{"id":"W4410151922","doi":"10.2214/ajr.25.32956","title":"Prompt Engineering for Large Language Models in Interventional Radiology","year":2025,"lang":"en","type":"article","venue":"American Journal of Roentgenology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canada Research Chairs; University of Toronto; University of New Brunswick","funders":"","keywords":"Medicine; Radiology; Medical physics; Interventional radiology","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":[0.003589744,0.000739592,0.0005569574,0.0006850087,0.0006834672,0.002869218,0.001246159,0.001638599,0.008445535],"category_scores_gemma":[0.02072822,0.0006370342,0.001289021,0.0005116686,0.001971582,0.003148325,0.003634434,0.002931948,0.002592911],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001206261,"about_ca_system_score_gemma":0.001682868,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001484605,"about_ca_topic_score_gemma":0.002399368,"domain_scores_codex":[0.9968952,0.001844645,0.0001769439,0.0003665526,0.0006103296,0.000106332],"domain_scores_gemma":[0.9915296,0.006944833,0.0002788185,0.0005960162,0.0005018718,0.000148901],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003714709,0.0001609802,0.001129381,0.001069397,0.00008083885,0.0008326675,0.004043583,0.196704,0.02401371,0.3883221,0.01616721,0.3671046],"study_design_scores_gemma":[0.00006108428,0.0001323792,0.0001940383,0.0001607405,0.00003311086,0.0004253359,0.0004416518,0.583903,0.009888745,0.3659419,0.03875418,0.00006380423],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003037294,0.0003079957,0.9915904,0.0006736186,0.00006386662,0.00009004524,0.0001070828,0.002142035,0.001987619],"genre_scores_gemma":[0.1918595,0.0006047882,0.8004677,0.0007179705,0.0001452113,0.000356642,0.0004676785,0.001122282,0.004258117],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008445535,"threshold_uncertainty_score":0.02825314,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04861748349655621,"score_gpt":0.405501091885598,"score_spread":0.3568836083890418,"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."}}