Caltech computational microscopy improves views of cancer cells

03 Jul 2024

Modifications to ptychographic technique make deeper, sharper images.

A project at Caltech has developed a route to high-resolution images of biological tissues while maintaining large field of view, in a variant of ptychographic microscopy.

Developed originally at Caltech in 2013, Fourier ptychographic microscopy (FPM) is a computational approach in which a set of full-field images acquired at different angles of illumination are synthesized into a wider numerical aperture, increasing the imaging resolution achieved.

Since then the ptychographic approach has been put to use in several ways, including as a route to lensless microscopy and for the imaging of tiny surface structures on semiconductor materials.

Some challenges have remained, however, including the technique's susceptibility to failure when dealing with excessive image aberrations, and the inherent possibility that the FPM operation never actually arrives at an optimal solution for its image reconstruction operation.

Changhuei Yang's Biophotonics Lab at Caltech has now improved matters, developing a new method that can outperform FPM in its ability to obtain images free of blurriness or distortion, while taking fewer measurements.

Published in Nature Communications, the APIC (for Angular Ptychographic Imaging with Closed-form) method could be valuable in areas like biomedical imaging, digital pathology and drug screening, according to Caltech.

"APIC has all the advantages of FPM without what could be described as its biggest weakness - namely that to arrive at a final image the FPM algorithm relies on starting at one or several best guesses, and then adjusting a bit at a time to arrive at its 'optimal' solution, which may not always be true to the original image," commented Caltech.

Better assessment of cancers using AI

Previous FPM methods have based this trial and error approach on a measurement of the phase of the returning light, and used that as a measure of the aberrations being introduced. But APIC avoids this source of inaccuracy by studying just the regions where the ptychographic data sets overlap; since the legitimate sample data should be the same here, the phase differences will be down to the aberrations alone.

"We arrive at a solution of the high-resolution complex field in a closed-form fashion, as we now have a deeper understanding of what the microscope captures, what we already know, and what we need to truly figure out," said Caltech's Ruizhi Cao. "So we don't need any iteration. We can basically guarantee that we are seeing the true final details of a sample."

In trials using both synthetic targets and stained breast cancer cells, the APIC method also allowed researchers to gather clear images over a large field of view without repeatedly refocusing the microscope, something that conventional FPM requires if the height of the sample varies by even tens of microns between sections.

Since computational microscopy techniques frequently involve stitching together more than 100 lower-resolution images to piece together the larger field of view, removing this refocusing makes APIC faster and prevents the possible introduction of human error, noted Caltech.

APIC could prove vital to the broader program of work underway in Changhuei Yang's lab, which is examining how best to optimize image data for input to AI applications.

"Recently my lab showed that AI can outperform expert pathologists at predicting metastatic progression from simple histopathology slides taken from lung cancer patients," said Yang. "That prediction ability is exquisitely dependent on obtaining uniformly in-focus and high-quality microscopy images, something that APIC is highly suited for."