{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T16:53:04Z","timestamp":1790787184119,"version":"4.1.0"},"reference-count":62,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T00:00:00Z","timestamp":1784851200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100011482","name":"Yantai Science and Technology Bureau","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100011482","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Oncol."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Accurate preoperative stratification of ground-glass nodules (GGNs) into invasive versus non-invasive adenocarcinomas is essential for optimizing surgical management and preventing overtreatment. Radiomics, deep learning, and hybrid models have shown potential for noninvasive histopathologic prediction, but their comparative diagnostic performance and generalizability remain unclear. This study systematically compares the diagnostic accuracy, methodological quality, and clinical applicability of radiomics, deep learning, and hybrid AI approaches for differentiating invasive from non-invasive lung adenocarcinomas manifesting as GGNs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A comprehensive search of PubMed, Embase, Cochrane Library, and Web of Science was conducted through 24 September 2025 (PROSPERO CRD42025115051). Twenty-two studies (9,768 nodules) met inclusion criteria, evaluating AI models with exact segmentation or embedded nodule localization and histopathologic confirmation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Hybrid models demonstrated higher pooled diagnostic performance (AUC 0.94, sensitivity 0.91 [95% CI 0.88\u20130.93], specificity 0.87 [0.83\u20130.90]) than deep learning (AUC 0.92) and radiomics (AUC 0.89). At a 20% pre-test probability, a positive hybrid AI result increased post-test probability of invasiveness to approximately 59%, while a negative result decreased it to approximately 5% (approximate pooled LR\u00a0+\u00a06.0; LR\u2013 0.22). However, effect estimates showed substantial heterogeneity (I\u00b2 &amp;gt; 60%) and publication bias suggested sensitivity inflation (3\u20135%). Most studies were single-center and retrospective, with 82% conducted in East Asia and limited external validation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Hybrid AI showed higher pooled accuracy compared with radiomics or deep learning alone, but the certainty of evidence is low\u2013moderate due to inconsistency, publication bias, and restricted population diversity. AI-based GGN risk stratification should currently complement, rather than replace, multidisciplinary clinical decision-making until validated in prospective multicenter trials.<\/jats:p>\n                    <jats:p>\n                      <jats:bold>Systematic review registration:<\/jats:bold>\n                      <jats:ext-link>https:\/\/www.crd.york.ac.uk\/prospero\/display_record.php?RecordID=CRD42025115051<\/jats:ext-link>\n                      , identifier CRD42025115051.\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fonc.2026.1752652","type":"journal-article","created":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T04:24:36Z","timestamp":1784867076000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Comparative diagnostic accuracy of radiomics, deep learning, and hybrid AI for invasiveness stratification of pulmonary ground-glass nodules: a systematic review and meta-analysis"],"prefix":"10.3389","volume":"16","author":[{"given":"Ning","family":"Dong","sequence":"first","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhongwei","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhenjie","family":"Cong","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xinru","family":"Ba","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jing","family":"Qin","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhuxiao","family":"Lin","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiaolin","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peng","family":"Liang","sequence":"additional","affiliation":[{"name":"Department of Radiology, Yantaishan Hospital","place":["Yantai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,24]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"12","DOI":"10.3322\/caac.21820","article-title":"Cancer statistics, 2024","volume":"74","author":"Siegel","year":"2024","journal-title":"CA Cancer J Clin"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.3322\/caac.21660","article-title":"Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries","volume":"71","author":"Sung","year":"2021","journal-title":"CA Cancer J Clin"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMoa1911793","article-title":"Reduced lung-cancer mortality with volume CT screening in a randomized trial","volume":"382","author":"de Koning","year":"2020","journal-title":"N