Automated Light-Sheet Microscopy Image Analysis ～Automatic Brain Region Segmentation and Improved Efficiency in Perivascular Environment Analysis～

Challenge

Lightsheet microscopy is a valuable approach for visualizing the entire mouse brain and observing the three-dimensional distribution of target substances within it, enabling the analysis of drug distribution at cellular resolution.

However, this three-dimensional data poses challenges for data analysis due to both the massive data volume and non-uniformity of fluorescence intensity.

In particular, manually extracting vascular structures from large-scale fluorescence images is highly time-consuming and thus requires automation, yet conventional luminance-based thresholding methods alone have proved inadequate.

LPIXEL’s Solution – AI Technology and Impact

LPIXEL, in collaboration with Daiichi Sankyo Co., Ltd., has developed a machine-learning–based method for extracting and analyzing fluorescently labeled vascular structures across the entire mouse brain from three-dimensional lightsheet microscopy images.

This includes an automated workflow that segments brain regions and quantifies the distinctive vascular features within each region.

Furthermore, we are currently developing methods for the three-dimensional analysis of the perivascular environment, including metrics such as the distance between blood vessels and surrounding cell nuclei.

Together, these innovations will enable highly detailed analyses of fluorescently labeled drug distribution and are expected to reveal mechanisms of action that were previously undetectable using conventional analytical approaches.