Predicting patient demographics from chest Radiographs with deep learning
Abstract
Background Deep learning models are increasingly informing medical decision making, for instance, in the detection of acute intracranial hemorrhage and pulmonary embolism. However, many models are trained on medical image databases that poorly represent the diversity of the patients they serve. In turn, many artificial intelligence models may not perform as well on assisting providers with important medical decisions for underrepresented populations.Purpose Assessment of the ability of deep learning models to classify the self-reported gender, age, self-reported ethnicity, and insurance status of an individual patient from a given chest radiograph.Methods Models were trained and tested with 55,174 radiographs in the MIMIC Chest X-ray (MIMIC-CXR) database. External validation data came from two separate databases, one from CheXpert and another from a multihospital urban health care system after institutional review board approval. Macro-averaged area under the curve (AUC) values were used to evaluate performance of models. Code used for this study is open-source and available at https://github.com/ai-bias/cxr-bias, and pixelstopatients.com/models/demographics.Results Accuracy of models to predict gender was nearly perfect, with 0.999 (95% confidence interval: 0.99-0.99) AUC on held-out test data and 0.994 (0.99-0.99) and 0.997 (0.99-0.99) on external validation data. There was high accuracy to predict age and ethnicity, ranging from 0.854 (0.80-0.91) to 0.911 (0.88-0.94) AUC, and moderate accuracy to predict insurance status, with AUC ranging from 0.705 (0.60-0.81) on held-out test data to 0.675 (0.54-0.79) on external validation data.Conclusions Deep learning models can predict the age, self-reported gender, self-reported ethnicity, and insurance status of a patient from a chest radiograph. Visualization techniques are useful to ensure deep learning models function as intended and to demonstrate anatomical regions of interest. These models can be used to ensure that training data are diverse, thereby ensuring artificial intelligence models that work on diverse populations.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
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?
Information on social determinants of health like area deprivation index are learned and encoded on chest X-rays.
“More concerning about these models is that information on social determinants of health like area deprivation index are learned and encoded on the CXRs.”
From AI pitfalls and what not to do: mitigating bias in AI (2023) · cites this paper as [29]
Chest X-rays can be used to predict demographic information including insurance, self-reported race, and age.
“CXRs can also be used to predict demographic information—including insurance, self-reported race, and age. 28,29”
From AI pitfalls and what not to do: mitigating bias in AI (2023) · cites this paper as [29]
Cites
Goldberger, A. L., Amaral, L. A. N., Glass, L., Hausdorff, J. M., Ivanov, P. C., Mark, R. G., Mietus, J. E., Moody, G. B., Peng, C.-K., & Stanley, H. E. (2000) · Circulation · doi:10.1161/01.cir.101.23.e215 · Cited by 14,809 · see citing works
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., et al. (2020) · Information Fusion · doi:10.1016/j.inffus.2019.12.012 · Cited by 10,423 · see citing works
Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning
Poplin, R., Varadarajan, A. V., Blumer, K., Liu, Y., McConnell, M. V., Corrado, G. S., Peng, L., & Webster, D. R. (2018) · Nature Biomedical Engineering · doi:10.1038/s41551-018-0195-0 · Cited by 1,865 · see citing works
MIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports
Johnson, A. E. W., Pollard, T. J., Berkowitz, S. J., Greenbaum, N. R., Lungren, M. P., Deng, C.-Y., Mark, R. G., & Horng, S. (2019) · Scientific Data · doi:10.1038/s41597-019-0322-0 · Cited by 1,816 · see citing works
The false hope of current approaches to explainable artificial intelligence in health care
Ghassemi, M., Oakden-Rayner, L., & Beam, A. L. (2021) · The Lancet Digital Health · doi:10.1016/s2589-7500(21)00208-9 · Cited by 1,724 · see citing works
Zech, J. R., Badgeley, M. A., Liu, M., Costa, A. B., Titano, J. J., & Oermann, E. K. (2018) · PLOS Medicine · doi:10.1371/journal.pmed.1002683 · Cited by 1,719 · see citing works
Receiver Operating Characteristic Curves and Their Use in Radiology
Obuchowski, N. A. (2003) · Radiology · doi:10.1148/radiol.2291010898 · Cited by 843 · see citing works
Meerwijk, E. L., & Sevelius, J. M. (2017) · American Journal of Public Health · doi:10.2105/ajph.2016.303578 · Cited by 800 · see citing works
Pisano, E. D., Zong, S., Hemminger, B. M., DeLuca, M. C., Johnston, R. E., Muller, K. E., Braeuning, M. P., & Pizer, S. M. (1998) · Journal of Imaging Informatics in Medicine · doi:10.1007/bf03178082 · Cited by 773 · see citing works
Machine Learning and Health Care Disparities in Dermatology
Adamson, A. S., & Smith, A. (2018) · JAMA Dermatology · doi:10.1001/jamadermatol.2018.2348 · Cited by 639 · see citing works
The concept of race and health status in America.
