{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,5]],"date-time":"2026-11-05T12:40:25Z","timestamp":1793882425126,"version":"4.7.0"},"reference-count":54,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T00:00:00Z","timestamp":1764892800000},"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:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Esophageal squamous cell carcinoma (ESCC) is difficult to detect early on white-light endoscopy (WLI) because lesions are subtle and artifacts (such as glare, bubbles, text, tools) mimic pathology.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This study benchmarked five object detectors including two You Only Look Once models (YOLOv5, YOLOv), Faster Region-based Convolutional Neural Networks (Faster R-CNN), Single Shot MultiBox Detector (SSD) and Real-time Detection Transformer (RT-DETR) on WLI dataset using harmonized training (from scratch, 150 epochs, identical hyperparameters) and two label configurations: a 4-label as major categories (SCC, Dysplasia, Bleeding, Inflammation) and an 11-label artifact. Evaluation used macro precision\/recall\/F1 at IoU 0.50 on a fixed 310-image test set.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Incorporating artifact classes improved overall macro metrics, with YOLOv5\/YOLOv8 providing the strongest performance in the 11-label scenarios, however, class-wise findings revealed persistent recall limitations for early disease. In the 11-label analysis, Dysplasia detection remained low (YOLOv5: 88\/201, 43.8%; YOLOv8: 82\/201, 40.8%), and SCC was only moderate (YOLOv5: 25\/44, 56.8%; YOLOv8: 24\/44, 54.5%). Confusion analyses showed that errors were dominated by non-detections (\u201cbackground\u201d) rather than misclassification with benign or artifact labels, while approximately one in five lesion predictions was a spurious unmatched false positive, implicating both sensitivity and specificity constraints.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These results indicate that labeling artifacts reduces non-lesion confusion but does not, by itself, recover subtle early lesions. Limitations include single-center, WLI-only data and training from scratch, future work should prioritize endoscopy-specific pretraining, explicit artifact suppression or joint segmentation, and external validation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fonc.2025.1707854","type":"journal-article","created":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T06:29:18Z","timestamp":1764916158000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":8,"title":["Investigating object detection errors in endoscopic imaging of esophageal SCC and dysplasia through precision\u2013recall analysis"],"prefix":"10.3389","volume":"15","author":[{"given":"Li-Jen","family":"Chang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kun-Hua","family":"Lee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Arvind","family":"Mukundan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Riya","family":"Karmakar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Achmad","family":"Bauravindah","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tsung-Hsien","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chien-Wei","family":"Huang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hsiang-Chen","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,12,5]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"649","DOI":"10.1053\/j.gastro.2022.05.054","article-title":"The global landscape of esophageal squamous cell carcinoma and esophageal adenocarcinoma incidence and mortality in 2020 and projections to 2040: new estimates from GLOBOCAN 2020","volume":"163","author":"Morgan","year":"2022","journal-title":"Gastroenterology"},{"key":"B2","article-title":"Survival rates for esophageal cancer | Esophageal cancer outlook","year":"2025"},{"key":"B3","article-title":"Esophageal dysplasia and chronic esophagitis: detection at upper gastrointestinal tract endoscopy","volume":"35","author":"Gr\u00fcber","year":"1998","journal-title":"Arquivos Gastroenterologia"},{"key":"B4","doi-asserted-by":"publisher","first-page":"72","DOI":"10.1177\/26345161231198626","article-title":"Screening for esophageal squamous cell carcinoma: A review","volume":"4","author":"Kodaganur Gopinath","year":"2024","journal-title":"Foregut"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.21037\/jtd.2019.06.50","article-title":"Accuracy of endoscopic ultrasound in esophageal cancer staging","volume":"11","author":"Krill","year":"2019","journal-title":"J Thorac