A pan-cancer spatial atlas from Asia Pacific reveals recurrent spatial theratopes

Cancer is rarely a single, uniform enemy. Inside a tumor, malignant cells do not exist in isolation; they are surrounded by a complex city of healthy cells, immune defenders, and structural support tissues. This surrounding environment, known as the tumor microenvironment, acts as a neighborhood where cells constantly interact, signal, and influence one another. Scientists have long known that the arrangement of these cells matters. Just as the layout of a city affects how traffic flows or how quickly a fire spreads, the physical organization of cells within a tumor can determine how fast the disease grows and how well it responds to treatment. However, until now, our maps of these cellular cities have been fragmented, often looking at only one type of cancer at a time or relying on data that mixes samples from different sources, making it difficult to see the bigger picture.

A new study from researchers across the Asia-Pacific region has changed this by creating a massive, unified map of nearly 400 tumors. This atlas covers 13 different types of tissue and 26 specific subtypes of cancer, bringing together data from institutions in Australia, Japan, and Taiwan. By using advanced technology to see exactly where every cell is located and what genes it is turning on, the team discovered that despite the vast differences between cancer types, these tumors share a surprising number of common neighborhood structures. Most importantly, they found that the specific arrangement of drug targets and immune cells in these neighborhoods can predict how a patient will respond to treatment, suggesting that the location of a target might be just as important as the target itself.

The researchers began by collecting tissue samples from patients with a wide variety of cancers, including those of the lung, liver, breast, and colon. They used a high-resolution imaging technique called Xenium to scan these tissues, looking at nearly 2,000 specific genes in almost 57 million individual cells. This allowed them to build a detailed, three-dimensional picture of the tumor's interior, identifying not just the cancer cells but also the immune cells, blood vessels, and support cells surrounding them. To ensure their findings were accurate, they also measured proteins in the same cells using a different method, confirming that the genetic signals they saw matched the physical proteins present. This rigorous approach ensured that the data was consistent and reliable, regardless of which hospital or country the sample came from.

One of the first things the team noticed was that while the cancer cells themselves looked very different depending on where they originated, the surrounding support cells often looked and acted the same. Immune cells and structural cells shared similar features across different types of cancer, suggesting that the body builds these neighborhoods using a common set of blueprints. However, the cancer cells remained distinct, retaining their unique identities based on their tissue of origin. This distinction is crucial because it means that while the "neighborhood" might be similar, the "residents" driving the disease are unique to each cancer type.

The researchers then grouped the cells based on who they were standing next to. They found that cells naturally clustered into eight distinct types of neighborhoods. Some neighborhoods were dominated by immune cells like T cells or B cells, while others were filled with support cells or blood vessels. A key discovery was that macrophages, a type of immune cell that acts as a cleaner and organizer, were the most consistent organizers across all these neighborhoods. Wherever macrophages were found, they seemed to shape the behavior of the cells around them, much like a community leader influencing the mood of a neighborhood. In some areas, these macrophages appeared to be helping the immune system fight the cancer, while in others, they seemed to be helping the cancer hide from the immune system.

Building on this understanding of neighborhoods, the team developed a new way to classify tumors based on "Spatial Theratopes." This concept looks at where specific drug targets are located relative to immune checkpoints. For example, many modern cancer drugs work by blocking a protein called PD-L1, which cancer cells use to turn off immune attacks. The researchers found that the effectiveness of these drugs depends heavily on whether the target is physically close to the immune cells it needs to engage. They identified four main patterns of these interactions. In one pattern, a drug target called VEGF-A, which helps tumors build blood vessels, was consistently found right next to PD-L1. This pairing was seen across many different types of cancer, suggesting that drugs targeting both of these together might work broadly, regardless of the tumor's origin.

The study also tested whether these neighborhood patterns could predict patient outcomes. In a group of patients with liver cancer who were treated with a combination of drugs targeting VEGF-A and PD-L1, the researchers found that the simple presence of the drug targets was not enough to guarantee success. Instead, the outcome depended on the surrounding neighborhood. Patients whose tumors had a high proportion of a specific support-cell neighborhood responded poorly to the treatment and saw their disease progress faster. In contrast, patients with a different neighborhood structure, rich in immune cell clusters, tended to do better. This suggests that looking at the map of the tumor, rather than just the presence of a single target, provides a much clearer picture of how a patient will respond.

These findings offer a new perspective on how to fight cancer. Rather than treating every tumor of a specific type the same way, doctors might one day be able to look at the spatial arrangement of cells to choose the best therapy. The study shows that the physical organization of a tumor is a powerful indicator of its behavior. By understanding these shared neighborhood structures, researchers can design better combinations of drugs that target the specific layout of the cancer. While the study is based on retrospective data and requires further testing in future clinical trials, it provides a strong foundation for a more precise, location-aware approach to cancer treatment. The atlas itself is now available to the scientific community, serving as a shared resource to help develop therapies that are tailored not just to the type of cancer, but to the unique neighborhood in which it lives.