Toward Generalizability in the Deployment of Artificial Intelligence in Radiology: Role of Computation Stress Testing to Overcome Underspecification
Open accessThe DOI matches its Crossref record; 76 of the 77 references it cites were found in a registry or an index.
Abstract
The clinical deployment of artificial intelligence (AI) applications in medical imaging is perhaps the greatest challenge facing radiology in the next decade. One of the main obstacles to the incorporation of automated AI-based decision-making tools in medicine is the failure of models to generalize when deployed across institutions with heterogeneous populations and imaging protocols. The most well-understood pitfall in developing these AI models is overfitting, which has, in part, been overcome by optimizing training protocols. However, overfitting is not the only obstacle to the success and generalizability of AI. Underspecification is also a serious impediment that requires conceptual understanding and correction. It is well known that a single AI pipeline, with prescribed training and testing sets, can produce several models with various levels of generalizability. Underspecification defines the inability of the pipeline to identify whether these models have embedded the structure of the underlying system by using a test set independent of, but distributed identically, to the training set. An underspecified pipeline is unable to assess the degree to which the models will be generalizable. Stress testing is a known tool in AI that can limit underspecification and, importantly, assure broad generalizability of AI models. However, the application of stress tests is new in radiologic applications. This report describes the concept of underspecification from a radiologist perspective, discusses stress testing as a specific strategy to overcome underspecification, and explains how stress tests could be designed in radiology—by modifying medical images or stratifying testing datasets. In the upcoming years, stress tests should become in radiology the standard that crash tests have become in the automotive industry. Keywords: Computer Applications-General, Informatics, Computer-aided Diagnosis © RSNA, 2021
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“A critical barrier to the clinical implementation of AI decision tools is their frequent failure to generalize across heterogeneous populations and imaging protocols (Eche et al., 2021).”
From Multimodal neuroimaging and AI integration in cognitive disorders: advances, challenges, and future directions for precision medicine (2026) · cites this paper as Eche et al., 2021
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Lecture Notes in Computer Science 1205
(1999) · Industrial Robot: An International Journal · doi:10.1108/ir.1999.04926fae.001 · Cited by 38,662
Federated Machine Learning: Concept and Applications
Yang, Q., Liu, Y., Chen, T., & Tong, Y. (2019) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3298981 · Cited by 6,322
Radiomics: the bridge between medical imaging and personalized medicine
Lambin, P., Leijenaar, R. T. H., Deist, T. M., Peerlings, J., de Jong, E. E. C., van Timmeren, J., Sanduleanu, S., Larue, R. T. H. M., Even, A. J. G., Jochems, A., et al. (2017) · Nature Reviews Clinical Oncology · doi:10.1038/nrclinonc.2017.141 · Cited by 5,688
Zwanenburg, A., Vallières, M., Abdalah, M. A., Aerts, H. J. W. L., Andrearczyk, V., Apte, A., Ashrafinia, S., Bakas, S., Beukinga, R. J., Boellaard, R., et al. (2020) · Radiology · doi:10.1148/radiol.2020191145 · Cited by 4,177
Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2016) · arXiv · doi:10.48550/arxiv.1612.01474 · Cited by 3,666
Neural Networks and the Bias/Variance Dilemma
Geman, S., Bienenstock, E., & Doursat, R. (1992) · Neural Computation · doi:10.1162/neco.1992.4.1.1 · Cited by 3,578
Shortcut learning in deep neural networks
Geirhos, R., Jacobsen, J.-H., Michaelis, C., Zemel, R., Brendel, W., Bethge, M., & Wichmann, F. A. (2020) · Nature Machine Intelligence · doi:10.1038/s42256-020-00257-z · Cited by 2,224
