{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T18:21:25Z","timestamp":1792088485988,"version":"4.4.0"},"reference-count":18,"publisher":"MDPI AG","issue":"13","license":[{"start":{"date-parts":[[2025,6,26]],"date-time":"2025-06-26T00:00:00Z","timestamp":1750896000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Diagnostics"],"abstract":"<jats:p>Background: Digital breast tomosynthesis (DBT) is a valuable imaging modality for breast cancer detection; however, its interpretation remains time-consuming and subject to inter-reader variability. This study aimed to develop and evaluate two deep learning (DL) models based on transfer learning for the binary classification of breast lesions (benign vs. malignant) using DBT images to support clinical decision-making and risk stratification. Methods: In this retrospective monocentric study, 184 patients with histologically or clinically confirmed benign (107 cases, 41.8%) or malignant (77 cases, 58.2%) breast lesions were included. Each case underwent DBT with a single lesion manually segmented for radiomic analysis. Two convolutional neural network (CNN) architectures\u2014ResNet50 and DenseNet201\u2014were trained using transfer learning from ImageNet weights. A 10-fold cross-validation strategy with ensemble voting was applied. Model performance was evaluated through ROC\u2013AUC, accuracy, sensitivity, specificity, PPV, and NPV. Results: The ResNet50 model outperformed DenseNet201 across most metrics. On the internal testing set, ResNet50 achieved a ROC\u2013AUC of 63%, accuracy of 60%, sensitivity of 39%, and specificity of 75%. The DenseNet201 model yielded a lower ROC\u2013AUC of 55%, accuracy of 55%, and sensitivity of 24%. Both models demonstrated relatively high specificity, indicating potential utility in ruling out malignancy, though sensitivity remained suboptimal. Conclusions: This study demonstrates the feasibility of using transfer learning-based DL models for lesion classification on DBT. While the overall performance was moderate, the results highlight both the potential and current limitations of AI in breast imaging. Further studies and approaches are warranted to enhance model robustness and clinical applicability.<\/jats:p>","DOI":"10.3390\/diagnostics15131631","type":"journal-article","created":{"date-parts":[[2025,6,26]],"date-time":"2025-06-26T11:15:23Z","timestamp":1750936523000},"page":"1631","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":5,"title":["Deep Learning with Transfer Learning on Digital Breast Tomosynthesis: A Radiomics-Based Model for Predicting Breast Cancer Risk"],"prefix":"10.3390","volume":"15","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2103-2856","authenticated-orcid":false,"given":"Francesca","family":"Galati","sequence":"first","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-6626-2478","authenticated-orcid":false,"given":"Roberto","family":"Maroncelli","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"},{"name":"Department of Experimental Medicine, Sapienza University of Rome, Viale Regina Elena, 324, 00161 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chiara","family":"De Nardo","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lucia","family":"Testa","sequence":"additional","affiliation":[{"name":"Department of Informatic, Automatic and Gestional Engineering (DIAG), Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gloria","family":"Barcaroli","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1781-3174","authenticated-orcid":false,"given":"Veronica","family":"Rizzo","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6595-1611","authenticated-orcid":false,"given":"Giuliana","family":"Moffa","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1432-2137","authenticated-orcid":false,"given":"Federica","family":"Pediconi","sequence":"additional","affiliation":[{"name":"Department of Radiological, Oncological and Pathological Sciences, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2025,6,26]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"358","DOI":"10.1053\/j.semnuclmed.2025.01.008","article-title":"AI in Breast Cancer Imaging: An Update and Future Trends","volume":"55","author":"Chen","year":"2025","journal-title":"Semin. Nucl. Med."},{"key":"ref_2","first-page":"e79177","article-title":"Leveraging Artificial Intelligence in Breast Cancer Screening and Diagnosis","volume":"17","author":"Hasan","year":"2025","journal-title":"Cureus"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"205","DOI":"10.3348\/jksr.2025.0011","article-title":"Clinical Application of Artificial Intelligence in Digital Breast Tomosynthesis","volume":"86","author":"Chang","year":"2025","journal-title":"J. Korean Soc. Radiol."