(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: Breast cancer is the second leading cause of cancer-related mortality among women after lung cancer. In general, women who are treated for localized breast cancer are expected to live longer, and thus it is imperative to maintain better quality of life for those patients. Cardiac toxicity associated with radiation therapy could reduce the quality of life after treatment, and therefore the ability of prediction would greatly increase the chances of having better outcome. In this preliminary work we hypothesize that imaging features, extracted from pre-treatment CT, can be used to predict the possibility of post-treatment heart toxicity.
Methods: We retrospectively studied seventeen patients who received standard radiation treatment for localized breast cancer. Eight of those patients developed post-radiotherapy cardiac toxicity. Delineation of left atrium, right atrium, left ventricle, right ventricle, left main coronary artery, left anterior descending coronary artery (LAD), right circumflex artery (RCA), right coronary artery, atrioventricular (AV) node, ascending aorta, and pulmonary artery were performed according to published guidelines. For each of the contoured structures 92 geometric, first-, second-, and third-order texture features were calculated for the planning CT. Odds ratio and p-value were estimated for each structure and pertinent imaging feature.
Results: Area under the receiver operating characteristic curve (AUC) was 0.875, and 0.917 for AV node and left atrium respectively. The first relevant variable was one texture feature for the AV node which was border-line significant but with large OR (OR: 38.78, p=0.07). The second relevant feature was the left atrium volume which was significant and again with large OR (OR: 39.89, p=0.037) .
Conclusion:
Post-treatment cardiac toxicity in breast cancer patients was associated with imaging features of the left atrium and the AV node. Future work would explore the expansion of the current database to include more patients, systemic therapy, and comorbidities.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
