AUTHOR=Wang Xiaowen , Wu Chunyue , Zhang Shudi , Yu Pengfei , Li Lu , Guo Chunming , Li Rui 

TITLE=A novel deep learning segmentation model for organoid-based drug screening

JOURNAL=Frontiers in Pharmacology

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.1080273

DOI=10.3389/fphar.2022.1080273

ISSN=1663-9812

ABSTRACT=Organoids are self-organized three-dimensional in vitro cell cultures grown derived from stem cells. It recapitulates organ development, tissue regeneration, and disease progression, and hence has broad applications in drug discovery. However, the limitations of effective graphic algorithms for organoid growth analysis have slowed the development of organoid-based drug screening. In this study, we took the advantage of bladder cancer organoid system and developed a deep learning model, the Res-double Dynamic Conv Attention U-Net (RDAU-Net) model, to improve the efficiency and accuracy of organoid-based drug screenings. In this RDAU-Net model, the dynamic convolution and attention modules were integrated. The feature-extracting capability of the encoder and the utilization of multi-scale information was substantially enhanced, and the semantic gap caused by skip connections has been filled, which substantially improved its anti-interference ability. A total of 200 images of bladder cancer organoids on culture day 1, 3, 5 and 7, with or without drug treatment, were employed for training and testing. Compared with the sample U-Net model, the segmentation indexes, such as Intersection over Union (IoU) and Dice Similarly Coefficient (Dice), in the RDAU-Net model have been elevated. In addition, this algorithm effectively prevented misidentification and misrecognition, while maintaining a smooth edge contour. Together, we proposed here a novel deep learning model which could significantly improve the efficiency and accuracy of high throughput drug screening and evaluation based on organoids.