Meeting Banner
Abstract #3364

Association of Lesion and Background Parenchyma in Diagnostically Challenging Breast Lesions

Ana Paula Klautau Leite1, 2, Melanie Freed1, Melanie Moccaldi3, Kai Tobias Block4, Amy Melsaether3, Alana Amarosa3, Sungheon Kim1, Linda Moy3

1Radiology, Center for Biomedical Imaging, New York Universiy School of Medicine, New York, United States; 2Radiology, Hospital das Clinicas- FMUSP, School of Medicine - University of Sao Paulo, Sao Paulo, SP, Brazil; 3Radiology, Cancer Institute, New York University School of Medicine, New York, United States; 4Radiology, NYU Langone Medical Center, New York, United States

The relationship between a lesion in the breast and its effect on contrast uptake kinetics in the breast parenchyma (PS) is unclear. This study investigated whether the quantitative analysis of MR contrast-enhancement kinetics of BP and lesions improves our ability to discriminate between benign and malignant tumors. We analyzed 102 lesions using a linear principal component analysis (PCA) and found that we can correctly identify benign lesions compared to malignant lesions by quantitatively comparing its enhancement with the adjacent BP. We almost reached statistical significance when performing the same analysis on malignant lesions.

Keywords

accuracy adjacent allows almost among amongst applied around artery association asymptomatic background benign biomedical biopsies biopsy block breast breasts cancer cancers challenging channel characteristics classified coil component conditions confidence consecutive considered considering contain contrast corresponds dashed datasets decompose delayed detailed device diagnosed diagnostic diagnostically discrimination divided drew dynamic eigenvalues eigenvectors electronic england enhancement entire existing fast feasibility fifth final findings five frame frames guided hence histology history house hundred improve indeterminate initial injection internal interpreting intervals investigate kinetic kinetics king known lesion lesions linear made malignant mammary manually maps measure measured measures medical medicine newly normalized objective others particularly patient patients paucity percent plot plots poorly positive post principal prior projection quantitative radial radiology reached reader recommended records regression resolution retrospectively review reviewed risk scatter school screening sensitivity side significance slice slices software specificity statistical studies suspicious system third treated treatment tumors type undergoing underwent unilateral unique unnecessary variable vibe women