K. K. Holli

Characterization of breast cancer types by texture analysis of magnetic resonance images.

RATIONALE AND OBJECTIVES This novel study aims to investigate texture parameters in distinguishing healthy breast tissue and breast cancer in breast magnetic resonance imaging (MRI). A specific aim was to identify possible differences in the texture characteristics of histological types (lobular and ductal) of invasive breast cancer and to determine the value of these differences for computer-assisted lesion classification. MATERIALS AND METHODS Twenty patients (mean age 50.6 + or - SD 10.6; range 37-70 years), with histopathologically proven invasive breast cancer (10 lobular and 10 ductal) were included in this preliminary study. The median MRI lesion size was 25 mm (range, 7-60 mm). The selected T1-weighted precontrast, post-contrast, and subtracted images were analyzed and classified with texture analysis (TA) software MaZda and additional statistical tests were used for testing the parameters separability. RESULTS All classification methods employed were able to differentiate between cancer and healthy breast tissue and also invasive lobular and ductal carcinoma with classification accuracy varying between 80% and 100%, depending on the used imaging series and the type of region of interest. We found several parameters to be significantly different between the regions of interest studied. The co-occurrence matrix based parameters proved to be superior to other texture parameters used. CONCLUSIONS The results of this study indicate that MRI TA differentiates breast cancer from normal tissue and may be able to distinguish between two histological types of breast cancer providing more accurate characterization of breast lesions thereby offering a new tool for radiological analysis of breast MRI.

Mild Traumatic Brain Injury: Tissue Texture Analysis Correlated to Neuropsychological and DTI Findings

RATIONALE AND OBJECTIVES The aim of this study was to evaluate whether texture analysis (TA) can detect subtle changes in cerebral tissue caused by mild traumatic brain injury (MTBI) and to determine whether these changes correlate with neuropsychological and diffusion tensor imaging (DTI) findings. MATERIALS AND METHODS Forty-two patients with MTBIs were imaged using 1.5T magnetic resonance imaging within 3 weeks after head injury. TA was performed for the regions corresponding to the mesencephalon, centrum semiovale, and corpus callosum. Using DTI, the fractional anisotropic and apparent diffusion coefficient values for the same regions were evaluated. The same analyses were performed on a group of 10 healthy volunteers. Patients also underwent a battery of neurocognitive tests within 6 weeks after injury. RESULTS TA revealed textural differences between the right and left hemispheres in patients with MTBIs, whereas differences were minimal in healthy controls. A significant correlation was found between scores on memory tests and texture parameters (sum of squares, sum entropy, inverse difference moment, and sum average) in patients in the area of the mesencephalon and the genu of the corpus callosum. Significant correlations were also found between texture parameters for the left mesencephalon and both fractional anisotropic and apparent diffusion coefficient values. CONCLUSIONS The data suggest that heterogeneous texture and abnormal DTI patterns in the area of the mesencephalon may be linked with verbal memory deficits among patients with MTBIs. Therefore, TA combined with DTI in patients with MTBIs may increase the ability to detect early and subtle neuropathologic changes.

MRI Texture Analysis in Multiple Sclerosis: Toward a Clinical Analysis Protocol

RATIONALE AND OBJECTIVES Magnetic resonance imaging (MRI)-based texture analysis has been shown to be effective in classifying multiple sclerosis lesions. Regarding the clinical use of texture analysis in multiple sclerosis, our intention was to show which parts of the analysis are sensitive to slight changes in textural data acquisition and which steps tolerate interference. MATERIALS AND METHODS The MRI datasets of 38 multiple sclerosis patients were used in this study. Three imaging sequences were compared in quantitative analyses, including a comparison of anatomical levels of interest, variance between sequential slices and two methods of region of interest drawing. We focused on the classification of white matter and multiple sclerosis lesions in determining the discriminatory power of textural parameters. Analyses were run with MaZda software for texture analysis, and statistical tests were performed for raw parameters. RESULTS MRI texture analysis based on statistical, autoregressive-model and wavelet-derived texture parameters provided an excellent distinction between the image regions corresponding to multiple sclerosis plaques and white matter or normal-appearing white matter with high accuracy (nonlinear discriminant analysis 96%-100%). There were no significant differences in the classification results between imaging sequences or between anatomical levels. Standardized regions of interest were tolerant of changes within an anatomical level when intra-tissue variance was tested. CONCLUSION The MRI texture analysis protocol with fixed imaging sequence and anatomical levels of interest shows promise as a robust quantitative clinical means for evaluating multiple sclerosis lesions.

Parkinson's disease: interhemispheric textural differences in MR images.