Dis"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.3748\/wjg.v27.i14.1392","article-title":"Artificial intelligence-assisted endoscopic detection of esophageal neoplasia in early stage: The next step","volume":"27","author":"Liu","year":"2021","journal-title":"World J Gastroenterol"},{"key":"B7","doi-asserted-by":"publisher","first-page":"54","DOI":"10.1186\/s12885-016-3011-9","article-title":"Narrow band imaging versus lugol chromoendoscopy to diagnose squamous cell carcinoma of the esophagus: a systematic review and meta-analysis","volume":"17","author":"Morita","year":"2017","journal-title":"BMC Cancer"},{"key":"B8","doi-asserted-by":"publisher","first-page":"12","DOI":"10.17632\/C7FJBXCGJ9.1","article-title":"Endoscopy artifact detection (EAD 2019) challenge dataset","author":"Ali","year":"2019","journal-title":"arXiv"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2021.102002","article-title":"Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy","volume":"70","author":"Ali","year":"2021","journal-title":"Med Image Anal"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2020.101900","article-title":"A deep learning framework for quality assessment and restoration in video endoscopy","volume":"68","author":"Ali","year":"2021","journal-title":"Med Image Anal"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1159\/000540251","article-title":"Endoscopic artificial intelligence for image analysis in gastrointestinal neoplasms","volume":"105","author":"Kikuchi","year":"2024","journal-title":"Digestion"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.1016\/S2589-7500(20)30160-6","article-title":"Artificial intelligence in medical imaging: switching from radiographic pathological data to clinically meaningful endpoints","volume":"2","author":"Oren","year":"2020","journal-title":"Lancet Digital Health"},{"key":"B13","doi-asserted-by":"publisher","first-page":"105111","DOI":"10.48550\/arXiv.2111.02398","article-title":"Transparency of deep neural networks for medical image analysis: A review of interpretability methods","volume":"140","author":"Salahuddin","year":"2021","journal-title":"Computers in Biology and Medicine"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1002\/ird3.70008","article-title":"Application of artificial intelligence in medical imaging: current status and future directions","volume":"3","author":"Yang","year":"2025","journal-title":"Iradiology"},{"key":"B15","article-title":"AI hasn\u2019t killed radiology, but it is changing it","author":"Serrano","year":"2025","journal-title":"Washington Post"},{"key":"B16","doi-asserted-by":"publisher","first-page":"30033","DOI":"10.1038\/s41598-024-81842-9","article-title":"CRH-YOLO for precise and efficient detection of gastrointestinal polyps","volume":"14","author":"Wan","year":"2024","journal-title":"Sci Rep"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0321092","article-title":"Early detection of esophageal cancer: Evaluating AI algorithms with multi-institutional narrowband and white-light imaging data","volume":"20","author":"Baik","year":"2025","journal-title":"PloS One"},{"key":"B18","doi-asserted-by":"publisher","first-page":"16907","DOI":"10.1038\/s41598-025-96804-y","article-title":"A monocular endoscopic image depth estimation method based on a window-adaptive asymmetric dual-branch Siamese network","volume":"15","author":"Chong","year":"2025","journal-title":"Sci Rep"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.3390\/cancers17122049","article-title":"Evaluation of spectral imaging for early esophageal cancer detection","volume":"17","author":"Chang","year":"2025","journal-title":"Cancers"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics15131686","article-title":"Optimizing esophageal cancer diagnosis with computer-aided detection by YOLO models combined with hyperspectral imaging","volume":"15","author":"Weng","year":"2025","journal-title":"Diagnostics"},{"key":"B21","doi-asserted-by":"publisher","first-page":"160","DOI":"10.1038\/s41746-025-01532-2","article-title":"Effect of computer aided detection system on esophageal neoplasm diagnosis in varied levels of endoscopists","volume":"8","author":"Li","year":"2025","journal-title":"NPJ Digital Med"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.3390\/ijms25021251","article-title":"Diagnostic Challenges during Inflammation and Cancer: Current Biomarkers and Future Perspectives in Navigating through the Minefield of Reactive versus Dysplastic and Cancerous Lesions in the Digestive System","volume":"25","author":"Pateras","year":"2024","journal-title":"Int J Mol Sci"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.3390\/s23020974","article-title":"Specular