研究人员使用来自瑞典的大量存档样本数据集，验证了GleasonAI（一款用于前列腺病理的人工智能模型）。该模型在ISUP分级中表现强劲，kappa得分为0.86，与经验丰富的病理学家相当。值得注意的是，GleasonAI在17年间收集的样本中均保持了稳定的性能，表明其对存档材料的差异具有鲁棒性。 AI
影响 证明了人工智能模型在多样化的医学存档数据中泛化的潜力，支持回顾性预后研究。
排序理由 该集群包含一篇详细介绍人工智能模型在特定医学应用中验证的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
Artificial intelligence (AI) is becoming a clinical tool for prostate pathology, but generalization across variations in sample preparation and preservation over prolonged time periods remains poorly understood. We evaluated GleasonAI, an end-to-end attention-based multiple insta…
arXiv:2605.02614v1 Announce Type: new Abstract: Artificial intelligence (AI) is becoming a clinical tool for prostate pathology, but generalization across variations in sample preparation and preservation over prolonged time periods remains poorly understood. We evaluated Gleason…
Artificial intelligence (AI) is becoming a clinical tool for prostate pathology, but generalization across variations in sample preparation and preservation over prolonged time periods remains poorly understood. We evaluated GleasonAI, an end-to-end attention-based multiple insta…