From PD-L1 to PSA: Virtual Clones and Cohorts in Action Across Cancer Types and Therapeutic Modalities

Add to your list(s) Download to your calendar using vCal

• Trachette Jackson (University of Michigan)
• Friday 19 September 2025, 09:30-10:30
• Seminar Room 1, Newton Institute.

If you have a question about this talk, please contact nobody.

OOEW07 - Mathematical Foundations of Oncological Digital Twins

The promise of oncological digital twins hinges on our ability to build virtual replicas of tumors and their microenvironments that faithfully forecast the diversity of patient-specific responses to various therapeutic options and inform clinical decision-making. In this talk, I present an integrative modeling framework that uses parameter identifiability and data-driven virtual tumor clones and virtual patient cohorts to enhance the predictive reliability of mathematical models in cancer treatment. Drawing on two recent studies—one focused on immune checkpoint inhibitor (ICI) therapy in bladder cancer, the other on androgen deprivation therapy (ADT) in prostate cancer—we demonstrate how virtual clones and cohorts can be used to test hypotheses, optimize therapies, and reveal information insights about the reliability of clinical biomarkers like immune cell infiltration and PSA levels.   Our bladder cancer study uses multiple in vivo datasets to construct virtual tumors under different immune conditions and treatment regimens. We show that models calibrated with insufficient data or based on unidentifiable parameters systematically overestimate tumor reduction and survival. Similarly, in prostate cancer, we construct virtual cohorts parameterized by clinical PSA trajectories and experimentally measured PSA expression rates to enhance the prediction of ADT failure. These virtual clones reveal that post-nadir PSA dynamics, rather than the initial decline, hold significant prognostic value. Together, these studies offer practical guidelines for developing data-driven digital twins. We highlight the importance of identifying minimal but informative datasets for calibration, assessing parameter identifiability, and tailoring virtual cohorts to reflect both parameter uncertainty and patient-specific biology. These insights chart a path toward translationally robust digital twins capable of forecasting individualized treatment outcomes, providing critical insights for optimizing treatment and monitoring schedules and adapting to emerging clinical data.  

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

• All CMS events
• Featured lists
• INI info aggregator
• Isaac Newton Institute Seminar Series
• School of Physical Sciences
• Seminar Room 1, Newton Institute
• bld31

Note that ex-directory lists are not shown.