Li, L., Qin, L., Xu, Z., Yin, Y., Wang, X., Kong, B., Bai, J., Lu, Y., Fang, Z., Song, Q., et al. (2020) · Radiology · doi:10.1148/radiol.2020200905 · Cited by 1,927
Neural Architecture Search: A Survey
Elsken, T., Metzen, J. H., & Hutter, F. (2018) · arXiv · doi:10.48550/arxiv.1808.05377 · Cited by 1,763
Preparing Medical Imaging Data for Machine Learning
Willemink, M. J., Koszek, W. A., Hardell, C., Wu, J., Fleischmann, D., Harvey, H., Folio, L. R., Summers, R. M., Rubin, D. L., & Lungren, M. P. (2020) · Radiology · doi:10.1148/radiol.2020192224 · Cited by 1,020
Research works printed but not linked: 2
- 73. Belkin M, Hsu D, Ma S, Mandal S. Reconciling modern machine learning practice and the bias-variance trade-off. ArXiv 1812.11118 [preprint] https:// arxiv.org/abs/1812.11118. Posted December 28, 2018. Accessed March 2021. · arXiv
- 77. Nakkiran P, Kaplun G, Bansal Y, Yang T, Barak B, Sutskever I. Deep double descent: where bigger models and more data hurt. ArXiv 1912.02292 [preprint] https://arxiv.org/abs/1912.02292. Posted December 4, 2019. Accessed March 2021. · arXiv
Cited by
Items 101–181 of 181
von Reppert, M., Chadha, S., Willms, K., Avesta, A., Maleki, N., Zeevi, T., Lost, J., Tillmanns, N., Jekel, L., Merkaj, S., et al. (2025) · American Journal of Neuroradiology · doi:10.3174/ajnr.a8805 · Cited by 6
Purkayastha, S., Shalu, H., Gutman, D. A., Holodny, A. I., Modak, S. I., Basu, E. M., Kushner, B. H., Kramer, K., Haque, S. S., & Stember, J. N. (2024) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-024-01165-z · Cited by 5
“Under-specification [6, 7] means multiple models can fit IID data.”
InfluentialA neural network to create super‐resolution MR from multiple 2D brain scans of pediatric patients
Benitez-Aurioles, J., Osorio, E. M. V., Aznar, M. C., van Herk, M., Pan, S., Sitch, P., France, A., Smith, E., & Davey, A. (2024) · Medical Physics · doi:10.1002/mp.17563 · Cited by 4
Shahrokhi, S., Oladosu, O., Tariq, R., & Zhang, Y. (2026) · Frontiers in Neuroinformatics · doi:10.3389/fninf.2026.1762794 · Cited by 4
“Routine MRI is typically more heterogeneous than streamlined datasets in both scanner type and acquisition parameters (Eche et al., 2021), requiring extra attention.”
Makowska, A., Ananthakrishnan, G., Christ, M., & Dehmer, M. (2025) · Healthcare · doi:10.3390/healthcare13040408 · Cited by 5
Agosti, E., Cuocolo, R., Mangili, M., Rampinelli, V., Veiceschi, P., Cappelletti, M., Panciani, P. P., Piazza, A., Bove, I., Solari, D., et al. (2025) · Journal of Neurological Surgery Part B Skull Base · doi:10.1055/a-2607-0735 · Cited by 4
Xu, G., Udupa, J. K., Yu, Y., Shao, H., Zhao, S., Liu, W., & Zhang, Y. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.133439 · Cited by 5
The Role of Health Data Utilities in Supporting Health AI.
Alkasir, A., Seidman, G., Ritzo, J., Bari, L., & Khurshid, A. (2024) · NEJM AI · doi:10.1056/aipc2400401 · Cited by 4
A modern radiologist’s guide to artificial intelligence
Kapur, J., Kelly, B. S., & Vega, R. (2026) · Pediatric Radiology · doi:10.1007/s00247-026-06542-0 · Cited by 3
Wearable Flexible Sensors for Cardiovascular Disease Monitoring
Xie, X., Qu, X., Zhou, B., Zhang, B., Wang, Q., Bian, H., Shao, J., Cai, Y., & Dong, X. (2026) · Advanced Materials · doi:10.1002/adma.202512939 · Cited by 9
unORANIC: Unsupervised Orthogonalization of Anatomy and Image-Characteristic Features
Doerrich, S., Salvo, F. D., & Ledig, C. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-45673-2_7 · Cited by 5
Кримовський, К., Мілещенко, А., & Бричко, Т. (2024) · Archive of Ophthalmology and Maxillofacial Surgery of Ukraine · doi:10.22141/aomfs.1.1.2024.9 · Cited by 3
Accelerating Cerebral Diagnostics with BrainFusion: A Comprehensive MRI Tumor Framework
Houmaidi, W., Sabiri, Y., Billah, S. E. M., & Abouaomar, A. (2025) · International Conference on Wireless Networks and Mobile Communications (WINCOM) · doi:10.1109/wincom65874.2025.11313453 · Cited by 3
“Despite these remarkable advancements, challenges remain in ensuring model generalizability across diverse datasets and improving interpretability for clinical adoption [18].”