Williams, D. R., Lavizzo-Mourey, R., & Warren, R. C. (1994) · PubMed Central · doi:10.13016/ltkr-rams · Cited by 533 · see citing works
Face Recognition Performance: Role of Demographic Information
Klare, B. F., Burge, M. J., Klontz, J. C., Vorder Bruegge, R. W., & Jain, A. K. (2012) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2012.2214212 · Cited by 516 · see citing works
Assessing gender bias in machine translation: a case study with Google Translate
Prates, M. O. R., Avelar, P. H., & Lamb, L. C. (2020) · Neural Computing and Applications · doi:10.1007/s00521-019-04144-6 · Cited by 410 · see citing works
CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparison
Irvin, J., Rajpurkar, P., Ko, M., Yu, Y., Ciurea-Ilcus, S., Chute, C., Marklund, H., Haghgoo, B., Ball, R., Shpanskaya, K., et al. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.3301590 · Cited by 392 · see citing works
How common is lntersex? A response to Anne Fausto‐Sterling
Sax, L. (2002) · The Journal of Sex Research · doi:10.1080/00224490209552139 · Cited by 347 · see citing works
CheXclusion: Fairness gaps in deep chest X-ray classifiers
Seyyed-Kalantari, L., Liu, G., McDermott, M., Chen, I. Y., & Ghassemi, M. (2020) · Biocomputing 2021 · doi:10.1142/9789811232701_0022 · Cited by 275 · see citing works
An algorithmic approach to reducing unexplained pain disparities in underserved populations
Pierson, E., Cutler, D. M., Leskovec, J., Mullainathan, S., & Obermeyer, Z. (2021) · Nature Medicine · doi:10.1038/s41591-020-01192-7 · Cited by 268 · see citing works
Flores, A. W., Bechtel, K., & Lowenkamp, C. T. (2016) · Federal probation · Cited by 244 · see citing works
Geographic Distribution of US Cohorts Used to Train Deep Learning Algorithms
Kaushal, A., Altman, R., & Langlotz, C. (2020) · JAMA · doi:10.1001/jama.2020.12067 · Cited by 239 · see citing works
Understanding artificial intelligence based radiology studies: What is overfitting?
Mutasa, S., Sun, S., & Ha, R. (2020) · Clinical Imaging · doi:10.1016/j.clinimag.2020.04.025 · Cited by 216 · see citing works
Equity in essence: a call for operationalising fairness in machine learning for healthcare
Wawira Gichoya, J., McCoy, L. G., Celi, L. A., & Ghassemi, M. (2021) · BMJ Health & Care Informatics · doi:10.1136/bmjhci-2020-100289 · Cited by 104 · see citing works
Not all biases are bad: equitable and inequitable biases in machine learning and radiology
Pot, M., Kieusseyan, N., & Prainsack, B. (2021) · Insights into Imaging · doi:10.1186/s13244-020-00955-7 · Cited by 73 · see citing works
Agénor, M., Murchison, G. R., Najarro, J., Grimshaw, A. A., Cottrill, A. A., Janiak, E., Gordon, A. R., & Charlton, B. M. (2021) · Sexual and Reproductive Health Matters · doi:10.1080/26410397.2021.1886395 · Cited by 57 · see citing works
Chatterjee, A. R., Somayaji, N. R., & Kabakis, I. M. (2019) · Stroke · doi:10.1161/str.50.suppl_1.wmp16 · Cited by 52 · see citing works
Radiology “forensics”: determination of age and sex from chest radiographs using deep learning
Yi, P. H., Wei, J., Kim, T. K., Shin, J., Sair, H. I., Hui, F. K., Hager, G. D., & Lin, C. T. (2021) · Emergency Radiology · doi:10.1007/s10140-021-01953-y · Cited by 41 · see citing works
Transgender health: on a world scale
Connell, R. W. (2021) · Health Sociology Review · doi:10.1080/14461242.2020.1868899 · Cited by 34 · see citing works
Prevalence of Psychosocial Distress in Transmen Seeking Gender-Affirming Mastectomy
Lane, M. E., Kirsch, M. J., Sluiter, E. C., Hamill, J. B., Ives, G. C., Gilman, R. H., Cederna, P. S., Kuzon, W. M., & Wilkins, E. G. (2020) · Plastic & Reconstructive Surgery · doi:10.1097/prs.0000000000007357 · Cited by 23 · see citing works
Unequal medicine harms: reflections on the experiences of an intersex physician
