[Paper Review] Quantitative cancer-immunity cycle modeling to optimize bevacizumab and atezolizumab combination therapy for advanced renal cell carcinoma

Lei Du, Chenghang Li+1 more|arXiv (Cornell University)|Jan 25, 2026

Cancer Immunotherapy and Biomarkers0 citations

TL;DR

The authors develop a Quantitative Cancer-Immunity Cycle (QCIC) model combining ODEs and stochastic methods to simulate RCC tumor-immune dynamics, generate a virtual patient cohort, and optimize bevacizumab–atezolizumab dosing with PK integration to predict treatment efficacy.

ABSTRACT

The incidence of advanced renal cell carcinoma(RCC) has been rising, presenting significant challenges due to the limited efficacy and severe side effects of traditional radiotherapy and chemotherapy. While combination immunotherapies show promise, optimizing treatment strategies remains difficult due to individual heterogeneity. To address this, we developed a Quantitative Cancer-Immunity Cycle (QCIC) model that integrates ordinary differential equations with stochastic modelling to quantitatively characterize and predict tumor evolution in patients with advanced RCC. By systematically integrating quantitative systems pharmacology principles with biological mechanistic knowledge, we constructed a virtual patient cohort and calibrated the model parameters using clinical immunohistochemistry data to ensure biological validity. To enhance predictive performance, we coupled the model with pharmacokinetic equations and defined the Tumor Response Index (TRI) as a quantitative metric of efficacy. Systematic analysis of the QCIC model allowed us to determine an optimal treatment regimen for the combination of bevacizumab and atezolizumab and identify tumor biomarkers with clinical predictive value. This study provides a theoretical framework and methodological support for precision medicine in the treatment of advanced RCC.

Motivation & Objective

• Motivate precision immunotherapy for advanced RCC by capturing tumor-immune interactions across multiple compartments.
• Develop a five-compartment QCIC framework linking immune generation, circulation, activation, effector action, and antigen transport.
• Calibrate the model with clinical immunohistochemistry data to ensure biological validity.
• Integrate pharmacokinetics to simulate drug disposition and define a Tumor Response Index (TRI) as a efficacy metric.
• Create a virtual patient cohort to explore fixed-dose versus adaptive dosing and biomarker prediction.

Proposed method

• Construct a five-compartment QCIC model (bone marrow/thymus, peripheral blood, tumor-draining lymph nodes, tumor microenvironment, lymphatic vessels).
• Describe eight cellular processes per cell type (source, differentiation, proliferation, transformation, migration, chemotaxis, killing, apoptosis) with detailed equations for each.
• Incorporate cytokine regulation via Michaelis–Menten and Hill kinetics and model tumor subpopulations with competitive dynamics under bevacizumab sensitivity.
• Calibrate 164 parameters through literature guidance, plausibility ranges, and clinical data fitting; perform sensitivity analysis to build a virtual patient cohort.
• Integrate a two-compartment pharmacokinetic model for atezolizumab and bevacizumab with impulse dosing at 3-week intervals and random parameter sampling to reflect heterogeneity.
• Define the Tumor Response Index (TRI) to quantify tumor burden changes and classify responses using RECIST-like thresholds.

Experimental results

Research questions

• RQ1 How can QCIC be calibrated to reflect RCC immune heterogeneity and predict treatment outcomes?
• RQ2 What dosing strategies (fixed vs adaptive) optimize combination therapy efficacy in virtual RCC patients?
• RQ3 Which tumor biomarkers emerge as predictive within the QCIC framework?
• RQ4 How well do QCIC-based virtual patients recapitulate clinical immunogenomic data and short-term outcomes?
• RQ5 Can PK variability be integrated to personalize bevacizumab–atezolizumab therapy?

Key findings

• The QCIC model reproduces clinical indicators of short-term outcomes for placebo, atezolizumab alone, and combination therapy within 95% confidence intervals.
• Virtual patients (10,000 per group; 100 sampled for analysis) yield TRI-based predictions aligned with RECIST-like classifications across treatments.
• Sensitivity analysis identifies key immune-heterogeneity parameters influencing late- and early-stage immune cell dynamics, guiding virtual patient generation.
• Pharmacokinetic modeling shows heterogeneity in central and peripheral drug concentrations consistent with population PK data, enabling personalized exposure assessments.
• The framework supports stratification into response subgroups and evaluation of fixed-dose versus adaptive dosing strategies using the virtual cohort.
• Simulation outputs (e.g., circular spider plots and waterfall plots) illustrate longitudinal treatment responses for virtual patients.
• JS and KL divergences indicate the virtual cohort closely matches distributions of CD8/CD4, CD4/Treg, and Treg/TAM from immunogenomic data.