{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T13:30:57Z","timestamp":1787923857023,"version":"build-2784847793"},"reference-count":63,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T00:00:00Z","timestamp":1787875200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Lung cancer results in roughly 1.8 million fatalities annually worldwide, with non-small cell lung cancer (NSCLC) comprising the majority of cases. Despite advancements in treatment, survival stratification remains challenging due to intratumoral heterogeneity inadequately captured by conventional descriptors. Standard radiomic and deep learning techniques regard imaging features as independent quantities, overlooking structured interactions between tumor characteristics. We evaluate whether structured proxy features can enhance multimodal NSCLC survival prediction by augmenting pretreatment computed tomography (CT) representations, radiomics, and clinical variables with six simulation-derived features designed to capture interactions between heterogeneity and morphology.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A radiomic-parameterized cellular automaton generates growth-rate and necrosis-ratio proxy features from baseline CT by using entropy and sphericity to compute low-dimensional proxy parameters. The imaging backbone is a Transformer-based Masked Autoencoder (TMAE), which was chosen after a systematic evaluation with alternative encoders within the same pipeline and provides attention-based visualizations that highlight tumor regions receiving higher model attention.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      On the public Lung1 cohort (\n                      <jats:italic>n<\/jats:italic>\n                      = 390), the primary four-modality fusion attained a C-index of 0.641 (iAUC 0.731, log-rank\n                      <jats:italic>p<\/jats:italic>\n                      &amp;lt; 0.001). The primary result compares favorably with prior multimodal results on Lung1 (C-index 0.631; iAUC 0.592 [15]) under a comparable evaluation protocol, while a separate exploratory coefficient-optimization analysis achieved a best observed C-index of 0.662 (iAUC 0.748).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These results indicate that, in addition to conventional radiomic, deep, and clinical representations within the Lung1 benchmark, simulation-derived proxy features may provide complementary predictive information within this fixed Lung1 benchmark. By integrating structured tumor-dynamics-inspired descriptors with modern volumetric CT representations, the framework provides a practical approach for retrospective relative risk ranking from routinely acquired pretreatment imaging and establishes a foundation for future repeated-split, external-cohort, and calibration studies.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2026.1865458","type":"journal-article","created":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T12:51:11Z","timestamp":1787921471000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Structured proxy features for multimodal NSCLC survival prediction from pretreatment CT"],"prefix":"10.3389","volume":"9","author":[{"given":"Huu Phong","family":"Nguyen","sequence":"first","affiliation":[{"name":"Department of Biomedical Informatics and Data Science, School of Medicine, The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Delower","family":"Hossain","sequence":"additional","affiliation":[{"name":"Department of Computer Science, The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ehsan","family":"Saghapour","sequence":"additional","affiliation":[{"name":"System Pharmacology and AI Research Center (SPARC), The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhandos","family":"Sembay","sequence":"additional","affiliation":[{"name":"System Pharmacology and AI Research Center (SPARC), The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jake Y.","family":"Chen","sequence":"additional","affiliation":[{"name":"Department of Biomedical Informatics and Data Science, School of Medicine, The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]},{"name":"Department of Computer Science, The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]},{"name":"System Pharmacology and AI Research Center (SPARC), The University of Alabama at Birmingham (UAB)","place":["Birmingham, AL, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,28]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"4006","DOI":"10.1038\/ncomms5644","article-title":"Decoding tumor phenotype by noninvasive imaging using a quantitative radiomics approach","volume":"5","author":"Aerts","year":"2014","journal-title":"Nat. Commun"},{"key":"B2","first-page":"6836","article-title":"\u201cVivit: