Optimal NIR fluorescence imaging window selection for high-NA wide-field microscopy

A research team from Zhejiang University has established a scientific framework for selecting the optimal near-infrared (NIR) fluorescence imaging window in wide-field microscopy, addressing a long-standing challenge in high-resolution biomedical optical imaging. The study, published in PhotoniX Life, proposes the novel concept of best-matched absorption as a universal standard to maximize image signal-to-background ratio (SBR) across diverse imaging conditions.

NIR fluorescence imaging, especially in the NIR-II window (900–1880 nm), is indispensable for in vivo biological imaging, tumor identification and image-guided surgery due to its deep tissue penetration and high contrast. However, existing imaging window selection methods are mostly designed for low-NA macroscopic systems, while high-NA wide-field microscopes collect far more scattered photons, leading to blurred images and poor contrast—this critical gap has limited the performance of high-resolution microscopic imaging.

To fill this gap, the team combined Monte Carlo simulations with a custom-built NIR fluorescence wide-field microscopy system, using indocyanine green (ICG) as a clinical fluorescent probe for systematic validation. The core findings include:

1. A clear best-matched absorption coefficient exists in high-NA microscopic imaging, which maximizes image peak signal-to-noise ratio (PSNR) by balancing tissue scattering and absorption.
2. Higher system NA, brighter fluorescent probes, and shallower imaging depth all induce a redshift of the optimal imaging window (from ~1200 nm to over 1400 nm for ICG).
3. Tailored window selection strategies are developed for three challenging scenarios: superficial signal occlusion, deep out-of-focus background, and targets extending beyond the depth of field.

The team verified the framework through in vitro phantom imaging and mouse tissue experiments. By switching to the optimal spectral window, the imaging PSNR was significantly improved, with clearer structural details and weaker background noise.

This research provides a practical, ready-to-use method for pre-selecting optimal NIR imaging wavelengths. It only requires minimal experimental fine-tuning to achieve the highest image quality, offering important theoretical guidance for NIR fluorescence wide-field microscopy in life science research and clinical imaging such as cerebrovascular disease diagnosis and tumor surgery guidance.