A new artificial intelligence tool called MISO (MultI-tropical smalloutdoors THEmicrophones) can detect cancer characteristics at the cellular level by examining data from extremely small pieces of tissue—some as small as 400 square micrometers, equivalent to the width of five human hairs. Developed by researchers at the Perelman School of Medicine at the University of Pennsylvania, the tool analyzes reams of data and can apply information to even the tiniest points in medical imaging. It could guide doctors to the individual treatments that work best for a variety of cancers, according to a new paper detailing MISO published today in Nature Methods.
Using MISO, researchers have discovered new information about a variety of different cancers using data and imaging from donated patient tissue, including:
- Bladder cancer: MISO identifies specialized group of cells responsible for forming tertiary lymphoid structures linked to improved responses to immunotherapy
- Stomach cancer: MISO distinguished cancer cells and the lining inside the tissue
- Colon cancer: MISO identified different subtypes of cancer cells that helped shed light on the various malignant cells that make up even a tumor
In addition, MISO was used to analyze the structures of non-cancerous brain tissues.
All of these achievements can guide better treatments, increase survival, and are very difficult, if not impossible, to do without an extremely powerful AI tool like MISO.
MISO was developed to work on “spatial polyomics”, a field of study in which researchers try to gain knowledge about different conditions by looking at the physical arrangement of tissue by looking at different “-omics” methods, such as transcriptional (study of gene expression), proteomic (proteins) and metabolomics (metabolites and their processes), among others.
“As the field of spatial ohmics advances, it has become possible to measure multiple ohmic modes from the same tissue slice, providing complementary information and offering a more comprehensive, insightful view,” said Mingyao Li, PhD, senior author of the study and professor of Biostatistics and Digital Pathology.
MISO addresses a huge data challenge by enabling simultaneous analysis of all spatial-ohmic modalities, as well as microscopic anatomy images when available. It is the only method that can handle datasets like these with hundreds of thousands of cells per sample.”
Mingyao Li, PhD, Professor of Biostatistics and Digital Pathology, University of Pennsylvania
When using spatial transcriptomics to examine an image, a single pixel in a single image contains 20,000 to 30,000 data points that need to be analyzed through the lens of -omics, and this number can double and triple if multiple -omics are considered. MRI and CT scans have only one data point (grayscale) per pixel for interpretation. Without some kind of AI tool to help them, doctors and researchers looking at medical images would almost never be able to get some of the insights that MISO can.
The latest effort in imaging -omics technology
MISO continues Li’s work to develop artificial intelligence imaging techniques capable of seeing what even trained humans cannot. Earlier this year, her team released a paper looking at iSTAR, a tool that probes genomics and found traces of cancer—and the body’s response to treatments for it—that would otherwise go unnoticed.
While MISO looks at a much larger range of data than iSTAR — some elements of which were even used to develop MISO-Li — he envisions both being useful, but in different areas. iSTAR is useful for increasing imaging sharpness and generating essentially spatio-ohmic data that MISO could then analyze, while MISO supports insight into finer topics (such as highly endothelial venule detection, a special group cells that recruit white blood cells to specific tissues).
Moving forward, the team hopes to bring together everything they’ve learned about spatial ohms and pathological imaging and improve MISO to be able to analyze multiple tissue samples simultaneously, exponentially increasing its output of findings.
Some data-like epigenetic marks (chemicals that regulate DNA but are influenced by the environment, not purely genetic)—have yet to be measured in large-scale ways, but MISO’s AI system allows it to “learn” as it processes information, making it so that it can identify this data as it becomes more available. “I expect that integrating these different types of data will allow MISO to provide deeper insights into various aspects of cellular behavior,” Li said.
Source:
