Columbia engineers develop first multi-organ chip to model cancer metastasis in human tissue

Two-thirds of cancer deaths trace back to a single, brutal fact: the disease has learned to travel. When a tumor breaks apart and sends cells coursing through the bloodstream to settle in distant organs—a process called metastasis—the body's defenses crumble. Drugs designed to stop this spread have failed more often than they've succeeded, largely because scientists have lacked a reliable way to watch it happen in human tissue. Animal models have taught us much, but a mouse is not a person, and cancer behaves differently in rodent bodies than it does in ours.

In August 2026, researchers at Columbia Engineering announced they had built something new: a multi-organ chip small enough to hold in your hand, engineered to replicate how cancer cells escape the bloodstream and colonize distant tissue. The device contains millimeter-sized sections of human bone and lung, connected by a vascular channel through which circulating breast cancer cells flow. It is, in essence, a working model of metastasis built from actual human cells.

Gordana Vunjak-Novakovic, a Columbia University professor and leader of the Laboratory for Stem Cells and Tissue Engineering, led the team that developed the platform. Her motivation was straightforward: the field needed a human-based alternative to animal studies. "The pressing need for developing human tissue models of metastasis has been a key motivation for our study," she said. The chip was designed to answer specific questions about how cancer cells breach the inner lining of blood vessels, survive in foreign tissue, and reprogram their new environment to support their own growth.

The engineering itself required precision. The bone and lung tissues were grown from induced pluripotent stem cells—adult cells reprogrammed to an embryonic state—using specialized scaffold-bioreactor systems. Each tissue compartment was optimized separately, then linked together through vascular circulation that mimicked the body's own blood flow. A selectively permeable endothelial barrier separated the tissue from the vascular channel, just as it does in living bodies. When the researchers introduced breast cancer cells into the circulation, the chip revealed patterns that matched what happens in patients: cancer cells that naturally gravitate toward bone showed stronger colonization of bone tissue and caused more pronounced bone degeneration, while cells that favor lung tissue caused greater disruption there and only modest bone colonization.

But the most striking discovery emerged from post-analysis of the engineered tissue. The cancer cells were not simply invading passive organs. They were actively conditioning the distant tissues before colonizing them, preparing the ground like an army laying siege. This phenomenon, called pre-metastatic niche formation, had been observed in patients but was difficult to study directly. The chip made it visible and measurable.

Vunjak-Novakovic emphasized what set this work apart: "The key advantages of this advanced model of metastasis are that it is human and can be patient-specific. It faithfully mimics some of the key aspects of human metastasis that are otherwise largely inaccessible for direct study." The implications ripple outward. As the FDA and NIH increasingly push for alternatives to animal testing in drug development, this chip represents a concrete example of what that shift looks like in practice. Ilaria Baldassarri, a PhD student and lead author of the study published in Science Translational Medicine, saw it as a turning point: engineered human tissues that could complement—and eventually replace—what animal models alone could teach.

The work was a collaboration across Columbia's cancer and engineering divisions, bringing together expertise in systems biology, pathology, and tissue engineering. What emerges is not just a better tool for understanding metastasis, but a template for how future drug development might work: testing compounds against human tissue in a controlled environment before they ever reach a patient. For a disease that kills by spreading, that could change everything.