In the landscape of cancer research, a quiet revolution is unfolding at the Institute for Systems Biology. There, a team of researchers has uncovered something remarkable: some lung tumors are not what they appear to be—and they can change their identity over time, slipping between cancer types in ways that have confounded scientists for years.
The findings, published in Cell Reports Medicine, focus on a rare but particularly aggressive form of lung cancer known as combined small-cell lung cancer, or cSCLC. These tumors contain a mix of both small-cell and non-small-cell cancer features, and doctors have historically treated them as standard small-cell lung cancer. The problem? Patients with these mixed tumors tend to fare worse—likely because the treatment didn't match the full picture of their disease.
Using cutting-edge spatial and single-cell genomic tools, Wei Wei, Ph.D., and colleagues at ISB, alongside collaborators at Fudan University, traced the origins of these puzzling tumors. Their conclusion challenges a long-held assumption. "We found that these tumors are not simply mixtures of different cancer types," Wei said. "They are dynamic systems, with cancer cells actively changing their identity."
The research revealed that cSCLC tumors arise from a single ancestral cell that gradually evolves, with its descendants shifting between identities as the cancer grows. Even more striking, about one-third of the SCLC-like tumor cells analyzed existed in hybrid states—carrying features of multiple cancer types simultaneously. Cancer, it seems, does not operate as a simple on-off switch but as a shifting continuum.
The team also mapped the tumor microenvironments and found that different regions within the same tumor created distinct neighborhoods. Some areas teemed with immune cells ready to fight; others were nearly sealed off, surrounded by dense bands of fibroblasts that may shield the cancer from attack.
But perhaps the most practical breakthrough from this work is a new diagnostic tool called cSCLC Detector—a four-gene test that can identify these mixed tumors more accurately than traditional methods. Because the small-cell and non-small-cell regions of cSCLC share early "trunk" mutations despite looking different under a microscope, a biopsy of just one section can still reveal clues about the whole tumor's identity. When tested on datasets of patients previously diagnosed with standard small-cell lung cancer, the tool flagged a substantially higher proportion of cases with combined features—suggesting the disease has been significantly underdiagnosed.
For patients, this work represents a shift in philosophy as much as science. "Cancer is not static," Wei said. "To treat it effectively, we need to understand how it evolves—not just what it is at a single point in time."