Goran Lovric from the PSI Center for Photon Science is combining artificial intelligence with synchrotron imaging to create three-dimensional virtual staining of tissue samples.
Photo Credit: © Paul Scherrer Institute PSI/Mahir Dzambegovic
Scientific Frontline: Extended "At a Glance" Summary: Virtual Tissue Staining in 3-D
The Core Concept: Virtual tissue staining in 3-D, pioneered through the VISTACT platform, is an AI-driven technique that applies traditional histological color markers to high-resolution, greyscale micro-computed tomography (µCT) scans. This enables the non-destructive, three-dimensional analysis of pathological tissue changes without the need to physically slice and chemically stain delicate samples.
Key Distinction/Mechanism: Traditional pathology relies on cutting tissue into ultra-thin, two-dimensional sections for manual staining and microscopic examination. In contrast, VISTACT utilizes high-resolution phase-contrast micro-CT paired with a conditional generative adversarial network. The AI automatically translates 3-D greyscale structural density data into the familiar diagnostic color contrasts used by pathologists (such as blue-violet for cell nuclei and pink for collagen).
Major Frameworks/Components:
- Phase-Contrast Micro-CT (PCµCT): An advanced imaging technique that captures highly detailed, three-dimensional structural data of soft tissues using X-ray phase shifts rather than simple density.
- Conditional Generative Adversarial Network: A specialized image-to-image machine learning model trained to link microscopic X-ray patterns with specific histological color profiles.
- Spatial Mapping Protocol: A multi-stage algorithmic process used to perfectly align delicate 2-D histological training sections within the comprehensive 3-D CT datasets to ensure accurate AI training.
Branch of Science: Pathology, Medical Physics, Artificial Intelligence (Machine Learning), Computed Tomography, and Histology.
Future Application: While currently in the proof-of-concept phase, this technology has the potential to drastically accelerate disease biomarker research and allow for the non-destructive clinical evaluation of complex 3-D tissue architectures, such as tumors or vascular lesions.
Why It Matters: By overcoming the two-dimensional and destructive limitations of 19th-century cellular pathology methods, virtual 3-D staining allows scientists to visualize whole disease morphologies in their complete spatial context while retaining the visual color cues essential for accurate medical diagnostics.
Rudolf Virchow fundamentally changed medicine when he formulated his cell theory of disease in the nineteenth century: diseases do not arise inexplicably within the organism, but rather in specific cells and tissues. To this day, pathology—the study of disease processes—is essentially based on the time-consuming examination of thin tissue sections, which are stained and then viewed under a microscope.
Now, an international research team at the Paul Scherrer Institute (PSI) has managed to overcome this two-dimensional limitation. Using high-resolution micro-computed tomography (µCT) and artificial intelligence, a group led by physicist Goran Lovric from the PSI Center for Photon Science generated virtual stains of tissue samples, so-called histological stains. This could potentially eliminate the need to prepare and stain ultrathin, delicate sections. “We have shown for the first time that a CT-based virtual stain can deliver results similar to conventional laboratory histology,” Lovric explains. “This could open up a wealth of clinical and scientific applications.”
Familiar Color Markers of Histology The researchers combined high-resolution phase-contrast micro-CT (PCµCT) with machine learning methods. The platform is called VISTACT—short for virtual staining of micro-computed tomography. While conventional computed tomography primarily measures differences in X-ray density, phase-contrast micro-CT utilizes additional X-ray information, thereby achieving significantly better visualization of soft tissue. This allows three-dimensional visualization of fine anatomical structures on the micrometer scale—so far, however, only in grayscale. In pathology, however, specialists are trained to interpret the typical color contrasts of conventional histological stains: cell nuclei appear blue-violet, collagen pink, and elastic fibers dark. Many of these visual reference points are lost in grayscale CT datasets.
“We therefore wanted to carry over the familiar color world of histology to three-dimensional CT data,” explains Lovric. To achieve this, the researchers trained a specialized AI model using pairs of real histological sections and their corresponding CT scans. In this way, the AI model learned which microscopic patterns typically receive which staining. It was then able to virtually stain new CT data—essentially an automatic translation between two image worlds.
More Precise Localization One crucial technical step was the precise mapping of the images. Histological sections are only a few micrometers thick and can easily become distorted during sectioning or mounting. In addition, it is essential to determine exactly where each section is located within the three-dimensional CT dataset. Lovric’s research group developed a multistage process that automatically identifies the corresponding layer and compares it with the histology data. According to the researchers, this spatial mapping is significantly more precise than previous standard methods.
To carry out the virtual staining, researchers used a so-called conditional generative adversarial network—a specialized AI model for image-to-image translation. With grayscale images from micro-CT scans as input, the model generated virtual histological specimens. Remarkably, the AI produced not merely coarse color areas but rather plausibly differentiated tissue components of various types: blood in the fine vessels appeared yellowish, collagen structures pink, and surfaces in the lungs gray to violet.
Lung Tissue Test Provides Proof of Concept The researchers tested their new method on lung tissue taken from individuals with pulmonary hypertension. This condition involves pathological remodeling of the pulmonary vessels. “We were able to map the altered vascular regions in three dimensions,” says Cristina Almagro-Pérez. She is the first author of the new publication and worked in Goran Lovric's group during her master’s thesis. She is now conducting research in the United States.
The new technique can be automated and can work significantly faster than the current method. However, it is not yet ready for routine use in hospitals: the necessary phase-contrast imaging was performed at the TOMCAT beamline of the Swiss Light Source (SLS), one of the large research facilities at PSI. The resulting volumes of data were enormous, and the resolution was often insufficient to depict individual cell nuclei reliably.
Furthermore, virtual histology remains a statistical reconstruction: the AI platform does not generate actual histological information, but rather plausible predictions based on the training data. Almagro-Pérez and Lovric emphasize that the procedure has not yet reached routine diagnostic quality. However, the “proof of concept” has been established, and the method is, in principle, applicable to the examination of various diseases. Particularly in examining tumors, vascular lesions, or complex tissue architectures, this form of nondestructive 3D pathology has the potential to accelerate research into disease biomarkers and thus open up new diagnostic perspectives in the long term.
More than 150 years after the advent of Virchow’s cellular pathology, histology might again be on the verge of a fundamental transformation.
Published in journal: Journal of the Royal Society Interface
Title: Histology-guided 3D virtual staining of microCT-imaged lung tissue via deep learning
Authors: Cristina Almagro-Pérez, Niccolò Peruzzi, Csaba Galambos, Andrew H. Song, Hans Brunnström, Kinga I. Gawlik, Marco Stampanoni, Karin Tran-Lundmark, Goran Lovric
Source/Credit: Paul Scherrer Institute | Werner Siefer
Edited by: Scientific Frontline
Reference Number: path061826_01
.png)
.jpg)