{"id":"W4409970142","doi":"10.7717/peerj-cs.2822","title":"Validation of automated paper screening for esophagectomy systematic review using large language models","year":2025,"lang":"en","type":"article","venue":"PeerJ Computer Science","topic":"Esophageal Cancer Research and Treatment","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ottawa Hospital; University of Calgary; University of Toronto; Dalhousie University; University of Ottawa","funders":"","keywords":"Esophagectomy; Computer science; Natural language processing; Medicine; Esophageal cancer; Internal 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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1997233,0.002703885,0.003701301,0.01716005,0.001528327,0.006415987,0.003066036,0.002420408,0.004331029],"category_scores_gemma":[0.5360563,0.001705395,0.009277467,0.008201742,0.001146254,0.00432236,0.004655143,0.001296227,0.001307129],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00358978,"about_ca_system_score_gemma":0.0162551,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005620276,"about_ca_topic_score_gemma":0.01128135,"domain_scores_codex":[0.8171112,0.1198597,0.03552106,0.01225356,0.01438597,0.0008684463],"domain_scores_gemma":[0.2967802,0.6273808,0.03142807,0.01982658,0.0232241,0.00136023],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.008867662,0.001090578,0.1730315,0.1350075,0.0316489,0.002374962,0.006918975,0.06506563,0.007608964,0.008039216,0.06721342,0.4931329],"study_design_scores_gemma":[0.008848724,0.002853646,0.08679558,0.02695723,0.02851628,0.003309191,0.002386604,0.7184426,0.01286699,0.03138191,0.07650039,0.00114083],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2585199,0.0307076,0.518733,0.01145379,0.001544769,0.03885697,0.08725218,0.0427882,0.01014368],"genre_scores_gemma":[0.438245,0.002374833,0.5115904,0.001843348,0.000267167,0.01938972,0.02511975,0.000588873,0.0005809095],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8002766,"threshold_uncertainty_score":0.9868836,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04105738923212442,"score_gpt":0.3866673114043312,"score_spread":0.3456099221722068,"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."}}