A Pilot Study on Patient-specific Computational Forecasting of Prostate Cancer Growth during Active Surveillance Using an Imaging-informed Biomechanistic Model

Guillermo Lorenzo ^1,2, Jon S Heiselman ^3,4, Michael A Liss ⁵

¹Department of Civil Engineering and Architecture, University of Pavia, Pavia, Italy.
²Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, Texas.
³Department of Biomedical Engineering, Vanderbilt University, Nashville, Tennessee.
⁴Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York.
⁵Department of Urology, The University of Texas Health Science Center at San Antonio, San Antonio, Texas.
⁶Vanderbilt Institute for Surgery and Engineering, Vanderbilt University, Nashville, Tennessee.
⁷Department of Neurological Surgery, Radiology, and Otolaryngology-Head and Neck Surgery, Vanderbilt University Medical Center, Nashville, Tennessee.
⁸School of Mechanical Engineering, Weldon School of Biomedical Engineering, and Purdue Institute for Cancer Research, Purdue University, West Lafayette, Indiana.
⁹Livestrong Cancer Institutes and Departments of Biomedical Engineering, Diagnostic Medicine, and Oncology, The University of Texas at Austin, Austin, Texas.
¹⁰Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, Texas.

⁰Department of Civil Engineering and Architecture, University of Pavia, Pavia, Italy.

Cancer Research Communications +

|

March 1, 2024

View abstract on PubMed

Summary

Personalized computational models predict prostate cancer growth using MRI data. This approach identifies higher-risk disease earlier, enabling tailored active surveillance (AS) plans for better patient outcomes.

Area Of Science

Urology
Computational Biology
Medical Imaging

Background

Active surveillance (AS) is a standard management strategy for low-to-intermediate risk prostate cancer.
Current AS relies on population-based monitoring, potentially delaying detection of tumor progression and limiting personalized management.
There is a need for patient-specific tools to optimize AS protocols and improve early detection of disease progression.

Purpose Of The Study

To develop and validate a personalized computational model for predicting prostate cancer growth.
To identify novel biomarkers for higher-risk prostate cancer within the AS cohort.
To assess the potential of the predictive model to enable earlier detection of disease progression and inform personalized AS plans.

Main Methods

A spatiotemporal biomechanistic model was developed and personalized using longitudinal multiparametric MRI (mpMRI) data from prostate cancer patients.
The model's ability to forecast global tumor burden was evaluated using concordance correlation coefficients.
A novel biomarker, mean tumor proliferation activity, was identified and used to develop a risk classifier via logistic regression.

Main Results

The personalized biomechanistic model accurately represented and forecasted global tumor burden in individual patients (CCC: 0.93-0.99).
A model-based biomarker (mean proliferation activity) was identified as indicative of higher-risk prostate cancer (P = 0.041).
The risk classifier achieved an AUC of 0.83, and coupling forecasts with the classifier enabled early identification of progression to higher-risk disease by over one year.

Conclusions

Personalized biomechanistic modeling of prostate cancer using mpMRI data can accurately predict tumor progression.
A novel biomarker and risk classifier show promise for identifying higher-risk disease.
This predictive technology offers a potential clinical decision-making tool for designing personalized AS plans, improving management of prostate cancer.