{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T15:15:08Z","timestamp":1783610108028,"version":"3.55.0"},"reference-count":39,"publisher":"MDPI AG","issue":"7","license":[{"start":{"date-parts":[[2026,7,8]],"date-time":"2026-07-08T00:00:00Z","timestamp":1783468800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Bioengineering"],"abstract":"<jats:p>Background: HER2 status guides targeted therapy in breast cancer but is currently determined by invasive biopsy. Imaging-based HER2 prediction from dynamic contrast-enhanced MRI (DCE-MRI) could provide a non-invasive adjunct decision-support signal, but published models are typically single-center with heterogeneous preprocessing that limits reproducibility. Methods: We trained a Triple-Head Dual-Attention ResNet (THDA-ResNet) that processes three DCE phases (pre-contrast, early post-contrast, and late post-contrast) on the multicenter BreastDCEDL dataset (n = 1149, I-SPY trials), and we compared it with Vision Transformer (ViT) and Convolutional Vision Transformer (CvT) baselines, all ImageNet-pretrained. We benchmarked 14 preprocessing strategies, with and without N4 bias-field correction. External validation used the independent BreastDCEDL_AMBL cohort (43 lesions). AUC confidence intervals used stratified bootstrap; model comparisons used DeLong\u2019s test. Results: THDA-ResNet achieved the highest AUC, 0.74 (95% CI 0.65\u20130.83), versus 0.66 for ViT and 0.63 for CvT, with the advantage reaching borderline significance over CvT (p=0.054) and not significant over ViT (p=0.14). At a threshold of 0.7, it retained discrimination (sensitivity 0.41, specificity 0.86), while transformers collapsed to near-trivial classifiers. External AUC was 0.66 (0.49\u20130.81). N4 correction did not improve performance. Conclusions: Attention-driven CNNs are a strong default for HER2 prediction from DCE-MRI on medium-sized cohorts, and N4 correction can be omitted, simplifying the pipeline.<\/jats:p>","DOI":"10.3390\/bioengineering13070788","type":"journal-article","created":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T14:17:51Z","timestamp":1783606671000},"page":"788","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Attention-Driven CNNs as a Strong Default for HER2 Prediction from DCE-MRI: A Comparison with Transformer Architectures"],"prefix":"10.3390","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3711-1655","authenticated-orcid":false,"given":"Naomi","family":"Fridman","sequence":"first","affiliation":[{"name":"Department of Industrial Engineering, Ariel University, Ariel 40700, Israel"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anat","family":"Goldstein","sequence":"additional","affiliation":[{"name":"Department of Industrial Engineering, Ariel University, Ariel 40700, Israel"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,7,8]]},"reference":[{"key":"ref_1","first-page":"229","article-title":"Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries","volume":"74","author":"Bray","year":"2024","journal-title":"CA Cancer J. Clin."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"672","DOI":"10.1148\/radiol.2443051661","article-title":"Current status of breast MR imaging part 2. Clinical applications","volume":"244","author":"Kuhl","year":"2007","journal-title":"Radiology"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"783","DOI":"10.1056\/NEJM200103153441101","article-title":"Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2","volume":"344","author":"Slamon","year":"2001","journal-title":"N. Engl. J. Med."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"2105","DOI":"10.1200\/JCO.2018.77.8738","article-title":"Human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology\/College of American Pathologists clinical practice guideline focused update","volume":"36","author":"Wolff","year":"2018","journal-title":"J. Clin. Oncol."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"100592","DOI":"10.1016\/j.ejro.2024.100592","article-title":"MRI-based machine learning radiomics for prediction of HER2 expression status in breast invasive ductal carcinoma","volume":"13","author":"Luo","year":"2024","journal-title":"Eur. J. Radiol. Open"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"2243","DOI":"10.17305\/bb.2025.12475","article-title":"Deep learning predicts HER2 status in invasive breast cancer from multimodal ultrasound and MRI","volume":"25","author":"Fan","year":"2025","journal-title":"Biomol. Biomed."