Paluh, J. L., Sunil, S., Rajeev, R., Chatterjee, A., Pilitsis, J. G., & Mukherjee, A. (2024) · Research Square · doi:10.21203/rs.3.rs-4087784/v1 · Cited by 4
Xiong, Z., Zhao, W., Tian, D., Zhang, J., He, Y., Qin, D., & Li, Z. (2023) · Journal of Cancer Research and Clinical Oncology · doi:10.1007/s00432-023-05105-2 · Cited by 2
“A study by Eche et al. (2021) further discussed stress test as a specific strategy to overcome the “underspecifica-tion” in radiomics research, which is a concept that hinders generalization (D'Amour et al. 2020).”
MethodRecent Advances in Artificial Intelligence for Dyslexia Detection: A Systematic Review
Pamungkas, Y., Rangkuti, R. Y., Karim, A., & Sangsawang, T. (2025) · International Journal of Robotics and Control Systems · doi:10.31763/ijrcs.v5i3.2057 · Cited by 3
The role of artificial intelligence in optimizing radiation dose in diagnostic imaging
Dehbaghi, H. A., & Khoshgard, K. (2026) · Health and Technology · doi:10.1007/s12553-025-01049-4 · Cited by 3
Ethical and Regulatory Consideration in AI-Based Medical Imaging
Nadeem, G., & Anis, M. I. (2025) · Springer eBooks · doi:10.1007/978-3-032-02963-8_10 · Cited by 4
On-Device Artificial Intelligence for Image Analysis in Resource-Limited Settings: A Review
Xie, C., An, J. E., Bae, H.-J., Yun, E. S., Ko, K.-C., Jeong, D. S., & Kwon, O. S. (2025) · Applied Science and Convergence Technology · doi:10.5757/asct.2026.35.1.1 · Cited by 2
Egbunu, A. S., & Okedoye, A. M. (2026) · Journal of Computing Theories and Applications · doi:10.62411/jcta.15367 · Cited by 3
Artificial intelligence and multiomics integration for Parkinson’s disease drug development
Lee, S., Kim, J., Jung, H., Kim, D., Cho, E., Kim, Y.-J., Ji, Y. R., Go, Y., Song, P., Oh, Y., et al. (2026) · Molecules and Cells · doi:10.1016/j.mocell.2026.100343 · Cited by 2
Syailendra, E. A., Arshad, H., Lopez‐Ramirez, F., Tixier, F., Kawamoto, S., Fishman, E. K., & Chu, L. C. H. (2026) · Korean Journal of Radiology · doi:10.3348/kjr.2026.0003 · Cited by 2
Dikici, E., Nguyen, X. V., Takacs, N., & Prevedello, L. M. S. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106901 · Cited by 3
“In their study, Eche et al. [16] identified the major causes of reduced generalizability in medical imaging systems as (1) overfitting (i.e., the AI-model learns unnecessary residual variations; information only applicable to training data [19]) and (2) underspecification (i.e., AI-model fails to…”
BackgroundDeep Learning and The Retina: A New Frontier in Multiple Sclerosis Diagnosis.
Abdul-Salam, S.-E., Florescu, R.-M., Sfredel, V., Alexandru, D., Şerbănescu, M.-S., & Rotaru-Zăvăleanu, A. D. (2025) · PubMed Central · doi:10.12865/chsj.51.01.03 · Cited by 3
“To address this limitation and improve model robustness, it is essential to work together to build large, diverse, and multicenter datasets [61].”