RATIONALE AND OBJECTIVES Early-stage diagnosis of Parkinsons disease (PD) is essential in making decisions related to treatment and prognosis. However, there is no specific diagnostic test for the diagnosis of PD. The aim of this study was to evaluate the role of texture analysis (TA) of magnetic resonance images in detecting subtle changes between the hemispheres in various brain structures in patients with early symptoms of parkinsonism. In addition, functional TA parameters for detecting textural changes are presented. MATERIALS AND METHODS Fifty-one patients with symptoms of PD and 20 healthy controls were imaged using a 3-T magnetic resonance device. Co-occurrence matrix-based TA was applied to detect changes in textures between the hemispheres in the following clinically interesting areas: dentate nucleus, basilar pons, substantia nigra, globus pallidus, thalamus, putamen, caudate nucleus, corona radiata, and centrum semiovale. The TA results were statistically evaluated using the Mann-Whitney U test. RESULTS The results showed interhemispheric textural differences among the patients, especially in the area of basilar pons and midbrain. Concentrating on this clinically interesting area, the four most discriminant parameters were defined: co-occurrence matrix correlation, contrast, difference variance, and sum variance. With these parameters, differences were also detected in the dentate nucleus, globus pallidus, and corona radiata. CONCLUSIONS On the basis of this study, interhemispheric differences in the magnetic resonance images of patients with PD can be identified by the means of co-occurrence matrix-based TA. The detected areas correlate with the current pathophysiologic and neuroanatomic knowledge of PD.

Detection of characteristic texture parameters in breast MRI

Breast cancer is the most common cancer in women. Breast MRI (BMRI) has emerged as a promising technique for detecting, diagnosing, and staging the condition. Automated image analysis aims to extract relevant information from MR images of the breast and improve the accuracy and consistency of image interpretation. Texture analysis (TA) is one possible means of detecting tissue features in biomedical images.

Texture Analysis of Corpus Callosum in Mild Traumatic Brain Injury Patients

Texture analysis (TA) is a quantitative approach for characterizing subtle changes in magnetic resonance (MR) images of different tissues. The aim of this study was to detect changes in tissue of corpus callosum (CC) in mild traumatic brain injury (MTBI) patients by the means of TA.

Software Phantoms for Texture Analysis

Texture analysis (TA) is a potential tool for analysis of medical images. It can be used for classification of pathological tissue. Co-occurrence matrix is one of the most promising TA methods. In this study we have analysed this method by using software phantoms instead of physical phantoms. This choice was made because the construction process of a real phantom is slow and only one set of parameter results is available. The software phantoms were implemented with Matlab program by constructing 16x16 matrices containing four different grey level values. Value from 0 to 1, 0 corresponding to black and 1 being white, was given for each matrix element to perform different texture patterns. Grey scale images were drawn from the matrices and the texture analysis was performed with MaZda. Software phantoms were proved to be an effective method to study the parameter value distribution because of the easy construction and modification of the matrices. However, more complex patterns should be used for further studies.

Manual Segmentation of Brain Tissue and Multiple Sclerosis Lesions for Texture Analysis

Magnetic resonance imaging based texture analysis has been shown effective on classifying multiple sclerosis lesions. Quantitative analysis of images contains several manual and automatic steps. Results of subtle changes between tissues may suffer from errors in analysis protocol. For the development of clinical analysis protocol we evaluate the potential of non-specialized medics to carry out image slice selection and manual segmentation of brain tissue for texture analysis purposes as an assistant for a specialist.

Texture Analysis as a Tool in Diagnosis of Parkinson’s Disease

Parkinson’s disease is the second most common neurodegenerative disorder of later life. Symptoms have proven to originate from the substantia nigra area. In our study, texture analysis was applied to 3-T magnetic resonance images of 28 patients with symptoms of parkinsonism. Of these, 14 had diagnosed Parkinson’s disease and 14 had no clinical confirmation for the particular disease. Areas of substantia nigra (pars reticulata, pars compacta and red nucleus), were analyzed with texture analysis software MaZda from both patient groups and results were further analyzed with the integrated software B11. Classification results show that substantia nigra areas of patients with and without diagnosis of Parkinson’s disease can be distinguished with high accuracy.

Use of Diffusion Tensor MR Imaging and 3D tractography software in mild traumatic brain injury

Patients with mild traumatic brain injury (MTBI) pose an important dilemma to daily radiological practice because of a negative conventional brain MRI. The aim was to analyze the diffusion tensor imaging (DTI) findings in the sub-acute stage (3 weeks from onset) in 10 MTBI patients and in 10 control subjects using a brain MRI (1.5 Tesla) and a novel DTI Task Card software. Average fractional anisotropy and apparent diffusion coefficient values were abnormal (FA=0.628 and ADC=7.298) and tractography images revealed discontinuation in long white matter tracts. The findings were significantly correlated to neuropsychological findings. Our findings suggest that cellular disturbance and breakdown is demonstrated in the white matter tracts in the sub-acute stage of TBI using DTI software and MRI.