基于序列图像模型的肝癌药效预测研究

吴 昊, 高 超
东南大学生物医学工程学院

摘要


本研究的目的是利用器官芯片培养的类器官图像数据与AI技术，深度学习模型，来预测用于治疗肝癌药物
的毒性。通过分析序列图像，使用Z-stack聚焦算法来叠加在不同时间段拍摄的同一细胞的图像，将处理后的图像数
据输入到自研的Stvit，Vivit，Vilt，Resnet3D等模型，对比在不同预训练模型上的学习效果及药物预测的推理测试，
旨在构建一个能够准确预测药物毒性的模型。

关键词


深度学习；药效预测；肝癌类器官

全文:

PDF


参考


[1]Zhang, L., Tan, J., Han, D., & Zhu, H. (2020).

Deep learning for drug discovery and cancer research:

Recent advances and future directions. Journal of Biomedical

Informatics, 112, 103598.

[2]Clevers, H., & Tuveson, D. A. (2019). Organoids in

cancer research: From modeling to drug screening. Nature

Reviews Cancer, 19(2), 65-78.

[3]Li, Y., Zhang, X., & Chen, Y. (2021). Multimodal

deep learning for biomedical data fusion: A review. IEEE

Transactions on Neural Networks and Learning Systems,

32(6), 1234-1248.

[4]Chalasani, N., & Björnsson, E. (2019). Drug-induced

liver injury: Mechanisms, biomarkers, and therapeutic

approaches.Hepatology, 70(1), 15-27.

[5]Drost, J., & Clevers, H. (2018). High-content

screening with organoids: A new era in drug discovery.Nature

Reviews Drug Discovery, 17(6), 407-420.

[6]Vaswani, A., Shazeer, N., & Parmar, N. (2021).

Transformer models for multimodal data fusion in healthcare.

Nature Machine Intelligence, 3(5), 401-410.

[7]Chen, H., Engkvist, O., & Wang, Y. (2020).

Applications of artificial intelligence in drug toxicity

prediction.Drug Discovery Today, 25(5), 817-825.

[8]Wang, Z., & Sun, Y. (2022). Integrating text

and image data for drug response prediction in cancer.

Bioinformatics, 38(8), 2201-2209.


Refbacks

  • 当前没有refback。