Assessing Outcomes in NSCLC: Radiomic analysis, kinetic analysis and circulating tumor DNA
Maniar, A. B., Wei, A. Z., Dercle, L., Bien, H., Fojo, T., Bates, S. E., & Schwartz, L. H. (2022) · Seminars in Oncology · doi:10.1053/j.seminoncol.2022.06.002 · Cited by 2
Self-citationArtificial Intelligence in Body Imaging: An Overview of Commercially Available Tools
Perchik, J. D., Rothenberg, S., Tridandapani, S., Perchik, L. M., & Smith, A. D. (2023) · Seminars in Roentgenology · doi:10.1053/j.ro.2023.02.007 · Cited by 2
Artificial intelligence in general internal medicine
Sherbini, A. E., Glicksberg, B. S., & Krittanawong, C. (2023) · Elsevier eBooks · doi:10.1016/b978-0-443-15688-5.00025-5 · Cited by 2
Detecting Interspecific Positive Selection Using Convolutional Neural Networks
West, C., Walker, C. R., Arasti, S., Vasilev, V., Xu, X., De Maio, N., & Goldman, N. (2025) · Molecular Biology and Evolution · doi:10.1093/molbev/msaf154 · Cited by 1
“…and Lempitsky 2015), there is a common struggle between creating models which perform well at a specific task but fail to generalize outside of it, versus models that offer wide general use but whose performance falls short for any specific subgroup (Eche et al. 2021; Huisman and Hannink 2023).”
Desiraju, B. K., Thiruvengadam, R., Wadhwa, N., Khurana, A., Papageorghiou, A. T., Noble, J. A., & Bhatnagar, S. (2024) · F1000Research · doi:10.12688/f1000research.152426.1 · Cited by 1
Predicting the generalization of computer aided detection (CADe) models for colonoscopy
Shor, J., McNeil, C., Intrator, Y., Ledsam, J. R., Yamano, H., Tsurumaru, D., Kayama, H., Hamabe, A., Ando, K., Ota, M., et al. (2024) · Discover Artificial Intelligence · doi:10.1007/s44163-024-00187-4 · Cited by 2
Shih, Y.-C., Ko, C., Wang, S.-Y., Chang, C.-Y., Lin, S.-S., Huang, C.-W., Cheng, M., Chen, C., & Wu, Y. (2025) · European Journal of Nuclear Medicine and Molecular Imaging · doi:10.1007/s00259-025-07243-w · Cited by 2
“To mitigate the risk of overfitting, besides selecting statistical artificial learning algorithms, the most reasonable and feasible approach is to include larger and more diverse samples from multiple centers in the training sets [26].”
Tulimilli, S. V., Karnik, M., Cockroft, S., Bettadapura, A. D. S., Nataraj, S. M., Mohideen, H. S., Dovat, S., Sharma, A., & Madhunapantula, S. V. (2026) · Oncology Reviews · doi:10.3389/or.2026.1760156 · Cited by 2
“AI models often perform well on the specific datasets they were trained and may not generalize well when applied to new datasets in different populations (103).”
Sung, C., Oh, J. S., Chae, S. Y., Lee, D. Y., Oh, M., Lee, S. S., Oh, S. J., Jo, S., Chung, S. J., & Kim, J. S. (2026) · PLoS ONE · doi:10.1371/journal.pone.0349375 · Cited by 1
Statie, M. D., Nieri, M., Rutili, V., Buffoni, L., Chicchi, L., Auconi, P., Giuntini, V., McNamara, J. A., & Franchi, L. (2026) · Orthodontics and Craniofacial Research · doi:10.1111/ocr.70142 · Cited by 1
Takala, J., Zanier, O., Blom, V., Kälin, V., Höbner, L. M., Klesel, C. M., Hostettler, I. C., Staartjes, V. E., & Korja, M. (2026) · Scientific Reports · doi:10.1038/s41598-026-54488-y · Cited by 1
Airway segmentation on CT – A systematic review of machine learning tools
Lsloum, N., Maiter, A., Alnasser, T. N., Albylwi, A., Alghamdi, K. A., Sharkey, M. J., Hokmabadi, A., Salehi, M., Dwivedi, K., Johns, C., et al. (2026) · European Journal of Radiology Open · doi:10.1016/j.ejro.2026.100764 · Cited by 1
“A key determinant of clinical applicability for AI-based airway segmentation models is their ability to generalise across different patient populations, imaging conditions, and healthcare settings [57].”
