Differential Diagnosis of Breast Cancers Using Texture Analysis of Magnetic Resonance Imaging

The purpose of this study was the differential diagnosis between the two types of invasive breast cancer (Lobular and Ductal) using MRI images in MATLAB software. In this perspective study ,20 patients with pathological confirmed invasive breast cancer (10 lobular carcinoma and 10 ductal carcinoma) were examined. The images that analyzed included T1-w before and after contrast injection then these two sets of images were subtracted, after that these subtraction images were imported to MATLAB software for expected texture analysis. 10 gray level co-occurrence matrix texture features were obtained from each lesion. These features were compared in two types of breast cancers by statistical analysis. The data of these patients were categorized in two subgroups included ductal and lobular carcinoma (mean age: 52.6 ±9.5). The two groups were matched for age and sex with each other. The texture analysis based on co-occurrence matrix of two types of breast cancers showed significantdifferences (p=0.05).The results of this study indicated that texture analysis of MRI images can distinguish between two types of breast cancers. By using these results, we can assess breast lesions decision more accurately. Thereby texture analysis of breast lesions offering a new tool for radiological analysis of breast MRI.

Keywords:

Breast cancer , MRI , MATLAB , Co-occurrence matrix