# Automated MRI-based 3D Contractility (Strain) Analysis for Detecting Subclinical Cardiotoxicity in Breast Cancer Patients Undergoing Chemotherapy

> **NIH NIH R21** · UNIVERSITY OF SOUTH ALABAMA · 2020 · $172,944

## Abstract

Project Summary
Chemotherapy has made remarkable advances in the treatment of solid malignancies which cure
millions of patients from breast cancer. However, the adverse effects of cardiotoxicity mediated
by dose-dependent chemotherapeutic agents (such as anthracyclines) limit the efficacy of these
therapies, causing left-ventricular (LV) dysfunction in the form of cardiomyopathies and even heart
failure. Emerging evidence now suggest that early, myocardial strain-based, subclinical detection
of cardiotoxicity may facilitate timely interventions in treatment management and therefore slow
the progress of LV dysfunction and lower the incidence of heart failure. Recent clinical studies also
show that the strain-based approach to detecting subclinical cardiotoxicity is superior to detections
based on measuring differentials in LV ejection fraction (LVEF).
The primary objective of this study is to create a highly automated, diagnostic application with an
intuitive user-interface for computation of MRI-based myocardial contractile metrics in the LV in
general, and to specifically use it in this study towards predicting the onset of subclinical
cardiotoxicity in breast cancer patients. A parallel and equally important goal will involve
demonstrating that regional contractile metrics (Lagrangian radial, circumferential and longitudinal
normal strains, circumferential-radial and circumferential-longitudinal shear strains as well as twist
and torsion) can predict cardiotoxicity prior to LVEF. Among the above mentioned 3D contractile
metrics monitored, there will be an emphasis on our main hypothesis that torsion, which
parameterizes the base-to-apex twisting motion of myofibers, is pivotal for indicating myocardial
dysfunction. The source of data for these metrics will be the cardiac motion (displacements)
recorded with the phase encoding MRI sequence of navigator-gated, spiral cine DENSE.
Ultimately, the originality of this study will lie in our ability to fully automate the contractility
measurement process including computations of 3D LV boundaries using a combination of image
quantization and phase-unwrapping the DENSE data, in addition to the pointwise computation of
the contractile metrics in a 3D myocardial grid of the patient’s LV using the DENSE displacements.
The surveillance in each patient will be conducted at baseline (initiation of chemotherapy) and
during regular follow-up investigations to determine the role that regional strain-based contractile
metrics may have in detecting subclinical cardiotoxicity prior to LVEF. If established, the proof of
concept for this early detection will be provided with intra-parametric and inter-parametric analysis
of variance models, within and between the regional contractile metrics and LVEF and their
correlations to clinical data conducted after two follow-up investigations performed on the enrolled
chemotherapy patients.
The ultimate goal is that this novel, contractility-based surveillance tool can provi...

## Key facts

- **NIH application ID:** 9973071
- **Project number:** 5R21EB028063-02
- **Recipient organization:** UNIVERSITY OF SOUTH ALABAMA
- **Principal Investigator:** Julia Kar
- **Activity code:** R21 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2020
- **Award amount:** $172,944
- **Award type:** 5
- **Project period:** 2019-07-05 → 2022-03-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/9973071

## Citation

> US National Institutes of Health, RePORTER application 9973071, Automated MRI-based 3D Contractility (Strain) Analysis for Detecting Subclinical Cardiotoxicity in Breast Cancer Patients Undergoing Chemotherapy (5R21EB028063-02). Retrieved via AI Analytics 2026-05-21 from https://api.ai-analytics.org/grant/nih/9973071. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*