Engl J Med"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMoa1209120","article-title":"Results of initial low-dose computed tomographic screening for lung cancer","volume":"368","author":"Church","year":"2013","journal-title":"N Engl J Med"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.2214\/AJR.10.5813","article-title":"Ground-glass nodules on chest CT as imaging biomarkers in the management of lung adenocarcinoma","volume":"196","author":"Goo","year":"2011","journal-title":"AJR Am J Roentgenol"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1097\/JTO.0b013e318206a221","article-title":"IASLC\/ATS\/ERS international multidisciplinary classification of lung adenocarcinoma","volume":"6","author":"Travis","year":"2011","journal-title":"J Thorac Oncol"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.3978\/j.issn.2072-1439.2015.08.27","article-title":"Growth of pure ground-glass lung nodule detected at computed tomography","volume":"7","author":"Aoki","year":"2015","journal-title":"J Thorac Dis"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.2017161659","article-title":"Guidelines for management of incidental pulmonary nodules detected on CT images: Fleischner Society 2017","volume":"284","author":"MacMahon","year":"2017","journal-title":"Radiology"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.21037\/tlcr.2018.07.04","article-title":"Ground-glass nodules of the lung in never-smokers and smokers: clinical and genetic insights","volume":"7","author":"Kobayashi","year":"2018","journal-title":"Transl Lung Cancer Res"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.21037\/jtd-22-583","article-title":"Risk of adenocarcinoma in suspicious ggo","volume":"14","author":"Roy","year":"2022","journal-title":"J Thorac Dis"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1245\/s10434-022-12621-x","article-title":"Long-term prognosis and prognostic indicators of stage IA lung adenocarcinoma","volume":"30","author":"Yotsukura","year":"2023","journal-title":"Ann Surg Oncol"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.1016\/S0140-6736(21)02333-3","article-title":"Segmentectomy versus lobectomy in small-sized peripheral non-small-cell lung cancer","volume":"399","author":"Saji","year":"2022","journal-title":"Lancet"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1245\/s10434-020-08241-y","article-title":"Clinicopathologic features and genetic alterations in adenocarcinoma in situ and minimally invasive adenocarcinoma","volume":"27","author":"Jia","year":"2020","journal-title":"Ann Surg Oncol"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1016\/j.jtho.2021.04.007","article-title":"Long-term prognosis of resected AIS and MIA","volume":"16","author":"Yotsukura","year":"2021","journal-title":"J Thorac Oncol"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0316-z","article-title":"A guide to deep learning in healthcare","volume":"25","author":"Esteva","year":"2019","journal-title":"Nat Med"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1496","DOI":"10.1016\/j.jtcvs.2021.02.010","article-title":"Predicting benign, preinvasive, and invasive lung nodules using ML","volume":"163","author":"Ashraf","year":"2022","journal-title":"J Thorac Cardiovasc Surg"},{"key":"B18","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1186\/s41747-018-0068-z","article-title":"Radiomics: facts and challenges","volume":"2","author":"Rizzo","year":"2018","journal-title":"Eur Radiol Exp"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2022.995870","article-title":"3D deep learning vs current methods for invasiveness prediction in lung adenocarcinoma","volume":"12","author":"Lv","year":"2022","journal-title":"Front Oncol"},{"key":"B20","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1038\/s41746-023-00866-z","article-title":"Ensemble deep learning for IA risk stratification","volume":"6","author":"Zhou","year":"2023","journal-title":"NPJ Dig Med"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-019-06628-4","article-title":"CT-based DL vs radiologists for IA differentiation","volume":"30","author":"Kim","year":"2020","journal-title":"Eur Radiol"},{"key":"B22","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1007\/s11547-023-01730-6","article-title":"Combined DL-radiomics-clinical model for lung nodules","volume":"129","author":"Lin","year":"2024","journal-title":"Radiol Med"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-024-10907-0","article-title":"Deep learning in pulmonary nodule detection and segmentation: a systematic review","volume":"35","author":"Gao","year":"2025","journal-title":"Eur Radiol"},{"key":"B24","doi-asserted-by":"publisher","first-page":"12688","DOI":"10.1038\/s41598-020-69534-6","article-title":"Radiomics feature reproducibility under segmentation variability","volume":"10","author":"Haarburger","year":"2020","journal-title":"Sci