{"id":"W4413316916","doi":"10.2196/70863","title":"A Multimodal Large Language Model as an End-to-End Classifier of Thyroid Nodule Malignancy Risk: Usability Study","year":2025,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Thyroid Cancer Diagnosis and Treatment","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Duke-NUS Medical School","keywords":"Thyroid nodules; Usability; Medicine; Artificial intelligence; Kappa; Nodule (geology); Risk stratification; Closed captioning; Computer science; Radiology; Machine learning; Thyroid; Medical physics; Internal medicine; Mathematics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01918548,0.001761046,0.0009313112,0.001084394,0.0004238896,0.002272473,0.001690917,0.001415784,0.00230166],"category_scores_gemma":[0.06622417,0.0004861083,0.001253711,0.0005426068,0.0004552347,0.002480461,0.001765276,0.001473829,0.0008373257],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001461249,"about_ca_system_score_gemma":0.001083519,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006052089,"about_ca_topic_score_gemma":0.005644555,"domain_scores_codex":[0.9911481,0.005901006,0.0007415413,0.001188412,0.0008241489,0.0001967681],"domain_scores_gemma":[0.9344498,0.05713408,0.0009268086,0.002630162,0.004163033,0.0006961324],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.01487942,0.009545763,0.1006567,0.005423485,0.002043701,0.001955295,0.016045,0.1004668,0.02938661,0.002553996,0.02120738,0.6958358],"study_design_scores_gemma":[0.0008462254,0.006982906,0.02059669,0.0005426211,0.0008697841,0.000731647,0.002340857,0.9452712,0.01134442,0.002021091,0.008199162,0.0002533459],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9233233,0.000707268,0.06446134,0.0005007422,0.0001459595,0.001393161,0.001175885,0.005986222,0.002306173],"genre_scores_gemma":[0.8947734,0.0003178191,0.09990074,0.000309876,0.00003556594,0.0009447684,0.001889667,0.0004359228,0.001392322],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01918548,"threshold_uncertainty_score":0.1014637,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03937921855738789,"score_gpt":0.4316845530668499,"score_spread":0.392305334509462,"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."}}