The researcher is aiming to improve the odds for these patients in two recent studies that are primarily concerned with understanding why and how people respond differently to diseases and therapies.

Researchers from Ahmet Coskun, a Georgia Tech and the findings were published in the journals NPJ Precision Oncology and iScience. "What we have learned is connectivity and communication between molecules and between cells is what really controls everything, regarding whether or not patients get healthy, or how they will respond to drugs," said Coskun, an assistant professor in the Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University.

The studies detail the development of tools and techniques to deeply explore the tumour microenvironment at the subcellular level, utilising the Coskun lab's expertise in combining multiplex cellular imaging methods with artificial intelligence.

"We are developing a better grasp of cellular signalling and decision making, and how it is coordinated in the tumour microenvironment, which can lead to better personalised, precision treatments for these patients," said Coskun, who is keenly interested in why some patients respond to groundbreaking immunotherapy drugs, and some don't.

With that in mind, his team developed SpatialVizScore, a new method they describe in NPJ Precision Oncology, to deeply study tumour immunology in cancer tissues and help identify which patients are more likely to respond to immunotherapy. It's a significant upgrade to the current standard methodology used by cancer physicians and researchers, Immunoscore.

Scoring Immunity

Immunoscore is used as a prognostic tool, measuring how well the body's immune cells surround and enter a tumour. It has shown promise in predicting a patient's risk of disease recurrence, a key step in developing a personalised treatment plan. A high score indicates better immune cell infiltration, while a low score indicates a greater risk for recurrence.

But immune cells are moving targets and exhibit a high level of molecular complexity that can't always be adequately captured by conventional Immunoscore methods. With SpatialVizScore, Coskun's team has given immunoscoring an expanded reach.

While the standard method looks at how T cells interact with tumours, Coskun's system looks at the interactions of additional immune cells, such as macrophages, which have two subtypes - M1 and M2, which often find themselves in conflict. M1 helps eliminate pathogens, while M2 can promote tumour growth.

Coskun's multiplex imaging system looks at all of that, visualising how these cells communicate and interact with each other, and with cancer cells, not just in and around the tumour, but across the tumour environment.

"Because cancer cells and immune cells are not always near each other, we visualise spatial connectivity, and we visualise communities of cells or neighbourhoods," Coskun said.

"But we're not only looking at a cancer cell interacting with an immune cell. We're looking at immune-immune interactions, too. By looking at the effects of these different interactions, we can explain the tumour, and we can come to develop a more comprehensive Immunocore."

Zooming In

In the iScience study, the team moves away from communities and neighbourhoods of cells. Instead, they zoom in on subcellular protein-protein interaction networks, which can affect signalling pathways in cancer -- when molecules in a cell work together to control a cell function, which can be cell division or cell death.

Each molecule activates another molecule, and the process is repeated along the "pathway" until the last molecule in line is activated and the cell function -- good or bad -- is carried out. Abnormal activation of a pathway can lead to cancer, but some drugs target specific molecules involved and can keep cancer cells from growing.

Coskun and his team are using their multiplex imaging tools and machine learning to probe protein-protein interactions to decipher the pathogenesis of signalling pathways that contribute to drug resistance in non-small-cell lung cancer.

"We can observe and map protein activity," said Coskun, whose team developed an imaging technique with subcellular resolution called rapid multiplexed immunofluorescence (RapMIF).

"Proteins are making the decisions that affect our cells," Coskun added. "Now we can see how they're communicating, how they're affecting what our cells ultimately do. It's a signalling discovery approach that can be used in designing precision therapies and ultimately help more patients who are battling cancer."