Magnetic Resonance Texture Analysis Research Articles

Texture Analysis Using Machine Learning-Based 3-T Magnetic Resonance Imaging for Predicting Pathological Response in Breast Cancer Patients Treated with Neoadjuvant Chemotherapy

Neoadjuvant chemotherapy (NAC) is the standard treatment for large operable or locally advanced breast cancer. Magnetic resonance imaging (MRI) is an important tool to evaluate tumor extent and response to treatment. In the future, surgery may not be required for patients showing pathologic complete response (pCR) after NAC. Therefore, early identification of such patients is crucial for improving and personalizing patient treatment. A total of 90 patients were identified who were diagnosed with invasive breast carcinoma and underwent breast MRI before and after NAC between August 2014 and May 2018. The patients were divided into training group (72 cases) and test group (18 cases). Immunohistochemical data were used to classify breast carcinomas into five molecular subtypes: Luminal A; Luminal B; Her-2 negative; Her-2 positive; or triple-negative and the corresponding surgical specimens were evaluated according to Residual Cancer Burden (RCB) criteria for pathological response. Statistical analyses were conducted using statistical software and statistical significance was defined as P < 0.05. A total of 90 patients with invasive breast carcinoma were enrolled in our study, including 72 patients in training group and 18 patients in test group. The accuracy of molecular subtypes is Luminal A (0.77); Luminal B (0.78); Her-2 negative (0.72); Her-2 positive (0.79); or triple-negative (0.87). Compared with other molecular subtypes, triple-negative show better prediction results (P= 0.028). The accuracy of pathological response is Grade 1 chemotherapy response (0.73); Grade 2 chemotherapy response (0.84); Grade 3 chemotherapy response (0.78); Grade 4 chemotherapy response (0.79); Grade 5 chemotherapy response (0.77). There was no significant difference in chemotherapy response between five groups. The results of this analysis suggested that three negative breast cancer has better prediction effect in magnetic resonance texture analysis for breast cancer patients receiving NAC. However, magnetic resonance texture analysis did not show statistical differences in chemotherapy response. A multicenter randomized controlled prospective trial may be needed.

Magnetic resonance texture analysis reveals stagewise nonlinear alterations of the frontal gray matter in patients with early psychosis.

Although gray matter (GM) abnormalities are present from the early stages of psychosis, subtle/miniscule changes may not be detected by conventional volumetry. Texture analysis (TA), which permits quantification of the complex interrelationship between contrasts at the individual voxel level, may capture subtle GM changes with more sensitivity than does volume or cortical thickness (CTh). We performed three-dimensional TA in nine GM regions of interest (ROIs) using T1 magnetic resonance images from 101 patients with first-episode psychosis (FEP), 85 patients at clinical high risk (CHR) for psychosis, and 147 controls. Via principal component analysis, three features of gray-level cooccurrence matrix - informational measure of correlation 1 (IMC1), autocorrelation (AC), and inverse difference (ID) - were selected to analyze cortical texture in the ROIs that showed a significant change in volume or CTh in the study groups. Significant reductions in GM volume and CTh of various frontotemporal regions were found in the FEP compared with the controls. Increased frontal AC was found in the FEP group compared to the controls after adjusting for volume and CTh changes. While volume and CTh were preserved in the CHR group, a stagewise nonlinear increase in frontal IMC1 was found, which exceeded both the controls and FEP group. Increased frontal IMC1 was also associated with a lesser severity of attenuated positive symptoms in the CHR group, while neither volume nor CTh was. The results of the current study suggest that frontal IMC1 may reflect subtle, dynamic GM changes and the symptomatology of the CHR stage with greater sensitivity, even in the absence of gross GM abnormalities. Some structural mechanisms that may contribute to texture changes (e.g., macrostructural cortical lamina, neuropil/myelination, cortical reorganization) and their possible implications are explored and discussed. Texture may be a useful tool to investigate subtle and dynamic GM abnormalities, especially during the CHR period.

Texture Analysis of the Apparent Diffusion Coefficient Focused on Contrast-Enhancing Lesions in Predicting Survival for Bevacizumab-Treated Patients with Recurrent Glioblastoma.

