Predicting tumour response to anti-PD-1 immunotherapy with computational modelling

Area of Science:

• Immunology
• Computational Biology
• Oncology

Background:

• Cancer immunotherapy, particularly with anti-programmed death-1 (anti-PD-1) antibodies, is advancing rapidly.
• Traditional trial-and-error methods for developing new treatments are time-consuming and expensive, creating a bottleneck in research.
• Computational modeling offers a complementary approach, but validation challenges limit its clinical application.

Purpose of the Study:

• To develop and validate a bottom-up deterministic computational model for simulating tumor response to anti-PD-1 antibody therapy.
• To identify potential biomarkers for predicting response to anti-PD-1 immunotherapy.
• To assess the influence of model parameters on treatment outcomes and explore biomarker interactions.

Main Methods:

• A bottom-up deterministic computational model was created with minimal, experimentally measurable parameters.
• The model was fitted to experimental data from B16-F10 melanoma in mice treated with anti-PD-1 antibodies.
• Model predictive accuracy was validated using two independent literature datasets; sensitivity analyses were performed.

Main Results:

• The model accurately simulated tumor growth curves, achieving mean relative deviations of 13%-20% compared to experimental data.
• Sensitivity studies revealed that Major Histocompatibility Complex (MHC) class I expression is a critical differentiator between responders and non-responders.
• MHC class I expression may influence the predictive power of current biomarkers like PD-1 ligand (PD-L1), with optimal response predicted at moderate PD-L1 levels.

Conclusions:

• Validated computational models can accelerate and guide cancer immunotherapy research.
• MHC class I expression is a promising biomarker for predicting response to anti-PD-1 therapy.
• Understanding biomarker interactions, such as MHC class I and PD-L1, is crucial for optimizing immunotherapy strategies.