通过分层培训确保模型公平性:TP53突变预测与雌激素受体分层在乳腺组织病理学中的突变
概括
医疗成像决策支持的AI模型可以过度适应诸如雌激素受体 (ER) 状态等偏见. 根据ER状态对训练数据进行分层,可以提高TP53突变预测模型的公平性和通用性.
科学领域:
- 人工智能在医学中的应用
- 医学图像分析 医学图像分析
- 基因组数据科学 基因组数据科学
背景情况:
- 人工智能 (AI) 模型越来越多地被用作医学成像中的决策支持系统.
- 一个关键的挑战是确保对数据集偏差和不平衡变量的模型稳定性,例如在雌激素受体 (ER) 阴性与ER阳性乳腺癌中TP53突变的不同患病率.
- 现有的模型往往忽略了这些内在的偏见,可能导致过拟合和减少概括性.
研究的目的:
- 调查训练用于TP53突变预测的AI模型是否过度适应雌激素受体 (ER) 状态.
- 评估按ER状态分层训练数据对模型性能,偏见,概括性和公平性的影响.
- 展示一种提高医疗图像分析中的AI模型稳定性的方法.
主要方法:
- 使用医学成像数据开发和培训用于TP53突变预测的AI模型.
- 对整个数据集进行训练的模型与按雌激素受体 (ER) 状态分层的模型进行比较.
- 评估不同子组的模型性能,偏见,可概括性和公平性.
主要成果:
- 训练用于TP53突变预测的AI模型在整个数据集上训练时显示过度适应雌激素受体 (ER) 状态.
- 基于ER状态的培训数据分层对所有子组都有好处.
- 经ER分层培训减少了偏见,并增加了预测模型的概括性和公平性.
结论:
- 医疗决策支持的AI模型需要严格测试数据集相关偏差.
- 根据ER状态等关键生物变量分层训练数据对于开发强大,可概括和公平的AI模型至关重要.
- 这种方法提高了AI驱动的TP53突变在乳腺癌的预测的可靠性.
相关概念视频
02:07Tumor Progression
Tumor progression is a phenomenon where the pre-formed tumor acquires successive mutations to become clinically more aggressive and malignant. In the 1950s, Foulds first described the stepwise progression of cancer cells through successive stages.
Colon cancer is one of the best-documented examples of tumor progression. Early mutation in the APC gene in colon cells causes a small growth on the colon wall called a polyp. With time, this polyp grows into a benign, pre-cancerous tumor. Further...
Colon cancer is one of the best-documented examples of tumor progression. Early mutation in the APC gene in colon cells causes a small growth on the colon wall called a polyp. With time, this polyp grows into a benign, pre-cancerous tumor. Further...
01:12Loss of Tumor Suppressor Gene Functions
Tumor suppressor genes are normal genes that can slow down cell division, repair DNA mistakes, or program the cells for apoptosis in case of irreparable damage. Hence, they play an essential role in preventing the proliferation of damaged cells.
When the tumor suppressor genes develop mutations or are lost, cells start growing out of control, leading to cancer. However, a single functional copy of the tumor suppressor gene is enough for the cells to maintain their normal functions and cell...
When the tumor suppressor genes develop mutations or are lost, cells start growing out of control, leading to cancer. However, a single functional copy of the tumor suppressor gene is enough for the cells to maintain their normal functions and cell...