a video vision transformer,\u201d","volume-title":"ICCV","author":"Arnab","year":"2021"},{"key":"B3","first-page":"15619","article-title":"\u201cSelf-supervised learning from images with a joint-embedding predictive architecture,\u201d","volume-title":"CVPR","author":"Assran","year":"2023"},{"key":"B4","doi-asserted-by":"publisher","first-page":"8730","DOI":"10.1038\/s41598-019-45117-y","article-title":"Prognostic impact of longitudinal monitoring of radiomic features in patients with advanced non-small cell lung cancer","volume":"9","author":"Bak","year":"2019","journal-title":"Sci. Rep"},{"key":"B5","doi-asserted-by":"publisher","first-page":"229","DOI":"10.3322\/caac.21834","article-title":"Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries","volume":"74","author":"Bray","year":"2024","journal-title":"CA Cancer J. Clin"},{"key":"B6","first-page":"6299","article-title":"\u201cQuo vadis, action recognition?\u201d","volume-title":"CVPR","author":"Carreira","year":"2017"},{"key":"B7","first-page":"584","article-title":"\u201cDino2-DR,\u201d","volume-title":"ICCV","author":"Cascone","year":"2025"},{"key":"B8","first-page":"1597","article-title":"\u201cA simple framework for contrastive learning of visual representations,\u201d","volume-title":"ICML","author":"Chen","year":"2020"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1045","DOI":"10.1007\/s10278-013-9622-7","article-title":"the cancer imaging archive (TCIA)","volume":"26","author":"Clark","year":"2013","journal-title":"J. Digital Imag"},{"key":"B10","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1111\/j.2517-6161.1972.tb00899.x","article-title":"Regression models and life-tables. J","volume":"34","author":"Cox","year":"1972","journal-title":"Royal Stat. Soc"},{"key":"B11","doi-asserted-by":"publisher","first-page":"73","DOI":"10.1016\/j.radonc.2019.08.003","article-title":"Is tumor sphericity an important prognostic factor in patients with lung cancer?","volume":"143","author":"Davey","year":"2020","journal-title":"Radiotherapy Oncol"},{"key":"B12","doi-asserted-by":"publisher","first-page":"112388","DOI":"10.1016\/j.jtbi.2026.112388","article-title":"Modeling tumor progression in heterogeneous microenvironments: a cellular automata approach","volume":"622","author":"Deng","year":"2026","journal-title":"J. Theor. Biol"},{"key":"B13","first-page":"2625","article-title":"\u201cLong-term recurrent convolutional networks for visual recognition and description,\u201d","volume-title":"CVPR","author":"Donahue","year":"2015"},{"key":"B14","doi-asserted-by":"publisher","first-page":"11929","DOI":"10.48550\/arXiv.2010.11929","article-title":"An image is worth 16 \u00d7 16 words: transformers for image recognition at scale","volume":"2010","author":"Dosovitskiy","year":"2020","journal-title":"arXiv"},{"key":"B15","doi-asserted-by":"publisher","first-page":"108496","DOI":"10.1016\/j.cmpb.2024.108496","article-title":"Integrating radiomic and 3D autoencoder-based features for Non-Small Cell Lung Cancer survival analysis","volume":"258","author":"Ferretti","year":"2025","journal-title":"Comp. Methods Programs Biomed"},{"key":"B16","volume-title":"Creation of MLPA: A Multi-Level Digital Twin Framework for Personalized Cancer Simulation and Treatment Optimization","author":"Gu","year":"2024"},{"key":"B17","first-page":"1188","article-title":"\u201cRadiomic feature stability analysis based on probabilistic segmentations,\u201d","volume-title":"ISBI","author":"Haarburger","year":"2020"},{"key":"B18","first-page":"1197","article-title":"\u201cImage-based survival prediction for lung cancer patients using CNNS,\u201d","volume-title":"ISBI","author":"Haarburger","year":"2019"},{"key":"B19","first-page":"6546","article-title":"\u201cCan spatiotemporal 3D cnns retrace the history of 2D cnns and imagenet?\u201d","volume-title":"CVPR","author":"Hara","year":"2018"},{"key":"B20","first-page":"272","article-title":"\u201cSwin unetr: swin transformers for semantic segmentation of brain tumors in MRI images,\u201d","volume-title":"MICCAI","author":"Hatamizadeh","year":"2021"},{"key":"B21","first-page":"574","article-title":"\u201cUnetr: transformers for 3d medical image segmentation,\u201d","volume-title":"WACV","author":"Hatamizadeh","year":"2022"},{"key":"B22","first-page":"9729","article-title":"\u201cMomentum contrast for unsupervised visual representation learning,\u201d","volume-title":"CVPR","author":"He","year":"2020"},{"key":"B23","first-page":"16000","article-title":"\u201cMasked autoencoders are scalable vision learners,\u201d","volume-title":"CVPR","author":"He","year":"2022"},{"key":"B24","first-page":"770","article-title":"\u201cDeep residual learning for image