Kramer, K. J., Runyan, A., Micks, E., Tamakuwala, S., Reid, M., Syed, S., Chao, C. R., & Recanati, M. (2021) · Sexual and Reproductive Health Matters · doi:10.1080/26410397.2021.1881211 · Cited by 3 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Impact of Artificial Intelligence on Dental Education: A Review and Guide for Curriculum Update
Thurzo, A., Strunga, M., Urban, R., Surovková, J., & Afrashtehfar, K. I. (2023) · Education Sciences · doi:10.3390/educsci13020150 · Cited by 345 · see citing works
“e have not yet fully recovered from the shock that an AI can recognize gender [42] and ethnicity [43] from an X-ray, nor from the fact that no one knows how the AI is capable of doing so, and here we are confronted with the possibility of semi-automatic communication of our office AI telemonitoring systems with the pa…”
BackgroundThe limits of fair medical imaging AI in real-world generalization
Yang, Y., Zhang, H., Gichoya, J. W., Katabi, D., & Ghassemi, M. (2024) · Nature Medicine · doi:10.1038/s41591-024-03113-4 · Cited by 285 · see citing works
“Recent studies have unveiled the surprising ability of deep models to predict demographic information, such as self-reported race 15 , sex and age 16 , from medical images, achieving performance far beyond that of radiologists.”
AI pitfalls and what not to do: mitigating bias in AI
Gichoya, J. W., Thomas, K., Celi, L. A., Safdar, N., Banerjee, I., Banja, J. D., Seyyed-Kalantari, L., Trivedi, H., & Purkayastha, S. (2023) · The British Journal of Radiology · doi:10.1259/bjr.20230023 · Cited by 277 · see citing works
“28,29 CXRs have also been used to predict biological age, cardiac disease outcomes and healthcare cost at 1, 3 and 5 years.”
Demographic bias in misdiagnosis by computational pathology models
Vaidya, A. J., Chen, R. J., Williamson, D. F. K., Song, A. H., Jaume, G., Yang, Y., Hartvigsen, T., Dyer, E., Lu, M. Y., Lipková, J., et al. (2024) · Nature Medicine · doi:10.1038/s41591-024-02885-z · Cited by 124 · see citing works
Marko, J. G. O., Neagu, D. C., & Anand, P. B. (2025) · BMC Medical Informatics and Decision Making · doi:10.1186/s12911-025-02884-1 · Cited by 120 · see citing works
“Studies show that AI models frequently rely on datasets that fail to reflect the diversity of global patient populations, particularly in areas like medical imaging [20–23].”
Risk of Bias in Chest Radiography Deep Learning Foundation Models
Glocker, B., Jones, C., Roschewitz, M., & Winzeck, S. (2023) · Radiology Artificial Intelligence · doi:10.1148/ryai.230060 · Cited by 90 · see citing works
“Use of foundation models in medical imaging may be of particular concern given the recently demonstrated ability of deep learning models to accurately recognize protected characteristics such as racial identity and other demographic information (13,14).”
BackgroundDemographic bias of expert-level vision-language foundation models in medical imaging
Yang, Y., Liu, Y., Liu, X., Gulhane, A., Mastrodicasa, D., Wu, W., Wang, E. J., Sahani, D. V., & Patel, S. (2025) · Science Advances · doi:10.1126/sciadv.adq0305 · Cited by 67 · see citing works
“Inspired by recent works on algorithmic encoding of demographic information by deep learning models 26–28 , we investigated whether the model encodes demographic information by examining the predictability of sensitive attributes by both the self-supervised foundation model and board-certified…”
Opportunistic detection of type 2 diabetes using deep learning from frontal chest radiographs
Pyrros, A. T., Borstelmann, S. M., Mantravadi, R., Zaiman, Z., Thomas, K., Price, B., Greenstein, E., Siddiqui, N., Willis, M., Shulhan, I., et al. (2023) · Nature Communications · doi:10.1038/s41467-023-39631-x · Cited by 60 · see citing works
“Previous studies have shown that convolutional neural networks can easily learn self-reported race and other sensitive attributes 23,24 .”
