Assessment of artificial intelligence-aided computed tomography in lung cancer screening

This study assesses the use of AI-aided CT with Careline Soft's AVIEW Metric software for lung cancer screening. Lung cancer is a leading cause of cancer deaths, making early detection crucial. AVIEW Metric's deep learning model was compared with radiologists' performance in detecting and classifying lung nodules. The study was conducted at Suez Canal University Hospital, implementing Lung-RADS protocols. Initial CAD system reviews showed high sensitivity (93.0%) and specificity (95.5%) with 93.6% accuracy in nodule detection. After radiologist review, sensitivity increased to 98.2%, with 95.5% specificity and 97.4% accuracy. For lung cancer screening, initial CAD results had 97.9% specificity and 96.9% accuracy. Radiologist-reviewed results showed excellent agreement (88.6%, κ = 0.951) with CAD for Lung-RADS 3 and 4 nodule categorization. The CAD system significantly improved detection and agreement in nodule categorization, enhancing lung cancer screening efficacy.