recognition,\u201d","volume-title":"CVPR","author":"He","year":"2016"},{"key":"B25","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1111\/j.0006-341X.2000.00337.x","article-title":"Time-dependent ROC curves for censored survival data and a diagnostic marker","volume":"56","author":"Heagerty","year":"2000","journal-title":"Biometrics"},{"key":"B26","doi-asserted-by":"publisher","first-page":"500","DOI":"10.1038\/s41568-018-0016-5","article-title":"Artificial intelligence in radiology","volume":"18","author":"Hosny","year":"2018","journal-title":"Nat. Rev. Cancer"},{"key":"B27","first-page":"1","article-title":"\u201chERG-LTN: a new paradigm for hERG cardiotoxicity prediction using neuro-symbolic and generative AI,\u201d","volume-title":"Proc. 47th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. EMBC","author":"Hossain","year":"2025"},{"key":"B28","first-page":"18213","article-title":"A study on neuro-symbolic artificial intelligence: healthcare perspectives","volume":"2503","author":"Hossain","year":"2025","journal-title":"arXiv"},{"key":"B29","volume-title":"SEER Cancer Statistics Review, 1975\u20132017","author":"Howlader","year":"2020"},{"key":"B30","first-page":"4700","article-title":"\u201cDensely connected convolutional networks,\u201d","volume-title":"CVPR","author":"Huang","year":"2017"},{"key":"B31","doi-asserted-by":"publisher","first-page":"74","DOI":"10.1038\/s41746-023-00811-0","article-title":"Self-supervised learning for medical image classification: a systematic review and implementation guidelines","volume":"6","author":"Huang","year":"2023","journal-title":"NPJ Digital Med"},{"key":"B32","doi-asserted-by":"publisher","first-page":"947","DOI":"10.1148\/radiol.2016152234","article-title":"Radiomics signature: a potential biomarker for the prediction of disease-free survival in early-stage NSCLC","volume":"281","author":"Huang","year":"2016","journal-title":"Radiology"},{"key":"B33","doi-asserted-by":"publisher","first-page":"200","DOI":"10.3390\/cancers18020200","article-title":"Prognosis from Pixels","volume":"18","author":"Ibrahim","year":"2026","journal-title":"Cancers"},{"key":"B34","doi-asserted-by":"publisher","first-page":"2121","DOI":"10.1056\/NEJMoa1616288","article-title":"Tracking the evolution of non-small-cell lung cancer","volume":"376","author":"Jamal-Hanjani","year":"2017","journal-title":"N Engl. J. Med"},{"key":"B35","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1006\/jtbi.2000.2000","article-title":"Simulated brain tumor growth dynamics using a three-dimensional cellular automaton","volume":"203","author":"Kansal","year":"2000","journal-title":"J. Theor. Biol"},{"key":"B36","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1186\/s12874-018-0482-1","article-title":"DeepSurv","volume":"18","author":"Katzman","year":"2018","journal-title":"BMC Med. Res. Methodol"},{"key":"B37","doi-asserted-by":"publisher","first-page":"84","DOI":"10.1145\/3065386","article-title":"Imagenet classification with deep convolutional neural networks","volume":"25","author":"Krizhevsky","year":"2012","journal-title":"NeurIPS"},{"key":"B38","doi-asserted-by":"publisher","first-page":"749","DOI":"10.1038\/nrclinonc.2017.141","article-title":"Radiomics: the bridge between medical imaging and personalized medicine","volume":"14","author":"Lambin","year":"2017","journal-title":"Nat. Rev. Clin. Oncol"},{"key":"B39","doi-asserted-by":"publisher","first-page":"441","DOI":"10.1016\/j.ejca.2011.11.036","article-title":"Radiomics: extracting more information from medical images using advanced feature analysis","volume":"48","author":"Lambin","year":"2012","journal-title":"European J. Cancer"},{"key":"B40","first-page":"3202","article-title":"\u201cVideo swin transformer,\u201d","volume-title":"CVPR","author":"Liu","year":"2022"},{"key":"B41","unstructured":"2024"},{"key":"B42","doi-asserted-by":"publisher","first-page":"101681","DOI":"10.1109\/ACCESS.2025.3573264","article-title":"Kinematic adaptive frame recognition (KAFR): a novel framework for video segmentation via frame similarity and surgical tool tracking","volume":"13","author":"Nguyen","year":"2025","journal-title":"IEEE Access"},{"key":"B43","first-page":"166","article-title":"\u201cAdvanced capsule networks via context awareness,\u201d","volume-title":"ICANN","author":"Nguyen","year":""},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-31332-6_45","article-title":"\u201cAn improvement for capsule networks using depthwise separable convolution,\u201d","author":"Nguyen","year":"","journal-title":"IbPRIA"},{"key":"B45","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41598-023-39744-9","article-title":"Video action recognition collaborative learning with dynamics via PSO-ConvNet