arXiv cs.AI· Md. Sadibul Hasan Sadib, Md. Mohayminul Mukit, Rahmatul Kabir Rasel Sarker, Tahmid Alam Tamim, Md. Monir Hossain Shimul·· 5 小时前AI 评分19
基于深度集成的甲状腺结节超声 ROI 分类校准选择性预测研究被作者撤稿
Calibrated Selective Prediction Using Deep Ensembles for ROI-Based Thyroid Nodule Ultrasound Classification Under Dataset Shift: A Retrospective Evaluation
AI 导读
作者 Sadibul Hasan Sadib 撤回了 arXiv 上关于甲状腺结节超声 ROI 分类的深度集成研究，称方法与分析存在重大错误，报告结果不可依赖。该研究原在 TN5000 上取得 AUC-ROC 0.9395、ECE 0.0088，但在外部数据集 TN3K 上 AUC-ROC 降至 0.7870、ECE 升至 0.1899，显示外部阈值可迁移性有限。作者表示未来或提交全面修订版本。
正文 · 原文
This paper has been withdrawn by Sadibul Hasan Sadib
No PDF available, click to view other formats
Abstract:Background: Deep learning models can classify thyroid nodules on ultrasound, but reliable clinical decision support also requires calibrated probabilities, uncertainty estimation, and selective referral, particularly under dataset shift.
Methods: We developed a calibrated deterministic five-member deep ensemble for ROI-based thyroid nodule classification and selective image-based triage. TN5000 was used for model development, five-fold cross-validation, member-wise vector-scaling calibration, and fold-specific threshold selection. TN3K served as an independent external dataset-shift evaluation. The framework used ConvNeXt-Tiny with squeeze-and-excitation attention, ensemble-mean malignancy probability, and mutual information (MI) as an ensemble-disagreement score. A three-tier policy assigned images to No-FNA suggestion, FNA recommendation, or radiologist review.
Results: On pooled out-of-fold TN5000 predictions, the ensemble achieved AUC-ROC 0.9395, AP 0.9715, ECE 0.0088, and Brier score 0.0813. At 50% nominal MI retention, 7.2% of cases received a No-FNA suggestion, 39.9% an FNA recommendation, and 52.9% radiologist review, with 98.3% No-FNA NPV and 99.83% malignancy capture. On TN3K, AUC-ROC decreased to 0.7870, AP to 0.7254, ECE increased to 0.1899, and Brier score to 0.2281. The frozen TN5000 policy assigned 83.7% to review, 1.0% to No-FNA, and 15.3% to FNA recommendation. No malignant image entered the No-FNA pathway, but FNA-recommendation PPV fell to 76.6%.
Conclusion: The framework showed strong internal discrimination and calibration, but limited external threshold transportability. Selective prediction may help identify images unsuitable for automated triage, but local recalibration, threshold validation, and prospective clinical evaluation are required before deployment.
| Comments: | Withdrawn by the authors. The methodology and analysis contain substantial errors, and as a result the reported results are incorrect and should not be relied upon. A thoroughly revised version may be submitted in the future |
| Subjects: | Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2607.12075 [eess.IV] |
| (or arXiv:2607.12075v2 [eess.IV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.12075 arXiv-issued DOI via DataCite |
Submission history
From: Sadibul Hasan Sadib [view email]
[v1]
Mon, 13 Jul 2026 18:50:57 UTC (1,803 KB)
[v2]
Tue, 27 Oct 2026 18:37:23 UTC (1 KB) (withdrawn)
来源：arXiv cs.AI · arxiv.org