Mehri-Kakavand, GhazalMdletshe, SibusisoWang, Alan2025-03-092025-03-092025-01(2025). Journal of Medical Radiation Sciences, Online Version of Record before inclusion in an issue.2051-3895https://hdl.handle.net/2292/71603<h4>Introduction</h4>Non-small cell lung cancer (NSCLC) is the leading cause of cancer-related mortality worldwide. Despite advancements in early detection and treatment, postsurgical recurrence remains a significant challenge, occurring in 30%-55% of patients within 5 years after surgery. This review analysed existing studies on the utilisation of artificial intelligence (AI), incorporating CT, PET, and clinical data, for predicting recurrence risk in early-stage NSCLCs.<h4>Methods</h4>A literature search was conducted across multiple databases, focusing on studies published between 2018 and 2024 that employed radiomics, machine learning, and deep learning based on preoperative positron emission tomography (PET), computed tomography (CT), and PET/CT, with or without clinical data integration. Sixteen studies met the inclusion criteria and were assessed for methodological quality using the METhodological RadiomICs Score (METRICS).<h4>Results</h4>The reviewed studies demonstrated the potential of radiomics and AI models in predicting postoperative recurrence risk. Various approaches showed promising results, including handcrafted radiomics features, deep learning models, and multimodal models combining different imaging modalities with clinical data. However, several challenges and limitations were identified, such as small sample sizes, lack of external validation, interpretability issues, and the need for effective multimodal imaging techniques.<h4>Conclusions</h4>Future research should focus on conducting larger, prospective, multicentre studies, improving data integration and interpretability, enhancing the fusion of imaging modalities, assessing clinical utility, standardising methodologies, and fostering collaboration among researchers and institutions. Addressing these aspects will advance the development of robust and generalizable AI models for predicting postsurgical recurrence risk in early-stage NSCLC, ultimately improving patient care and outcomes.Print-ElectronicItems in ResearchSpace are protected by copyright, with all rights reserved, unless otherwise indicated. Previously published items are made available in accordance with the copyright policy of the publisher.https://researchspace.auckland.ac.nz/docs/uoa-docs/rights.htmPET/CTartificial intelligencenon‐small cell lung cancerradiomicsrecurrence prediction32 Biomedical and Clinical Sciences3202 Clinical Sciences3211 Oncology and CarcinogenesisMachine Learning and Artificial IntelligenceBiomedical ImagingBioengineeringData ScienceNetworking and Information Technology R&D (NITRD)PreventionClinical ResearchLungCancerLung Cancer4.2 Evaluation of markers and technologies4.1 Discovery and preclinical testing of markers and technologiesGeneric health relevance3 Good Health and Well BeingJournal Article10.1002/jmrs.860Copyright: The authors39844750 (pubmed)2051-3909Attribution 4.0 International