Genialis and Inventia Life Science Collaborate to Advance Patient-Relevant Pancreatic Cancer Models
Genialis, a therapeutic intelligence company, and Inventia Life Science, a developer of advanced 3D cell model technologies, have announced a collaboration aimed at improving how new pancreatic cancer treatments are evaluated and matched to patients who may be most likely to benefit.
The companies are combining Genialis’ artificial intelligence-powered patient insights with Inventia Life Science’s RASTRUM 3D cell model platform to develop human-relevant models of pancreatic ductal adenocarcinoma (PDAC). The objective is to create experimental models that more closely correspond to the biological characteristics of identifiable patient populations and can be used to investigate potential therapeutic strategies.
Initial findings from the collaboration will be presented at the Pancreatic Cancer Action Network (PanCAN) 2026 Scientific Summit in San Diego. The work illustrates an approach in which patient-derived molecular information is used to identify clinically meaningful disease states, which are then recreated in laboratory models for experimental testing.
The collaboration is focused on addressing one of the longstanding challenges in oncology drug development: determining whether findings generated in preclinical models are likely to translate to specific groups of patients.
Connecting Patient Biology With Preclinical Research
Pancreatic cancer remains a complex area of therapeutic development, and advances in targeted treatment strategies have increased interest in identifying the biological characteristics that distinguish patients who may respond to particular interventions.
Within pancreatic cancer, PDAC is the most common form and is characterized by a complex tumor microenvironment. Cancer cells interact with surrounding stromal components, including fibroblasts and other cell types, which can influence tumor behavior, treatment response and disease progression.
Traditional laboratory models may not always reproduce the complexity or biological diversity observed in human tumors. Genialis and Inventia are therefore seeking to create models that reflect clinically observed tumor biology more closely.
Inventia’s RASTRUM platform uses 3D bioprinting technology to produce reproducible pancreatic cancer models in which tumor and stromal components can be controlled and perturbed independently. This provides researchers with an experimental system for studying interactions between cancer cells and components of the surrounding tumor environment.
Genialis adds an analytical and computational layer to this process through its Genialis Supermodel, a large molecular model trained using a proprietary harmonized and curated transcriptomics dataset.
The Supermodel is used to characterize biological states and connect experimental model profiles with molecular states observed in patients. Through this combination, researchers can move between clinical observations and laboratory experiments in a more systematic manner.
Identifying Patients With Poor-Prognosis PDAC
As part of the study being presented at the PanCAN Scientific Summit, Genialis analyzed molecular information from 644 pancreatic cancer patient tumors included in the PanCAN SPARK “Know Your Tumor” cohort.
The analysis focused on the composition and activity of the tumor microenvironment and was used to identify PDAC patient populations associated with poor prognosis.
Rather than treating pancreatic cancer as a biologically uniform disease, the approach seeks to identify distinct molecular and cellular states within the patient population.
Understanding these differences can be important for drug development because a treatment that affects a particular biological pathway may not have the same relevance across all patients. Identifying the molecular characteristics associated with specific clinical outcomes can therefore help researchers develop hypotheses about which patients might benefit from particular therapeutic approaches.
Following identification of the patient profiles, Genialis mapped those biological states to tumor-fibroblast co-culture models generated using the RASTRUM platform.
This creates an experimental bridge between patient data and laboratory testing. Researchers can use the resulting models to investigate biological mechanisms associated with poor outcomes and explore potential therapeutic strategies in a controlled setting.
Modeling the Tumor Microenvironment
The tumor microenvironment is an important consideration in pancreatic cancer research because malignant cells exist within a complex network of surrounding cells and extracellular components.
Fibroblasts, for example, can interact with cancer cells and contribute to the biological characteristics of the tumor environment. Reproducing these interactions in laboratory systems may provide additional information compared with models that focus solely on isolated cancer cells.
RASTRUM is designed to generate 3D cell models in which tumor and stromal components can be independently controlled. This allows researchers to investigate how different components of the tumor environment contribute to disease biology and how those components respond to therapeutic interventions.
By mapping RASTRUM models to molecular profiles identified from patient tumors, the collaboration seeks to make these experimental systems more directly relevant to clinically observed disease states.
The objective is not simply to build a more complex model, but to build a model that corresponds to a specific biological question arising from patient data.
Using AI to Connect Experimental and Clinical Data
Genialis’ contribution centers on the use of computational biology and artificial intelligence to interpret molecular information and identify clinically meaningful biological states.
The company’s Supermodel is based on a proprietary harmonized and curated transcriptomics dataset. It is used to characterize molecular profiles and help connect experimental observations with patient biology.
In the current collaboration, the platform was applied to pancreatic cancer data from the PanCAN SPARK “Know Your Tumor” cohort. The analysis helped identify patient profiles associated with poor prognosis based on tumor microenvironment composition and activity.
Those profiles could then be compared with the molecular characteristics of the RASTRUM-generated models.
This creates a framework in which researchers can begin with observations from real patients, identify relevant biological characteristics, recreate those characteristics in an experimental system and then test therapeutic hypotheses.
“This collaboration closes an important gap in drug discovery by connecting what we learn from patients with what we can test in the lab,” said Cameron Ferris, PhD, Chief Executive Officer and Co-Founder of Inventia Life Science.
Ferris said Genialis can identify biologically meaningful states in real tumors, while RASTRUM provides a system for modeling and perturbing those states at scale.
The combination, he said, can provide a way to generate and test therapeutic hypotheses that are grounded in the biology of the patients who may ultimately receive those treatments.
Creating a Translational Feedback Loop
A key element of the collaboration is the creation of what the companies describe as a translational loop.
