Colorectal cancer (CRC) has emerged as one of the most pressing public health crises of the 21st century. Once viewed primarily as a disease of aging, it is now the leading cause of cancer-related mortality among adults under 50 in the United States and the second leading cause of cancer death overall. Despite the rapid evolution of oncology, the most promising frontier—immunotherapy—has remained frustratingly inaccessible for the vast majority of patients.
Dr. Karin Pelka, an investigator at the J. David Gladstone Institutes and a recipient of the prestigious Cancer Research Institute (CRI) Technology Impact Award, is challenging this status quo. By applying cutting-edge artificial intelligence to the microscopic "molecular conversations" occurring within tumor environments, Dr. Pelka is attempting to decode why most colorectal tumors remain invisible to the immune system, and more importantly, how to change that.
The Landscape: Why Immunotherapy Struggles with Colorectal Cancer
To understand the scope of the challenge, one must first understand how immunotherapy functions. Drugs known as checkpoint inhibitors work by "releasing the brakes" on the immune system, allowing T cells to recognize and destroy cancerous invaders. However, this mechanism relies on the tumor being "visible."
In cancers like melanoma or non-small cell lung cancer, the tumor cells are often laden with DNA mutations. These mutations create "neoantigens"—abnormal proteins that appear foreign to the immune system, effectively acting as a red flag that attracts immune cells. In contrast, most colorectal cancers are "immunologically cold." They do not accumulate enough mutations to signal their presence effectively, allowing them to hide in plain sight.
The 5% Miracle
The exception to this rule lies in a small subset of patients—roughly 5% to 10%—whose tumors exhibit a "mismatch repair deficiency" (dMMR). These tumors are highly unstable and mutation-rich, making them prime targets for immunotherapy.
In a landmark 2022 clinical trial, the results were nothing short of historic: patients with locally advanced rectal cancer who were treated with immunotherapy saw their tumors disappear entirely. This success allowed these patients to avoid the grueling "trifecta" of traditional treatment: chemotherapy, radiation, and surgery. For this small group, the cure was complete and the side effects minimal.
"The vast majority of colon cancer falls into this relatively large area of tumors where we haven’t quite figured out yet how to use the immune system to fight them," Dr. Pelka notes. Her mission is to translate that 5% miracle into a standard of care for the remaining 90% of patients.
Chronology of a Scientific Shift
The journey toward understanding the tumor microenvironment has been a decades-long progression, marked by distinct technological leaps:
- 1990s–2000s: The Genomic Era. Early research focused on identifying the specific genetic mutations that cause colon cancer. While this led to targeted therapies, it proved insufficient as cancer cells rapidly evolved resistance to these drugs.
- 2010s: The Immunotherapy Breakthrough. The advent of checkpoint inhibitors revolutionized cancer care, yet it quickly became clear that these drugs were not a "one-size-fits-all" solution.
- 2020–2022: The Clinical Proof-of-Concept. The dMMR rectal cancer trials confirmed that if the immune system is properly engaged, surgery and radiation can potentially be avoided. This provided the "north star" for researchers like Dr. Pelka.
- 2024–Present: The AI and Spatial Transcriptomics Era. Dr. Pelka’s current research utilizes high-resolution imaging and machine learning to map the spatial architecture of tumors, marking a move from studying which genes are active to where and how they interact within the tumor.
Mapping the "Immune Hubs"
The core of Dr. Pelka’s recent work involves identifying "immune hubs"—highly structured clusters within a tumor where killer T cells, cancer cells, and structural tissue engage in complex molecular signaling.
Think of these hubs as the front lines of a battlefield. In some instances, the communication within these hubs leads to the destruction of the tumor. In others, the cancer successfully hijacks the conversation, signaling the immune cells to stand down or even support the tumor’s growth.
The Predictive Power of Structure
Dr. Pelka’s team discovered that the presence of these immune hubs in a specific, activated state is a powerful predictor of how a patient will respond to immunotherapy. Remarkably, these hubs are not unique to colorectal cancer; they have been observed in lung cancer and melanoma as well. This suggests that the "social network" of a tumor follows universal biological rules. The current research focus is to determine what prevents these hubs from forming—or remaining active—in patients who currently fail to respond to immunotherapy.
Teaching AI to Read the Language of Cells
The complexity of these molecular interactions is staggering. To parse through millions of data points, Dr. Pelka’s lab is leveraging Geneformer, a foundational AI model developed in collaboration with Dr. Christina Theodoris.
Much like a Large Language Model (LLM) learns the structure of human language, Geneformer was trained on tens of millions of human cell profiles to understand the "grammar" of gene activity. By applying this to the spatial transcriptomics of a tumor, Dr. Pelka can identify "central regulators"—the master molecular switches that dictate whether a cancer cell remains visible to the immune system.
By mapping these switches across the physical landscape of the tumor, the researchers aim to identify new therapeutic targets. The goal is to develop drugs that can "flip" these switches, turning an "immunologically cold" tumor into a "hot" one that the immune system can finally identify and eliminate.
Implications: A New Era for Early-Onset Patients
The urgency of this work cannot be overstated. With a shifting demographic of CRC patients—many of whom are under 50 and facing long lives ahead—the limitations of traditional therapies are increasingly untenable.
The Case for Screening
While Dr. Pelka works on the next generation of immunotherapy, the medical community emphasizes that the most powerful tool currently available is prevention. Colorectal cancer typically begins as benign polyps, which take 10 to 15 years to develop into cancer. Regular screenings, starting at age 45, allow for the removal of these polyps before they ever become malignant.
The Vision of a "Living Therapeutic"
For those already diagnosed, the appeal of immunotherapy lies in its potential to act as a "living therapeutic." Unlike fixed chemical drugs, which cancer cells can eventually learn to circumvent, an active immune response can adapt. If the cancer evolves, the immune system evolves with it, potentially providing long-term, durable protection.
Reflecting on the support from the Cancer Research Institute, Dr. Pelka emphasizes the importance of funding high-risk, high-reward research. "The CRI Technology Impact Award is a perfect example of bringing together new technologies with the problem of getting cancer immunotherapy to work," she says. "Having a foundation willing to take a bet early on—when the science is not yet ready to be employed widely—is often research that is hard to fund. But that is exactly where the needle moves."
As Dr. Pelka and her colleagues continue to decode the tumor’s secrets, they bring us closer to a future where colorectal cancer is no longer a death sentence, but a manageable condition—or better yet, a memory. The work is difficult, the data is vast, and the stakes are the highest they have ever been, but the path toward a universal immunotherapy for colorectal cancer is finally being mapped.
