High-throughput widefield fluorescence imaging of 3D samples using deep learning for 2D projection image restoration

Abstract

Fluorescence microscopy has become a core tool for visualizing and quantifying the spatial and temporal dynamics of complex biological processes. Thanks to its low cost and ease-of-use, widefield fluorescent imaging remains one of the most widely used fluorescence microscopy imaging modalities. To accomplish imaging of 3D samples, conventional fluorescence imaging entails acquiring a sequence of 2D images spaced along the z-dimension, typically called a z-stack. Oftentimes, the next step is to project the 3D volume into a single 2D image, as 3D image data can be cumbersome to manage and challenging to analyze and interpret, effectively limiting the utlity of z-dimensional information. Furthermore, z-stack acquisition is often time-consuming and consequently may induce photodamage to the biological sample, which are both major hurdles for its application in experiments that require high-throughput, such as drug screening. As an alternative to z-stacks, axial sweep acquisition schemes have been proposed to circumvent these drawbacks and offers potential of 100-fold faster image acquisition for 3D-samples compared to z-stack acquisition but unfortunately results in blurry, low-quality raw 2D z-projected images. We propose a novel workflow to combine axial z-sweep acquisition with deep learning-based image restoration, ultimately enabling high-throughput and high-quality imaging of complex 3D-samples using 2D projection images. To demonstrate the capabilities of our proposed workflow, we apply it to live-cell imaging of 3D tumor spheroids and find we can produce high-fidelity images appropriate for quantitative analysis. Therefore, we conclude that combining axial z-sweep image acquisition with deep learning-based image restoration enables high-throughput and high-quality fluorescence imaging of complex 3D biological samples.