潘达-PLUS:从PANDA挑战中改善了前列腺整体幻灯片图像的数据集,带有像素级专家注释
Spencer Hopson1, Carson Mildon1, Corbyn Kubalek1
1Department of Physics and Astronomy, Brigham Young University, Provo, UT 84602, United States of America.
Journal of pathology informatics
|February 2, 2026
概括
一个新的数据集,PANDA-PLUS,提供了改进的像素级注释,用于使用人工智能 (AI) 检测前列腺癌. 这一数据集提高了培训数据的质量,解决了当前用于病理学的全幻灯片图像 (WSI) 分析的局限性.
科学领域:
- 数字病理学和计算分析
- 医学成像和人工智能 医学成像和人工智能
- 瘤学和癌症诊断的研究.
背景情况:
- 人工智能 (AI) 使用整片图像 (WSI) 显示前列腺癌检测的前景,可能缓解病理学家短缺并提高诊断一致性.
- 目前用于WSI分类的AI工具受到培训数据的可用性和质量的限制.
- 像Kaggle的PANDA挑战这样的现有大规模数据集使用幻灯片级标签,这可以引入噪音并降低模型可靠性,而具有详细注释的较小数据集需要大量的劳动力来创建.
研究的目的:
- 介绍PANDA-PLUS,这是一个新的数据集,来自PANDA数据集的像素级注释,旨在改善前列腺癌病理学的AI模型培训.
- 呈现一个高效的注释管道,旨在尽量减少病理学家所需的时间承诺.
- 对PANDA-PLUS与原始PANDA数据集进行比较分析,以评估分级准确性和标签噪声.
主要方法:
- 开发PANDA-PLUS,一个数据集,包括546个WSIs与增强的像素级注释.
- 实施注释管道,以简化病理学家的数据标签过程.
- 使用格里森分数和ISUP等级进行比较分析,在各种权重方案下使用Cohen的kappa (κ) 和PABAK等协议指标.
主要成果:
- 与PANDA数据集相比,PANDA-PLUS的分级始终较低,在更高的癌症等级中存在显著的分歧.
- 对不同注释细度的分析 (幻灯片级别与补丁级别) 揭示了分级比例和图像评分中的潜在扭曲.
- PANDA-PLUS数据集识别了原始PANDA数据集中的潜在分级不准确性,突出了标签噪声和注释细粒度的问题.
结论:
- 潘达-PLUS提供了一个有价值的资源,改进了注释细粒度和减少了标签噪声,作为前列腺癌AI开发的潘达标签的优越替代品.
- 数据集和注释管道在为前列腺癌病理学中的临床AI应用创建高质量的公共数据集方面取得了重大进展.
- 这些发现强调了注释质量对人工智能模型性能和数字病理学可靠性的关键影响.
相关概念视频
Genome Annotation and Assembly
21.0K
The genome refers to all of the genetic material in an organism. It can range from a few million base pairs in microbial cells to several billion base pairs in many eukaryotic organisms. Genome assembly refers to the process of taking the DNA sequencing data and putting it all back together in a correct order to create a close representation of the original genome. This is followed by the identification of functional elements on the newly assembled genome, a process called genome annotation.
21.0K
Design Example: Designing Water Slide
629
When designing a water slide, controlling the speed of water flow is crucial for rider safety while maintaining an exciting experience. As water flows down the slide, gravity causes it to accelerate, with its speed at the bottom depending on the height from which it starts. The higher the slide, the more potential energy the water has at the top, which is converted into kinetic energy as it descends, increasing its speed.
Bernoulli's principle determines the water's velocity along the slide....
Bernoulli's principle determines the water's velocity along the slide....
629
Leveling Effect
1.4K
In acid-base chemistry, the leveling effect refers to the limitation imposed by the solvent on the strength of acids and bases in solution. When a base stronger than the solvent's conjugate base is used, it deprotonates the solvent until the base is entirely consumed, making it ineffective against weaker acids. Conversely, an acid stronger than the solvent's conjugate acid protonates the solvent until the acid is depleted, rendering it ineffective against weaker bases. Essentially, the...
1.4K
High-Level and Low-Level Awareness
757
Controlled processes in human consciousness represent high-alert mental states where individuals deliberately focus their attention on achieving specific goals. Controlled processes can be seen in situations like mastering new technology, where a person might become so absorbed that they ignore surrounding distractions. Such processes involve selective attention, requiring one to concentrate on particular elements of experience while disregarding others. These are governed by executive...
757
Levels of Organization
140.8K
Biological organization is the classification of biological structures, ranging from atoms at the bottom of the hierarchy to the Earth's biosphere. Each level of the hierarchy represents an increase in complexity that builds upon the previous level.
Molecules Are Composed of Atoms, and Biomolecules Are Assembled from Molecules:
The most basic levels include atoms, molecules, and biomolecules. Atoms, the smallest unit of ordinary matter, are composed of a nucleus and electrons. Molecules...
Molecules Are Composed of Atoms, and Biomolecules Are Assembled from Molecules:
The most basic levels include atoms, molecules, and biomolecules. Atoms, the smallest unit of ordinary matter, are composed of a nucleus and electrons. Molecules...
140.8K
Fermi Level
1.8K
The Fermi-Dirac function is represented by an S-shaped curve indicating the probability of an energy state being occupied by an electron at a given temperature. The Fermi level is the energy level at which there is a fifty percent chance of finding an electron, and it is positioned between the lower-energy valence band and the higher-energy conduction band.
At absolute zero temperature, electrons fill all energy states up to the Fermi level, leaving upper states empty. As the temperature rises,...
At absolute zero temperature, electrons fill all energy states up to the Fermi level, leaving upper states empty. As the temperature rises,...
1.8K