Colacci, M., Huang, Y. Q., Postill, G., Zhelnov, P., Fennelly, O., Verma, A. A., Straus, S. E., & Tricco, A. C. (2024) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2024.111606 · Cited by 56 · see citing works
Implications of predicting race variables from medical images
Zou, J., Gichoya, J. W., Ho, D. E., & Obermeyer, Z. (2023) · Science · doi:10.1126/science.adh4260 · Cited by 30 · see citing works
Patient Perspectives on Artificial Intelligence in Radiology
Kitts, A. B. (2023) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2023.05.017 · Cited by 27 · see citing works
Will AI Replace Physicians in the Near Future? AI Adoption Barriers in Medicine
Obuchowicz, R., Piórkowski, A., Nurzyǹska, K., Obuchowicz, B., Strzelecki, M., & Bielecka, M. (2026) · Diagnostics · doi:10.3390/diagnostics16030396 · Cited by 26 · see citing works
“In a similar manner, demographic attributes—including patient race—can be inferred from chest X-rays, despite the fact that these images are primarily acquired to assess lung parenchyma [14,15].”
Pitfalls and Best Practices in Evaluation of AI Algorithmic Biases in Radiology
Yi, P. H., Bachina, P., Bharti, B., Garin, S. P., Kanhere, A., Kulkarni, P., Li, D. K. B., Parekh, V. S., Santomartino, S. M., Moy, L. P., et al. (2025) · Radiology · doi:10.1148/radiol.241674 · Cited by 28 · see citing works
Doo, F. X., & McGinty, G. B. (2023) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2023.06.014 · Cited by 23 · see citing works
Deep learning-based age estimation from chest CT scans
Azarfar, G., Ko, S., Adams, S., & Babyn, P. (2023) · International Journal of Computer Assisted Radiology and Surgery · doi:10.1007/s11548-023-02989-w · Cited by 19 · see citing works
Baughan, N. M., Whitney, H. M., Drukker, K., Sahiner, B., Hu, T., Kim, G. H. J., McNitt‐Gray, M. F., Myers, K. J., & Giger, M. L. (2023) · Journal of medical imaging · doi:10.1117/1.jmi.10.6.064501 · Cited by 16 · see citing works
Clunie, D., Flanders, A. E., Taylor, A. J., Erickson, B., Bialecki, B., Brundage, D., Gutman, D. A., Prior, F., Seibert, J. A., Perry, J. H., et al. (2023) · PubMed · doi:10.48550/arxiv.2303.10473 · Cited by 21 · see citing works
Larson, D. B., Doo, F. X., Allen, B., Mongan, J., Flanders, A. E., & Wald, C. (2024) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2024.01.024 · Cited by 14 · see citing works
Giouroukou, K., Kostas Marias, A., Tsiknakis, M., & Klontzas, M. E. (2025) · Radiology Artificial Intelligence · doi:10.1148/ryai.250273 · Cited by 12 · see citing works
Re-identification of patients from imaging features extracted by foundation models
Nebbia, G., Kumar, S., McNamara, S. M., Bridge, C. P., Campbell, J. P., Chiang, M. F., Mandava, N., Singh, P., & Kalpathy–Cramer, J. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-01801-0 · Cited by 11 · see citing works
“These resultsareinlinewithpreviousstudies:genderAUC-ROC>99% 52,53 , race AUC-ROC> 98% 35,53 , and age AUC-ROC > 88% 52 .”
