of scans meeting the modern quantitative image standards required for optimizing algorithm performance. Newer databases with improved quality are needed for successful development of artificial intelligence algorithms.

Asset Subtitle

Yongjie Wang, China

Meta Tag

Speaker Yongjie Wang, China

Topic Early Detection and Screening - Biomarkers

Keywords

National Lung Screening Trial

NLST

CT images

lung nodule measurement

lung nodule detection

lung nodule diagnosis

Quantitative Imaging Biomarker Alliance

QIBA

slice thickness

lung cancer screening CT protocols