In an effort to produce effective targeted therapies for cancer, scientists have to isolate the genetic and phenotypic characteristics of cancer cells, each inside and across different tumors, because those differences impact how tumors reply to treatment.
A part of this work requires a deep understanding of the RNA or protein molecules each cancer cell expresses, where it’s situated within the tumor, and what it looks like under a microscope.
Traditionally, scientists have checked out a number of of those facets individually, but now a brand new deep learning AI tool, CellLENS (Cell Local Environment and Neighborhood Scan), fuses all three domains together, using a mix of convolutional neural networks and graph neural networks to construct a comprehensive digital profile for each single cell. This enables the system to group cells with similar biology — effectively separating even those who appear very similar in isolation, but behave in another way depending on their surroundings.
The study, published recently in , details the outcomes of a collaboration between researchers from MIT, Harvard Medical School, Yale University, Stanford University, and University of Pennsylvania — an effort led by Bokai Zhu, an MIT postdoc and member of the Broad Institute of MIT and Harvard and the Ragon Institute of MGH, MIT, and Harvard.
Zhu explains the impact of this recent tool: “Initially we’d say, oh, I discovered a cell. This is named a T cell. Using the identical dataset, by applying CellLENS, now I can say it is a T cell, and it’s currently attacking a selected tumor boundary in a patient.
“I can use existing information to raised define what a cell is, what’s the subpopulation of that cell, what that cell is doing, and what’s the potential functional readout of that cell. This method could also be used to discover a brand new biomarker, which provides specific and detailed details about diseased cells, allowing for more targeted therapy development.”
It is a critical advance because current methodologies often miss critical molecular or contextual information — for instance, immunotherapies may goal cells that only exist on the boundary of a tumor, limiting efficacy. Through the use of deep learning, the researchers can detect many alternative layers of data with CellLENS, including morphology and where the cell is spatially in a tissue.
When applied to samples from healthy tissue and several other sorts of cancer, including lymphoma and liver cancer, CellLENS uncovered rare immune cell subtypes and revealed how their activity and site relate to disease processes — similar to tumor infiltration or immune suppression.
These discoveries could help scientists higher understand how the immune system interacts with tumors and pave the best way for more precise cancer diagnostics and immunotherapies.
“I’m extremely excited by the potential of latest AI tools, like CellLENS, to assist us more holistically understand aberrant cellular behaviors inside tissues,” says co-author Alex K. Shalek, the director of the Institute for Medical Engineering and Science (IMES), the J. W. Kieckhefer Professor in IMES and Chemistry, and an extramural member of the Koch Institute for Integrative Cancer Research at MIT, in addition to an Institute member of the Broad Institute and a member of the Ragon Institute. “We are able to now measure an incredible amount of data about individual cells and their tissue contexts with cutting-edge, multi-omic assays. Effectively leveraging that data to nominate recent therapeutic leads is a critical step in developing improved interventions. When coupled with the correct input data and careful downsteam validations, such tools promise to speed up our ability to positively impact human health and wellness.”
