Managing Cell and Gene Therapy Risk in Oncology Using Evidium's World Models for Healthcare

For a patient with relapsed or refractory cancer, a cell or gene therapy like CAR-T can mean the difference between running out of options and getting another shot at remission. That promise is real, and it's why cell and gene therapies (CGTs) have become one of the most closely watched developments in oncology.

It's also why getting the decision right matters so much, both for the patient in front of you and for the organization managing that risk. Oncology remains one of the most complex and costly areas of healthcare, and CGTs add a new layer of financial complexity on top of their clinical promise. The therapy alone can cost between $400,000 and $500,000, and total cost of care often exceeds $1 million once hospitalization and associated services are included. At the same time, the number of people living with and beyond cancer keeps growing, with projections exceeding 22 million survivors in the U.S. by 2035.

Behind every one of those numbers is a patient and a family making a high-stakes decision, often with less clarity than they deserve about how that decision is likely to play out. That's the gap Evidium is built to close.

What “Computational” Actually Means

Right now, most of what we know about cancer care, treatment guidelines, clinical trial results, published research, lives as language: text written for a person to read, not for software to reason over. A clinician can read an NCCN guideline and know what it implies for the patient sitting across from them. A computer cannot, at least not on its own.

Evidium changes that. We take medical knowledge, guidelines, literature, real-world evidence, and turn it into a structured, computable map: the clinical states a patient can be in (post-CAR-T remission, biomarker-positive relapse, and so on), the transitions between those states, and the causal relationships that drive them, why one patient responds and another doesn't. Because that map is computable, it can be searched, simulated, and applied directly to an individual patient's data.

This is a different kind of model than the ones people are used to hearing about. LLMs compute language: given some text, they predict the next word. Evidium computes what causes what. Feed the model a specific patient's clinical and claims history, and it projects that patient onto the map to forecast their likely path forward, which treatments they're likely to need, how they're likely to respond, and what it's likely to cost. The output is a concrete, evidence-grounded forecast for that one patient, not a generic statement about “patients like them.”

Why Claims Data Alone Isn't Enough

Most current approaches to understanding healthcare risk lean heavily on claims data, a structured record of services delivered and payments made. That's useful for understanding utilization and cost, but it's retrospective: it shows what already happened without explaining why. It doesn't capture the laboratory results, biomarker data, or clinical reasoning that actually explain how a patient's disease is progressing. And it's often delayed and fragmented across systems, which matters a great deal when a treatment decision can't wait.

Understanding what's really happening with a patient means engaging with the clinical detail behind the care, not just the bill for it. Where clinical data is available, Evidium incorporates it directly. Where it isn't, which is most of the time, Evidium infers the clinical picture from claims, mapping care patterns against established guidelines such as NCCN for cancer to reconstruct what likely happened to the patient and project what's likely to come next, including future treatment states and associated cost.

Applying This to Cell and Gene Therapy Decisions

CGTs call for a different kind of decision-making, one built around the patient in front of you rather than the average patient in a clinical trial.

First, organizations need to model how a therapy willactually perform once it leaves the controlled conditions of a clinical trial. By combining clinical evidence with real-world data, Evidium can simulate how anew treatment is likely to affect a specific population: which patients are likely to qualify, when the therapy is likely to be introduced, and how total cost of care may change under different scenarios. Even small shifts ineligibility, timing, or cost assumptions can meaningfully change the financial picture, and scenario modeling lets stakeholders see that before the costs materialize.

Second, precision has to happen at the level of the individual patient. Two patients with the same diagnosis can respond very differently depending on their biomarkers, genetics, and treatment history. By analyzing integrated clinical and claims data alongside real-world evidence, Evidium can identify which specific patients are most likely to benefit from a given therapy, supporting more individualized treatment decisions, improving outcomes, and reducing unnecessary use of high-cost interventions.

Third, these decisions have to be viewed over a longer time horizon than the initial invoice. CGTs carry significant upfront cost, but they may reduce future treatment needs or improve long-term outcomes. Understanding that trade-off means modeling total cost of care over multiple years, not just immediate spend, so clinical value and financial sustainability can be weighed side by side.

The Patient Behind the Number

Zoom out far enough, and a portfolio of oncology risk is really just thousands of individual patient stories, seen all at once. Evidium's world models are built so that zooming out never means losing sight of the person behind each number. The same map that forecasts one patient's path, whether they respond to CAR-T, what comes next if they don't, what it's likely to cost, also aggregates into a clearer, evidence-grounded picture of risk across an entire population.

As CGTs continue to advance, the organizations that manage this risk well will be the ones that can connect clinical innovation to real-world impact for the patient it's meant to help, and to the financial outcome that follows from it. That's what Evidium's world models are built todo.

Contact us today to see how Evidium's World Models for Healthcare can help you understand and manage oncology risk, turning oncology complexity into actionable, evidence-grounded insight for every patient.