The ‘Gooey Cake’ of Cancer: How Spatial Biology Could Transform Clinical Predictions

When Alex Xu began his career as an engineer, he built devices and injected materials into cells. But today, his tools are not microfluidic chips, they're microscopes and mass spectrometers that map how cells talk to one another inside tumors.

At a recent biomedical engineering seminar held in Stony Brook University's Javits Center, special guest Xu, now an assistant professor of bioengineering at the University of Maryland and a former project scientist at Cedars-Sinai, described how his career shift reshaped his approach to medicine. His new focus is finding biomarkers, which are measurable biological features that reveal a patient's risk or treatment outcome.

"We're asking, is there a single thing we can measure that tells us the most about what's going on in a patient?" Xu explained.

In his lecture, Xu covered his multiple research studies. While explaining all of his projects, he used the same analogy throughout: the complexity of human tissue is like a pineapple fruit cake. You have the pineapple chunks, the cherries and the gooey matter around it. All of these different components coexist together as spatial biology. The "flavors" depend on how each piece is arranged.

"Spatial biology is making a huge dent on the current state of tissue measurements," said Xu. "For the future, the excitement comes from getting a clearer picture of human biology than ever before, and it will be on us to develop new computational tools, new experimental models, and new engineered therapies that take full advantage of this knowledge."

Spatial biology combines imaging and molecular mapping to measure where cells are located and how they interact, revealing a "geography" of tissues that can explain why some patients respond to treatment while others do not.

Xu's first case study focused on ovarian cancer. This form of cancer is a disease that is often diagnosed late and known for its complex tissue structure. His team analyzed tumor samples from 42 patients using imaging mass cytometry, a cutting-edge technique that stains tissues with metal-tagged antibodies and vaporizes them to detect dozens of proteins at once.

At first, the results were disappointing. Measurements of cell types and counts didn't correlate with patient outcomes. But when the team dug deeper, quantifying not only which cells were present but also wherethey were, patterns began to emerge.

By developing mathematical "spatial metrics" to describe how immune cells, tumors, and fibroblasts clustered, Xu's group finally uncovered correlations between tissue layout and early relapse. They also identified plasma cells as potential indicators of disease recurrence, results later published in Science Advances (2024).

Xu extended his analogy, within the fruitcake lies the "gooey" part - which represents the interactions between tumor and immune cells. These molecular exchanges, he argues, may hold the key to predicting how a patient's cancer behaves.

In studies of Hodgkin Lymphoma, Xu's team mapped thousands of immune-cell interactions and found that certain spatial arrangements, where tumor cells are surrounded by dense clusters of T-cells, were linked to how patients responded to chemotherapy. The findings, published in JCO (2024)and Nature Biomedical Engineering (2025), suggest that cell-to-cell communication networks could serve as powerful biomarkers for treatment outcomes.

"It's not only about the type of fruit or where it's placed," Xu said. "It's about what happens in between, the communication that makes the whole thing hold together."

Xu's group is working to make the technology clinically practical. Using machine-learning models, they reduced 35 protein markers down to a six-marker panel measurable on standard hospital equipment.

Students and faculty from all over Stony Brook University filled the lecture hall.

One student, a biomedical engineering major at Stony Brook, noted that spatial biology can be used to analyze cancer tissue and figure out how likely it is to relapse or what treatments would be effective.

Xu's work represents a new kind of clinical engineering, blending precision measurement with metaphors, aiming to turn spatial data into tools doctors can use in practicality.

- Angelina Livigni