Rep"},{"key":"B25","doi-asserted-by":"publisher","first-page":"240","DOI":"10.1186\/s12880-024-01421-2","article-title":"Radiomics-based ML to predict invasiveness in GGO pulmonary adenocarcinoma","volume":"24","author":"Bin","year":"2024","journal-title":"BMC Med Imaging"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.21037\/tlcr-24-1039","article-title":"Development and validation of growth prediction models for multiple pulmonary ground-glass nodules based on CT features, radiomics, and deep learning","volume":"14","author":"Cui","year":"2025","journal-title":"Transl Lung Cancer Res"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.21037\/qims-24-1839","article-title":"Diagnostic accuracy of DL for invasiveness assessment of GGNs: systematic review and meta-analysis","volume":"15","author":"Wu","year":"2025","journal-title":"Quant Imaging Med Surg"},{"key":"B28","first-page":"97","article-title":"Searching and selecting studies","volume-title":"Cochrane handbook for DTA","author":"Spijker","year":"2023"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.7326\/0003-4819-155-8-201110180-00009","article-title":"QUADAS-2: a revised quality assessment tool for DTA studies","volume":"155","author":"Whiting","year":"2011","journal-title":"Ann Intern Med"},{"key":"B30","first-page":"249","article-title":"Undertaking meta-analysis","volume-title":"Cochrane handbook for DTA","author":"Takwoingi","year":"2023"},{"key":"B31","doi-asserted-by":"publisher","first-page":"3","DOI":"10.4078\/jrd.2018.25.1.3","article-title":"Process of conducting DTA meta-analyses","volume":"25","author":"Lee","year":"2018","journal-title":"J Rheum Dis"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1002\/cam4.6402","article-title":"Transfer learning for subtype discrimination in GGN adenocarcinoma","volume":"12","author":"Fu","year":"2023","journal-title":"Cancer Med"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2021.725599","article-title":"Serial non-contrast CT-based DL for invasiveness prediction","volume":"11","author":"Xu","year":"2021","journal-title":"Front Oncol"},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2020.00418","article-title":"Fusion of DL and radiomics features for IA in GGNs","volume":"10","author":"Xia","year":"2020","journal-title":"Front Oncol"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1002\/ima.22614","article-title":"Segmentation and classification of ground glass nodule on CT images","volume":"31","author":"Wang","year":"2021","journal-title":"Int J Imaging Syst Technol"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2021.700158","article-title":"DL for Malignancy and invasiveness in subsolid nodules","volume":"11","author":"Shen","year":"2021","journal-title":"Front Oncol"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.2214\/AJR.23.29674","article-title":"Lung-PNet for IA diagnosis in pure GGNs","volume":"222","author":"Qi","year":"2024","journal-title":"AJR Am J Roentgenol"},{"key":"B38","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15293","article-title":"IMAL-Net: interpretable multi-task attention learning network for invasive lung adenocarcinoma screening in CT images","volume":"48","author":"Wang","year":"2021","journal-title":"Med Phys"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.21037\/jtd-20-2981","article-title":"Peritumoral radiomics improves prediction of invasiveness in ground-glass nodules","volume":"13","author":"Wang","year":"2021","journal-title":"J Thorac Dis"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-020-07620-z","article-title":"DL-based differentiation of IA from pre\/minimally invasive lesions","volume":"31","author":"Park","year":"2021","journal-title":"Eur Radiol"},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.14132187","article-title":"Texture analysis of part-solid GGNs to differentiate invasiveness","volume":"273","author":"Chae","year":"2014","journal-title":"Radiology"},{"key":"B42","doi-asserted-by":"publisher","DOI":"10.1155\/2018\/6803971","article-title":"Radiomics model to predict IA of GGNs","volume":"2018","author":"Xue","year":"2018","journal-title":"BioMed Res Int"},{"key":"B43","doi-asserted-by":"publisher","first-page":"26","DOI":"10.1016\/j.ejrad.2019.05.012","article-title":"Radiomics model for invasiveness of solitary part-solid nodules","volume":"117","author":"Weng","year":"2019","journal-title":"Eur J Radiol"},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejrad.2019.01.021","article-title":"Development and validation of a radiomics nomogram for identifying invasiveness of pulmonary adenocarcinomas appearing as subcentimeter ground-glass opacity nodules","volume":"112","author":"Zhao","year":"2019","journal-title":"Eur J Radiol"},{"key":"B45","doi-asserted-by":"publisher","DOI":"10.1016\/j.crad.2020.10.005","article-title":"A