Dontchos, B. N., Dodelzon, K., Bhole, S., Edmonds, C. E., Mullen, L. A., Parikh, J. R., Daly, C., Epling, J. A., Christensen, S., & Grimm, L. J. (2025) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2025.05.001 · Cited by 9 · see citing works
The Potential for Using ChatGPT to Improve Imaging Appropriateness
Varney, E. T., & Lee, C. I. (2023) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2023.06.005 · Cited by 7 · see citing works
Kocher, M., & Lee, C. I. (2022) · Journal of the American College of Radiology · doi:10.1016/j.jacr.2022.07.021 · Cited by 5 · see citing works
Ranganathan, S., Srinivasan, K., Sriramakrishnan, P., & Kalaiselvi, T. (2025) · International Journal of Imaging Systems and Technology · doi:10.1002/ima.70233 · Cited by 3 · see citing works
Burns, J. L., Zaiman, Z., VanSchaik, J. T., Luo, G., Peng, L., Price, B., Mathias, G., Mittal, V. A., Sagane, A., Tignanelli, C. J., et al. (2023) · Journal of medical imaging · doi:10.1117/1.jmi.10.6.061106 · Cited by 2 · see citing works
Kavandi, H., Kulkarni, P., Garin, S. P., Bachina, P., Parekh, V. S., & Yi, P. H. (2024) · American Journal of Roentgenology · doi:10.2214/ajr.24.31963 · Cited by 2 · see citing works
Yadlapalli, P., & Bhavana, D. (2024) · Scalable Computing Practice and Experience · doi:10.12694/scpe.v25i1.2320 · Cited by 1 · see citing works
“A three-dimensional neural network design, known as a deep convolutional neural network (DCNN), is utilized to concurrently process the Red, Green, and Blue components of an image [9,10].”
MethodCarrillo-Larco, R. M. (2024) · Pediatric Cardiology · doi:10.1007/s00246-024-03561-2
“Existing studies have successfully used ML classifiers to identify traits, such as sex, age, and race from chest X-rays [8–11] and fundus images [12–19].”
InfluentialPredicting Patient Self-reported Race From Skin Histological Images with Deep Learning
Chen, S., Verma, R., Clare, K., Jegminat, J., Alleva, E., Huang, K., Veremis, B. M., Fuchs, T. J., & Campanella, G. S. V. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-05870-6_11
“While histological images may not encode demographic signals as strongly as radiological images [13, 14], DL models can still predict self-reported race, likely by leveraging morphological shortcuts such as epidermal structures in skin slides.”
InfluentialRetinaGuard: Obfuscating Retinal Age in Fundus Images for Biometric Privacy Preserving
Luo, Z., Liu, C., Xiao, D., Yu, Z. T., Wang, Y., & Zhu, T. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356446
Kamel, K. M. M. E., Tawfik, D., Saged, N. M., El-Khoribi, R. A.-W. A., Mohamed, M. A. K., Hassan, D. A. E., & Ghaleb, S. S. (2026) · Egyptian Journal of Forensic Sciences · doi:10.1186/s41935-026-00567-0
“Their model heatmaps focused on the level of breast shadows ( Adleberg et al. 2022).”
Improving racial fairness in brain age models using style-transfer synthesisPreprint
Banchieri, V., Zemlyanker, D., Iglesias, J. E., & Initiative, T. A. D. N. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.09.16.751796
“…neural‐networks (CNNs) have been shown to predict protected attributes (characteristics that can be used to dis‐ criminate individuals (Zhang and Wu 2017), such as race, sex, and age) directly from medical images (Gichoya et al. 2022, Duffy et al. 2022, Adleberg et al. 2022, Yang et al. 2024).”
- Authors
- Adleberg, J., Wardeh, A., Doo, F. X., Marinelli, B., Cook, T. S., Mendelson, D. S., & Kagen, A.
- Affiliations
- Mount Sinai Health System (US)
- Upstate University Hospital (US)
- University of Pennsylvania (US)
- Icahn School of Medicine at Mount Sinai (US)
- Corresponding author
- Adleberg, Jason
- Venue
- Journal of the American College of Radiology
- Publisher
- Elsevier BV
- Published
- 2022-10
- Volume
- 19
- Issue
- 10
- Pages
- 1151-1161
- ISSN
- 1546-1440 (ISSN-L)
- OpenAlex
- W4291002614
- Cited by
- 32
- References
- 28
- Citations by year
- 2026: 3 · 2025: 9 · 2024: 7 · 2023: 12 · 2022: 1
- FWCI
- 2.19
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Journal of the American College of Radiology (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
- Social Sciences › Arts and Humanities › Archeology › Forensic Anthropology and Bioarchaeology Studies
- Keywords
- chest radiograph, gender classification, age estimation, ethnicity prediction, insurance status, algorithmic bias, underrepresented groups, MIMIC-CXR, CheXpert, external validation, model interpretability, training data diversity
- MeSH terms
- Deep Learning, Artificial Intelligence, Ethnicity, Humans, Radiography, Radiography, Thoracic
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