Shomurodov, K., Mirkhusanova, R., Cahyanto, A., Хуршидовна, Ш. Г., Abdukhalik-Zade, N., Рашидов, Р. А. Ў., Rafikov, K., & Tuygunov, N. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01883-z · Cited by 1
He, Z., Zhu, T., Luo, X., Qiu, Y., Zhang, M., Xu, X., & Guo, Q. (2026) · Journal of Artificial Intelligence and Information · doi:10.66069/ojspub.27450843 · Cited by 1
“Comparative evaluation should address clinical effectiveness, after which post-deployment monitoring should cover data and performance drift, subgroup differences, user response, and actual quality outcomes [2,4,39 – 41].”
Lin, X.-X., Qian, Y., Zhou, S.-H., Han-bin, S., Zhang, Y., Huang, J.-W., Chen, Y., & Tian, N.-F. (2026) · Radiology · doi:10.1148/radiol.253931 · Cited by 1
Bone Age Prediction under Stress
Faghani, S., & Erickson, B. J. (2024) · Radiology Artificial Intelligence · doi:10.1148/ryai.240137 · Cited by 1
Advances in Immunotherapy and Efficacy Prediction for Gastric Cancer
陈, 进. (2024) · Advances in Clinical Medicine · doi:10.12677/acm.2024.1492556 · Cited by 1
Commentary: Artificial intelligence-based screening of retina
Bansal, M., & Rangarajan, K. (2022) · Indian Journal of Ophthalmology · doi:10.4103/ijo.ijo_3097_21 · Cited by 1
“In this context, it is important that the model be generalizable, and stress testing to limit underspecification is important in evaluating the AI algorithm to be deployed.[3] Second, a strong referral system needs to be built that is capable of treating patients once detected, without a significant lag.”
BackgroundSIENNA: Generalizable Lightweight Machine Learning Platform for Brain Tumor DiagnosticsPreprint
Sunil, S., Rajeev, R. S., Chatterjee, A., Pilitsis, J. G., Mukherjee, A., & Paluh, J. L. (2024) · medRxiv · doi:10.1101/2024.04.03.24305210 · Cited by 1
“This was sufficient to reduce spatial intensity differences between skull and/or brain tissues in these regions that can mislead the model into learning incorrect patterns that can result in overfitting irrelevant features [72].”
BackgroundSutton, C., Prowse, J., Elshehaly, M., & Randell, R. (2025) · medRxiv · doi:10.1101/2025.07.24.25332132 · Cited by 2
Detecting interspecific positive selection using convolutional neural networks
West, C., Walker, C. R., Arasti, S., Vasilev, V., Xu, X., De Maio, N., & Goldman, N. (2024) · doi:10.1101/2024.11.18.624117 · Cited by 1
Personalizing endocrine therapy for breast cancer using MRI
Ragusi, M. A. A.-D. D. (2022) · Utrecht University Repository (Utrecht University) · doi:10.33540/1248
Rudzite, L. (2023) · International Comparative Jurisprudence · doi:10.13165/j.icj.2023.06.010
Multi-Modality Knowledge Transfer for Generalizable Chest Radiograph Interpretation
Hou, X., Li, Y., Mei, Y., Liu, D., & Wang, S. (2024) · International Conference on Biomedical and Bioinformatics Engineering (ICBBE) · doi:10.1145/3707127.3707128
Revolutionizing stock market trading with artificial intelligence
Kotte, G. (2025) · INTERNATIONAL JOURNAL OF CREATIVE RESEARCH THOUGHTS · doi:10.56975/ijcrt.v13i2.277612
Gupta, D. (2025) · IEEE Conference on Artificial Intelligence (CAI) · doi:10.1109/cai64502.2025.00089
“There is also the challenge of overfitting, where models trained on limited data fail to generalize across diverse clinical scenarios [30].”