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"19034","DOI":"10.1038\/s41598-025-04136-8","article-title":"Histopathological image based breast cancer diagnosis using deep learning and bio inspired optimization","volume":"15","author":"Thatha","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"e24094","DOI":"10.1016\/j.heliyon.2024.e24094","article-title":"Enhancing breast cancer classification via histopathological image analysis: Leveraging self-supervised contrastive learning and transfer learning","volume":"10","author":"Ashraf","year":"2024","journal-title":"Heliyon"},{"key":"ref_9","unstructured":"Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., and Gelly, S. (2021, January 3\u20137). An image is worth 16x16 words: Transformers for image recognition at scale. Proceedings of the International Conference on Learning Representations (ICLR), Virtual."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Wu, H., Xiao, B., Codella, N., Liu, M., Dai, X., Yuan, L., and Zhang, L. (2021, January 10\u201317). CvT: Introducing convolutions to Vision Transformers. Proceedings of the IEEE\/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada.","DOI":"10.1109\/ICCV48922.2021.00009"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"264","DOI":"10.1038\/s41597-026-06589-6","article-title":"BreastDCEDL: A standardized deep learning-ready breast DCE-MRI dataset of 2070 patients","volume":"13","author":"Fridman","year":"2026","journal-title":"Sci. Data"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Wang, F., Jiang, M., Qian, C., Yang, S., Li, C., Zhang, H., Wang, X., and Tang, X. (2017, January 21\u201326). Residual attention network for image classification. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.","DOI":"10.1109\/CVPR.2017.683"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"10334","DOI":"10.1038\/s41598-022-14372-x","article-title":"A ResNet attention model for classifying mosquitoes from wing-beating sounds","volume":"12","author":"Wei","year":"2022","journal-title":"Sci. Rep."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"196647","DOI":"10.1109\/ACCESS.2024.3520608","article-title":"Breast cancer classification with enhanced interpretability: DALAResNet50 and DT Grad-CAM","volume":"12","author":"Liu","year":"2024","journal-title":"IEEE Access"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"17","DOI":"10.1615\/CritRevBiomedEng.2024053351","article-title":"Channel-specific and spatial residual attention network for medical image denoising","volume":"52","author":"Hu","year":"2024","journal-title":"Crit. Rev. Biomed. Eng."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"3099","DOI":"10.62347\/WKNS8633","article-title":"Quality assessment of abdominal CT images: An improved ResNet algorithm with dual-attention mechanism","volume":"16","author":"Zhu","year":"2024","journal-title":"Am. J. Transl. Res."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"107660","DOI":"10.1016\/j.cmpb.2023.107660","article-title":"ResNet and its application to medical image processing: Research progress and challenges","volume":"240","author":"Xu","year":"2023","journal-title":"Comput. Methods Programs Biomed."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"102313","DOI":"10.1016\/j.media.2021.102313","article-title":"ResGANet: Residual group attention network for medical image classification and segmentation","volume":"76","author":"Cheng","year":"2022","journal-title":"Med. Image Anal."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"1072","DOI":"10.1002\/(SICI)1522-2594(199912)42:6<1072::AID-MRM11>3.0.CO;2-M","article-title":"On standardizing the MR image intensity scale","volume":"42","author":"Nyul","year":"1999","journal-title":"Magn. Reson. Med."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"1310","DOI":"10.1109\/TMI.2010.2046908","article-title":"N4ITK: Improved N3 bias correction","volume":"29","author":"Tustison","year":"2010","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"663","DOI":"10.1148\/radiol.12110748","article-title":"Locally advanced breast cancer: MR imaging for prediction of response to neoadjuvant chemotherapy\u2014Results from ACRIN 6657\/I-SPY TRIAL","volume":"263","author":"Hylton","year":"2012","journal-title":"Radiology"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"1045","DOI":"10.1007\/s10278-013-9622-7","article-title":"The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository","volume":"26","author":"Clark","year":"2013","journal-title":"J. Digit. Imaging"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"303","DOI":"10.1007\/s12609-019-00334-2","article-title":"I-SPY 2: A neoadjuvant adaptive clinical trial designed to improve outcomes in high-risk breast cancer","volume":"11","author":"Wang","year":"2019","journal-title":"Curr. Breast Cancer Rep."},{"key":"ref_24","unstructured":"Fridman, N., and Goldstein, A. (2025). Transformer classification of breast lesions: The BreastDCEDL_AMBL benchmark dataset and 0.92 AUC baseline. arXiv."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1352","DOI":"10.1016\/j.acra.2020.05.040","article-title":"Radiomics signatures based on multiparametric MRI for the preoperative prediction of the HER2 status of patients with breast cancer","volume":"28","author":"Zhou","year":"2021","journal-title":"Acad. Radiol."