{"id":"W4399781895","doi":"10.1148/radiol.233117","title":"Large Language Models for Automated Synoptic Reports and Resectability Categorization in Pancreatic Cancer","year":2024,"lang":"en","type":"article","venue":"Radiology","topic":"Radiology practices and education","field":"Medicine","cited_by":75,"is_retracted":false,"has_abstract":true,"ca_institutions":"Princess Margaret Cancer Centre; Toronto General Hospital; University of Toronto; University Health Network","funders":"","keywords":"Medicine; Categorization; Cancer; Radiology; Pancreatic cancer; Natural language processing; Medical physics; Artificial intelligence; Oncology; General surgery; Internal medicine","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005377516,0.00007778228,0.0002314756,0.0001122051,0.00003225105,0.00001036637,0.00001814618,0.000136608,0.00003528197],"category_scores_gemma":[0.0003374266,0.00006442116,0.00002450505,0.0001407978,0.00004169535,0.0001202543,0.000008339133,0.0001199315,0.000001582356],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00012549,"about_ca_system_score_gemma":0.0001606555,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003248365,"about_ca_topic_score_gemma":0.0001283397,"domain_scores_codex":[0.9991715,0.00007907786,0.0002241566,0.0002903162,0.00003494717,0.0002000138],"domain_scores_gemma":[0.9994572,0.0002686456,0.00004656623,0.0001496682,0.00003108434,0.00004686366],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002198699,0.0009777831,0.7291552,0.008216474,0.001264244,0.00244804,0.061838,0.007475342,0.08010554,0.05709557,0.03390698,0.01531813],"study_design_scores_gemma":[0.0007393702,0.0002985826,0.1145403,0.0001324256,0.0002461228,0.002157702,0.0009108907,0.8753196,0.0001476236,0.003160717,0.002195616,0.0001509986],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9864915,0.00847018,0.001252671,0.002355326,0.0004840601,0.0004903841,0.000007831516,0.0001614513,0.0002866174],"genre_scores_gemma":[0.9982041,0.0004710132,0.0004320781,0.0001321621,0.000163075,0.0001901247,0.0001029649,0.00001355355,0.0002909408],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8678443,"threshold_uncertainty_score":0.2627019,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02171129996342566,"score_gpt":0.3599718246146089,"score_spread":0.3382605246511833,"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."}}