Glioblastoma often recurs after treatment. Bevacizumab increases progression-free survival in some patients with recurrent glioblastoma. Identifying pretreatment predictors of survival can help clinical decision making. Magnetic resonance texture analysis (MRTA) quantifies macroscopic tissue heterogeneity indirectly linked to microscopic tissue properties. We investigated the usefulness of MRTA in predicting survival in patients with recurrent glioblastoma treated with bevacizumab. We evaluated retrospective longitudinal data from 33 patients (20 men; mean age 56 ± 13 years) who received bevacizumab on the first recurrence of glioblastoma. Volumes of contrast-enhancing lesions segmented on postcontrast T1-weighted sequences were co-registered on apparent diffusion coefficient maps to extract 107 radiomic features. To assess the performance of textural parameters in predicting progression-free survival and overall survival, we used receiver operating characteristic curves, univariate and multivariate regression analysis, and Kaplan-Meier plots. Longer progression-free survival (>6 months) and overall survival (>1 year) were associated with lower values of major axis length (MAL), a lower maximum 2D diameter row (m2Ddr), and higher skewness values. Longer progression-free survival was also associated with higher kurtosis, and longer overall survival with higher elongation values. The model combining MAL, m2Ddr, and skewness best predicted progression-free survival at 6 months (AUC 0.886, 100% sensitivity, 77.8% specificity, 50% PPV, 100% NPV), and the model combining m2Ddr, elongation, and skewness best predicted overall survival (AUC 0.895, 83.3% sensitivity, 85.2% specificity, 55.6% PPV, 95.8% NPV). Our preliminary analyses suggest that in patients with recurrent glioblastoma pretreatment, MRTA helps to predict survival after bevacizumab treatment.

Stroke risk study based on deep learning-based magnetic resonance imaging carotid plaque automatic segmentation algorithm.

The primary factor for cardiovascular disease and upcoming cardiovascular events is atherosclerosis. Recently, carotid plaque texture, as observed on ultrasonography, is varied and difficult to classify with the human eye due to substantial inter-observer variability. High-resolution magnetic resonance (MR) plaque imaging offers naturally superior soft tissue contrasts to computed tomography (CT) and ultrasonography, and combining different contrast weightings may provide more useful information. Radiation freeness and operator independence are two additional benefits of M RI. However, other than preliminary research on MR texture analysis of basilar artery plaque, there is currently no information addressing MR radiomics on the carotid plaque. For the automatic segmentation of MRI scans to detect carotid plaque for stroke risk assessment, there is a need for a computer-aided autonomous framework to classify MRI scans automatically. We used to detect carotid plaque from MRI scans for stroke risk assessment pre-trained models, fine-tuned them, and adjusted hyperparameters according to our problem. Our trained YOLO V3 model achieved 94.81% accuracy, RCNN achieved 92.53% accuracy, and MobileNet achieved 90.23% in identifying carotid plaque from MRI scans for stroke risk assessment. Our approach will prevent incorrect diagnoses brought on by poor image quality and personal experience. The evaluations in this work have demonstrated that this methodology produces acceptable results for classifying magnetic resonance imaging (MRI) data.

MRI whole-lesion texture analysis on ADC maps for the prognostic assessment of ischemic stroke

BackgroundThis study aims is to explore whether it is feasible to use magnetic resonance texture analysis (MRTA) in order to distinguish favorable from unfavorable function outcomes and determine the prognostic factors associated with favorable outcomes of stroke.MethodsThe retrospective study included 103 consecutive patients who confirmed unilateral anterior circulation subacute ischemic stroke by computed tomography angiography between January 2018 and September 2019. Patients were divided into favorable outcome (modified Rankin scale, mRS ≤ 2) and unfavorable outcome (mRS > 2) groups according to mRS scores at day 90. Two radiologists manually segmented the infarction lesions based on diffusion-weighted imaging and transferred the images to corresponding apparent diffusion coefficient (ADC) maps in order to extract texture features. The prediction models including clinical characteristics and texture features were built using multiple logistic regression. A univariate analysis was conducted to assess the performance of the mean ADC value of the infarction lesion. A Delong’s test was used to compare the predictive performance of models through the receiver operating characteristic curve.ResultsThe mean ADC performance was moderate [AUC = 0.60, 95% confidence interval (CI) 0.49–0.71]. The texture feature model of the ADC map (tADC), contained seven texture features, and presented good prediction performance (AUC = 0.83, 95%CI 0.75–0.91). The energy obtained after wavelet transform, and the kurtosis and skewness obtained after Laplacian of Gaussian transformation were identified as independent prognostic factors for the favorable stroke outcomes. In addition, the combination of the tADC model and clinical characteristics (hypertension, diabetes mellitus, smoking, and atrial fibrillation) exhibited a subtly better performance (AUC = 0.86, 95%CI 0.79–0.93; P > 0.05, Delong’s).ConclusionThe models based on MRTA on ADC maps are useful to evaluate the clinical function outcomes in patients with unilateral anterior circulation ischemic stroke. Energy obtained after wavelet transform, kurtosis obtained after Laplacian of Gaussian transform, and skewness obtained after Laplacian of Gaussian transform were identified as independent prognostic factors for favorable stroke outcomes.

