{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006616242,0.0006660421,0.0007329936,0.0009837722,0.0003600017,0.001197038,0.001194265,0.001398475,0.0006416643],"category_scores_gemma":[0.02273467,0.0005252202,0.0008994285,0.0007212029,0.0005881848,0.0009641898,0.001379523,0.001162754,0.0003308078],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001611959,"about_ca_system_score_gemma":0.001169013,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01505926,"about_ca_topic_score_gemma":0.01906405,"domain_scores_codex":[0.9968373,0.00130509,0.0002129754,0.0009111607,0.0005437532,0.000189779],"domain_scores_gemma":[0.9789031,0.01526256,0.00131048,0.001770766,0.002449947,0.0003031654],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","study_design_scores_codex":[0.00391002,0.000379344,0.04800434,0.0004451203,0.001476225,0.0001164567,0.0006686085,0.6062473,0.01491447,0.0009298654,0.004527858,0.3183804],"study_design_scores_gemma":[0.00005658482,0.00022869,0.02385341,0.00004483204,0.0001676817,0.0001814221,0.00008943968,0.9647239,0.008582489,0.001189075,0.0008357257,0.00004678113],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8261014,0.002419964,0.1634578,0.000374839,0.0002166356,0.000122122,0.001069113,0.003460784,0.002777417],"genre_scores_gemma":[0.9792401,0.0001449764,0.01825812,0.0001218691,0.00002579856,0.00002799401,0.0009904276,0.0003607123,0.0008299818],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01505926,"threshold_uncertainty_score":0.03499043,"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."}}