文献速递医学人工智能与信息学基础与转化研究

人工智能用于前列腺癌诊断与Gleason分级:PANDA挑战赛

Artificial intelligence for diagnosis and Gleason grading of prostate cancer: the PANDA challenge.
PubMed 文献 · Nat Med · 2022-01-13

PANDA挑战赛是迄今最大的组织病理学竞赛,1,290名开发者使用10,616例数字化前列腺活检开发可重复的Gleason分级AI算法。提交的多种算法在美国和欧洲外部验证集上与专家泌尿病理医师的一致性分别为0.862(二次加权κ,95%CI 0.840-0.884)和0.868(95%CI 0.835-0.900)。算法在不同患者人群、实验室和参考标准间成功泛化,作者认为需在前瞻性临床试验中评估基于AI的Gleason分级。

要点

  • PANDA挑战赛使用10,616例数字化前列腺活检,1,290名开发者参与。
  • 美国外部验证集上算法与专家泌尿病理医师一致性为0.862(二次加权κ,95%CI 0.840-0.884)。
  • 欧洲外部验证集上一致性为0.868(95%CI 0.835-0.900)。

核心结论

  • 多种提交算法在独立跨洲队列上达到病理医师级性能,与专家泌尿病理医师一致性美国集为0.862(95%CI 0.840-0.884)、欧洲集为0.868(95%CI 0.835-0.900)。
  • 算法在不同患者人群、实验室和参考标准间成功泛化,作者认为需在前瞻性临床试验中评估基于AI的Gleason分级。

研究设计与样本

  • PANDA挑战赛为迄今最大的组织病理学竞赛,1,290名开发者参与,使用10,616例数字化前列腺活检。
  • 提交算法在独立跨洲队列上验证,对算法开发者完全盲法。

主要发现

  • 美国外部验证集上算法与专家泌尿病理医师一致性为0.862(二次加权κ,95%CI 0.840-0.884)。
  • 欧洲外部验证集上一致性为0.868(95%CI 0.835-0.900)。
  • 多种算法方法在不同患者人群、实验室和参考标准间实现成功泛化。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

Artificial intelligence (AI) has shown promise for diagnosing prostate cancer in biopsies. However, results have been limited to individual studies, lacking validation in multinational settings. Competitions have been shown to be accelerators for medical imaging innovations, but their impact is hindered by lack of reproducibility and independent validation. With this in mind, we organized the PANDA challenge-the largest histopathology competition to date, joined by 1,290 developers-to catalyze development of reproducible AI algorithms for Gleason grading using 10,616 digitized prostate biopsies. We validated that a diverse set of submitted algorithms reached pathologist-level performance on independent cross-continental cohorts, fully blinded to the algorithm developers. On United States and European external validation sets, the algorithms achieved agreements of 0.862 (quadratically weighted κ, 95% confidence interval (CI), 0.840-0.884) and 0.868 (95% CI, 0.835-0.900) with expert uropathologists. Successful generalization across different patient populations, laboratories and reference standards, achieved by a variety of algorithmic approaches, warrants evaluating AI-based Gleason grading in prospective clinical trials.

来源:Nat Med,查看原文。版权归原作者所有。

引用

  1. Bulten W, Kartasalo K, Chen PC, et al. Artificial intelligence for diagnosis and Gleason grading of prostate cancer: the PANDA challenge. Nat Med. 2022. PMID: 35027755. doi:10.1038/s41591-021-01620-2

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。