{"id":"W4396780833","doi":"10.1016/j.cell.2024.03.035","title":"Analysis of 3D pathology samples using weakly supervised AI","year":2024,"lang":"en","type":"article","venue":"Cell","topic":"AI in cancer detection","field":"Computer Science","cited_by":94,"is_retracted":false,"has_abstract":false,"ca_institutions":"Health Sciences Centre; University of Toronto; Sunnybrook Health Science Centre","funders":"DOD Prostate Cancer Research Program; National Center for Advancing Translational Sciences; National Human Genome Research Institute; Clarendon Fund; National Institutes of Health; Johns Hopkins University; UK Research and Innovation; University of Washington; National Institute of Biomedical Imaging and Bioengineering; University of Southampton; Canary Foundation; National Institute of General Medical Sciences; Massachusetts General Hospital; Harvard University; U.S. Department of Defense; National Cancer Institute; Brigham and Women's Hospital; U.S. Department of Veterans Affairs","keywords":"Digital pathology; Biology; Virtual microscopy; Histopathology; Computer science; Artificial intelligence; Pathology; Medicine","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":[0.001065881,0.0009222602,0.0008673663,0.00277871,0.0004634261,0.00189411,0.00108923,0.001110436,0.001796435],"category_scores_gemma":[0.002533265,0.0004211435,0.0009767797,0.001437335,0.0005701084,0.0006552075,0.001222075,0.0007939587,0.001503276],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003952149,"about_ca_system_score_gemma":0.0006152678,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001953406,"about_ca_topic_score_gemma":0.003149434,"domain_scores_codex":[0.9991387,0.0001364148,0.000052344,0.0001899829,0.0003526262,0.0001299192],"domain_scores_gemma":[0.9984842,0.0005329419,0.0001144108,0.0002441828,0.0005332759,0.00009112724],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007036904,0.0002340142,0.01925685,0.0004414925,0.0001984066,0.0007167945,0.0003500644,0.04726879,0.3448027,0.003045178,0.005052424,0.5779296],"study_design_scores_gemma":[0.00001201073,0.0001100327,0.01037843,0.00002135182,0.00006239402,0.0006746149,0.0001723974,0.9144281,0.06687966,0.003006281,0.004224537,0.00003025049],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1384618,0.0008341119,0.8523909,0.0003747409,0.0001038202,0.0001765014,0.0006196187,0.004182952,0.002855541],"genre_scores_gemma":[0.6021466,0.0004712253,0.389709,0.0002851658,0.0001263175,0.0001729175,0.002511689,0.0003787579,0.004198341],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00277871,"threshold_uncertainty_score":0.006009638,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03118363842706094,"score_gpt":0.2756412212527695,"score_spread":0.2444575828257086,"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."}}