{"id":"W4207027034","doi":"10.1101/2022.01.22.22269695","title":"Segmentation stability of human head and neck cancer medical images for radiotherapy applications under de-identification conditions: benchmarking data sharing and artificial intelligence use-cases","year":2022,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Medical Imaging and Analysis","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"National Institute of Biomedical Imaging and Bioengineering; National Institute of Dental and Craniofacial Research; University of Texas MD Anderson Cancer Center; University of Texas Health Science Center at Houston; Center for Clinical and Translational Sciences, University of Texas Health Science Center at Houston; Division of Mathematical Sciences; National Cancer Institute; National Institutes of Health; National Science Foundation","keywords":"Computer science; Artificial intelligence; Segmentation; Benchmarking; Medical imaging; Medical physics; Deep learning; Head and neck cancer; Medicine; Radiation therapy; 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":[0.0009022842,0.0001425969,0.0002571435,0.0001158765,0.0001742346,0.0001125474,0.0003430518,0.00008555887,0.0003931742],"category_scores_gemma":[0.0002029158,0.0001559996,0.00004016857,0.0001261787,0.0001734314,0.0001386627,0.0002574644,0.0002727384,1.336652e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009643972,"about_ca_system_score_gemma":0.00005579938,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003763229,"about_ca_topic_score_gemma":0.0001455766,"domain_scores_codex":[0.9985368,0.00005602694,0.0005020716,0.0005021626,0.0002583952,0.0001445952],"domain_scores_gemma":[0.9987972,0.0003640321,0.0001305742,0.0005558532,0.00005301796,0.00009935093],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002136577,0.0003529365,0.71899,0.00510047,0.001040771,0.000005899641,0.001857076,0.02384203,0.09212047,0.00258743,0.0005864039,0.1534951],"study_design_scores_gemma":[0.0001922599,0.00002535985,0.05635199,0.0003961501,0.0004118488,0.000006798848,0.0007591036,0.9201071,0.01169958,0.009307864,0.0002881545,0.0004538528],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6620818,0.001407091,0.3353503,0.0002242656,0.00008663198,0.0003452313,0.0004444154,0.00005702084,0.00000323456],"genre_scores_gemma":[0.9917184,0.002304492,0.003675313,0.00003287802,0.0001502207,0.0006217667,0.001465272,0.00002543622,0.000006188627],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.896265,"threshold_uncertainty_score":0.636148,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1428809440247545,"score_gpt":0.4051903173518598,"score_spread":0.2623093733271054,"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."}}