Transformer","volume":"13","author":"Nguyen","year":"2023","journal-title":"Sci. Rep"},{"key":"B46","doi-asserted-by":"publisher","first-page":"994","DOI":"10.1109\/TMI.2023.3326188","article-title":"Slim UNETR","volume":"43","author":"Pang","year":"2024","journal-title":"IEEE Trans. Med. Imag"},{"key":"B47","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1007\/s13755-025-00404-z","article-title":"Predicting lung cancer survival with attention-based CT slices combination","volume":"14","author":"Paolo","year":"2026","journal-title":"Health Inform. Sci. Sys"},{"key":"B48","doi-asserted-by":"publisher","first-page":"158","DOI":"10.1038\/s41698-024-00656-0","article-title":"Radiology and multi-scale data integration for precision oncology","volume":"8","author":"Paverd","year":"2024","journal-title":"NPJ Precision Oncol"},{"key":"B49","doi-asserted-by":"publisher","first-page":"1607","DOI":"10.1007\/s12530-024-09581-w","article-title":"Self-supervised learning for medical image analysis: a comprehensive review","volume":"15","author":"Rani","year":"2024","journal-title":"Evol. Sys."},{"key":"B50","doi-asserted-by":"publisher","first-page":"6923","DOI":"10.3390\/app14166923","article-title":"A genetic programming approach to radiomic-based feature construction for survival prediction in NSCLC","volume":"14","author":"Scalco","year":"2024","journal-title":"Appl. Sci"},{"key":"B51","first-page":"15","article-title":"Practical bayesian optimization of machine learning algorithms","volume":"25","author":"Snoek","year":"2012","journal-title":"NeurIPS"},{"key":"B52","doi-asserted-by":"publisher","first-page":"10078","DOI":"10.52202\/068431-0732","article-title":"Videomae: masked autoencoders are data-efficient learners for self-supervised video pre-training","volume":"35","author":"Tong","year":"2022","journal-title":"NeurIPS"},{"key":"B53","first-page":"4489","article-title":"\u201cLearning spatiotemporal features with 3d convolutional networks,\u201d","volume-title":"ICCV","author":"Tran","year":"2015"},{"key":"B54","doi-asserted-by":"publisher","first-page":"1155","DOI":"10.1109\/ACCESS.2017.2778011","article-title":"Action recognition in video sequences using deep bi-directional LSTM with CNN features","volume":"6","author":"Ullah","year":"2017","journal-title":"IEEE Access"},{"key":"B55","doi-asserted-by":"publisher","first-page":"1105","DOI":"10.1002\/sim.4154","article-title":"On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data","volume":"30","author":"Uno","year":"2011","journal-title":"Stat. Med"},{"key":"B56","doi-asserted-by":"publisher","first-page":"e104","DOI":"10.1158\/0008-5472.CAN-17-0339","article-title":"Computational radiomics system to decode the radiographic phenotype","volume":"77","author":"van Griethuysen","year":"2017","journal-title":"Cancer Res"},{"key":"B57","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1186\/s13244-020-00887-2","article-title":"Radiomics in medical imaging\u2014\u2018how-to' guide and critical reflection","volume":"11","author":"van Timmeren","year":"2020","journal-title":"Insights Imag"},{"key":"B58","first-page":"63","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"NeurIPS"},{"key":"B59","first-page":"14549","article-title":"\u201cVideomae v2: scaling video masked autoencoders with dual masking,\u201d","volume-title":"CVPR","author":"Wang","year":"2023"},{"key":"B60","doi-asserted-by":"publisher","first-page":"2","DOI":"10.1016\/j.radonc.2018.10.027","article-title":"Vulnerabilities of radiomic signature development: the need for safeguards","volume":"130","author":"Welch","year":"2019","journal-title":"Radiotherapy Oncol"},{"key":"B61","doi-asserted-by":"publisher","first-page":"4851","DOI":"10.48550\/arXiv.1712.04851","article-title":"Rethinking spatiotemporal feature learning for video understanding","volume":"1712","author":"Xie","year":"2017","journal-title":"arXiv"},{"key":"B62","doi-asserted-by":"publisher","first-page":"03605","DOI":"10.48550\/arXiv.2203.03605","article-title":"DINO: DETR with improved denoising anchor boxes","volume":"2203","author":"Zhang","year":"2022","journal-title":"arXiv"},{"key":"B63","first-page":"1","article-title":"\u201cSelf pre-training with masked autoencoders for medical image classification and segmentation,\u201d","volume-title":"ISBI","author":"Zhou","year":"2023"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1865458\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T12:51:22Z","timestamp":1787921482000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1865458\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,28]]},"references-count":63,"alternative-id":["10.3389\/frai.2026.1865458"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1865458","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,28]]},"article-number":"1865458"}}