In conventional drug development, discoveries made using laboratory models are eventually tested in clinical trials. However, there can be uncertainty about how closely the biology of a preclinical model corresponds to the disease biology of the patients enrolled in a clinical study.
The Genialis and Inventia approach attempts to address this disconnect by beginning with patient data.
Clinical molecular information is used to identify relevant biological states and patient populations. These findings are then used to inform the development or selection of laboratory models. Researchers can subsequently perturb those models to investigate potential therapeutic approaches.
The resulting experimental findings can then be considered in the context of the patient populations from which the original biological hypotheses were derived.
This type of iterative process could potentially provide drug developers with additional information when making decisions about therapeutic mechanisms, combinations and patient selection.
The approach is particularly relevant for pharmaceutical companies developing targeted treatments, where identifying the appropriate patient population can be critical to clinical development.
Supporting Drug Development Decisions
Rafael Rosengarten, PhD, Chief Executive Officer and Co-Founder of Genialis, highlighted the challenge pharmaceutical companies face when attempting to determine whether preclinical findings will translate into patient outcomes.
“Our pharma partners need to make critical decisions about which patients to enroll and which combinations to pursue, yet historically, whether a preclinical result would translate to patients often wasn’t clear until Phase 2,” Rosengarten said.
According to Rosengarten, Inventia’s RASTRUM platform provides a model in which therapeutic hypotheses can be experimentally tested, while the Genialis Supermodel can help identify the patients most likely to benefit.
Patient selection is increasingly important in precision oncology because cancers that appear similar based on anatomical location can have substantially different molecular characteristics. Identifying those differences before clinical testing may help researchers develop more targeted hypotheses regarding treatment response.
The collaboration therefore combines computational patient stratification with 3D experimental modeling to investigate whether specific biological states can be reproduced and therapeutically manipulated.
Building More Human-Relevant Models
The companies’ work reflects a broader movement within drug development toward more human-relevant preclinical systems.
Traditional two-dimensional cell cultures can provide valuable information about cancer biology, but they may not fully capture the spatial relationships and cellular interactions present in tumors. Three-dimensional models can provide researchers with additional opportunities to investigate those interactions.
At the same time, computational analysis of patient-derived molecular data can help researchers understand which biological characteristics are most relevant to actual disease outcomes.
Combining these approaches could help make experimental models more closely aligned with clinical biology.
For pancreatic cancer research, this may be particularly valuable given the complex interactions between tumor cells and the surrounding microenvironment. By controlling tumor and stromal components within RASTRUM models, researchers can systematically investigate these relationships.
The collaboration with Genialis adds another dimension by helping determine which molecular states should be modeled based on patient data.
Initial Results to Be Presented at PanCAN Summit
The first results from the collaboration are being presented at the PanCAN 2026 Scientific Summit in San Diego, providing researchers and the pancreatic cancer community with an opportunity to examine the approach.
The study uses patient data from 644 pancreatic cancer tumors in the PanCAN SPARK “Know Your Tumor” cohort to identify PDAC populations associated with poor prognosis. Genialis then mapped these profiles to tumor-fibroblast co-culture models produced with the RASTRUM platform.
The resulting models are intended to provide a basis for investigating the biological characteristics associated with poor outcomes and evaluating potential therapeutic strategies.
The findings are an early demonstration of the companies’ strategy, and additional research will be required to establish how effectively the models predict clinical responses or support therapeutic development.
Nevertheless, the collaboration provides a framework for integrating patient-derived molecular information, AI-driven biological analysis and advanced 3D experimental systems.
Potential Applications Beyond a Single Model
While the initial work focuses on PDAC and tumor-fibroblast interactions, the underlying strategy could potentially support broader translational research programs.
The ability to identify patient populations computationally and then reproduce relevant biological states experimentally may be useful for pharmaceutical developers investigating targeted therapies, combination treatments and biomarker-defined populations.
For drug developers, the ultimate objective is to generate evidence that can inform decisions before therapies reach later-stage clinical development. By linking experimental systems to identifiable patient biology, the collaboration seeks to reduce uncertainty around whether a particular preclinical observation is relevant to the patients being targeted.
The Genialis and Inventia partnership will continue to explore this connection between clinical data and experimental models, with pancreatic cancer serving as an initial application.
Through the combination of Genialis’ Supermodel and AI-enabled patient insights with Inventia’s RASTRUM 3D modeling platform, the companies are developing an approach intended to bring patient biology closer to the laboratory.
The collaboration illustrates an emerging model for translational research in which patient-derived molecular data can guide the construction of human-relevant experimental systems, allowing researchers to test therapeutic hypotheses against biological states that have already been observed in patients.
As the work progresses, further validation will be needed to determine how these models perform in therapeutic research and whether they can help improve patient selection, treatment development and clinical translation in pancreatic cancer.
About Genialis
Genialis provides therapeutic intelligence with its Supermodel of cancer biology. We develop and validate clinically actionable AI models informed by the world’s most ethnographically diverse cancer data sets to predict patient responses and guide treatment decisions for targeted inhibitors, immunotherapies, and other emerging therapeutic classes. Genialis is trusted by pharma and diagnostics partners, and together, we are transforming medicine through data. The company is headquartered in Houston, Boston and Ljubljana.
About Inventia Life Science
Inventia Life Science develops advanced 3D cell culture solutions that help researchers generate reproducible, biologically relevant in vitro models at scale. Built around the RASTRUM platform, Inventia combines precision drop-on-demand bioprinting, tunable synthetic matrices, workflow-driven software, and scientific expertise to support disease modeling, mechanistic biology, and drug discovery. To learn more, visit inventialifescience.com.