Quantitative cancer-immunity cycle modeling for predicting disease progression in advanced metastatic colorectal cancer

Abstract

AbstractPatients diagnosed with advanced metastatic colorectal cancer (mCRC) often exhibit heterogeneous disease progression and face poor survival prospects. In order to comprehensively analyze the varied treatment responses among individuals and the challenge of tumor recurrence resistant to drugs in advanced mRCR, we developed a novel quantitative cancer-immunity cycle (QCIC) model. The proposed model was meticulously crafted utilizing a blend of differential equations and randomized modeling techniques to quantitatively elucidate the intricate mechanisms governing the cancer-immunity cycle and forecast tumor dynamics under different treatment modalities. Furthermore, by integrating diverse clinical datasets and rigorous model analyses, we introduced two pivotal concepts: the treatment response index (TRI) and the death probability function (DPF). These concepts are crucial tools for translating model predictions into clinically relevant evaluation indexes. Using virtual patient technology, we extrapolated tumor predictive biomarkers from the model to predict survival outcomes for mCRC patients. Our findings underscore the significance of tumor-infiltrating CD8+CTL cell density as a key predictive biomarker for short-term treatment responses in advanced mCRC while emphasizing the potential predict value of the tumor-infiltrating CD4+Th1/Treg ratio in determining patient survival. This study presents a pioneering methodology bridging the divide between diverse clinical data sources and the generation of virtual patients, offering invaluable insights into understanding inter-individual treatment variations and forecasting survival outcomes in mCRC patients.Author summaryThis study introduces a sophisticated modeling approach to delineate the intricate dynamics of tumor-immune interactions within the cancer-immunity cycle. Utilizing a multi-compartmental, multi-scale, multi-dimensional differential equation model, we quantified the complex interplay between tumor cells, immune cells, cytokines, and chemokines. By integrating virtual patient technology, we have conductedin silicoclinical trials that accurately predict disease progression across multiple treatment modalities for cancer patients, particularly in advanced mCRC. Through the combination of differential equations and randomized modeling, we effectively captured the diverse treatment responses and clinical manifestations of drug-resistant tumor recurrence. Furthermore, we explored the pivotal role of predictive biomarkers in determining patients’ survival prognosis, offering insights for tailored therapeutic strategies. Importantly, our computational framework holds promise for the extension to the investigation of patients with other tumor types, thus contributing to the personalized investigation of patients with other tumor types and cancer care.