{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T14:20:39Z","timestamp":1784211639217,"version":"3.55.0"},"reference-count":33,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,12,21]],"date-time":"2023-12-21T00:00:00Z","timestamp":1703116800000},"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. Pediatr."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>In the field of pediatric trauma computer-aided detection (CADe) and computer-aided diagnosis (CADx) systems have emerged offering a promising avenue for improved patient care. Especially children with wrist fractures may benefit from machine learning (ML) solutions, since some of these lesions may be overlooked on conventional X-ray due to minimal compression without dislocation or mistaken for cartilaginous growth plates. In this article, we describe the development and optimization of AI algorithms for wrist fracture detection in children.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>A team of IT-specialists, pediatric radiologists and pediatric surgeons used the freely available GRAZPEDWRI-DX dataset containing annotated pediatric trauma wrist radiographs of 6,091 patients, a total number of 10,643 studies (20,327 images). First, a basic object detection model, a You Only Look Once object detector of the seventh generation (YOLOv7) was trained and tested on these data. Then, team decisions were taken to adjust data preparation, image sizes used for training and testing, and configuration of the detection model. Furthermore, we investigated each of these models using an Explainable Artificial Intelligence (XAI) method called Gradient Class Activation Mapping (Grad-CAM). This method visualizes where a model directs its attention to before classifying and regressing a certain class through saliency maps.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Mean average precision (mAP) improved when applying optimizations pre-processing the dataset images (maximum increases of <jats:inline-formula><mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"IM1\"><mml:mo>+<\/mml:mo><\/mml:math><\/jats:inline-formula>25.51% mAP@0.5 and <jats:inline-formula><mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"IM2\"><mml:mo>+<\/mml:mo><\/mml:math><\/jats:inline-formula>39.78% mAP@[0.5:0.95]), as well as the object detection model itself (maximum increases of <jats:inline-formula><mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"IM3\"><mml:mo>+<\/mml:mo><\/mml:math><\/jats:inline-formula>13.36% mAP@0.5 and <jats:inline-formula><mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"IM4\"><mml:mo>+<\/mml:mo><\/mml:math><\/jats:inline-formula>27.01% mAP@[0.5:0.95]). Generally, when analyzing the resulting models using XAI methods, higher scoring model variations in terms of mAP paid more attention to broader regions of the image, prioritizing detection accuracy over precision compared to the less accurate models.<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion<\/jats:title><jats:p>This paper supports the implementation of ML solutions for pediatric trauma care. Optimization of a large X-ray dataset and the YOLOv7 model improve the model\u2019s ability to detect objects and provide valid diagnostic support to health care specialists. Such optimization protocols must be understood and advocated, before comparing ML performances against health care specialists.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fped.2023.1291804","type":"journal-article","created":{"date-parts":[[2023,12,21]],"date-time":"2023-12-21T04:40:43Z","timestamp":1703133643000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":27,"title":["Development and optimization of AI algorithms for wrist fracture detection in children using a freely available dataset"],"prefix":"10.3389","volume":"11","author":[{"given":"Tristan","family":"Till","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sebastian","family":"Tschauner","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Georg","family":"Singer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Klaus","family":"Lichtenegger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Holger","family":"Till","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2023,12,21]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"6816","DOI":"10.1007\/s00330-021-07811-2","article-title":"AI-based detection and classification of distal radius fractures using low-effort data labeling: evaluation of applicability and effect of training set size","volume":"31","author":"Tobler","year":"2021","journal-title":"Eur Radiol"},{"key":"B2","doi-asserted-by":"publisher","first-page":"6006","DOI":"10.1038\/s41598-021-85570-2","article-title":"Critical evaluation of deep neural networks for wrist fracture detection","volume":"11","author":"Raisuddin","year":"2021","journal-title":"Sci Rep"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1005099","DOI":"10.3389\/fped.2022.1005099","article-title":"Pediatric radius torus fractures in x-rays-how computer vision could render lateral projections obsolete","volume":"10","author":"Janisch","year":"2022","journal-title":"Front Pediatr"},{"key":"B4","doi-asserted-by":"publisher","first-page":"183","DOI":"10.1177\/0009922810384725","article-title":"Most frequently missed fractures in the emergency department","volume":"50","author":"Mounts","year":"2011","journal-title":"Clin Pediatr (Phila)"},{"key":"B5","doi-asserted-by":"publisher","first-page":"107293","DOI":"10.1109\/ACCESS.2022.3209825","article-title":"A review on machine learning styles in computer vision\u2014techniques and future directions","volume":"10","author":"Mahadevkar","year":"2022","journal-title":"IEEE Access"},{"key":"B6","volume-title":"Probabilistic machine learning: an introduction","author":"Murphy","year":"2022"},{"key":"B7","author":"Redmon","year":""},{"key":"B8","author":"Wang","year":"2022"},{"key":"B9","author":"Silva","year":""},{"key":"B10","doi-asserted-by":"publisher","first-page":"76","DOI":"10.5206\/uwomj.v86i2.2060","article-title":"Machine