Assessing the variability of cardiac elasticity imaging to identify subclinical therapy related cardiac degradation in premenopausal women with breast cancer

Abstract Introduction: The Increasing incidence of breast cancer in younger women and significant survival gains underscore the importance of understanding cardiotoxicities of anti-neoplastic therapies. Current methods for assessing cardiac decline such as left ventricle ejection fraction and strain imaging are too coarse to identify pre-clinical cardiotoxicity, and emerging methods are cumbersome and/or expensive, complicating clinical utility. We have developed a model-based elasticity imaging method for assessing direct, functional mechanical stiffness of the left ventricle (LV) to indicate early cardiac function degradation which we used to assess cardiovascular changes in premenopausal women with breast cancer and further validated in a healthy cohort. Methods: Our method utilizes CINE cardiac MR images to produce LV elasticity maps in an inverse framework by comparing observed deformation of the LV estimated through non-rigid registration of cardiac MR time frames to biomechanical model estimated deformation. We examined our model-based framework in a cohort of pre-menopausal women either undergoing ovarian function suppression concurrent with aromatase inhibitors (OFS+AI) for HR+ breast cancer or triple negative breast cancer (TNBC) patients after completion of chemotherapy. Elasticity maps of the LV at initial and 3-6 month follow-up were compared to determine regional LV stiffening. To determine expected variability of LV regional stiffening, we conducted a reproducibility study in a healthy cohort by comparing elasticity maps based on scan/re-scan images acquired within the same session. Results: Elasticity maps showed significant stiffness changes in the HR+ cohort compared to the comparator TNBC cohort in longitudinal, radial, and shear directions (Table 1). LVEF and strain image assessment indicated no significant differences. Using the same model-based elasticity methodology we calculated global differences in resultant elasticity maps of healthy participants between scan and re-scan images and determined significant changes in the HR+ cohort greater than expected from methodological variability. Average global changes in stiffness in the HR+ cohort exceeded expected method variation with values 0.2944, 0.344, and 0.8199 for shear, radial, and longitudinal modulus respectively. (Table 2). Conclusion: This work demonstrates our ability to identify changes in LV mechanical elastic modulus using cardiac MR images, where clinical strain and LVEF analysis yielded inconclusive results. This study shows statistically significant differences in LV stiffness in a cohort of women undergoing OFS+AI therapy for HR+ breast cancer when compared to a comparator TNBC cohort. Stiffness changes exceed expected methodological variability, determined from a separate healthy cohort comparing scan/re-scan CINE cardiac MR images. This cardiac elasticity imaging method shows promise as an early indicator of cardiac function decline which might thereby allow for early cardioprotective interventions. Table 1.Mann-Whitney test p-values comparing changes in average relative moduli between HR-positiveBasal AnteriorMid AnteriorApical AnteriorApical InferiorMid InferiorBasal InferiorGlobalLongitudinal Modulus0.1570.2740.5880.036*0.011*0.0670.210Radial Modulus0.008*0.014*0.3500.5880.2410.2100.588Shear Modulus0.003*0.2740.9990.3110.4850.003*0.134 Table 2.Reproducibility statistics for global elasticity changes within same-session acquisitionsRadial ModulusCircumferential ModulusShear ModulusRepeatability0.1330.1070.07895% CI0.0330.0260.019wCV0.2190.2110.283 Citation Format: Caroline E Miller, Jennifer H Jordan, Emily Douglas, Katherine Ansley, Alexandra Thomas, Jared Weis. Assessing the variability of cardiac elasticity imaging to identify subclinical therapy related cardiac degradation in premenopausal women with breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-02-07.