{"id":"W4416781075","doi":"10.1016/j.esmorw.2025.100597","title":"401P Persistent bias in head and neck segmentation: Evidence from MedSAM2","year":2025,"lang":"en","type":"article","venue":"ESMO Real World Data and Digital Oncology","topic":"Advanced Radiotherapy Techniques","field":"Physics and Astronomy","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Princess Margaret Cancer Centre","funders":"","keywords":"Head and neck; Segmentation; Head (geology); Radiation therapy; Foundation (evidence)","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":[0.0000794708,0.0001149524,0.0002349557,0.00009202662,0.0000493288,0.00008202491,0.0001798214,0.00002476072,0.00004330252],"category_scores_gemma":[0.00001554457,0.0001049968,0.00002207123,0.0001787014,0.0001429715,0.0006520862,0.0002502419,0.0001108022,0.000001518701],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005856696,"about_ca_system_score_gemma":0.00006771021,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001469622,"about_ca_topic_score_gemma":0.0009903021,"domain_scores_codex":[0.9992051,0.00003319009,0.0002029615,0.0003546582,0.00005714904,0.0001468999],"domain_scores_gemma":[0.9991636,0.0004054477,0.00006914857,0.000294572,0.00001390589,0.00005334216],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00008904218,0.00008039832,0.675815,0.000005679392,0.00005209415,0.000006913418,0.0001385801,0.000002854413,0.00007512452,0.002123605,0.001735072,0.3198757],"study_design_scores_gemma":[0.004009086,0.0009047224,0.534121,0.0006269458,0.0001187212,0.000005723035,0.001644907,0.00271587,0.0008101733,0.01967835,0.4346016,0.0007628953],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9460881,0.00142884,0.0107568,0.002878563,0.0001462807,0.0004307039,0.000733282,0.00007000581,0.03746742],"genre_scores_gemma":[0.9881679,0.0003958047,0.009421997,0.0002882212,0.00009325772,0.00002644085,0.0005499952,0.000008758363,0.001047572],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4328665,"threshold_uncertainty_score":0.4281645,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07141697942490335,"score_gpt":0.3852661163662907,"score_spread":0.3138491369413874,"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."}}