Introduction to the PDX Models in Gynecologic Cancers
Gynecologic cancers including ovarian and cervical cancers remain challenging, with high relapse and chemoresistance rates (five year survival: ovarian ~48%, cervical ~66%). Tumor heterogeneity limits the effectiveness of conventional therapies.
Patient-Derived Xenograft (PDX) models overcome these limitations by transplanting human tumor tissue into immunodeficient mice, preserving molecular, genetic, histological, and drug-response characteristics. PDX models are widely used to study drug resistance, tumor relapse, metastasis, and therapeutic efficacy, particularly in ovarian and cervical cancers, supporting personalized treatment strategies and improving preclinical research translation to the clinic.
Advantages of PDX Models
While cancer cell lines (CCLs) are derived from patient tumors, extensive in vitro culture leads to genetic and phenotypic homogeneity, causing a loss of the original tumor’s heterogeneity and microenvironment. Consequently, CCL-derived xenografts often fail to predict drug responses accurately, contributing to the low clinical approval rate of anti-cancer agents (~5%). Drugs targeting specific molecules in homogeneous CCLs may not replicate the complex responses seen in heterogeneous patient tumors.
Patient-Derived Xenograft (PDX) models overcome these limitations by implanting patient tumor tissue directly into immunodeficient mice without prior in vitro culture. PDXs retain the genetic, epigenetic, and histological characteristics of the primary tumor, including the surrounding human stroma. Tumors grow in a physiologically relevant microenvironment, preserving oxygen, nutrient, and hormone levels similar to the patient’s tumor site.
As a result, PDX models accurately recapitulate patient-specific drug responses, enabling preclinical testing of multiple therapies on expanded xenografts from a single biopsy. They are widely applied in breast, prostate, pancreatic, colorectal, lung, and gynecologic cancers to study drug safety, efficacy, and personalized therapeutic strategies.
Drug Screening and Biomarker Development with PDX Models
Patient-Derived Xenograft (PDX) models have been extensively used in both retrospective studies and preclinical trials across multiple cancers, including breast, renal, non-small cell lung (NSCLC), head and neck squamous cell carcinoma, colorectal, and pancreatic cancers. Drug response rates in PDX models closely mirror clinical outcomes for both targeted therapies and conventional cytotoxins. For example, renal cell cancer PDXs respond to sirolimus, sunitinib, and dovitinib but not to erlotinib, consistent with patient data. Similarly, responses to cetuximab, MEK inhibitors, and PI3K/mTOR inhibitors in PDX models reflect clinical trial results, and standard chemotherapies such as paclitaxel, carboplatin, gemcitabine, and 5-fluorouracil show comparable efficacy in PDXs and patients.
PDX models also serve as predictive platforms for prospective clinical trials, saving time and resources. For instance, pancreatic cancer PDXs showed no response to metformin, aligning with negative clinical trial outcomes.
Beyond drug testing, PDX models are instrumental in biomarker discovery and resistance studies. The correlation between PDX and patient responses allows identification of predictive markers of drug sensitivity or resistance. Heterogeneous PDXs enable linking molecular characteristics to therapeutic outcomes, such as gemcitabine resistance in pancreatic cancer or vemurafenib resistance in melanoma, where mutant BRAF elevation was identified as a key resistance mechanism. Comparative proteomic and genomic analyses between sensitive and resistant PDXs further facilitate the discovery of prognostic and predictive biomarkers.
Co-clinical trials involve conducting preclinical studies in parallel with clinical trials. Patient-specific clinical, biological, and pharmacologic data including somatic mutations, germline variants, drug responsiveness, imaging, transcriptomics, and proteomicare integrated to identify biomarkers predicting treatment response. In this strategy, a PDX model is generated from a patient enrolled in a clinical trial and treated with the same therapeutic regimen to monitor concordance with clinical outcomes.
This approach allows screening for prognostic biomarkers, investigation of underlying drug response mechanisms, and the proposal of novel combination therapies. For instance, Heid et al. performed co-clinical assessment of tumor cellularity in pancreatic cancer, while Owonikoko et al. showed that PDX models accurately replicated clinical outcomes in a phase II trial of arsenic trioxide for relapsed small cell lung cancer. Although currently limited in use, co-clinical trials are expected to expand, offering a faster, personalized, and preclinically validated pathway for drug development.
Advances in cancer genomics and molecular profiling have enabled precision medicine, aiming to deliver the right drug, to the right patient, at the right time.Unlike conventional chemotherapy, precision oncology integrates the tumor’s genomic and molecular characteristics with targeted therapies or immunotherapeutics to maximize efficacy while minimizing adverse effects. Patients are stratified into subpopulations based on genomic profiling, allowing treatments to be tailored to specific genetic subgroups.
PDX models are ideally suited for precision medicine as they preserve the genetic and phenotypic features of the original tumor, including intratumoral heterogeneity, and accurately represent patient-specific molecular profiles. For example, integrated PDX platforms have identified individualized therapeutic vulnerabilities in pancreatic cancer, testing over 500 single and combination regimens based on exome sequencing of patient tumors. Similarly, PDX models have guided genomic-informed therapy in advanced breast cancer and evaluated anti-cancer agents in ovarian cancer, while adaptations in leukemia studies have considered cross species cytokine interactions. Overall, PDX models serve as a robust preclinical platform for implementing precision medicine, with the growing power of next-generation sequencing (NGS) and bioinformatics further enhancing their predictive potential.
PDX models offer a powerful preclinical platform by preserving tumor heterogeneity and patient-specific molecular features, yet several challenges remain. Efficient engraftment requires fresh, representative tumor tissue, sometimes limited to small specimens, and collaboration with pathologists and surgeons. The choice of implantation strategy is critical: orthotopic models accurately mimic the tumor microenvironment but are technically demanding, whereas subcutaneous models are simpler and faster but less physiologically relevant. Establishment typically takes 2–8 months, which may limit real-time personalized therapy applications, and engraftment failure remains significant for some tumor types. Conventional PDX models rely on immunodeficient mice, restricting immune-oncology studies, though humanized PDX models offer a potential solution. Additionally, repeated passages can replace human stroma with murine elements, and development costs are substantial. Despite these limitations, PDX models remain invaluable for studying cancer biology, drug response, and patient-specific therapies.