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"529","DOI":"10.1007\/s10549-024-07589-z","article-title":"Traditional versus modern approaches to screening mammography: A comparison of computer-assisted detection for synthetic 2D mammography versus an artificial intelligence algorithm for digital breast tomosynthesis","volume":"210","author":"Bahl","year":"2025","journal-title":"Breast Cancer Res. Treat."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"e180096","DOI":"10.1148\/ryai.2019180096","article-title":"Improving Accuracy and Efficiency with Concurrent Use of Artificial Intelligence for Digital Breast Tomosynthesis","volume":"1","author":"Conant","year":"2019","journal-title":"Radiol. Artif. Intell."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"420","DOI":"10.1007\/s12672-025-02170-6","article-title":"The value of predicting breast cancer with a DBT 2.5D deep learning model","volume":"16","author":"Niu","year":"2025","journal-title":"Discov. Oncol."},{"key":"ref_7","first-page":"206","article-title":"Breast cancer detection and classification with digital breast tomosynthesis: A two-stage deep learning approach","volume":"31","author":"Alashban","year":"2025","journal-title":"Diagn. Interv. Radiol."},{"key":"ref_8","first-page":"795","article-title":"Artificial intelligence in breast cancer imaging: Risk stratification, lesion detection and classification, treatment planning and prognosis\u2014A narrative review","volume":"3","author":"Caloro","year":"2022","journal-title":"Explor. Target. Antitumor Ther."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"102049","DOI":"10.1016\/j.media.2021.102049","article-title":"Applying deep learning in digital breast tomosynthesis for automatic breast cancer detection: A review","volume":"71","author":"Bai","year":"2021","journal-title":"Med. Image Anal."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"111457","DOI":"10.1016\/j.ejrad.2024.111457","article-title":"Artificial Intelligence for breast cancer detection: Technology, challenges, and prospects","volume":"175","author":"Sechopoulos","year":"2024","journal-title":"Eur. J. Radiol."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"e230318","DOI":"10.1148\/ryai.230318","article-title":"Impact of AI for Digital Breast Tomosynthesis on Breast Cancer Detection and Interpretation Time","volume":"6","author":"Park","year":"2024","journal-title":"Radiol. Artif. Intell."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"682","DOI":"10.21147\/j.issn.1000-9604.2021.06.05","article-title":"Deep learning model improves radiologists\u2019 performance in detection and classification of breast lesions","volume":"33","author":"Sun","year":"2021","journal-title":"Chin. J. Cancer Res."},{"key":"ref_13","first-page":"3958","article-title":"Impact of real-life use of artificial intelligence as support for human reading in a population-based breast cancer screening program with mammography and tomosynthesis","volume":"34","author":"Brehl","year":"2024","journal-title":"Eur. Radiol."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"303","DOI":"10.1016\/j.media.2016.07.007","article-title":"Large scale deep learning for computer aided detection of mammographic lesions","volume":"35","author":"Kooi","year":"2017","journal-title":"Med. Image Anal."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"273","DOI":"10.21037\/qims-23-846","article-title":"2.5D transfer deep learning model for segmentation of contrast-enhancing lesions on brain magnetic resonance imaging of multiple sclerosis and neuromyelitis optica spectrum disorder","volume":"14","author":"Huang","year":"2024","journal-title":"Quant. Imaging Med. Surg."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"World Medical Association (2013). World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects. JAMA, 310, 2191\u20132194.","DOI":"10.1001\/jama.2013.281053"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"75","DOI":"10.1186\/s13244-023-01415-8","article-title":"CheckList for EvaluAtion of Radiomics research (CLEAR): A step-by-step reporting guideline for authors and reviewers endorsed by ESR and EuSoMII","volume":"14","author":"Kocak","year":"2023","journal-title":"Insights Imaging"},{"key":"ref_18","unstructured":"D\u2019Orsi, C.J., Sickles, E.A., Mendelson, E.B., and Morris, E.A. (2013). ACR BI-RADS\u00ae Atlas, Breast Imaging Reporting and Data System, American College of Radiology. [5th ed.]."}],"container-title":["Diagnostics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-4418\/15\/13\/1631\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T17:59:30Z","timestamp":1760032770000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-4418\/15\/13\/1631"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,26]]},"references-count":18,"journal-issue":{"issue":"13","published-online":{"date-parts":[[2025,7]]}},"alternative-id":["diagnostics15131631"],"URL":"https:\/\/doi.org\/10.3390\/diagnostics15131631","relation":{},"ISSN":["2075-4418"],"issn-type":[{"value":"2075-4418","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,6,26]]}}}