Skip to Main Content
An official website of the United States government

Parameter Calibration for a Modified KELIM Model

Principal Investigator

Name
Benedetto Piccoli

Degrees
Ph.D.

Institution
Rutgers University - Camden

Position Title
Joseph and Loretta Lopez Chair Professor of Mathematics

Email
piccoli@camden.rutgers.edu

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-1887

Initial CDAS Request Approval
Apr 23, 2025

Title
Parameter Calibration for a Modified KELIM Model

Summary
The KELIM (K-ELIMination rate) Model is a mathematical model that is used to estimate the responsiveness of ovarian carcinomas to chemotherapy in a neo-adjuvant setting. The KELIM model uses longitudinal CA-125 measurements to inform the timing of a de-bulking surgery. Additionally, the model can be applied in an adjuvant setting to determine if the ovarian cancer is responding to adjuvant therapies. This project aims to fit the standard KELIM model parameters for CA-125 dynamics in ovarian cancer patients across different patient populations. This project will also examine a modified version of the standard KELIM model that introduces explicit equations for tumor size, motivated by the goal of creating a KELIM model that is more mechanistically motivated. We aim to fit model parameters to the modified KELIM model in order to compare to the standard KELIM model.

Aims

- Using longitudinal CA-125 measurements from the Ovarian cancer dataset, we will fit the standard KELIM (K-ELIMination rate) model parameters
- We may examine the dependence of model parameters on secondary data types, such as demographic information (such as age)
- We will propose a modification of the KELIM model that introduces additional model equations corresponding to compartments for tumor size
- Using the same longitudinal CA-125 measurements from the first bullet, we aim to fit the modified model parameters
- Compare the performance of the modified KELIM model to the standard KELIM model
- Compare the each models' robustness to "secondary" data types, such as demographic data (eg: age)

Collaborators

Christopher Denaro
Kazi Tanzina Begum
Precious Akinde
Nikita Yazvinskyi
Ryan Weightman