Physics-informed information field theory for predictive medicine

Interdisciplinary Areas: Engineering-Medicine

Project Description

In the predictive medicine of the future, patients will be monitored by non-invasive medical imaging, the imaging data will be sent to a predictive computational model, and the model will forecast the evolution of the disease and recommend a personalized treatment, providing also the uncertainty in the predictions and treatment recommendations. This project aims at exercising this paradigm in the context of localized prostate cancer. This is enabled by a unique combination of prostate cancer modeling efforts [1,2], state-of-the-art prostate cancer imaging data [3], and the use of the newly developed Physics Informed Information Field Theory (PIFT) [4].

The uniqueness and potential of the project are justified as follows: Tumor growth models involve unknown parameters that need to be estimated from data at multiple time points, but such data rarely exists because most solid tumors are surgically removed right after their detection. However, many localized prostate tumors are not removed, providing a unique opportunity for serial imaging, model calibration, and validation. Finally, PIFT is the ideal mathematical framework to quantify the uncertainty in the model predictions because it is robust even when the model is inaccurate.

The paradigm proposed here may be extended to other diseases.

Start Date

Spring 2025

Postdoc Qualifications

• The candidate is expected to have earned a Ph.D. in Mechanical Engineering, Physics, Applied Mathematics, Computer Science, or other related fields.
• Strong programming skills.
• Basic knowledge of finite elements and uncertainty quantification.
• Knowledge of advanced Bayesian methods (Markov chain Monte Carlo, variational inference) and probabilistic programming frameworks (pyro, pymc) is preferred but not required.
• Excellent analytical and communication skills.

Co-advisors

Hector Gomez https://engineering.purdue.edu/gomez/
Ilias Bilionis https://predictivesciencelab.org/

Bibliography

[1] G. Lorenzo, M. Scott, K. Tew, T.J.R. Hughes, Y.J. Zhang, L. Liu, G. Vilanova, H. Gomez, Tissue-scale, personalized modeling and simulation of prostate cancer growth, Proceedings of the National Academy of Sciences, 113(48), E7663-E7671, 2016.
[2] G. Lorenzo, T.J.R. Hughes, P. Dominguez-Frojan, A. Reali, H. Gomez, Computer simulations suggest that prostate enlargement due to benign prostatic hyperplasia mechanically impedes prostate cancer growth, Proceedings of the National Academy of Sciences, 116(4), 1152-1161, 2019.
[3] C.H Feng, C.C Conlin, K.B. Batra, A.E. Rodríguez-Soto, R. Karunamuni, A. Simon, J. Kuperman, R. Rakow-Penner, M.E Hahn, A.M Dale, T.M Seibert, Voxel-level classification of prostate cancer using a four-compartment restriction spectrum imaging model, Journal of Magnetic Resonance Imaging, 54(3), 2021.
[4] A. Alberts, I. Bilionis, Physics-informed information field theory for modeling physical systems with uncertainty quantification, Journal of Computational Physics, 486, 112100, 2023.