CAR T Cell Therapy Response Prediction Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CAR T cell therapies for Large B cell Lymphoma (LBCL) face challenges in achieving long-term durable remission, particularly for patients refractory to first or second-line therapies, with limited understanding of the mechanistic relationships between tumor size, T cell expansion, and lymphodepletion in CAR T cell therapy outcomes.
Innovation Solution
A computer-implemented method and system that generates a model to simulate dynamics and interactions among normal T cells, engineered CAR T cells, and tumor cells, using cell population compartments modeled with continuous-time birth and death stochastic processes and deterministic mean-field equations, to predict patient responses to immune-based or targeted therapies, such as CAR T cell infusion, by analyzing pre- and post-treatment data including tumor volume, lymphocyte counts, and CAR T cell populations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If CAR T cell therapy is administered to patients with large tumor mass, then the therapy can target more tumor cells, but the efficacy is reduced due to insufficient T cell expansion and tumor escape
Solution Approach 1:
The patent applies preliminary action by performing lymphodepletion chemotherapy before CAR T cell infusion. This pre-treatment reduces the tumor burden and creates a more favorable immunological environment, allowing the subsequently administered CAR T cells to expand more effectively and achieve better anti-tumor activity. The conditional step of reducing tumor mass prior to CAR T administration directly addresses the contradiction by preparing the system in advance to handle larger initial tumor loads.
2Quantity of substance
If intensive lymphodepletion chemotherapy is administered to enhance CAR T cell expansion, then T cell proliferation is improved, but patient toxicity increases
Solution Approach 1:
The patent applies partial action by using moderate lymphodepletion chemotherapy rather than intensive regimens. The conditional steps specify administering lymphodepletion at controlled doses that are sufficient to reduce immunosuppression and enhance CAR T cell expansion, but not so intense as to cause severe toxicity. This balanced approach achieves the necessary T cell proliferation while limiting harmful effects on the patient.
3Measurement precision
If multiple cell population parameters are monitored to improve prediction accuracy, then the model precision increases, but the complexity of data collection and analysis increases
Solution Approach 1:
The patent applies segmentation by dividing the complex immune system into distinct cell population compartments (naive T cells, memory T cells, effector T cells, tumor cells). Each compartment is modeled separately with its own differential equations, allowing the system to track multiple parameters simultaneously while maintaining mathematical tractability. This compartmentalization enables accurate prediction of therapy response by capturing the dynamics of each cell type without creating an intractably complex model.
Data Source
AI summary
Methods, systems, and computer-readable media for predicting a patient's response to immune based or target therapy are described herein. An example computer-implemented method includes generating a model configured to represent dynamics and interactions among normal T cells, engineered cells, and tumor cells, where the model includes a plurality of cell population compartments. The computer-implemented method also includes receiving pre-treatment patient data for a cancer patient, and receiving post-treatment patient data for the cancer patient. Each of the pre-treatment patient data and the post-treatment patient data includes a measure of at least one of tumor volume, total lymphocytes, memory T cells, memory engineered cells, tumor killing cells, or antigen-presenting tumor cells. The computer-implemented method further includes quantitatively predicting the cancer patients response to the immune-based or targeted therapy using the model, the pre-treatment patient data, and the post-treatment patient data.