Mbasso, W. F. (2025) · Jordan Journal of Applied Science - Natural Science Series · doi:10.35192/jjoas-n.v19i1.2218
Cuocolo, R., Stanzione, A., Ponsiglione, A., Ugga, L., & Koçak, B. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-29242-2.00012-4
The critical role of artificial intelligence in biomaterials design and development
Zarrintaj, P., Yazdi, M. K., Bencherif, S. A., Saeb, M. R., & Mozafari, M. (2025) · Elsevier eBooks · doi:10.1016/b978-0-323-95464-8.00015-3
Rekinen, S. (2023) · Aaltodoc (Aalto University)
Rudzite-Celmina, L. (2025) · DSpace repository (University of Tartu)
Interpretable and Automated Bias Detection for AI in Healthcare
Alexiev, C. (2024) · DSpace@MIT (Massachusetts Institute of Technology) · Cited by 2
Song, L. (n.d.) · PolyU Institutional Research Archive (Hong Kong Polytechnic University)
Sreejith, R., Ramasamy, R. K., Mohd-Isa, W.-N., & Abdullah, J. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-96-9048-0_39
Anjankar, A. P., Basheeruddin, M., Dhok, A., Lambe, S., Kushwaha, A. K., & Ingole, R. P. (2025) · DMIHER International Conference on Artificial Intelligence in Healthcare, Education and Industry (IDICAIHEI) · doi:10.1109/idicaihei65991.2025.11378808
“[19], [20] V.”
Zeng, Y., & Sun, P. (2026) · Neuro-Oncology · doi:10.1093/neuonc/noag046
Comparison of Deep Learning Architectures for Automated Breast Ultrasound Report Generation
Khatoon, S., Mahdaoui, A., & Mahmood, A. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-3-032-20318-2_6
Asghar, F., Rehman, R. U., & Ali, M. H. (2026) · Journal of Gastroenterology and Hepatology · doi:10.1111/jgh.70382
“Firstly, the algorithm should undergo rigorous external, multicenter validation using independent data-sets to eliminate overfitting and confirm reproducibility across different equipment, operators, and populations [2].”
郑, 家. (2026) · Advances in Clinical Medicine · doi:10.12677/acm.2026.1641644
Clinical AI/ML Model Testing and Validation
Pandey, A., Gupta, P., & Sehgal, N. K. (2026) · Springer eBooks · doi:10.1007/978-3-032-10808-1_6
Moassefi, M., Decker, P. A., Conte, G. M., Kosel, M., Molinaro, A. M., Niederschweiberer, M. A., Nikanpour, Y., Ruff, M. W., Burns, T. C., Farooq, U., et al. (2026) · Cancer Research Communications · doi:10.1158/2767-9764.crc-25-0710
Artificial intelligence in ovarian cancer: advancing in precision diagnosis and clinical management
Shao, M., Wang, T., Ji, L., Xu, L., Zhang, Y., Wang, D. Q., Jia, C., Chen, L., Zhang, H., Yan, W., et al. (2026) · Frontiers in Immunology · doi:10.3389/fimmu.2026.1795144
Mau, B., Garg, S. K., Harper, S. B., Gomes, R., Rolli, A. C., & Chaudhry, R. (2026) · Mayo Clinic Proceedings Digital Health · doi:10.1016/j.mcpdig.2026.100372
“17 The effectiveness of our approach is closely tied to the quality, diversity, and domain alignment of the training dataset.”
Eriksson, M., Czene, K., Scott, C. G., Stoddard, S. M., Smith, H., Hall, P. N., & Vachon, C. M. (2026) · Science Translational Medicine · doi:10.1126/scitranslmed.ady7414
Roßkopf, J., Yazilitas, A., Elfeel, S., Baumgartner, T. J., Franz, A. M., Mueller, R., Ertl, M., Schmitz, B., & Althaus, K. (2026) · Neuroradiology · doi:10.1007/s00234-026-04046-0
“Overfitting remains a relevant concern for AI systems in radiology [26 ].”
