Breastis :|a software tool for flexible breast MRI analysis

Abstract

Magnetic resonance imaging of breast provides valuable information about breast tissue composition. A common breast MRI protocol may include dynamic contrast enhanced (DCE) MRI, diffusion weighted MRI (DWI) and proton MR spectroscopic imaging (1H-MRSI). There have been several software tools that can analyze each of these data types separately. In this study, a flexible and open-source postprocessing software called ‘BreastIS ’was developed to analyze DCE-MRI, DWI, and 1H-MRSI and store them in a database for further exploration. BreastIS image processing software was implemented using MATLAB and the graphical user interface was developedusingMATLABGUIandJava. ThesoftwarecouldberunonWindows, Mac and Linux computers. A retrospective study was conducted to test the suitability of the analysis algorithms of BreastIS tool for the use with clinical dataset. DCE-MRI data of 16 and DWI data of 6 breast cancer subjects were analyzed with BreastIS. For DCE-MRI analysis, the semi-quantitative parameters such as early and maximum percentage enhancement, signal uptake pattern and maximum pixel intensity were calculated for healthy and tumor regions of each subject. For DWI analysis, mean and maximum apparent diffusion coefficient (ADC) values were calculated for tumor and healthy regions of each subject. A Mann-Whitney rank sum test with Bonferroni multiple comparison correction was applied to find statistically significant differences betweenhealthyandtumorregions. Maximumandearlypercentageenhancementsand maximum pixel intensity were higher (P<0.001), and mean ADC values were lower in tumor regions (P=0.002). The DCE-MRI signal uptake pattern displayed wash-out in tumor regions. The sample analysis results indicated the suitability and usability of BreastIS tool for analysis of clinical breast MRI datasets.|Keywords : Breast cancer, MRI, post-processing, software analysis tool development.