reflections detection and removal for endoscopic images based on brightness classification","volume":"23","author":"Nie","year":"2023","journal-title":"Sensors (Basel Switzerland)"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1055\/a-0855-3532","article-title":"A deep neural network improves endoscopic detection of early gastric cancer without blind spots","volume":"51","author":"Wu","year":"2019","journal-title":"Endoscopy"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2210.15033","article-title":"Multi-scale structural-aware exposure correction for endoscopic imaging","author":"Garcia-Vega","year":"2022","journal-title":"IEEE 20th International Symposium on Biomedical Imaging (ISBI)"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics13010067","article-title":"A comparison of techniques for class imbalance in deep learning classification of breast cancer","volume":"13","author":"Walsh","year":"2023","journal-title":"Diagnostics"},{"key":"B27","doi-asserted-by":"publisher","first-page":"e190413","DOI":"10.48550\/arXiv.1909.12475","article-title":"Hidden stratification causes clinically meaningful failures in machine learning for medical imaging","volume":"2","author":"Oakden-Rayner","year":"2019","journal-title":"Radiology: Artificial Intelligence"},{"key":"B28","doi-asserted-by":"publisher","first-page":"918","DOI":"10.1038\/s41597-025-05251-x","article-title":"Polyp-size: A precise endoscopic dataset for AI-driven polyp sizing","volume":"12","author":"Song","year":"2025","journal-title":"Sci Data"},{"key":"B29","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1038\/s41746-024-01270-x","article-title":"A scoping review of reporting gaps in FDA-approved AI medical devices","volume":"7","author":"Muralidharan","year":"2024","journal-title":"NPJ Digital Med"},{"key":"B30","volume-title":"LabelImg is a graphical image annotation tool and label object bounding boxes in images 1.8.6","year":""},{"key":"B31","doi-asserted-by":"publisher","first-page":"64","DOI":"10.1016\/j.inffus.2022.11.023","article-title":"Fusion of colour contrasted images for early detection of oesophageal squamous cell dysplasia from endoscopic videos in real time","volume":"92","author":"Gao","year":"2023","journal-title":"Inf Fusion"},{"key":"B32","doi-asserted-by":"publisher","first-page":"22441","DOI":"10.1038\/s41598-024-72312-3","article-title":"Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration","volume":"14","author":"Fujii","year":"2024","journal-title":"Sci Rep"},{"key":"B33","doi-asserted-by":"publisher","first-page":"306","DOI":"10.1186\/s12880-024-01479-y","article-title":"Enhancing image-based diagnosis of gastrointestinal tract diseases through deep learning with EfficientNet and advanced data augmentation techniques","volume":"24","author":"Zubair Rahman","year":"2024","journal-title":"BMC Med Imaging"},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.1007\/s40747-023-01271-5","article-title":"Multi-classification deep learning models for detection of ulcerative colitis, polyps, and dyed-lifted polyps using wireless capsule endoscopy images","volume":"10","author":"Malik","year":"2024","journal-title":"Complex Intelligent Syst"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.3390\/app132212325","article-title":"Investigating effective data augmentation techniques for accurate gastric classification in the development of a deep learning-based computer-aided diagnosis system","volume":"13","author":"Park","year":"2023","journal-title":"Appl Sci"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2303.01178","article-title":"Augmenting medical imaging: A comprehensive catalogue of 65 techniques for enhanced data analysis","author":"Cossio","year":"2023","journal-title":"arXiv"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.1016\/j.health.2024.100340","article-title":"A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions","volume":"5","author":"Islam","year":"2024","journal-title":"Healthcare Analytics"},{"key":"B38","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2407.20892","article-title":"What is YOLOv5: A deep look into the internal features of the popular object detector","volume":"12","author":"Khanam","year":"2024","journal-title":"arXiv"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1911.08287","article-title":"Distance-ioU loss: faster and better learning for bounding box regression","volume":"34","author":"Zheng","year":"2019","journal-title":"AAAI Conference on Artificial Intelligence (AAAI 2020)"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.3390\/make5040083","article-title":"A