A PDX model, short for patient-derived xenograft, is a living cancer research tool created by transplanting a piece of a human tumor directly into an immunodeficient mouse. The mouse’s weakened immune system allows the human tissue to survive and grow, producing a replica of the patient’s original cancer that researchers can study and test drugs against. PDX models are widely considered one of the best ways to bridge the gap between lab research and actual clinical treatment.
How a PDX Model Is Created
The process starts with a tumor sample taken from a cancer patient during surgery or biopsy. That tissue is then implanted into a specially bred mouse that lacks a functioning immune system. Without an immune response, the mouse’s body won’t reject the foreign human tissue, and the tumor can establish itself and grow.
One key difference between PDX models and older methods is that there’s no lab culture step in between. Traditional xenograft models use cancer cell lines, which are cells that have been grown in dishes, sometimes for years, before being injected into mice. That long stretch of artificial growth changes the cells. PDX skips this entirely. The tumor fragment goes from the patient into the mouse as quickly as possible, preserving the original architecture, the mix of different cell types, and the interactions between them.
The first mouse to receive the implant is called F0. Once the tumor grows, researchers can transfer pieces of it into additional mice, creating subsequent generations labeled F1, F2, F3, and so on. This lets a single patient’s tumor supply enough material for extensive drug testing across dozens of animals.
Why PDX Models Matter More Than Cell Lines
Cancer cells grown in a flat dish behave differently from cancer cells inside a living body. They lose their three-dimensional structure, their blood supply networks, and much of the genetic diversity that makes real tumors so difficult to treat. PDX models solve many of these problems. Studies show they retain the tissue structure, blood vessel architecture, and molecular characteristics of the original patient tumor. That complexity is what makes a tumor respond to one drug and resist another, so preserving it gives researchers far more realistic test results.
This matters most for drug development. A therapy that kills cancer cells in a dish may fail completely in a human patient because the dish didn’t replicate the real conditions. PDX models offer a much closer match. Clinical trials have confirmed that drug response rates observed in PDX models correlate significantly with actual patient outcomes, with one study reporting an 88% objective effective rate when PDX-guided treatments were applied clinically.
Personalized Treatment and Drug Screening
One of the most promising uses of PDX models is in personalized medicine. The idea is straightforward: take a patient’s tumor, grow it in several mice, and test different drugs on each one. Whichever drug works best in the mice becomes the recommended treatment for that specific patient. Researchers sometimes call these “avatar” models because the mice stand in for the patient.
This approach is especially valuable when standard treatments have failed and oncologists need to identify an alternative. PDX models allow teams to test and rank potential therapies before administering them to the patient, essentially creating a personalized drug screening platform. For example, researchers have used PDX models to identify that certain repurposed drugs can block specific cancer-driving proteins, opening treatment options that wouldn’t have been considered through standard protocols. In colorectal cancer, PDX testing has shown that combining two classes of targeted drugs can be effective for tumors with specific genetic mutations, helping clinicians tailor therapy based on the patient’s mutational profile.
Engraftment Rates Vary by Cancer Type
Not every tumor sample successfully grows in a mouse. The success rate, called the engraftment rate, varies widely depending on the type of cancer. Aggressive, fast-growing tumors tend to engraft more reliably. Breast cancer, for instance, has historically been one of the most challenging types to establish as a PDX, with reported success rates ranging from just 8% to 77%. That wide range reflects differences in tumor subtype, the amount of tissue available, and the specific mouse strain used.
This variability is one reason PDX models aren’t yet a routine clinical tool for every patient. When engraftment fails, there’s no model to test, and the time and resources invested are lost.
Limitations of PDX Models
The biggest limitation is the mouse’s missing immune system. Because the mouse must be immunodeficient to accept human tissue, PDX models can’t be used to test immunotherapies, which are drugs that work by activating the patient’s own immune cells to attack cancer. Since immunotherapy has become one of the most important treatment categories in oncology, this is a significant blind spot.
Researchers are working around this by creating “humanized” mice, animals that receive transplanted human immune cells alongside the tumor. These models allow some degree of immune interaction, but they’re more complex and expensive to produce, and they don’t perfectly replicate a full human immune system.
Cost and time are also real barriers. PDX models require specialized animal facilities, trained staff, and months of waiting for tumors to grow through multiple passages. For a patient with an aggressive cancer that needs treatment decisions within weeks, a PDX model may simply take too long to be useful. The extensive use of animals also raises ethical concerns that limit how widely these models can be scaled.
Finally, as tumors pass through multiple generations of mice, they undergo what’s called mouse-specific evolution. The human tumor cells survive, but the surrounding support tissue (blood vessels, connective tissue) gradually gets replaced by mouse equivalents. Over many passages, this can subtly shift the tumor’s behavior away from the original patient’s cancer.
PDX Models vs. Patient-Derived Organoids
A newer alternative to PDX is the patient-derived organoid, or PDO. Organoids are tiny three-dimensional clusters of tumor cells grown in a dish using specialized growth factors and scaffolding materials. They capture some of the architectural and molecular features of the original tumor without requiring any animals.
The trade-offs are clear. PDO models are cheaper, faster, and can be scaled up for high-throughput drug screening, where dozens or hundreds of compounds are tested simultaneously. They also avoid the ethical issues of animal use. However, they lack the blood vessel networks, surrounding tissue complexity, and immune interactions that PDX models provide. This means organoids can’t be used to study drugs that target blood vessel growth or immune pathways.
For straightforward drug sensitivity testing in precision oncology, some researchers now argue that organoids may be the better first choice given their lower cost and faster turnaround. PDX models remain superior when the research question requires a full living system, particularly for studying how tumors evolve, resist treatment, or interact with their surrounding environment over time.
Where PDX Models Fit in Cancer Research
PDX models sit in a specific niche: they’re more realistic than cell lines or organoids, but slower and more expensive. They’re less convenient than computer models, but far more biologically accurate. Their greatest strength is fidelity to the original patient’s tumor, which makes them uniquely suited for testing targeted therapies, studying drug resistance, and guiding treatment decisions when standard options have been exhausted.
The concept dates back to the 1960s and 1970s, when researchers first implanted human tumor tissue into nude mice (a strain born without a thymus and therefore lacking key immune cells). Since then, advances in mouse genetics have produced increasingly immunodeficient strains that accept human tissue more reliably, expanding the range of cancers that can be modeled this way. Today, large PDX biobanks catalog hundreds of tumor models across cancer types, giving researchers a library of living tumors to draw from without needing to start fresh from a patient sample each time.