Making More Accurate, Personalized Cancer Treatment Decisions With Clinically Relevant PDX Models

In the past few decades, researchers and clinicians have gained a deep understanding of the biology behind tumor growth and metastasis and the clinical course for many different types of cancers. Despite these advancements, cancers are still treated in a rudimentary way: Through trial-and-error. For rare, recalcitrant, or metastasized cancers the situation is more dire. With little clinical data, doctors may have no information to base their treatment decisions. Given these knowledge gaps, wouldn’t it be powerful to accurately predict your tumor’s progression, metastasis, and treatment response?


Forecasting all of these clinical characteristics has been the core goal of personalized medicine. And there have been some incredible successes. Many clinical aspects of cancer can be, in part, informed through the detection of specific DNA mutations, RNAs, or proteins in tumors. These biological markers, collectively called biomarkers, can be used to diagnose specific cancer types or stages, estimate clinical outcomes, and select appropriate treatments.1


But there are still some challenges with using biomarkers for such clinical applications.2 Chief among them, says Drs. Hannah Wise and David Solit of Memorial Sloan Kettering Cancer Center in their preview in Cancer Cell is that “...current tumor profiling panels identify genomic alterations that robustly predict for drug response in only a minority of cancer patients.”2 Also, many patients have multiple biomarkers that aren’t always analyzed but can have dramatic effects on treatment response.2 


Overall, biomarkers provide an incomplete view: An out-of-focus, static snapshot of a patient’s cancer and its potential. The heterogeneity of cancer, within an individual’s tumor and even across individuals with similar cancer types, makes this lack of detail and dynamics they provide all the more precarious. Every picture of a tumor can be fuzzy and incomplete in a different way. For patients, caregivers, and doctors, this makes personalized clinical decisions difficult to make confidently. 


Clarifying Personalized Medicine With Patient-Derived Xenografts (PDX) Models

Over the past few decades, researchers, drug developers, and clinicians have been developing methods for getting a crisper, clearer picture of how an individual’s cancer will behave in the future. PDX models are one such method and allow doctors to mimic many aspects of cancer, including growth rate, metastatic potential, and treatment response. 


The method uses special immunodeficient mice to grow tumor cells that have been isolated from a patient’s tumor. Once introduced to the mouse, a process called engraftment, tumor cells can grow as they would on or in a patient’s body. Clinicians can use the engrafted tumor to test individual treatments or treatment combinations and use the data to make clinical decisions. It’s like making a photocopy of your tumor, preserving many of your personal, clinically relevant details.


Today, the method is poised to change how clinicians individualize cancer treatment decisions. A recent review of the scientific literature shows how clinically-relevant these models can be with colon, pancreatic, brain, breast, lung, skin, and ovarian cancers.3 For instance, in a publication from researchers at MD Anderson Cancer Center, PDX models were established from 25 patients with breast cancer, 24 of which had triple-negative breast cancer (TNBC). TNBC is infamously aggressive and its genetic heterogeneity makes it difficult to develop personalized treatments. The research group was able to establish “PDXs [that] capture the molecular and phenotypic heterogeneity of TNBC” and test several approved and exploratory therapeutics.4 They conclude, “the heterogeneity of TNBC clearly adds to the challenge of precision oncology; however our models provide a representative panel for testing of novel agents and biomarker-sensitivity associations,” demonstrating the clear clinical utility and potential of PDX models in highly variable cancer types.4


There are still barriers to the widespread adoption of PDX models and a need for standardization. The above-mentioned review outlines some core areas in the PDX methodology that require further optimization. Let’s discuss some of these challenges.

Optimizing PDX Procedures

There are two primary ways to introduce tumor cells into a PDX mouse model. Tumor samples can be injected just beneath the surface of the skin, called subcutaneous PDX, or in the same organ or tissue as it occurs in your body, called orthotopic PDX or O-PDX. For many tumor types, O-PDX answers the questions posed at the beginning of this post: It accurately replicates tumor progression, metastatic potential, and treatment responses seen in the clinic.


Yet subcutaneous methods are more common than O-PDX models as it’s easier to measure tumor growth. While this is great for laboratory technicians developing a PDX model, it makes the data collected less clinically accurate or applicable to you as a patient. 


There has been a lot of discussion about subcutaneous vs orthotopic methodologies in the scientific literature, with several studies confirming O-PDX’s ability to more accurately recapitulate clinically relevant details. A recent editorial, published in the British Journal of Cancer, sings the praises of the method, proposing, “...the use of xenografts to expand the current limited basis for drug testing in cancer patients, minimizing the need for expensive and prolonged randomized controlled trials.”5 Many primary studies have reinforced this assertion in many different cancer types, including high-risk neuroblastoma, rare Rhabdomyoscarcoma, and others.6,7


Part 2 of this blog series will discuss the subcutaneous PDX vs. O-PDX debate in further clinical detail, so be sure to subscribe to our blog and get it delivered straight to your inbox!


