CT Simulation is Useful for Quantifying Pre-Treatment Coronary Artery Calcification in Breast Cancer Patients

Abstract

Women diagnosed with breast cancer often receive oncologic treatments like chemotherapy and radiation that increase their risk of having a significant cardiac event. If high-cardiovascular-risk women can be identified before treatment, measures could be taken to reduce cardiac toxicity resulting in a large impact on cardiac- and all-cause mortality. In addition to traditional Framingham risk factors, coronary artery calcium (CAC) score gives important prognostic information for risk of cardiac event. While a cardiac-cycle gated CT scan is the standard imaging modality for CAC quantification, non-gated scans have shown good concordance with gated scans and other CV risk factors. Herein, we undertake a survey of breast cancer patients to determine the distribution of cardiovascular risk using CAC score, and quantify the reproducibility of CAC score calculated using CT simulation scans for breast cancer patients. The 100 most recent patients carrying a diagnosis of breast cancer and having CT simulation scans (one primary and a subsequent scan for boost plan) within 2 weeks and 2 months apart were included. Patients’ CT simulation scans were screened for CAC in a treatment planning system. CT window was set between 100-120 Hounsfield units. For patients’ whose CT simulation scans were negative for calcium during screening, a CAC score of 0 was recorded. For any patient with CAC in any of his/her scans, were exported as DICOM files and imported into InTuition for manual CAC score calculation using the Agatston method. Briefly, the density (in Hounsfield units) of a lesion in a coronary artery is converted to a score. That score is then weighted by the area of the lesion in square millimeters. CAC score risk stratification was performed using previously defined Agatston score ranges, and concordance assessed between scans for each patient. There were 47 left breast, 48 right breast, and 5 bilateral breast cancer cases. 73% of patients had a CAC score of zero on both scans. The agreement of the absolute CAC score between CT sim scans of the same individual was high (Rˆ2 = 0.98). The CAC risk group concordance was 90%. Calculating a CAC from a CT simulation scan is feasible and has good replicability on multiple scans. Future directions for this research include improved treatment decision-making by leveraging this radiographic data to personalize treatments with more conformal therapies to avoid heart dose for women at higher baseline risk of CVD.