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Xu, A., Chu, X., Zhang, S., Zheng, J., Shi, D., Lv, S., Li, F., and Weng, X. (2022). Development and validation of a clinicoradiomic nomogram to assess the HER2 status of patients with invasive ductal carcinoma. BMC Cancer, 22.","DOI":"10.1186\/s12885-022-09967-6"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"16012","DOI":"10.1038\/npjbcancer.2016.12","article-title":"Quantitative MRI radiomics in the prediction of molecular classifications of breast cancer subtypes in the TCGA\/TCIA data set","volume":"2","author":"Li","year":"2016","journal-title":"NPJ Breast Cancer"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"5162","DOI":"10.1002\/mp.12453","article-title":"A deep feature fusion methodology for breast cancer diagnosis demonstrated on three imaging modality datasets","volume":"44","author":"Antropova","year":"2017","journal-title":"Med. Phys."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"7269","DOI":"10.1002\/mp.17266","article-title":"Identifying radiogenomic associations of breast cancer based on DCE-MRI by using Siamese Neural Network with manufacturer bias normalization","volume":"51","author":"Chen","year":"2024","journal-title":"Med. Phys."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"276","DOI":"10.1007\/s10278-019-00179-2","article-title":"Predicting breast cancer molecular subtype with MRI dataset utilizing convolutional neural network algorithm","volume":"32","author":"Ha","year":"2019","journal-title":"J. Digit. Imaging"},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"32","DOI":"10.1007\/s11307-025-01981-x","article-title":"Deep radiogenomics sequencing for breast tumor gene-phenotype decoding using dynamic contrast magnetic resonance imaging","volume":"27","author":"Shiri","year":"2025","journal-title":"Mol. Imaging Biol."},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"e03925","DOI":"10.1002\/advs.202503925","article-title":"A dynamic contrast-enhanced MRI-based Vision Transformer model for distinguishing HER2-zero, -low, and -positive expression in breast cancer and exploring model interpretability","volume":"12","author":"Zhang","year":"2025","journal-title":"Adv. Sci."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1186\/s13058-025-02118-2","article-title":"AI-driven MRI biomarker for triple-class HER2 expression classification in breast cancer: A large-scale multicenter study","volume":"27","author":"Wong","year":"2025","journal-title":"Breast Cancer Res."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"84","DOI":"10.1007\/s10916-024-02105-8","article-title":"Comparison of vision transformers and convolutional neural networks in medical image analysis: A systematic review","volume":"48","author":"Takahashi","year":"2024","journal-title":"J. Med. Syst."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"3248","DOI":"10.1007\/s10278-024-01322-4","article-title":"Systematic review of hybrid vision transformer architectures for radiological image analysis","volume":"38","author":"Kim","year":"2025","journal-title":"J. Imaging Inform. Med."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"87","DOI":"10.1109\/42.668698","article-title":"A nonparametric method for automatic correction of intensity nonuniformity in MRI data","volume":"17","author":"Sled","year":"1998","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"105093","DOI":"10.1016\/j.compbiomed.2021.105093","article-title":"A U-Net ensemble for breast lesion segmentation in DCE MRI","volume":"140","author":"Khaled","year":"2022","journal-title":"Comput. Biol. Med."},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S., and Sun, J. (2015, January 7\u201313). Delving deep into rectifiers: Surpassing human-level performance on ImageNet classification. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Santiago, Chile.","DOI":"10.1109\/ICCV.2015.123"},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach","volume":"44","author":"DeLong","year":"1988","journal-title":"Biometrics"}],"container-title":["Bioengineering"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2306-5354\/13\/7\/788\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T14:37:47Z","timestamp":1783607867000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2306-5354\/13\/7\/788"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,8]]},"references-count":39,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2026,7]]}},"alternative-id":["bioengineering13070788"],"URL":"https:\/\/doi.org\/10.3390\/bioengineering13070788","relation":{},"ISSN":["2306-5354"],"issn-type":[{"value":"2306-5354","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,8]]}}}