{"id":"W4407719815","doi":"10.1177/22925503251315489","title":"Beyond the Surface: Assessing GPT-4's Accuracy in Detecting Melanoma and Suspicious Skin Lesions From Dermoscopic Images","year":2025,"lang":"en","type":"article","venue":"Plastic Surgery","topic":"Cutaneous Melanoma Detection and Management","field":"Medicine","cited_by":13,"is_retracted":false,"has_abstract":true,"ca_institutions":"Princess Margaret Cancer Centre; University of Toronto","funders":"","keywords":"Melanoma; Melanoma diagnosis; Artificial intelligence; Dermatology; Medicine; Radiology; Nuclear medicine; Computer science; Cancer research","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":[0.0003346603,0.0001753944,0.0003517229,0.0002349453,0.0002173301,0.0001246852,0.00006424254,0.00007905077,0.0000972531],"category_scores_gemma":[0.004274627,0.0001348634,0.00006989569,0.0003847775,0.0000736185,0.00008364387,0.0001097102,0.0002825928,0.00001263467],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000895168,"about_ca_system_score_gemma":0.0001239622,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005202746,"about_ca_topic_score_gemma":0.0002627303,"domain_scores_codex":[0.9987304,0.0001150388,0.0003669616,0.0003214876,0.0001768258,0.0002892372],"domain_scores_gemma":[0.9803646,0.0191798,0.00009590823,0.0002553718,0.00003340815,0.00007090225],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000332533,0.0003228732,0.6346064,0.0003868533,0.0003995703,0.001690023,0.0007559339,0.0002682857,0.02900345,0.0001252573,0.006935807,0.325173],"study_design_scores_gemma":[0.0008942586,0.00003514758,0.9820463,0.0006432991,0.0002372053,0.0001320065,0.002053799,0.002727713,0.007791218,0.0003318455,0.002860179,0.0002469951],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9935967,0.0009068825,0.001053548,0.000870451,0.002043772,0.0002147855,0.000004649188,0.00008380137,0.001225406],"genre_scores_gemma":[0.9989903,0.0001081366,0.0002120705,0.0002877051,0.00009628084,0.00001606978,0.000009461122,0.00001744811,0.0002625092],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3474399,"threshold_uncertainty_score":0.549957,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01594977064736323,"score_gpt":0.2751757988256273,"score_spread":0.259226028178264,"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."}}