{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,18]],"date-time":"2026-09-18T14:01:46Z","timestamp":1789740106761,"version":"4.0.1"},"reference-count":36,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T00:00:00Z","timestamp":1784678400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T00:00:00Z","timestamp":1784678400000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["J Applied Clin Med Phys"],"published-print":{"date-parts":[[2026,8]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Accurate and rapid diagnosis of meniscal tears is crucial for effective management of sports\u2010related injuries and degenerative knee disorders. Magnetic resonance imaging (MRI) is widely used for meniscus evaluation; however, manual interpretation is time\u2010consuming and subject to inter\u2010observer variability. Automated and reliable classification systems may therefore support clinical decision\u2010making.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>This study aims to develop a robust and interpretable deep learning framework for the automatic four\u2010class classification of meniscal conditions using MRI images.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A total of 2,000 knee MRI images obtained from a tertiary care university hospital were categorized into four classes: Grade I, Grade II, Grade III, and normal meniscus. A Random Search\u2010based Neural Architecture Search (RS\u2010NAS) strategy was used to generate task\u2010specific CNN architectures. The class probability outputs of the top\u2010performing NAS\u2010derived CNN models were combined using a late\u2010fusion stacking strategy, with ElasticNet employed as the meta\u2010learner. The proposed framework was evaluated using both a single\u2010run assessment and a repeated validation protocol consisting of 10 independent runs with 5\u2010fold cross\u2010validation. Performance was assessed using ACC, F1\u2010score, MCC, AUC, and PR\u2010AUC. Statistical comparisons and explainability analyses, including Grad\u2010CAM and ElasticNet feature importance, were also performed.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In the single\u2010run evaluation, the ElasticNet\u2010based stacking model achieved an ACC of 0.9300, F1\u2010score of 0.9304, and AUC of 0.9913. Under the repeated 10\u2010run 5\u2010fold cross\u2010validation protocol, the proposed model obtained an ACC of 0.9083 \u00b1 0.0019, F1\u2010score of 0.9084 \u00b1 0.0018, MCC of 0.8783 \u00b1 0.0024, AUC of 0.9883 \u00b1 0.0008, and PR\u2010AUC of 0.9719 \u00b1 0.0020. The proposed stacking framework outperformed the best NAS CNN, DenseNet121\u2010TL, averaging ensemble, and majority voting ensemble across the main evaluation metrics. Grad\u2010CAM visualizations and feature importance analysis further indicated that the model relied on clinically meaningful image regions and informative base\u2010model predictions.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The integration of RS\u2010NAS with ElasticNet\u2010based late\u2010fusion stacking provides a stable, high\u2010performing, and interpretable framework for four\u2010class meniscus MRI classification. Although the single\u2010run evaluation showed higher peak performance, the repeated cross\u2010validation results provide a more reliable estimate of model robustness and generalization. Future studies should validate the proposed framework using independent multi\u2010center datasets and different imaging protocols.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/acm2.70711","type":"journal-article","created":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T08:45:49Z","timestamp":1784709949000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["A stacking based deep learning framework integrating random search neural architecture search for meniscus tear diagnosis"],"prefix":"10.1002","volume":"27","author":[{"given":"Ebubekir","family":"Seyyarer","sequence":"first","affiliation":[{"name":"Department of Computer Engineering Van Y\u00fcz\u00fcnc\u00fc Y\u0131l University  Van T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hasan","family":"Gen\u00e7","sequence":"additional","affiliation":[{"name":"Radiology Clinic Elaz\u0131\u011f Fethi Sekin City Hospital  Elaz\u0131\u011f T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Faruk","family":"Ayata","sequence":"additional","affiliation":[{"name":"Department of Computer Technologies Van Y\u00fcz\u00fcnc\u00fc Y\u0131l University  Van