{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,8]],"date-time":"2026-09-08T05:10:33Z","timestamp":1788844233571,"version":"build-2803163510"},"reference-count":50,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,8]],"date-time":"2026-09-08T00:00:00Z","timestamp":1788825600000},"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. Oncol."],"abstract":"<jats:p>Computer-aided diagnosis (CADx) for gastrointestinal (GI) endoscopy increasingly depends on deep models trained end-to-end on raw images. However, raw images are often unavailable in legacy clinical systems or privacy-sensitive settings. This work presents a methodological study on multi-class GI disease classification from pre-extracted handcrafted descriptors on the public Kvasir benchmark, rather than a clinically validated deployment system. We propose a Descriptor-Attention Mixture-of-Experts (DA-MoE) model tailored to this descriptor-only scenario. DA-MoE first projects heterogeneous descriptors (JCD, Tamura, ColorLayout, EdgeHistogram, AutoColorCorrelogram, and PHOG) into a shared token space and applies transformer-style self-attention for descriptor-level fusion. A descriptor-aware mixture-of-experts classifier then performs sample-adaptive expert routing on the fused representation. Under a unified evaluation protocol using accuracy (ACC), macro F1, and Matthews correlation coefficient (MCC), with five-fold stratified cross-validation and repeated random splits for robustness assessment, DA-MoE achieves 77.4% accuracy and an MCC of 0.75 on the held-out test split, outperforming strong feature-based baselines including a residual multi-layer perceptron (Res-MLP). Ablation studies, hyperparameter analysis, MoE routing interpretability, full per-class metrics, and clustering analysis further show that DA-MoE produces more compact and better-separated representations in descriptor embedding space. These findings support DA-MoE as a practical component of feature-based GI CADx pipelines under data-sharing constraints, but prospective validation on raw endoscopic data remains necessary.<\/jats:p>","DOI":"10.3389\/fonc.2026.1911603","type":"journal-article","created":{"date-parts":[[2026,9,8]],"date-time":"2026-09-08T04:26:15Z","timestamp":1788841575000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["DA-MoE: descriptor-attention mixture-of-experts for multi-class gastrointestinal disease classification"],"prefix":"10.3389","volume":"16","author":[{"given":"Yiliu","family":"Xu","sequence":"first","affiliation":[{"name":"Guangxi University of Chinese Medicine","place":["Nanning, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lingling","family":"Liu","sequence":"additional","affiliation":[{"name":"Hubei University of Chinese Medicine","place":["Wuhan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Meiwen","family":"Tang","sequence":"additional","affiliation":[{"name":"Guangxi University of Chinese Medicine","place":["Nanning, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,8]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1069","DOI":"10.1053\/j.gastro.2018.06.037","article-title":"Deep learning localizes and identifies polyps in real time with 96% accuracy in screening colonoscopy","volume":"155","author":"Urban","year":"2018","journal-title":"Gastroenterology"},{"key":"B2","doi-asserted-by":"crossref","first-page":"599","DOI":"10.1136\/gutjnl-2017-314547","article-title":"Realtime differentiation of adenomatous and hyperplastic diminutive colorectal polyps during colonoscopy using a novel computer-aided diagnostic system: A prospective study","volume":"68","author":"Byrne","year":"2019","journal-title":"Gut"},{"key":"B3","doi-asserted-by":"crossref","DOI":"10.1145\/3193289","article-title":"Kvasir: A multi-class image dataset for computer aided gastrointestinal disease detection","author":"Pogorelov","year":"2017"},{"key":"B4","doi-asserted-by":"crossref","first-page":"283","DOI":"10.1038\/s41597-020-00622-y","article-title":"Hyperkvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy","volume":"7","author":"Borgli","year":"2020","journal-title":"Sci Data"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1109\/76.927424","article-title":"Color and texture descriptors","volume":"11","author":"Manjunath","year":"2001","journal-title":"IEEE Trans Circuits Syst Video Technol"},{"key":"B6","doi-asserted-by":"crossref","DOI":"10.1109\/TMI.2016.2528162","article-title":"Deep convolutional neural networks for computer-aided detection: Cnn architectures, dataset characteristics and transfer learning","volume":"35","author":"Shin","year":"2016","journal-title":"IEEE Trans Med Imaging"},{"key":"B7","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.media.2017.07.005","article-title":"A survey on deep learning in medical image analysis","volume":"42","author":"Litjens","year":"2017","journal-title":"Med Image Anal"},{"key":"B8","first-page":"1","article-title":"An image is worth 16x16 words: Transformers for image recognition at