{"id":"W3023773746","doi":"10.1016/j.media.2020.101714","title":"The reliability of a deep learning model in clinical out-of-distribution MRI data: A multicohort study","year":2020,"lang":"en","type":"article","venue":"Medical Image Analysis","topic":"MRI in cancer diagnosis","field":"Medicine","cited_by":171,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Institute on Aging; Canadian Institutes of Health Research; National Institutes of Health; IXICO; H. Lundbeck A/S; Servier; Karolinska Institutet; Alzheimerfonden; Swedish Brain Power; Hjärnfonden; Vetenskapsrådet; Eisai; Stiftelsen Olle Engkvist Byggmästare; Genentech; Stiftelsen för Strategisk Forskning; Center for Innovative Medicine; Northern California Institute for Research and Education; Stockholms Läns Landsting; DoD Alzheimer's Disease Neuroimaging Initiative; Pfizer; Biogen; BioClinica; Nvidia; University of Southern California; U.S. Department of Defense; Eli Lilly and Company; Bristol-Myers Squibb; Meso Scale Diagnostics; Alzheimer's Disease Neuroimaging Initiative; Novartis Pharmaceuticals Corporation; Alzheimer's Association; Åke Wiberg Stiftelse","keywords":"Artificial intelligence; Computer science; Reliability (semiconductor); Deep learning; Convolutional neural network; Protocol (science); Machine learning; Neuroradiologist; Neuroimaging; Medical physics; Pattern recognition (psychology); Magnetic resonance imaging; Medicine; Pathology; Radiology","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.005142118,0.0001321115,0.0009649502,0.00006642455,0.00004865484,0.00001079993,0.0005306237,0.0001294066,0.0002044135],"category_scores_gemma":[0.01951907,0.00008986767,0.0003165038,0.001243503,0.0004369591,0.00009387089,0.000412549,0.0007684733,0.000007988167],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006477437,"about_ca_system_score_gemma":0.0001997015,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003989371,"about_ca_topic_score_gemma":0.0008256701,"domain_scores_codex":[0.995982,0.0005631977,0.001459406,0.0005866549,0.001181504,0.0002272278],"domain_scores_gemma":[0.9966168,0.001437261,0.000299814,0.001043761,0.0002789497,0.0003234314],"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":[0.0002620619,0.001335287,0.9785652,0.00007075765,0.0009654597,0.00005494041,0.000828448,0.002303418,0.00002038166,0.000002101244,0.001334495,0.01425747],"study_design_scores_gemma":[0.00130025,0.0002653491,0.1328674,0.0000266522,0.002971399,2.498609e-7,0.0008582703,0.8611294,0.00003562405,0.000009272422,0.0004690067,0.00006711841],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7885395,0.0002940293,0.1911619,0.01899629,0.00007581964,0.0006958004,0.00006160379,0.0000415279,0.0001335305],"genre_scores_gemma":[0.9968489,0.0008575664,0.001691557,0.0002425738,0.00009778951,0.00002959737,0.0002083627,0.00001043405,0.00001326217],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.858826,"threshold_uncertainty_score":0.9887399,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07275462749978208,"score_gpt":0.4260945749124501,"score_spread":0.353339947412668,"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."}}