Seghier, M. L. (2026) · International Journal of Imaging Systems and Technology · doi:10.1002/ima.70393
Chen, Y., Chen, X., Yue, B., Zhong, X., Zhang, H., Chen, X., Chen, X., Dai, Z., & Yang, Z. (2026) · Frontiers in Oncology · doi:10.3389/fonc.2026.1824808
“Although radiomics may extract additional information (20, 35, 39), concerns regarding feature stability, reproducibility, and clinical interpretability persist (17, 38).”
Guellec, B. L., & Chassagnon, G. (2026) · Radiology Artificial Intelligence · doi:10.1148/ryai.260590
Khorram‐Manesh, A., & Carlström, E. (2026) · BMC Emergency Medicine · doi:10.1186/s12873-026-01699-1
“This broader approach reinforces that disaster medicine intersects with emergency medicine and public health through shared concerns such as surge sustainability, equity in tri-age, population risk management, and cross-sector coordination [3, 54–57].”
Clinical application of artificial intelligence in corneal transplantation
Rad, N. R. (2026) · Medical Hypothesis Discovery & Innovation in Optometry · doi:10.51329/mehdioptometry249
“Smaller datasets may increase the risk of overfitting and limit generalizability [44–46], whereas larger datasets support more robust model training, validation, and external applicability [47, 48].”
Sugawara, H., Kato, S., & Takada, A. (2026) · Skeletal Radiology · doi:10.1007/s00256-026-05335-8
“Furthermore, mitigating the risk of overfitting and ensuring robust generalizability demand large validation cohorts [4].”
The promise of quantitative approaches to computed tomography imaging in pulmonary sarcoidosis
Lippitt, W., & Carlson, N. E. (2026) · Current Opinion in Pulmonary Medicine · doi:10.1097/mcp.0000000000001301
Rashidisabet, H., Chan, R. V. P., Vajaranant, T. S., & Yi, D. (2026) · Translational Vision Science & Technology · doi:10.1167/tvst.15.9.3
“19 For example, fundus images collected at two different hospitals may represent distinct domains, often referred to as a source domain (used for training) and a target domain (used for deployment), due to differences in camera types, imaging workflows, and population characteristics.”
DWI/FLAIR Mismatch An Automated, Continuous & Clinically Viable Approach
Johansen, J. (2026) · Research at the University of Copenhagen (University of Copenhagen)
Wang, L., Shi, Y., Deng, Z., Wang, Y., Righi, D. R. D., Javed, S., Qureshi, T. A., Yang, J. D., Xie, Y., Pandol, S. J., et al. (2026) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2026.103535
Artificial intelligence in medical image processing and translational science
Kashyap, K., & Zujam, D. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-40338-5.00004-4
Imaging as a Surgical Decision System in Modern Gynecologic Oncology
Yordanov, A. D., Dimitrova, V., Kostov, S. G., Hasan, I., Shemeri, S. E., & Tsoneva, E. (2026) · Diagnostics · doi:10.3390/diagnostics16213666
Metadata
- Authors
- Eche, T., Schwartz, L. H., Mokrane, F.-Z., & Dercle, L.
- Affiliations
- NewYork–Presbyterian Hospital (US)
- Venue
- Radiology: Artificial Intelligence
- Published
- 2021-11-01
- Volume
- 3
- Issue
- 6
- Pages
- e210097-e210097
- ISSN
- 2638-6100 (ISSN-L)
- PMC
- PMC8637230
- OpenAlex
- W3210784269
- Cited by
- 163
- References
- 71
- Citations by year
- 2026: 33 · 2025: 40 · 2024: 47 · 2023: 30 · 2022: 13
- FWCI
- 6.56
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Locations
- Radiology Artificial Intelligence (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Funders
- Alain Rahmouni
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiology practices and education
- Keywords
- underspecification, stress testing, generalizability, AI deployment, diagnostic imaging, overfitting, model evaluation, test set design, artificial intelligence in radiology, computer-aided diagnosis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.