scale","author":"Dosovitskiy","year":"2021"},{"key":"B9","first-page":"1","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","author":"Shazeer","year":"2017"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1007\/BF00994018","article-title":"Support-vector networks","volume":"20","author":"Cortes","year":"1995","journal-title":"Mach Learn"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.1967.1053964","article-title":"Nearest neighbor pattern classification","volume":"13","author":"Cover","year":"1967","journal-title":"IEEE Trans Inf Theory"},{"key":"B12","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach Learn"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1013203451","article-title":"Greedy function approximation: A gradient boosting machine","volume":"29","author":"Friedman","year":"2001","journal-title":"Ann Stat"},{"key":"B14","first-page":"23","article-title":"A decision-theoretic generalization of on-line learning and an application to boosting","author":"Freund","year":"1997"},{"key":"B15","volume-title":"Pattern Recognition and Machine Learning","author":"Bishop","year":"2006"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1126\/science.1127647","article-title":"Reducing the dimensionality of data with neural networks","volume":"313","author":"Hinton","year":"2006","journal-title":"Science"},{"key":"B17","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1023\/A:1007379606734","article-title":"Multitask learning","volume":"28","author":"Caruana","year":"1997","journal-title":"Mach Learn"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1055\/a-1010-5705","article-title":"A technical review of artificial intelligence as applied to gastrointestinal endoscopy: clarifying the terminology","volume":"7","author":"Probst","year":"2019","journal-title":"Endoscopy Int Open"},{"key":"B19","article-title":"Artificial intelligence for endoscopy in gastroenterology: A review","volume":"7","author":"R\u00fcdiger","year":"2019","journal-title":"United Eur Gastroenterol J"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2020.101923","article-title":"Deep learning to find colorectal polyps in colonoscopy: A systematic literature review","volume":"108","author":"Machi-Fl\u00f3rez","year":"2020","journal-title":"Artif Intell Med"},{"key":"B21","first-page":"5998","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"B22","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV48922.2021.00986","article-title":"Swin transformer: Hierarchical vision transformer using shifted windows","author":"Liu","year":"2021"},{"key":"B23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.48550\/arXiv.2101.03961","article-title":"Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity","volume":"23","author":"Fedus","year":"2022","journal-title":"J Mach Learn Res"},{"key":"B24","article-title":"Transunet: Transformers make strong encoders for medical image segmentation","volume-title":"Arxiv Preprint Arxiv:2102.04306","author":"Chen","year":"2021"},{"key":"B25","article-title":"Some methods for classification and analysis of multivariate observations","author":"MacQueen","year":"1967"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1093\/oso\/9780199546329.003.0006","article-title":"Visualizing data using t-sne","volume":"9","author":"van der Maaten","year":"2008","journal-title":"J Mach Learn Res"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2008.239","article-title":"Learning from imbalanced data","volume":"21","author":"He","year":"2009","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.1186\/s12864-019-6413-7","article-title":"The advantages of the matthews correlation coefficient (mcc) over f1 score and accuracy in binary classification evaluation","volume":"21","author":"Chicco","year":"2020","journal-title":"BMC Genomics"},{"key":"B29","volume-title":"Endoscopic Image Super-Resolution Algorithm Using Edge and Disparity Awareness","author":"Hayat","year":"2023"},{"key":"B30","first-page":"1","article-title":"HAMSRNethybrid attention multiscale super-resolution network for endoscopic images","author":"Hayat","year":"2025"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2021.09.015","article-title":"An adaptive federated learning scheme with differential privacy preserving","volume":"127","author":"Wu","year":"2022","journal-title":"Future Gener Comput Syst"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.2196\/83743","article-title":"A bilayer feature fusion framework for pan-cancer survival prediction based on multihead attention and adaptive differential privacy: Model development and validation study","volume":"14","author":"Chen","year":"2026","journal-title":"JMIR Med Inf"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.1109\/TCBB.2022.3155774","article-title":"IPP: An intelligent privacy-preserving scheme for detecting interactions in genome association studies","volume":"20","author":"Wang","year":"2023","journal-title":"IEEE\/ACM Trans Comput Biol Bioinf"},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.1007\/s10586-017-1691-9","article-title":"A differential