{"id":"W4407170397","doi":"10.3389/fonc.2025.1456563","title":"Improving diagnostic precision in thyroid nodule segmentation from ultrasound images with a self-attention mechanism-based Swin U-Net model","year":2025,"lang":"en","type":"article","venue":"Frontiers in Oncology","topic":"Thyroid Cancer Diagnosis and Treatment","field":"Medicine","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Nodule (geology); Segmentation; Thyroid nodules; Ultrasound; Mechanism (biology); Thyroid; Medicine; Computer science; Artificial intelligence; Radiology; Computer vision; Internal medicine; Biology; Physics","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":[0.0002280159,0.000215591,0.0004886431,0.0004157187,0.00005198515,0.00002226261,0.00009593159,0.0001860571,0.00001777972],"category_scores_gemma":[0.0001636443,0.0001875295,0.00006141853,0.000347281,0.00004320879,0.0001215766,0.00002920205,0.0002287895,0.000004456006],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001911337,"about_ca_system_score_gemma":0.000485477,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007402949,"about_ca_topic_score_gemma":0.0009716897,"domain_scores_codex":[0.9984696,0.0001298002,0.0003821895,0.0005168226,0.0001948705,0.0003067002],"domain_scores_gemma":[0.9989419,0.0005315935,0.0001378883,0.000265072,0.0000563652,0.00006717029],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00251964,0.004052366,0.8698205,0.0001462083,0.000280248,0.0003529324,0.0009440284,0.01086345,0.04416398,0.0001642138,0.0117821,0.05491027],"study_design_scores_gemma":[0.06694155,0.006752656,0.6148108,0.001813547,0.001649563,0.00002298989,0.003589218,0.1718499,0.1184595,0.0126,0.0005449806,0.0009653369],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7696002,0.001622898,0.2252675,0.001091533,0.0005831512,0.001334609,0.00003869322,0.0000703071,0.000391127],"genre_scores_gemma":[0.8738114,0.000486169,0.1241373,0.0005949028,0.0000382166,0.0006219117,0.0001943291,0.00002230886,0.00009340804],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2550098,"threshold_uncertainty_score":0.7647232,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.005160837039135114,"score_gpt":0.259532120545762,"score_spread":0.2543712835066269,"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."}}