Mapping Tumor Evolution to Improve Treatment

An Interview with Sohrab Shah, PhD | MSK SPECTRUM Co-chair

Summary

Ovarian cancer emerges from a constantly evolving ecosystem of cells. The MSK SPECTRUM project is one of the most ambitious efforts to map that evolution in the clinical setting by studying patients over time. This work brings together many groups and specialists at Memorial Sloan Kettering to combine single-cell genomics from samples collected throughout patients’ treatments with the mathematical modeling of evolutionary processes to understand how tumors adapt, resist therapy, and interact with the immune system.

 

One key finding was that treatment-resistant cancer cells may already be present and capable of evading detection by the immune system from the very start, which can drive recurrence after initial treatment. This challenges traditional assumptions about cancer progression and underscores the need for treatment that adapts to address the different types of cells present in the tumor as the cancer evolves. SPECTRUM is also paving the way for earlier detection and smarter, more personalized treatment, with the goal of improving long-term outcomes for patients.

 

Interview

Q: What is the MSK SPECTRUM study trying to accomplish?

A: The goal of the SPECTRUM study is to better understand how ovarian cancer tumors change over time, from initial diagnoses, during treatment, and at recurrence. To do this, we collect blood and tissue samples at each of these stages and use new technologies that allow us to study cancer at the level of individual cells.

 

This approach gives us a much clearer and more detailed picture of what’s happening inside a tumor and how that evolves over time and in response to treatment. The goal is to move beyond static descriptions of cancer and toward a framework where we can understand and eventually predict its behavior.

 

Q: What makes this research different from past studies?

A: In the past, most studies looked at tumors in bulk, meaning they averaged all the cells together and treated the tumor as a single entity. But tumors are actually made up of many different types of cells, each with their own characteristics, each affecting the tumor. By studying individual cells, SPECTRUM allows researchers to see that diversity in a way that wasn’t possible before.

 

On top of that, SPECTRUM is longitudinal, taking samples over time, and it is multi-modal, meaning that we integrate bulk whole-genome sequencing, gene expression profiling and DNA sequencing in single cells, imaging, and clinical data. And OCRA-supported projects are helping expand the research beyond single-cell data, including work using advanced imaging of diseased tissues and cells and AI-driven analysis, which can scale insights across larger patient populations.

 

That level of resolution and depth of study lets us see the evolutionary structure of the tumor in a way that was previously impossible. It reveals how different groups of cancer cells behave, how they respond to treatment, and how they interact with the immune system, which is critical for figuring out why some treatments work and others don’t.

 

Q: One of your key findings involves "whole-genome doubling." What is that, and why does it matter?

A: Whole-genome doubling is a process where a cancer cell duplicates all of its DNA, ending up with twice the normal number of chromosomes. This is something that happens quite frequently in ovarian cancer, and it has profound consequences. It increases genomic instability, which means cells can accumulate mutations more rapidly, giving the tumor more raw material for evolution—more ways to adapt to therapies.

 

What’s especially important is that this isn’t just a one-time event. The study found that genome doubling is happening continuously as the tumor evolves, and that different cells within the same tumor can have different amounts of DNA. This ongoing process can make tumors more adaptable and may contribute to treatment resistance and disease progression.

 

Q: How does genome doubling affect the immune system?

A: This is one of the more surprising aspects of the work. The study found that tumors with more genome-doubled cells tend to create an environment that suppresses the immune system. Normally, the body has ways of recognizing and attacking cancer cells, but these tumors seem to turn down those signals.

 

In particular, we observed changes in a pathway involving a protein called STING, which helps activate immune responses. When this pathway is suppressed, the immune system is less able to detect and respond to the cancer, which may explain why some ovarian cancers do not respond well to current immunotherapies. Understanding this mechanism may help us design effective immunotherapies.

 

Q: What did the study reveal about how ovarian cancer evolves over time?

A: One of the key findings is that ovarian cancer is constantly changing. By analyzing tumors at single-cell resolution, we can reconstruct when key events in the tumor’s evolution occur, such as genome doubling, and how they shape the trajectory of the disease.

 

At the time of diagnosis, tumors are already made up of many different groups (or clones) of cancer cells.  When treatment begins, many of these clones are eliminated, but not all of them. The cells that survive are often those that were already present and resistant to treatment from the beginning. Over time, these resistant cells can grow and lead to recurrence. This means that resistance isn’t always something that develops later; it can be present from the start, just in a small number of cells.

 

Q: How could this change the way we treat ovarian cancer?