privacy DNA motif finding method based on closed frequent patterns","volume":"22","author":"Wu","year":"2019","journal-title":"Cluster Comput"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2026.115578","article-title":"EFSRNet: Multi-scale exposure normalization and dual-branch aggregation for overexposed face super-resolution","volume":"339","author":"Yang","year":"2026","journal-title":"Knowledge-Based Syst"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1016\/j.ymssp.2026.114123","article-title":"Unsupervised deep learning framework for damage identification under ambient excitations: trait of damage localization and demonstrative applications","volume":"250","author":"Huang","year":"2026","journal-title":"Mech Syst Signal Process"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.1016\/j.aej.2026.05.026","article-title":"An intelligent recognition method for narcotic and psychotropic drugs in complex scenes based on multi-stage image enhancement and optimized YOLOv8","volume":"147","author":"Liu","year":"2026","journal-title":"Alexandria Eng J"},{"key":"B38","doi-asserted-by":"publisher","DOI":"10.1002\/advs.75815","article-title":"Interpretable machine learning framework for Nb\u2013Si based alloy design with enhanced fracture toughness","author":"Chen","year":"2026","journal-title":"Adv Sci"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2025.110654","article-title":"Integration of neural networks in brain\u2013 computer interface applications: Research frontiers and trend analysis based on Python","volume":"151","author":"Chen","year":"2025","journal-title":"Eng Appl Artif Intell"},{"key":"B40","doi-asserted-by":"crossref","DOI":"10.1109\/SCSET55041.2022.00060","article-title":"A new forest fire risk rating forecast model based on XGBoost","author":"Yan","year":"2022"},{"key":"B41","doi-asserted-by":"crossref","DOI":"10.1109\/MLBDBI54094.2021.00015","article-title":"Forest fire risk forecast method with pseudo label based on semi-supervised learning","author":"Wei","year":"2021"},{"key":"B42","doi-asserted-by":"publisher","DOI":"10.3390\/electronics13040785","article-title":"A circularly polarized complementary antenna with substrate integrated coaxial line feed for X-band applications","volume":"13","author":"Ji","year":"2024","journal-title":"Electronics"},{"key":"B43","doi-asserted-by":"crossref","DOI":"10.1109\/APUSNCURSINRSM.2017.8073094","article-title":"Array-antenna decoupling surfaces for quasi-Yagi antenna arrays","author":"Wei","year":"2017"},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1002\/mop.33306","article-title":"Decoupling and bandwidth equalization for MIMO antennas by using novel network design","volume":"64","author":"Wu","year":"2022","journal-title":"Microwave Opt Technol Lett"},{"key":"B45","doi-asserted-by":"publisher","DOI":"10.1109\/TIE.2026.3695241","article-title":"Multi-agent dynamic surface consensus control of dual-motor steer-by-wire system under communication faults","author":"Xu","year":"2026","journal-title":"IEEE Trans Ind Electron"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.1016\/j.cie.2023.109509","article-title":"Sustainable interior design: A new approach to intelligent design and automated manufacturing based on Grasshopper","volume":"183","author":"Chen","year":"2023","journal-title":"Comput Ind Eng"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.1177\/00368504251360892","article-title":"Exercise\u2019s protective role in chronic obstructive pulmonary disease via modulation of M1 macrophage phenotype through the miR-124-3p\/ERN1 axis","volume":"108","author":"Zeng","year":"2025","journal-title":"Sci Prog"},{"key":"B48","doi-asserted-by":"publisher","first-page":"156","DOI":"10.1186\/s12931-023-02461-y","article-title":"Regulation of whole-transcriptome sequencing expression in COPD after personalized precise exercise training: A pilot study","volume":"24","author":"Liu","year":"2023","journal-title":"Respir Res"},{"key":"B49","doi-asserted-by":"crossref","DOI":"10.2139\/ssrn.6449738","article-title":"Beyond external constraints: The missing dimension of AI governance","volume-title":"Ssrn","author":"Chen","year":"2026"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.1109\/TLT.2024.3390593","article-title":"A tutorial-generating method for autonomous online learning","volume":"17","author":"Wu","year":"2024","journal-title":"IEEE Trans Learn Technol"}],"container-title":["Frontiers in Oncology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2026.1911603\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,8]],"date-time":"2026-09-08T04:26:19Z","timestamp":1788841579000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2026.1911603\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,8]]},"references-count":50,"alternative-id":["10.3389\/fonc.2026.1911603"],"URL":"https:\/\/doi.org\/10.3389\/fonc.2026.1911603","relation":{},"ISSN":["2234-943X"],"issn-type":[{"value":"2234-943X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,8]]},"article-number":"1911603"}}