Advantages of integrating artificial intelligence and spectral CT for lung nodule classification and prognostic...

The accurate diagnosis of lung nodules remains a significant challenge in clinical practice due to their diverse and often nonspecific imaging characteristics. This limitation underscores the need for more advanced analytical approaches. The present review aims to summarise and discuss the advancements and applications of integrating artificial intelligence (AI) with spectral computed tomography (CT) for diagnosing lung nodules with diverse characteristics.

This narrative review sourced literature from PubMed/MEDLINE, Web of Science, and Google Scholar (2010-2025) using keywords "spectral CT", "pulmonary nodule", and "artificial intelligence". Inclusion criteria focused on studies applying spectral CT and/or AI to lung nodule characterization. Two reviewers independently screened and selected studies, with a third resolving discrepancies. A total of 25 studies were included for analysis.

This review highlights the advances in applying this dual strategy to the multiparametric analysis of pulmonary nodules. Studies indicate that combining the rich parametric information provided by spectral CT [e.g., iodine concentration (IC), spectral curves] with AI's powerful pattern recognition and quantitative analysis capabilities can significantly enhance diagnostic efficacy for pulmonary nodules exhibiting diverse characteristics (e.g., varying sizes, densities, locations). This integrated approach demonstrates considerable potential for improving diagnostic accuracy in lung nodules. It significantly enhances diagnostic efficacy for nodules exhibiting diverse characteristics (e.g., varying size, density, and location). This combined methodology shows significant promise in improving the accuracy of benign-malignant differentiation and prognosis prediction.

The synergistic application of AI and energy-spectrum CT is recognized as an emerging frontier in pulmonary nodule diagnosis. This dual-strategy approach overcomes the limitations of traditional imaging and single-technology methods, providing a more comprehensive and reliable tool for the precise identification, qualitative diagnosis, and prognostic assessment of pulmonary nodules. It demonstrates significant clinical value and broad application prospects.