{"id":"W4415481727","doi":"10.1101/2025.10.21.25338475","title":"Digital Registrar: A Schema-First Framework for Multi-Cancer Privacy-Preserving Pathology Abstraction via Local LLMs","year":2025,"lang":"","type":"preprint","venue":"medRxiv","topic":"Artificial Intelligence in Healthcare","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"AUG Signals (Canada)","funders":"","keywords":"Reuse; Abstraction; Population; Cancer registry; Completeness (order theory); Cancer","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":[0.006449217,0.001149938,0.0008103649,0.00167776,0.0007199802,0.004446583,0.004030737,0.001274436,0.007931763],"category_scores_gemma":[0.01437527,0.001172218,0.003868107,0.001254546,0.001522353,0.005860147,0.008834953,0.002753348,0.005295861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001918868,"about_ca_system_score_gemma":0.00447134,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007038353,"about_ca_topic_score_gemma":0.008239207,"domain_scores_codex":[0.9963071,0.0009028923,0.000488088,0.0009225482,0.001125999,0.0002532254],"domain_scores_gemma":[0.9937068,0.001505119,0.000348435,0.003474202,0.0007083097,0.0002572508],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0009672838,0.0003787139,0.01908309,0.001798613,0.0006859173,0.00142883,0.004055646,0.1328514,0.02593533,0.2279979,0.109656,0.4751613],"study_design_scores_gemma":[0.0001165061,0.000130904,0.001919039,0.0003263067,0.0001852418,0.0006529873,0.000577287,0.5660776,0.0396908,0.1468966,0.2432506,0.0001760676],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003018504,0.0001499528,0.9010128,0.0004754959,0.00006847003,0.0002291669,0.002577597,0.09046505,0.002002902],"genre_scores_gemma":[0.1338162,0.0004070248,0.8214843,0.001283459,0.00007631288,0.0005751,0.0199541,0.01751306,0.004890436],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007931763,"threshold_uncertainty_score":0.03410715,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.216441883731342,"score_gpt":0.4895632775683701,"score_spread":0.273121393837028,"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."}}