{"id":"W2986021933","doi":"10.1016/j.radonc.2019.10.019","title":"Comparing deep learning-based auto-segmentation of organs at risk and clinical target volumes to expert inter-observer variability in radiotherapy planning","year":2019,"lang":"en","type":"article","venue":"Radiotherapy and Oncology","topic":"Advanced Radiotherapy Techniques","field":"Physics and Astronomy","cited_by":241,"is_retracted":false,"has_abstract":false,"ca_institutions":"Saskatchewan Cancer Agency; BC Cancer Agency","funders":"","keywords":"Contouring; Segmentation; Medicine; Radiation therapy; Artificial intelligence; Deep learning; Hausdorff distance; Observer (physics); Radiation treatment planning; Nuclear medicine; Computer science; Pattern recognition (psychology); Medical physics; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009736989,0.0002285932,0.0007258167,0.0001276361,0.00008281416,0.00001721351,0.0001380387,0.0001497323,0.0005574644],"category_scores_gemma":[0.00002007967,0.0002194001,0.00008520405,0.0001493763,0.0001553383,0.0001276214,0.00003401839,0.0003418234,7.959258e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003908955,"about_ca_system_score_gemma":0.00005340679,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002560548,"about_ca_topic_score_gemma":0.0000420195,"domain_scores_codex":[0.9978821,0.0006995777,0.0005615109,0.0004842075,0.00009333657,0.0002792422],"domain_scores_gemma":[0.9985899,0.0007487972,0.0002887586,0.0002312662,0.00003470127,0.0001065753],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0005905254,0.0002147125,0.8945981,0.00001553083,0.0001073009,0.000001219656,0.002550924,0.003636177,0.00954131,0.00007563464,0.00003827807,0.08863031],"study_design_scores_gemma":[0.01084625,0.00389341,0.5931534,0.0001247332,0.00003099902,0.000005770525,0.0006310096,0.3197567,0.007082963,0.0004848011,0.06320206,0.00078797],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6150814,0.0007211202,0.3835068,0.00008186907,0.000101025,0.0004014345,0.000003972216,0.00003717268,0.00006521856],"genre_scores_gemma":[0.89456,0.0003597849,0.1044888,0.0003004011,0.000127315,0.00003999413,0.00001435381,0.00003972318,0.00006962365],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3161205,"threshold_uncertainty_score":0.8946878,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01725662655470704,"score_gpt":0.3434558950437536,"score_spread":0.3261992684890465,"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."}}