Abstract
Lung cancer is one of the most lethal cancers worldwide, largely due to its late diagnosis and progression before symptoms manifest. Lung nodules, detected primarily through CT imaging, are a key indicator of early lung cancer, and timely identification is critical for effective intervention. Recently, artificial intelligence (AI), particularly deep learning algorithms like convolutional neural networks (CNNs), has shown potential for enhancing the detection accuracy of these nodules. This systematic review compares the diagnostic accuracy, sensitivity, specificity, and time efficiency of AI algorithms and radiologists in lung nodule detection on CT scans. We conducted an exhaustive search across PubMed, Cochrane Library, IEEE Xplore, and Embase, covering studies published between 2010 and 2023. Fifty studies meeting inclusion criteria were analyzed, focusing on performance metrics and the algorithms used. Results indicated that AI models achieved higher sensitivity, especially with nodules <6mm, and reduced detection times; however, specificity remained variable. This study underscores AI’s role in advancing early lung cancer detection but highlights the need for integration strategies, ethical frameworks, and further clinical trials.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call