← 研踪 PaperTrailarXiv:2611.15370 · v1
Supervision-Aware Evaluation of MedGemma for Reference-Assisted Radiotherapy Contour Ranking
首次提交 2026-11-11 19:50
版本更新 2026-11-11 19:50
雷达观测 2026-11-13 11:11
摘要
Radiotherapy contour ranking can support quality assurance, but vision-language model (VLM) gains may reflect supervision, score construction, available inputs, or representation access. We evaluated frozen MedGemma 1.5 in a single-institution study, a 12-patient held-out HaN-Seg test, five-fold cross-fitting on 30 non-test HaN-Seg patients, and transfer without refitting to 120 SegRap2023 patients. On held-out HaN-Seg, a supervised readout of the final prompt state exceeded a supervised readout of the model's yes/no output probabilities by +0.3047 pooled average precision (AP; 95% patient-bootstrap interval [+0.2822, +0.3269]); the advantage persisted against a parameter-count-matched nonlinear control. A readout of projected visual tokens exceeded the final-state readout in cross-fitting and transfer (+0.0706 AP averaged equally across patients on SegRap; 95% interval [+0.0607, +0.0804]), while a structured panel-feature readout remained competitive with visual tokens. The anatomy-conditioned readout did not demonstrate an advantage over the native-feature comparator. These results motivate visual-token and structured-feature controls in medical-imaging VLM evaluations; outcomes were computational or synthetic, not clinical acceptability.
作者备注：5 pages, 1 figure, 3 tables. Submitted to IEEE ISBI 2027
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-11 19:50 · 最新观测版本
Supervision-Aware Evaluation of MedGemma for Reference-Assisted Radiotherapy Contour Ranking
Radiotherapy contour ranking can support quality assurance, but vision-language model (VLM) gains may reflect supervision, score construction, available inputs, or representation access. We evaluated frozen MedGemma 1.5 in a single-institution study, a 12-patient held-out HaN-Seg test, five-fold cross-fitting on 30 non-test HaN-Seg patients, and transfer without refitting to 120 SegRap2023 patients. On held-out HaN-Seg, a supervised readout of the final prompt state exceeded a supervised readout of the model's yes/no output probabilities by +0.3047 pooled average precision (AP; 95% patient-bootstrap interval [+0.2822, +0.3269]); the advantage persisted against a parameter-count-matched nonlinear control. A readout of projected visual tokens exceeded the final-state readout in cross-fitting and transfer (+0.0706 AP averaged equally across patients on SegRap; 95% interval [+0.0607, +0.0804]), while a structured panel-feature readout remained competitive with visual tokens. The anatomy-conditioned readout did not demonstrate an advantage over the native-feature comparator. These results motivate visual-token and structured-feature controls in medical-imaging VLM evaluations; outcomes were computational or synthetic, not clinical acceptability.
查看此版本解读 →查看此版本原文 ↗