2023 World Conference on Lung Cancer (Posters)-EP04.01. AI-based Detection on Low-Dose CT: A Focus on Augmenting Model Performance

Pdf Summary

This presentation discusses the use of artificial intelligence (AI) in the detection of lung nodules on low-dose chest computed tomography (CT) scans. The study aims to improve the performance of AI models in detecting these nodules to assist radiologists in their workload. <br /><br />Lung cancer is a leading cause of cancer-related deaths worldwide, and CT scans have proven to be effective in screening and diagnosing the disease. However, radiologists face the challenge of identifying small nodules within 3D volume images. The advent of low-dose CT screening has been beneficial in reducing mortality rates for lung cancer.<br /><br />The researchers collected 970 chest CT scans, including normal and abnormal datasets, to evaluate the performance of an AI-based CAD system designed using deep learning technology. Additionally, a reader study involving three expert radiologists was conducted to assess the system's clinical usefulness as a second reader. The analysis was performed using the Receiver-Operating Characteristic (ROC) method.<br /><br />The CAD system achieved a sensitivity of 0.72, specificity of 0.70, and accuracy of 0.71 per nodule. In specific cases, the sensitivity increased to 0.76, specificity to 0.81, and accuracy to 0.78. The results of the second reader study showed a significant improvement in detection performance when using the AI system in conjunction with human readers.<br /><br />In conclusion, the deep learning-based CAD system demonstrated excellent performance in detecting lung nodules on low-dose CT scans and acted as a valuable second reader. This system has the potential to enhance the detection performance and workload of radiologists in the field of lung cancer screening and early detection.

Asset Subtitle

Soo-Youn Ham

Meta Tag

Speaker Soo-Youn Ham

Topic Screening & Early Detection: Biomarkers/Imaging Technology

Keywords

artificial intelligence

AI

detection

lung nodules

low-dose chest computed tomography

CT scans

radiologists

workload

lung cancer

CAD system