comprehensive review of yolo architectures in computer vision: from yolov1 to yolov8 and yolo-nas","volume":"5","author":"Terven","year":"2023","journal-title":"Mach Learn Knowledge Extraction"},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2006.04388","article-title":"Generalized focal loss: learning qualified and distributed bounding boxes for dense object detection","author":"Li","year":"2020","journal-title":"Neural Information Processing Systems (NeurIPS 2020)"},{"key":"B42","article-title":"YOLOv8 ultralytics","year":"2023"},{"key":"B43","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1506.01497","article-title":"Faster R-CNN: towards real-time object detection with region proposal networks","volume":"39","author":"Ren","year":"2016","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)"},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1155\/2021\/2144472","article-title":"Single shot multibox detector automatic polyp detection network based on gastrointestinal endoscopic images","volume":"2021","author":"Chen","year":"2021","journal-title":"Comput Math Methods Med"},{"key":"B45","doi-asserted-by":"publisher","first-page":"124","DOI":"10.1186\/s12911-023-02235-y","article-title":"Machine learning applications for early detection of esophageal cancer: a systematic review","volume":"23","author":"Hosseini","year":"2023","journal-title":"BMC Med Inf Decision Making"},{"key":"B46","doi-asserted-by":"publisher","first-page":"36","DOI":"10.53469\/jtpes.2023.03(12).06","article-title":"Detection of esophageal cancer lesions based on CBAM faster R-CNN","volume":"3","author":"Liu","year":"2023","journal-title":"J Theory Pract Eng Sci"},{"key":"B47","doi-asserted-by":"publisher","first-page":"21","DOI":"10.48550\/arXiv.1512.02325","article-title":"SSD: single shot multiBox detector","author":"Liu","year":"2016","journal-title":"European conference on computer vision"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2021.106172","article-title":"Esophageal cancer detection based on classification of gastrointestinal CT images using improved Faster RCNN","volume":"207","author":"Chen","year":"2021","journal-title":"Comput Methods Programs Biomedicine"},{"key":"B49","doi-asserted-by":"publisher","DOI":"10.1007\/s11548-019-01914-4","article-title":"Early esophageal adenocarcinoma detection using deep learning methods","volume":"14","author":"Ghatwary","year":"2019","journal-title":"Int J Comput Assisted Radiol Surg"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.3390\/cancers13184593","article-title":"Hyperspectral imaging combined with artificial intelligence in the early detection of esophageal cancer","volume":"13","author":"Tsai","year":"2021","journal-title":"Cancers"},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2304.08069","article-title":"DETRs beat YOLOs on real-time object detection","author":"Zhao","year":"2024"},{"key":"B52","doi-asserted-by":"publisher","DOI":"10.3389\/frai.2025.1529814","article-title":"Precision enhancement in wireless capsule endoscopy: a novel transformer-based approach for real-time video object detection","volume":"8","author":"Habe","year":"2025","journal-title":"Front Artif Intell"},{"key":"B53","doi-asserted-by":"publisher","DOI":"10.21037\/aoe-2020-33","article-title":"Development of artificial intelligence for the detection and staging of esophageal cancer","volume":"6","author":"Tokai","year":"2023","journal-title":"Ann Esophagus"},{"key":"B54","doi-asserted-by":"publisher","first-page":"25","DOI":"10.1016\/j.gie.2018.07.037","article-title":"Diagnostic outcomes of esophageal cancer by artificial intelligence using convolutional neural networks","volume":"89","author":"Horie","year":"2019","journal-title":"Gastrointestinal Endoscopy"}],"updated-by":[{"DOI":"10.3389\/fonc.2025.1773175","type":"correction","label":"Correction","source":"publisher","updated":{"date-parts":[[2026,1,6]],"date-time":"2026-01-06T00:00:00Z","timestamp":1767657600000}}],"container-title":["Frontiers in Oncology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2025.1707854\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,6]],"date-time":"2026-01-06T15:11:35Z","timestamp":1767712295000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2025.1707854\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,5]]},"references-count":54,"alternative-id":["10.3389\/fonc.2025.1707854"],"URL":"https:\/\/doi.org\/10.3389\/fonc.2025.1707854","relation":{},"ISSN":["2234-943X"],"issn-type":[{"value":"2234-943X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,5]]},"article-number":"1707854"}}