{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T05:20:48Z","timestamp":1787721648296,"version":"build-2784847793"},"reference-count":45,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T00:00:00Z","timestamp":1787702400000},"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. Med."],"abstract":"<jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>Discriminative information in ultrasound images is often manifested as local boundary changes, internal echo heterogeneity and differences in fine-scale structure. However, it is susceptible to interference from speckle and uncorrelated background responses. This paper proposed the Statistical Echo Evidence and Detail Network (SEED-Net) for ultrasound lesion classification across multiple organ-specific datasets.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>SEED-Net comprises Foreground Tissue Gate, Guided Detail Unit, Channel Recalibration and Statistical Echo Evidence Head, which are used, respectively, for modulating the response of foreground tissue, modelling residuals of local structure and detail, recalibrating channel features, and fusing statistical evidence from medium and deep layers. Each prior branch uses a near-zero gate so that its initial contribution to the residual backbone is approximately zero. The model was trained on three ultrasound datasets, TN5000, SMC-LUD and PU2756, for the classification of ultrasound images of thyroid nodules, liver lesions and lung tumours, respectively. All experiments utilised whole-image input, random initialisation and independent training with five random seeds, and the model was compared with convolutional networks, visual transformers, hybrid architectures and visual state-space models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>SEED-Net achieved accuracy values of 0.8702\u202f\u00b1\u202f0.0055, 0.9851\u202f\u00b1\u202f0.0075, and 0.6054\u202f\u00b1\u202f0.0086 on TN5000, SMC-LUD, and PU2756, respectively, with corresponding Macro F1-scores of 0.8384\u202f\u00b1\u202f0.0048, 0.9850\u202f\u00b1\u202f0.0075, and 0.5785\u202f\u00b1\u202f0.0229. It achieved the highest or joint-highest Macro F1-score among the evaluated learning-based models on all three datasets, although its absolute performance on PU2756 remained limited. Ablation experiments demonstrate that the modelling of local details in the backbone and the statistical evidence classification head play complementary roles. Grad-CAM visualisations showed more compact activation patterns around lesion-related regions in the displayed examples, providing qualitative observations of the model responses.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>By integrating local detail modelling with multi-layer statistical evidence, SEED-Net achieves a favourable balance between performance and computational complexity in ultrasound classification tasks across different organs and varying levels of difficulty.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fmed.2026.1942733","type":"journal-article","created":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T04:49:47Z","timestamp":1787719787000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["SEED-Net: statistical echo evidence and detail modeling for ultrasound lesion classification across multiple organ-specific datasets"],"prefix":"10.3389","volume":"13","author":[{"given":"Qiong","family":"Liu","sequence":"first","affiliation":[{"name":"School of Medical Imaging, Jiangsu Medical College","place":["Yancheng, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chaofan","family":"Li","sequence":"additional","affiliation":[{"name":"Affiliated Hospital 6 of Nantong University, Yancheng Third People's Hospital","place":["Yancheng, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,26]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"1398393","DOI":"10.3389\/fphy.2024.1398393","article-title":"A survey on deep learning in medical ultrasound imaging","volume":"12","author":"Song","year":"2024","journal-title":"Front Phys"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1186\/s42492-023-00147-2","article-title":"Application and prospects of AI-based radiomics in ultrasound diagnosis","volume":"6","author":"Zhang","year":"2023","journal-title":"Vis Comput Ind Biomed Art"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"677","DOI":"10.1016\/j.ultrasmedbio.2022.11.003","article-title":"Ultrasound signal processing: from models to deep learning","volume":"49","author":"Luijten","year":"2023","journal-title":"Ultrasound Med Biol"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"54310","DOI":"10.1109\/ACCESS.2021.3071301","article-title":"Deep learning in medical ultrasound image analysis: a review","volume":"9","author":"Wang","year":"2021","journal-title":"IEEE Access"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"2361","DOI":"10.3390\/s25082361","article-title":"Deep learning-based medical ultrasound image and video segmentation methods: overview, frontiers, and challenges","volume":"25","author":"Xiao","year":"2025","journal-title":"Sensors (Basel)"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"542","DOI":"10.3390\/diagnostics14050542","article-title":"Ultrasound image analysis with vision transformers","volume":"14","author":"Vafaeezadeh","year":"2024","journal-title":"Diagnostics (Basel)"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"103031","DOI":"10.52202\/079017-3273","article-title":"VMamba: visual state space