radiomics study to predict invasive pulmonary adenocarcinoma appearing as pure ground-glass nodules","volume":"76","author":"Cai","year":"2021","journal-title":"Clin Radiol"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.1097\/MD.0000000000016119","article-title":"DL for invasiveness prediction: preliminary study","volume":"98","author":"Yanagawa","year":"2019","journal-title":"Med (Balt)"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejrad.2021.110041","article-title":"DL to distinguish MIA vs IA in subsolid nodules","volume":"145","author":"Chen","year":"2021","journal-title":"Eur J Radiol"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-025-13447-9","article-title":"Deep learning and radiomics fusion for predicting the invasiveness of lung adenocarcinoma within ground glass nodules","volume":"15","author":"Sun","year":"2025","journal-title":"Sci Rep"},{"key":"B49","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2023.1208758","article-title":"Dual-energy CT radiomics for invasiveness prediction","volume":"13","author":"Zheng","year":"2023","journal-title":"Front Oncol"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.21037\/jtd-24-27","article-title":"Radiomics of invasion and instability in GGNs","volume":"16","author":"Zhang","year":"2024","journal-title":"J Thorac Dis"},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejro.2025.100680","article-title":"Ct radiomics ml for ia prediction in ggns","volume":"15","author":"Chen","year":"2025","journal-title":"Eur J Radiol Open"},{"key":"B52","doi-asserted-by":"publisher","DOI":"10.1016\/j.jclinepi.2019.12.020","article-title":"GRADE guidelines: study design, risk of bias, and indirectness in DTA","volume":"122","author":"Sch\u00fcnemann","year":"2020","journal-title":"J Clin Epidemiol"},{"key":"B53","doi-asserted-by":"publisher","DOI":"10.21037\/tlcr-20-370","article-title":"Multi-task learning and deep radiomics for IA risk","volume":"9","author":"Wang","year":"2020","journal-title":"Transl Lung Cancer Res"},{"key":"B54","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMoa2212083","article-title":"Lobar vs sublobar resection in stage ia nsclc","volume":"388","author":"Altorki","year":"2023","journal-title":"N Engl J Med"},{"key":"B55","doi-asserted-by":"publisher","first-page":"99","DOI":"10.1007\/s13139-017-0512-7","article-title":"Radiomics as quantitative imaging biomarker","volume":"52","author":"Park","year":"2018","journal-title":"Nucl Med Mol Imaging"},{"key":"B56","doi-asserted-by":"publisher","first-page":"97","DOI":"10.17691\/stm2021.13.2.11","article-title":"Radiomics and digital texture analysis in oncology: review","volume":"13","author":"Litvin","year":"2021","journal-title":"Sovrem Tekhnologii Med"},{"key":"B57","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1186\/s13244-020-00887-2","article-title":"Radiomics in medical imaging: guide and reflection","volume":"11","author":"van Timmeren","year":"2020","journal-title":"Insights Imaging"},{"key":"B58","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1080\/23808993.2019.1585805","article-title":"Dl and radiomics in precision medicine","volume":"4","author":"Parekh","year":"2019","journal-title":"Expert Rev Precis Med Drug Dev"},{"key":"B59","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.221843","article-title":"Combining dl and handcrafted radiomics in contrast-enhanced mammograms","volume":"307","author":"Beuque","year":"2023","journal-title":"Radiology"},{"key":"B60","doi-asserted-by":"publisher","first-page":"065008","DOI":"10.1088\/2057-1976\/ac8e6f","article-title":"Radiomics software comparison: ibsi-compliance and concordance","volume":"8","author":"Paquier","year":"2022","journal-title":"BioMed Phys Eng Expr"},{"key":"B61","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics13172760","article-title":"Ai integration in medical imaging: redefining radiology","volume":"13","author":"Najjar","year":"2023","journal-title":"Diagn (Basel)"},{"key":"B62","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2023-078378","article-title":"Tripod+ai statement: updated guidance for reporting clinical prediction models that use regression or machine learning","volume":"385","author":"Collins","year":"2024","journal-title":"BMJ"}],"container-title":["Frontiers in Oncology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2026.1752652\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T04:24:38Z","timestamp":1784867078000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2026.1752652\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,24]]},"references-count":62,"alternative-id":["10.3389\/fonc.2026.1752652"],"URL":"https:\/\/doi.org\/10.3389\/fonc.2026.1752652","relation":{},"ISSN":["2234-943X"],"issn-type":[{"value":"2234-943X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,24]]},"article-number":"1752652"}}