{"id":"W2921197949","doi":"10.1001/jamanetworkopen.2019.0442","title":"Comparison of Artificial Intelligence Techniques to Evaluate Performance of a Classifier for Automatic Grading of Prostate Cancer From Digitized Histopathologic Images","year":2019,"lang":"en","type":"article","venue":"JAMA Network Open","topic":"AI in cancer detection","field":"Computer Science","cited_by":111,"is_retracted":false,"has_abstract":true,"ca_institutions":"Richmond Hospital; Vancouver Coastal Health; Simon Fraser University; BC Cancer Agency; Vancouver General Hospital; University of British Columbia","funders":"","keywords":"McNemar's test; Artificial intelligence; Prostatectomy; Prostate cancer; Classifier (UML); Convolutional neural network; Grading (engineering); Computer science; Medicine; Pattern recognition (psychology); Machine learning; Cancer; Mathematics; Statistics; Internal medicine","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":[],"consensus_categories":[],"category_scores_codex":[0.001119899,0.0001502951,0.0005903223,0.00008976835,0.00005663756,0.00008711562,0.00128636,0.00007914314,0.00002852243],"category_scores_gemma":[0.00005587018,0.0001350237,0.0000732699,0.0004549137,0.00007649465,0.0005866108,0.0005195056,0.0001199613,0.000004633571],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00013414,"about_ca_system_score_gemma":0.000135015,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002014695,"about_ca_topic_score_gemma":0.00002694175,"domain_scores_codex":[0.9981655,0.0001192142,0.0007811118,0.0003895088,0.0002872018,0.00025742],"domain_scores_gemma":[0.9981027,0.0002869191,0.0007563169,0.0005543741,0.0002562561,0.00004341145],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005251966,0.00009097883,0.03487925,0.0002586607,0.00003768906,4.29039e-7,0.00197797,0.01411034,0.0203383,0.000873878,0.00136346,0.9255438],"study_design_scores_gemma":[0.000184801,0.0008844024,0.01211347,0.0008880792,0.0000261382,7.614086e-7,0.000070626,0.5691553,0.4084015,0.007764962,0.0003151007,0.0001948593],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8354849,0.0001849859,0.1611034,0.0003138615,0.0004748501,0.002020012,0.00002440201,0.00007060466,0.0003229738],"genre_scores_gemma":[0.8643624,0.00003525747,0.1350522,0.00005773375,0.00006730959,0.000336919,0.000003181686,0.00001240675,0.00007258725],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.925349,"threshold_uncertainty_score":0.5506108,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0690392427908858,"score_gpt":0.3695125131974473,"score_spread":0.3004732704065615,"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."}}