T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,7,22]]},"reference":[{"key":"e_1_2_11_2_1","doi-asserted-by":"publisher","DOI":"10.46810\/tdfd.1722759"},{"key":"e_1_2_11_3_1","doi-asserted-by":"publisher","DOI":"10.1140\/epjp\/s13360\u2010025\u201007148\u20105"},{"key":"e_1_2_11_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.measurement.2025.117294"},{"key":"e_1_2_11_5_1","doi-asserted-by":"publisher","DOI":"10.2196\/53616"},{"key":"e_1_2_11_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2023.105894"},{"key":"e_1_2_11_7_1","doi-asserted-by":"publisher","DOI":"10.3389\/fmed.2024.1505692"},{"key":"e_1_2_11_8_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2021.3054390"},{"key":"e_1_2_11_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2017.07.005"},{"key":"e_1_2_11_10_1","doi-asserted-by":"publisher","DOI":"10.1146\/annurev\u2010bioeng\u2010071516\u2010044442"},{"key":"e_1_2_11_11_1","doi-asserted-by":"publisher","DOI":"10.3389\/fbioe.2021.747217"},{"key":"e_1_2_11_12_1","doi-asserted-by":"publisher","DOI":"10.1186\/s40001\u2010022\u201000883\u2010w"},{"key":"e_1_2_11_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/ksa.12369"},{"key":"e_1_2_11_14_1","doi-asserted-by":"publisher","DOI":"10.31436\/ijpcc.v10i1.460"},{"key":"e_1_2_11_15_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00256\u2010020\u201003410\u20102"},{"key":"e_1_2_11_16_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11042\u2010023\u201015793\u20100"},{"key":"e_1_2_11_17_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00330\u2010023\u201009483\u20106"},{"key":"e_1_2_11_18_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11042\u2010023\u201017092\u20100"},{"key":"e_1_2_11_19_1","unstructured":"ChenL\u2010C CollinsMD ZhuY et\u00a0al 2018.Searching for efficient multi\u2010scale architectures for dense image prediction. In:Proceedings of the 32nd International Conference on Neural Information Processing Systems NIPS\u201918 Curran Associates Inc.:8713\u20138724. doi:10.48550\/arXiv.1809.04184"},{"key":"e_1_2_11_20_1","doi-asserted-by":"crossref","unstructured":"ZophB VasudevanV ShlensJ LeQV.Learning transferable architectures for scalable image recognition. In:Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).2018:8697\u20138710.","DOI":"10.1109\/CVPR.2018.00907"},{"key":"e_1_2_11_21_1","doi-asserted-by":"crossref","unstructured":"RealE AggarwalA HuangY LeQV.Regularized evolution for image classifier architecture search. In:Proceedings of the AAAI Conference on Artificial Intelligence.2019;33(1):4780\u20134789. doi:10.1609\/aaai.v33i01.33014780","DOI":"10.1609\/aaai.v33i01.33014780"},{"key":"e_1_2_11_22_1","first-page":"281","article-title":"Random search for hyper\u2010parameter optimization","volume":"13","author":"Bergstra J","year":"2012","journal-title":"J Mach Learn Res"},{"key":"e_1_2_11_23_1","doi-asserted-by":"publisher","DOI":"10.31590\/ejosat.802174"},{"key":"e_1_2_11_24_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS SunJ (2015).Deep residual learning for image recognition. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). doi:10.1109\/CVPR.2016.90","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_11_25_1","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics13223419"},{"key":"e_1_2_11_26_1","doi-asserted-by":"publisher","DOI":"10.1007\/s42044\u2010024\u201000216\u20106"},{"key":"e_1_2_11_27_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.procs.2024.04.022"},{"key":"e_1_2_11_28_1","doi-asserted-by":"crossref","unstructured":"HeW ZhouT XiangY LinY HuJ BaoR.Deep learning in image classification: evaluating VGG19's performance on complex visual data. arXiv.2024. doi:10.48550\/arXiv.2412.20345","DOI":"10.1109\/NNICE64954.2025.11064681"},{"issue":"4","key":"e_1_2_11_29_1","first-page":"215","article-title":"VGG19 demonstrates the highest accuracy rate in a nine\u2010class wound classification task among various deep learning networks: a pilot study","volume":"33","author":"Lee J\u2010H","year":"2024","journal-title":"Wounds"},{"key":"e_1_2_11_30_1","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab041"},{"key":"e_1_2_11_31_1","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbaa125"},{"key":"e_1_2_11_32_1","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab172"},{"key":"e_1_2_11_33_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0893\u20106080(05)80023\u20101"},{"key":"e_1_2_11_34_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jhydrol.2018.04.008"},{"key":"e_1_2_11_35_1","doi-asserted-by":"publisher","DOI":"10.46465\/endustrimuhendisligi.811183"},{"key":"e_1_2_11_36_1","doi-asserted-by":"publisher","DOI":"10.25092\/baunfbed.1299583"},{"key":"e_1_2_11_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ipm.2009.03.002"}],"container-title":["Journal of Applied Clinical Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.70711","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/acm2.70711","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.70711","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,18]],"date-time":"2026-09-18T13:06:39Z","timestamp":1789736799000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/acm2.70711"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,22]]},"references-count":36,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2026,8]]}},"alternative-id":["10.1002\/acm2.70711"],"URL":"https:\/\/doi.org\/10.1002\/acm2.70711","archive":["Portico"],"relation":{},"ISSN":["1526-9914","1526-9914"],"issn-type":[{"value":"1526-9914","type":"print"},{"value":"1526-9914","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,22]]},"assertion":[{"value":"2026-02-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-02","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-22","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70711"}}