{"id":"W4409283616","doi":"10.1038/s41586-025-08869-4","title":"Towards accurate differential diagnosis with large language models","year":2025,"lang":"en","type":"article","venue":"Nature","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":173,"is_retracted":false,"has_abstract":true,"ca_institutions":"Google (Canada)","funders":"","keywords":"McNemar's test; Differential diagnosis; Test (biology); Medicine; Process (computing); Computer science; Medical physics; Differential (mechanical device); Pathology; Statistics; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"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":[],"consensus_categories":[],"category_scores_codex":[0.005226974,0.001536589,0.001275784,0.002168673,0.000507602,0.003028597,0.001955518,0.001665741,0.003221124],"category_scores_gemma":[0.03247626,0.0007948126,0.001834907,0.0008762514,0.0008799839,0.003101278,0.003823372,0.002372494,0.001983153],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00139128,"about_ca_system_score_gemma":0.002204831,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005277595,"about_ca_topic_score_gemma":0.006302256,"domain_scores_codex":[0.9948925,0.003047994,0.0003063753,0.0007712584,0.0008024252,0.0001794771],"domain_scores_gemma":[0.9728885,0.02312734,0.0009787421,0.001243059,0.001399145,0.0003632344],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002012994,0.0005810796,0.01912323,0.0009020427,0.0006162656,0.001147296,0.001022347,0.3621025,0.01350686,0.01966686,0.01732685,0.5619916],"study_design_scores_gemma":[0.00009853064,0.00008886487,0.0005300212,0.00004617919,0.00005396099,0.0001564889,0.0001205733,0.9690461,0.002560093,0.0251232,0.002152118,0.00002397894],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0516976,0.0006975167,0.9337943,0.002241801,0.00009143017,0.0003262321,0.001528349,0.007961528,0.00166123],"genre_scores_gemma":[0.4257755,0.0003764472,0.5673155,0.00111209,0.0001287868,0.0004861999,0.002918115,0.0003333862,0.001554114],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005277595,"threshold_uncertainty_score":0.0276432,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05125932754292,"score_gpt":0.4184976255190135,"score_spread":0.3672382979760935,"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."}}