Achieving Clinically Relevant Timelines

Establishing a PDX model can take as long as 8 months, depending on how successful tumor growth is in the immunodeficient mouse models used.3 For many cancers, that’s simply too long to wait. Tumors can often fail to engraft, another barrier to the widespread use of PDX models. 


Yet, O-PDX methodology may hasten these timelines providing actionable insights for patients and clinicians: A study by Kamili et al. demonstrated that “orthotopic engraftment was more efficient than subcutaneous or intramuscular engraftment” and it “...allows more rapid model development, increasing the likelihood of developing an avatar model within a clinically useful timeframe.”6


Minimizing Changes to the Tumor Microenvironment

The area around a tumor, also known as the tumor microenvironment, plays an important role in the clinical progression of an individual’s cancer. There are typically specialized immune cells, a blood supply, and structural support that can affect how a tumor grows and responds to treatment. 


With PDX models, tumors can sometimes be passaged multiple times in or on a mouse, which can cause progressive changes in the tumor microenvironment.3 The tumor is infiltrated by special mouse cells and structural elements from the original tumor are replaced by the mouse counterparts. Minimizing the number of passages and therefore, the changes in the tumor microenvironment can be critical to accurately capturing the behavior of your tumor. 


The mouse models used for PDX testing are also immunodeficient and are missing many of the cells present in our immune system. This too can affect the area around a tumor and change how it would behave in your body. It also limits the types of therapies that can be tested in PDX models. For instance, immunotherapies, which help to activate immune cells to target and destroy cancer cells, cannot be tested in PDX models.


O-PDX: A Pathway Forward for personalized medicine

PDX, and O-PDX specifically, offers a clearer path forward for personalized medicine. At Certis, we collect clinically relevant data on your tumor, using O-PDX as our chosen methodology. “This technology has been around for decades,” explains Certis Oncology’s CEO, Peter Ellman, “and it has evolved from rudimentary research to an advanced technique with clinical relevance in individualized therapy. Yes, there are still challenges, but the bottom line is that O-PDX histology, pathology, and drug sensitivity or resistance have extremely high concordance with what is seen in the clinic.”


Our O-PDX studies typically take 4 to 5 months, enabling you to begin or continue your current treatment, while we do our study initiation and start our pharmacology studies. At study completion, we provide your treating oncologist with a final report, empowering your doctor to make an informed and accurate decision about the next step in treatment. “If your oncologist is weighing 6 different potential treatment regimens, we can develop O-PDX models, test each treatment, and be precise about choosing the best therapy available,” explains Ellman. 


To ensure we collect the most meaningful data and reduce the changes in the tumor microenvironment, we only passage your tumor once before starting our O-PDX studies. This minimizes tumor infiltration by mouse cells and other structural factors that can change how your tumor behaves.


In our next blog post, we’ll take a deep dive into the scientific studies that have compared subcutaneous PDX to O-PDX in a variety of cancer contexts. 


Until then, you can learn more about our approach to O-PDX and the successes we’ve had with patients: 


  1. Tumor Markers: National Cancer Institute website: Published May 6, 2019. Accessed March 7th, 2021. 
  2. Wise HC, Solit DB. Precision Oncology: Three Small Steps Forward. Cancer Cell. 2019;35(6):825-826.
  3. Yoshida GJ. Applications of patient-derived tumor xenograft models and tumor organoids. J Hematol Oncol. 2020;13(1):4. 
  4. Evans KW, Yuca E, Akcakanat A, et al. A Population of Heterogeneous Breast Cancer Patient-Derived Xenografts Demonstrate Broad Activity of PARP Inhibitor in BRCA1/2 Wild-Type Tumors. Clin Cancer Res. 2017;23(21):6468-6477. 
  5. Bhimani J, Ball K, Stebbing J. Patient-derived xenograft models—the future of personalised cancer treatment. Br J Cancer. 2020;122:601–602.
  6. Kamili A, Gifford AJ, Li N, et al. Accelerating development of high-risk neuroblastoma patient-derived xenograft models for preclinical testing and personalised therapy. Br J Cancer. 2020;122:680–691.
  7. Extensive, 10-Arm O-PDX Study Predicts Clear “Therapeutic Winners” for Pediatric Patient. Certis Oncology website: Published January 8, 2021. Accessed March 7th, 2021.

About the Author:

Elie Diner has a Ph.D. in bioengineering and 12 years of research experience in microbiology, synthetic biology, and immunology. During his time at the bench, he developed a passion for effective science communication and eventually transitioned into a career as a professional science and content writer. He's authored 12 peer-reviewed scientific publications and numerous blogs, whitepapers, and e-books for life science companies. 

Back to Feed

Share This