learning in medicine","volume":"86","author":"Gui","year":"2017","journal-title":"Univ West Ont Med J"},{"key":"B11","doi-asserted-by":"publisher","first-page":"102470","DOI":"10.1016\/j.media.2022.102470","article-title":"Explainable artificial intelligence (XAI) in deep learning-based medical image analysis","volume":"79","author":"van der Velden","year":"2022","journal-title":"Med Image Anal"},{"key":"B12","doi-asserted-by":"publisher","first-page":"52","DOI":"10.3390\/jimaging6060052","article-title":"Explainable deep learning models in medical image analysis","volume":"6","author":"Singh","year":"2020","journal-title":"J Imaging"},{"key":"B13","author":"Selvaraju","year":""},{"key":"B14","author":"Till","year":""},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-022-01328-z","article-title":"A pediatric wrist trauma x-ray dataset (GRAZPEDWRI-DX) for machine learning","author":"Nagy","year":"2022","journal-title":"Sci Data"},{"key":"B16","doi-asserted-by":"publisher","first-page":"6999","DOI":"10.1109\/TNNLS.2021.3084827","article-title":"A survey of convolutional neural networks: analysis, applications,, prospects","volume":"33","author":"Li","year":"2022","journal-title":"IEEE Trans Neural Netw Learn Syst"},{"key":"B17","doi-asserted-by":"publisher","first-page":"022022","DOI":"10.1088\/1742-6596\/1168\/2\/022022","article-title":"An overview of overfitting, its solutions","volume":"1168","author":"Ying","year":"2019","journal-title":"J Phys Conf Ser"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.3390\/molecules26041111","article-title":"Effect of dataset size and train\/test split ratios in QSAR\/QSPR multiclass classification","volume":"26","author":"R\u00e1cz","year":"2021","journal-title":"Molecules"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.3390\/agronomy13030887","article-title":"Improving deep learning classifiers performance via preprocessing and class imbalance approaches in a plant disease detection pipeline","volume":"13","author":"Ojo","year":"2023","journal-title":"Agronomy"},{"key":"B20","doi-asserted-by":"publisher","first-page":"52818","DOI":"10.1109\/ACCESS.2022.3174583","article-title":"Optimizing the hyperparameter tuning of YOLOv5 for underwater detection","volume":"10","author":"Isa","year":"2022","journal-title":"IEEE Access"},{"key":"B21","author":"Henderson","year":""},{"key":"B22","doi-asserted-by":"publisher","first-page":"19","DOI":"10.1097\/TA.0b013e3181f8a903","article-title":"Evaluation of fractures in children and adolescents in a Level I Trauma Center in Austria","volume":"71","author":"Schalamon","year":"2011","journal-title":"J Trauma"},{"key":"B23","doi-asserted-by":"publisher","first-page":"843","DOI":"10.1016\/j.rcl.2019.02.009","article-title":"Frequently missed fractures in pediatric trauma: a pictorial review of plain film radiography","volume":"57","author":"George","year":"2019","journal-title":"Radiol Clin North Am"},{"key":"B24","doi-asserted-by":"publisher","first-page":"425","DOI":"10.1016\/j.cps.2019.02.012","article-title":"Pediatric hand and wrist fractures","volume":"46","author":"Liao","year":"2019","journal-title":"Clin Plast Surg"},{"key":"B25","doi-asserted-by":"publisher","first-page":"1675","DOI":"10.1016\/j.jacr.2021.07.026","article-title":"Pediatric radiologist workforce shortage: action steps to resolve","volume":"18","author":"Farmakis","year":"2021","journal-title":"J Am Coll Radiol"},{"key":"B26","doi-asserted-by":"publisher","first-page":"112821","DOI":"10.1016\/j.eswa.2019.112821","article-title":"A systematic survey of computer-aided diagnosis in medicine: past and present developments","volume":"138","author":"Yanase","year":"2019","journal-title":"Expert Syst Appl"},{"key":"B27","doi-asserted-by":"publisher","first-page":"933","DOI":"10.1016\/j.media.2012.02.005","article-title":"Machine learning and radiology","volume":"16","author":"Wang","year":"2012","journal-title":"Med Image Anal"},{"key":"B28","doi-asserted-by":"publisher","first-page":"647","DOI":"10.1007\/s10140-023-02154-5","article-title":"AI tools in Emergency Radiology reading room: a new era of Radiology","volume":"30","author":"Dundamadappa","year":"2023","journal-title":"Emerg Radiol"},{"key":"B29","doi-asserted-by":"publisher","first-page":"e231114","DOI":"10.1148\/radiol.231114","article-title":"The future of AI and informatics in radiology: 10 predictions","volume":"309","author":"Langlotz","year":"2023","journal-title":"Radiology"},{"key":"B30","doi-asserted-by":"publisher","first-page":"20230299","DOI":"10.1259\/bjr.20230299","article-title":"Towards AI-augmented radiology education: a web-based application for perception training in chest X-ray nodule detection","volume":"96","author":"Borgbjerg","year":"2023","journal-title":"Br J Radiol"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacr.2023.09.011","article-title":"Enhancing patient communication with Chat-GPT in radiology: evaluating the efficacy and readability of answers to common imaging-related questions","author":"Gordon","year":"2023","journal-title":"J Am Coll Radiol"},{"key":"B32","author":"Lai","year":""},{"key":"B33","doi-asserted-by":"publisher","first-page":"265","DOI":"10.1016\/b978-0-12-821259-2.00014-4","article-title":"Artificial intelligence in radiology","author":"Jin","year":"2021","journal-title":"Chapter 14 - Artificial intelligence in medicine"}],"container-title":["Frontiers in Pediatrics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fped.2023.1291804\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,12,21]],"date-time":"2023-12-21T04:40:46Z","timestamp":1703133646000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fped.2023.1291804\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,12,21]]},"references-count":33,"alternative-id":["10.3389\/fped.2023.1291804"],"URL":"https:\/\/doi.org\/10.3389\/fped.2023.1291804","relation":{},"ISSN":["2296-2360"],"issn-type":[{"value":"2296-2360","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,12,21]]},"article-number":"1291804"}}