{"id":"W4411117613","doi":"10.1101/2025.06.06.25328913","title":"Lack of children in public medical imaging data points to growing age bias in biomedical AI","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"Hospital for Sick Children","funders":"National Institute of Biomedical Imaging and Bioengineering; Advanced Research Projects Agency; National Institutes of Health; Alliance de recherche numérique du Canada; Hospital for Sick Children","keywords":"Representation (politics); Health care; Population; Equity (law); Medical imaging; Medicine; Public health; Data science; Artificial intelligence; Computer science; Political science; Pathology; Environmental health","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.05908312,0.0004769033,0.001067739,0.005423367,0.001002184,0.004002855,0.002179119,0.001558037,0.00813409],"category_scores_gemma":[0.3197789,0.0005903823,0.002130922,0.007295457,0.002191385,0.005094322,0.003158459,0.002760245,0.001638377],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001714677,"about_ca_system_score_gemma":0.00385247,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008123164,"about_ca_topic_score_gemma":0.01086289,"domain_scores_codex":[0.9457971,0.029019,0.007441572,0.00674779,0.01030834,0.0006861047],"domain_scores_gemma":[0.4832755,0.4145998,0.04336533,0.02862282,0.02857652,0.001559997],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001009681,0.00008383139,0.4934148,0.01305059,0.002996843,0.0007324753,0.002953882,0.002240136,0.001407284,0.02196696,0.2270089,0.2331345],"study_design_scores_gemma":[0.0001694679,0.0003006625,0.5057102,0.0251348,0.003251464,0.003685899,0.004164387,0.00495466,0.007329957,0.05297568,0.3919758,0.0003470678],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.2209057,0.1150934,0.1075037,0.2254155,0.005292655,0.0005987665,0.2976587,0.001358004,0.02617365],"genre_scores_gemma":[0.7883599,0.02117066,0.04334362,0.05197898,0.003546125,0.0009883867,0.08699129,0.0009515289,0.002669527],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9409169,"threshold_uncertainty_score":0.3124651,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3220359159637519,"score_gpt":0.4868946938124666,"score_spread":0.1648587778487147,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}