{"id":"W4408043079","doi":"10.1177/08465371251323124","title":"Evaluating Adherence to Canadian Radiology Guidelines for Incidental Hepatobiliary Findings Using RAG-Enabled LLMs","year":2025,"lang":"en","type":"article","venue":"Canadian Association of Radiologists Journal","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University; University of Toronto","funders":"","keywords":"Medicine; Readability; Guideline; Comprehension; Pathology","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.005671315,0.0002149109,0.0004891927,0.001187743,0.0007983088,0.0001781011,0.00120133,0.0003224121,0.00003945405],"category_scores_gemma":[0.02137011,0.0002269512,0.0001621237,0.0008091857,0.00004124948,0.0002745797,0.000071783,0.0005372131,0.000006538586],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.008362989,"about_ca_system_score_gemma":0.009920412,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.5046063,"about_ca_topic_score_gemma":0.6895952,"domain_scores_codex":[0.9966792,0.0006165504,0.0009854665,0.0004339875,0.0003073401,0.0009773861],"domain_scores_gemma":[0.9953198,0.000705446,0.0005916902,0.0003436807,0.002144202,0.0008951113],"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":[0.00002863457,0.00001216282,0.7935113,0.00008411222,0.0001879769,0.00006445064,0.001127439,0.03325871,0.001077804,0.009095748,0.1258041,0.03574749],"study_design_scores_gemma":[0.003073192,0.001408938,0.2311439,0.0007954488,0.0001109334,0.0009177875,0.0005183163,0.7202855,0.0002900602,0.01090257,0.02938226,0.001171143],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8453841,0.0006992829,0.1022318,0.04444549,0.005058468,0.001273838,0.0001172325,0.00007507481,0.0007147312],"genre_scores_gemma":[0.8627675,0.00002848906,0.1267,0.009306069,0.0003534271,0.00003252366,0.00001787824,0.00001824479,0.0007758705],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6870268,"threshold_uncertainty_score":0.9956924,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1122429857990092,"score_gpt":0.4238276171464156,"score_spread":0.3115846313474065,"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."}}