model","volume":"37","author":"Liu","year":"2024","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"7714","DOI":"10.1088\/1361-6560\/aa82ec","article-title":"A deep learning framework for supporting the classification of breast lesions in ultrasound images","volume":"62","author":"Han","year":"2017","journal-title":"Phys Med Biol"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"746","DOI":"10.1002\/mp.13361","article-title":"Breast mass classification in sonography with transfer learning using a deep convolutional neural network and color conversion","volume":"46","author":"Byra","year":"2019","journal-title":"Med Phys"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"106300","DOI":"10.1016\/j.ultras.2020.106300","article-title":"A generic deep learning framework to classify thyroid and breast lesions in ultrasound images","volume":"110","author":"Zhu","year":"2021","journal-title":"Ultrasonics"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"110335","DOI":"10.1016\/j.ejrad.2022.110335","article-title":"Deep convolutional neural network for classification of thyroid nodules on ultrasound: comparison of the diagnostic performance with that of radiologists","volume":"152","author":"Kim","year":"2022","journal-title":"Eur J Radiol"},{"key":"ref12","first-page":"1","volume-title":"2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC)","author":"Ferreira","year":"2023"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"1419","DOI":"10.3390\/bioengineering10121419","article-title":"Multi-instance classification of breast tumor ultrasound images using convolutional neural networks and transfer learning","volume":"10","author":"Ciobotaru","year":"2023","journal-title":"Bioengineering (Basel)"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"103924","DOI":"10.1016\/j.bspc.2022.103924","article-title":"Intelligent diagnosis algorithm for thyroid nodules based on deep learning and statistical features","volume":"78","author":"Yu","year":"2022","journal-title":"Biomed Signal Process Control"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"230","DOI":"10.1186\/s12880-024-01404-3","article-title":"Revolutionizing breast ultrasound diagnostics with EfficientNet-B7 and explainable AI","volume":"24","author":"Latha","year":"2024","journal-title":"BMC Med Imaging"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"1849727","DOI":"10.3389\/fmed.2026.1849727","article-title":"An interpretable deep concatenated architecture for osteoporosis detection using enhanced knee radiographs","volume":"13","author":"Kaur","year":"2026","journal-title":"Front Med"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"1851416","DOI":"10.3389\/fncom.2026.1851416","article-title":"AI-driven neuroanalytic modeling for mental health: multichannel CNN-based autism spectrum disorder detection via facial pattern analysis","volume":"20","author":"Kaur","year":"2026","journal-title":"Front Comput Neurosci"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"544","DOI":"10.1186\/s12885-025-13917-3","article-title":"Thyroid nodule classification in ultrasound imaging using deep transfer learning","volume":"25","author":"Xu","year":"2025","journal-title":"BMC Cancer"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"104","DOI":"10.1016\/j.acra.2023.08.006","article-title":"Distilling knowledge from an ensemble of vision transformers for improved classification of breast ultrasound","volume":"31","author":"Zhou","year":"2024","journal-title":"Acad Radiol"},{"key":"ref20","doi-asserted-by":"crossref","first-page":"246","DOI":"10.1109\/SITIS57111.2022.00027","volume-title":"2022 16th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS)","author":"Saad","year":"2022"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"125009","DOI":"10.1016\/j.eswa.2024.125009","article-title":"Swin transformer-based fork architecture for automated breast tumor classification","volume":"256","author":"\u00dczen","year":"2024","journal-title":"Expert Syst Appl"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"525","DOI":"10.1016\/j.ultrasmedbio.2024.11.014","article-title":"Interactively fusing global and local features for benign and malignant classification of breast ultrasound images","volume":"51","author":"Wang","year":"2025","journal-title":"Ultrasound Med Biol"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"104101","DOI":"10.1016\/j.medengphy.2024.104101","article-title":"SRT: Swin-residual transformer for benign and malignant nodules classification in thyroid ultrasound images","volume":"124","author":"Huang","year":"2024","journal-title":"Med Eng Phys"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"1679310","DOI":"10.3389\/frai.2025.1679310","article-title":"Early-fusion hybrid CNN-transformer models for multiclass ovarian tumor ultrasound classification","volume":"8","author":"Garcia-Atutxa","year":"2025","journal-title":"Front Artif Intell"},{"key":"ref25","first-page":"13001","volume-title":"2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","author":"Zhao","year":"2024"},{"key":"ref26","first-page":"148","volume-title":"Deep Breast Workshop on AI and Imaging for Diagnostic and Treatment Challenges in Breast Care","author":"Nasiri-Sarvi","year":"2024"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"3633","DOI":"10.3390\/electronics14183633","article-title":"UltrascanNet: a mamba-inspired hybrid backbone for breast ultrasound classification","volume":"14","author":"Laicu-Hausberger","year":"2025","journal-title":"Electronics (Basel)"},{"key":"ref28","doi-asserted-by":"publisher","first-page":"1686248","DOI":"10.3389\/fendo.2025.1686248","article-title":"Thyroid