RSR-MSI: Reference-Based Super-Resolution for Mass Spectrometry Imaging of Tissues and Single Cells

Area of Science:

• Analytical Chemistry
• Biotechnology
• Microscopy

Background:

• High-spatial-resolution mass spectrometry imaging (MSI) is crucial for visualizing molecular distributions in cells and tissues.
• Current methods face limitations in acquisition time and throughput due to increased pixel density for higher resolution.
• Existing deep learning approaches for MSI often require extensive training datasets or paired images, which are not always available.

Purpose of the Study:

• To introduce a novel method, reference-based super-resolution for mass spectrometry imaging (RSR-MSI), to reconstruct high-resolution MSI data.
• To leverage optical microscopy images as reference frames for extracting texture information.
• To develop an image-specific super-resolution network for MSI data enhancement.

Main Methods:

• The RSR-MSI method integrates ion intensity data from low-resolution MS images with texture information from reference optical images.
• An image-specific super-resolution network was developed to reconstruct high-resolution MS images.
• The method was validated using biological tissues and single cells, utilizing only a single low-resolution MS image and a reference optical image.

Main Results:

• Successfully reconstructed high-resolution mass spectrometry imaging data with rich chemical and textural details for biological samples.
• Achieved a significant reduction in data acquisition time (by an order of magnitude) compared to routine pixel-by-pixel scanning.
• Demonstrated high spatial resolution at the subcellular level for single-cell MSI without specialized equipment modifications.

Conclusions:

• The RSR-MSI method offers a viable solution for achieving high-spatial-resolution and high-throughput mass spectrometry imaging.
• This approach enables subcellular resolution MSI, advancing cellular biology research.
• The technique enhances MSI capabilities using existing instrumentation, reducing the need for custom modifications.