Evaluating the Accuracy of Artificial Intelligence Models for Early Lung Cancer Detection: Evidence From a Systematic Review

Lung cancer remains the leading cause of cancer-related mortality worldwide, with early detection critical for improving outcomes. While low-dose computed tomography (CT) screening has demonstrated mortality benefits, its implementation is constrained by high false-positive rates and inter-observer variability. Artificial intelligence (AI) has emerged as a promising tool to enhance diagnostic accuracy. This systematic review evaluates the accuracy of AI models for early lung cancer detection. A systematic search was conducted across PubMed, Scopus, Embase, ACM Digital Library, and IEEE Xplore for studies published between January 2021 and December 2025 following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Studies evaluating AI models for early lung cancer detection using imaging modalities were included. Methodological quality was assessed using the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2) tool. Ten studies met the inclusion criteria. AI models, predominantly convolutional neural networks, demonstrated high diagnostic accuracy across CT, chest X-ray, and histopathology imaging. Accuracy ranged from 75.2% to 99.17%, with sensitivities between 80% and 100% and specificities from 58.7% to 98.03%. Several studies reported AI performance comparable to or exceeding that of radiologists. Ensemble and hybrid models consistently outperformed single-architecture approaches. Quality assessment revealed moderate to high methodological quality overall. AI models achieve high diagnostic accuracy for early lung cancer detection, with performance often comparable to that of radiologists. AI is best positioned as a complementary tool to augment human expertise. Future research should prioritize prospective designs, external validation, and standardized reporting.