intelligent diagnosis based on THMSNet","volume":"16","author":"Rao","year":"2025","journal-title":"Front Endocrinol (Lausanne)"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"101750","DOI":"10.1016\/j.imu.2026.101750","article-title":"Breast cancer classification in ultrasound imaging using mamba: an efficient state-space model for computer-aided diagnosis","volume":"62","author":"Aniq","year":"2026","journal-title":"Inform Med Unlocked"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"1201","DOI":"10.1007\/s11760-025-04756-7","article-title":"MambaCA-net: a hybrid dual-stream architecture integrating temporal and spatial features for breast ultrasound analysis","volume":"19","author":"He","year":"2025","journal-title":"Signal Image Video Process"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"8733","DOI":"10.1007\/s00330-021-07850-9","article-title":"Joint segmentation and classification of hepatic lesions in ultrasound images using deep learning","volume":"31","author":"Ryu","year":"2021","journal-title":"Eur Radiol"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"108427","DOI":"10.1016\/j.patcog.2021.108427","article-title":"Segmentation information with attention integration for classification of breast tumor in ultrasound image","volume":"124","author":"Luo","year":"2022","journal-title":"Pattern Recogn"},{"key":"ref33","first-page":"344","volume-title":"International Conference on Medical Image Computing and Computer-Assisted Intervention","author":"Shareef","year":"2023"},{"key":"ref34","doi-asserted-by":"publisher","first-page":"205021","DOI":"10.1088\/1361-6560\/acfdf0","article-title":"Automatic ultrasound diagnosis of thyroid nodules: a combination of deep learning and KWAK TI-RADS","volume":"68","author":"Zhang","year":"2023","journal-title":"Phys Med Biol"},{"key":"ref35","doi-asserted-by":"publisher","first-page":"105522","DOI":"10.1016\/j.ijmedinf.2024.105522","article-title":"Segmentation-based BI-RADS ensemble classification of breast tumours in ultrasound images","volume":"189","author":"Bobowicz","year":"2024","journal-title":"Int J Med Inform"},{"key":"ref36","doi-asserted-by":"publisher","first-page":"107346","DOI":"10.1016\/j.cmpb.2023.107346","article-title":"MsGoF: breast lesion classification on ultrasound images by multi-scale gradational-order fusion framework","volume":"230","author":"Zhong","year":"2023","journal-title":"Comput Methods Prog Biomed"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"144","DOI":"10.1002\/mp.15341","article-title":"FMRNet: a fused network of multiple tumoral regions for breast tumor classification with ultrasound images","volume":"49","author":"Cui","year":"2022","journal-title":"Med Phys"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"38732","DOI":"10.1109\/ACCESS.2024.3374380","article-title":"Incorporating tumor edge information for fine-grained BI-RADS classification of breast ultrasound images","volume":"12","author":"Xu","year":"2024","journal-title":"IEEE Access"},{"key":"ref39","doi-asserted-by":"publisher","first-page":"2819","DOI":"10.1016\/j.ultrasmedbio.2020.06.015","article-title":"Attention-enriched deep learning model for breast tumor segmentation in ultrasound images","volume":"46","author":"Vakanski","year":"2020","journal-title":"Ultrasound Med Biol"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"103453","DOI":"10.1016\/j.media.2025.103453","article-title":"Automated ultrasonography of hepatocellular carcinoma using discrete wavelet transform-based deep-learning neural network","volume":"101","author":"Rhyou","year":"2025","journal-title":"Med Image Anal"},{"key":"ref41","doi-asserted-by":"publisher","first-page":"14300","DOI":"10.1038\/s41598-025-99009-5","article-title":"Variational mode-directed deep learning framework for breast lesion classification using ultrasound imaging","volume":"15","author":"Saini","year":"2025","journal-title":"Sci Rep"},{"key":"ref42","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TPAMI.2020.2974833","article-title":"Deep CNNs meet global covariance pooling: better representation and generalization","volume":"43","author":"Wang","year":"2021","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"ref43","doi-asserted-by":"publisher","first-page":"1437","DOI":"10.1038\/s41597-025-05757-4","article-title":"TN5000: an ultrasound image dataset for thyroid nodule detection and classification","volume":"12","author":"Zhang","year":"2025","journal-title":"Sci Data"},{"key":"ref44","doi-asserted-by":"publisher","first-page":"649","DOI":"10.1038\/s41597-026-07023-7","article-title":"SMC-LUD: large-scale B-mode liver ultrasound dataset for hepatocellular carcinoma and hemangioma classification","volume":"13","author":"Tak","year":"2026","journal-title":"Sci Data"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-026-07715-0","article-title":"A pulmonary ultrasound dataset for tumor segmentation and classification","author":"Jiang","year":"2026","journal-title":"Sci Data"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1942733\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T04:50:03Z","timestamp":1787719803000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1942733\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,26]]},"references-count":45,"alternative-id":["10.3389\/fmed.2026.1942733"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1942733","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,26]]},"article-number":"1942733"}}