Scholarlink

<p>Deep learning models for medical image analysis often fail in clinical deployment due to domain shift from varied acquisition hardware and protocols. We present a comprehensive evaluation of various domain generalization (DG) techniques to mitigate this performance degradation. We evaluate six DG algorithms against a baseline on two distinct tasks using large, multi-institutional datasets: grading prostate cancer aggressiveness from MRI using the ProstateNet dataset and assessing breast density from mammograms using the DMIST dataset, using a leave-one-domain-out protocol. Our results show that DG methods, particularly those that explicitly regularize the learning process, improve out-of-domain generalization, but do not fully close the gap with in-domain performance. On the ProstateNet dataset, the FISH algorithm achieved the highest average out-of-domain AUROC (0.678), a statistically significant improvement over the baseline (0.613). We observed similar trends on the DMIST dataset. These findings underscore the necessity of incorporating DG strategies to develop clinically deployable AI models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Domain Generalization Mitigates Scanner-Induced Domain Shift in Medical Imaging

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月7日

查看全文
  • Dagoberto Pulido-Arias,
  • Mason C. Cleveland,
  • Jay Patel,
  • Zhewei Wang,
  • Yu Leng,
  • Tiago Filipe Goncalves,
  • Maggie Yao,
  • Albert E. Kim,
  • Daniele Regge,
  • Kostas Marias,
  • Manolis Tsiknakis,
  • Henkjan Huisman,
  • Nikolaos Papanikolaou,
  • Jayashree Kalpathy-Cramer,
  • Christopher P. Bridge

摘要

Deep learning models for medical image analysis often fail in clinical deployment due to domain shift from varied acquisition hardware and protocols. We present a comprehensive evaluation of various domain generalization (DG) techniques to mitigate this performance degradation. We evaluate six DG algorithms against a baseline on two distinct tasks using large, multi-institutional datasets: grading prostate cancer aggressiveness from MRI using the ProstateNet dataset and assessing breast density from mammograms using the DMIST dataset, using a leave-one-domain-out protocol. Our results show that DG methods, particularly those that explicitly regularize the learning process, improve out-of-domain generalization, but do not fully close the gap with in-domain performance. On the ProstateNet dataset, the FISH algorithm achieved the highest average out-of-domain AUROC (0.678), a statistically significant improvement over the baseline (0.613). We observed similar trends on the DMIST dataset. These findings underscore the necessity of incorporating DG strategies to develop clinically deployable AI models.