Preliminary Application of a Multitaper Generalized Spectrum Approach to Quantify Scatterer Nonrandomness in Breast Lesions
I Rosado-Mendez*, H Gerges-Nasief, L Carlson, S Kohn, T Hall, J Zagzebski, University of Wisconsin-Madison, Madison, WIWE-E-134-1 Wednesday 2:00PM - 3:50PM Room: 134
Purpose: The generalized spectrum (GS) has been used to quantify nonrandomness of scattering sources and to characterize breast and liver tissues. In this work, we implement a multitaper GS approach creating parametric images of GS features that describe the degree of nonrandom structure in breast lesions.
Methods: This IRB-approved, HIPAA compliant study recruited subjects who were scheduled for core biopsy of a suspicious breast mass. Subjects were scanned with a Siemens Acuson S2000 using a linear array transducer at 15MHz. Radiofrequency (RF) echo data were obtained in radial (parallel to breast lactiferous ducts) and antiradial planes. Cases (n=34) were selected based on the availability of biopsy results and a lesion size of at least 3mm in both planes. The multitaper GS was obtained by computing the correlation matrix of the Fourier transform of RF signal segments tapered by Prolate Spheroidal Sequences. Nonrandomness was quantified by parameterizing the collapsed average of the GS. A region of interest (ROI) within each lesion and each plane was defined,and the mean and standard deviation of the nonrandomness within it were computed.
Results: Cases were grouped into: fibroadenoma (n=12, median diameter (MD)=7.6mm),other benign(n=11, MD=8.2mm), invasive ductal carcinoma (IDC) (n=5, MD=6.3mm), and other malignant(n=6, MD=6.9mm). The IDC nonrandomness inter-case median of the mean and standard deviation was 64% and 59% of the values from other lesions, and the fractional difference in the standard deviation of the nonrandomness between the radial and antiradial views for the IDC cases was 59% smaller than the values for other lesions.
Conclusion: Preliminary results indicate that the spatial distribution of scatterers within IDCs is more homogeneously random compared to other lesions. However, inter-case variability was comparable to the median values. Future work will expand the population and use beam steered acquisitions to test for anisotropy in these features.
Funding Support, Disclosures, and Conflict of Interest: This work was supported, in part, by NIH (grants R01CA111289, R21HD061896, R21HD063031, and R01HD072077) and the Consejo Nacional de Ciencia y Tecnologia of Mexico (Reg. 206414).