A: These findings suggest that treatment may need to become more personalized and adaptable over time. Instead of basing decisions only on what the tumor looks like at diagnosis, doctors may eventually be able to track how the cancer is evolving and adjust treatment accordingly. In one example from the study, a patient benefited from a targeted therapy because doctors were able to identify the specific cancer cells driving her disease after initial treatment. This approach of treating what’s actually present at a given moment could lead to more effective and longer-lasting responses.

 

The SPECTRUM study is also helping us address an important question: how can we design treatments that limit the tumor’s ability to adapt? Approaches may include using strategic, sequential therapies, combining treatments from the start, or targeting key evolutionary processes like genome doubling. Ultimately, the goal is to develop diagnostic and treatment strategies that are tailored not just to the tumor's current state, but to its likely future trajectory.

 

Q: What role does technology play in this research?

A: This type of research generates enormous amounts of data, so advanced computational tools and modeling approaches are essential. Researchers use sophisticated models to analyze information from thousands of individual cells and identify meaningful patterns. Computation is a kind of “microscope” that allows scientists to make sense of complex data.

 

Without these tools, it would be nearly impossible to uncover the insights that come from studying cancer at this level of detail. We want to move toward predictive models that can help us anticipate how a tumor will evolve under a given treatment so we can adjust clinical management accordingly.

 

Q: What is the long-term vision for this work?

A: One of the biggest goals is to detect ovarian cancer earlier. Right now, many patients are diagnosed at a later stage, when the disease has already become very complex and harder to treat. We are now studying very early changes in the fallopian tubes, where many ovarian cancers begin. By understanding these early steps of cancer formation, we hope to identify biomarkers that could lead to earlier detection and, ultimately, better outcomes.

 

We are also expanding beyond ovarian cancer to include other gynecologic cancers, such as endometrial and cervical cancers. By building on the same infrastructure and approach, we aim to apply these high-resolution, single-cell, and spatial technologies more broadly, helping to uncover shared and distinct mechanisms across gynecologic malignancies.

 

Q: How do you hope other researchers will use this work?

A: A core principle of the SPECTRUM effort, aligned with OCRA’s broader goals, is open data sharing so that researchers around the world can build on it. The dataset is incredibly rich, and the team believes we’ve only begun to scratch the surface of what can be learned. By making the data widely available, we hope to accelerate discovery across the field and enable new insights that could lead to better treatments for patients. This approach is already having an impact, with other groups leveraging the data to support new studies and advance understanding beyond the original work. It’s why we’re excited to be getting in on the ground floor with our data in OCRA’s Discovery Lab.

 

The combination of deep biological data, collaborative infrastructure, and open access reflects a model of “team science” that is helping move the field forward more quickly. It is work that would not have been possible without OCRA’s critical funding and partnership that helped enable and sustain the research effort.

 

What This Means for Patients

• Ovarian cancer is not a static disease driven by one type of cell; it’s made up of many different cell types that can change over time.
• Some cancer cells may already be resistant to treatment from the start, which can lead to tumor recurrence later.
• New mathematical models and AI-based technologies are helping doctors better understand these changes, which may lead to more personalized and adaptive treatment strategies.
• Tumors can sometimes hide from the immune system, helping explain why immunotherapy doesn’t work for everyone.
• This research is paving the way for earlier detection and smarter treatment approaches, with the goal of improving long-term outcomes for patients.

 

 

 

Related Information

Explore the dataset in the Discovery Lab.

 

MSK SPECTRUM: https://componcmsk.org/msk-spectrum/

 

McPherson A, Vázquez-García I, et al. Ongoing genome doubling shapes evolvability and immunity in ovarian cancer. Nature, 2025. doi: 10.1038/s41586-025-09240-3.

 

Williams M, Vázquez-García I, et al. Tracking clonal evolution during treatment in ovarian cancer using cell-free DNA. Nature, 2025. doi: 10.1038/s41586-025-09580-0.

 

Vázquez-García I, Uhlitz F, et al. Ovarian cancer mutational processes drive site-specific immune evasion. Nature, 2022. doi: 10.1038/s41586-022-05496-1.

 

2022 OCRA Collaborative Research Development Grant – Microsoft AI for Health. “Inferring Mutational Processes and Patient Stratification From Standard-of-Care Clinical Imaging”. Granted to Sohrab Shah, PhD. This grant was made possible in part by a generous donation from The Edmée Firth Fund for Research in Ovarian Cancer (EFFROC). https://ocrahope.org/research/current-grantees/sohrab-shah-phd/

 

2019 Collaborative Research Development Grant. "Profiling co-evolution of ovarian cancer and its immune microenvironment". Granted to Sohrab Shah, PhD.