Magnetic resonance imaging texture analysis for the evaluation of viable ovarian tissue in patients with ovarian endometriosis: a retrospective case-control study.

BackgroundTexture analysis has been used as a method for quantifying image properties based on textural features. The aim of the present study was to evaluate the usefulness of magnetic resonance imaging (MRI) texture analysis for the evaluation of viable ovarian tissue on the perfusion map of ovarian endometriosis.MethodsTo generate a normalized perfusion map, subtracted T1-weighted imaging (T1WI), T1WI and contrast-enhanced T1W1 with sequences were performed using the same parameters in 25 patients with surgically confirmed ovarian endometriosis. Integrated density is defined as the sum of the values of the pixels in the image or selection. We investigated the parameters for texture analysis in ovarian endometriosis, including angular second moment (ASM), contrast, correlation, inverse difference moment (IDM), and entropy, which is equivalent to the product of area and mean gray value.ResultsThe perfusion ratio and integrated density of normal ovary were 0.52±0.05 and 238.72±136.21, respectively. Compared with the normal ovary, the affected ovary showed significant differences in total size (p<0.001), fractional area ratio (p<0.001), and perfusion ratio (p=0.010) but no significant differences in perfused tissue area (p=0.158) and integrated density (p=0.112). In comparison of parameters for texture analysis between the ovary with endometriosis and the contralateral normal ovary, ASM (p=0.004), contrast (p=0.002), IDM (p<0.001), and entropy (p=0.028) showed significant differences. A linear regression analysis revealed that fractional area had significant correlations with ASM (r2=0.211), IDM (r2=0.332), and entropy (r2=0.289).ConclusionMRI texture analysis could be useful for the evaluation of viable ovarian tissues in patients with ovarian endometriosis.

Survival prediction in glioblastoma on post-contrast magnetic resonance imaging using filtration based first-order texture analysis: Comparison of multiple machine learning models.

Magnetic resonance texture analysis (MRTA) is a relatively new technique that can be a valuable addition to clinical and imaging parameters in predicting prognosis. In the present study, we investigated the efficacy of MRTA for glioblastoma survival using T1 contrast-enhanced (CE) images for texture analysis. We evaluated the diagnostic performance of multiple machine learning models based on first-order histogram statistical parameters derived from T1-weighted CE images in the survival stratification of glioblastoma multiforme (GBM). Retrospective evaluation of 85 patients with GBM was performed. Thirty-six first-order texture parameters at six spatial scale filters (SSF) were extracted on the T1 CE axial images for the whole tumor using commercially available research software. Several machine learning classification models (in four broad categories: linear, penalized linear, non-linear, and ensemble classifiers) were evaluated to assess the survival prediction performance using optimal features. Principal component analysis was used prior to fitting the linear classifiers in order to reduce the dimensionality of the feature inputs. Fivefold cross-validation was used to partition the data iteratively into training and testing sets. The area under the receiver operating characteristic curve (AUC) was used to assess the diagnostic performance. The neural network model was the highest performing model with the highest observed AUC (0.811) and cross-validated AUC (0.71). The most important variable was the age at diagnosis, with mean and mean of positive pixels (MPP) for SSF = 0 being the second and third most important, followed by skewness for SSF = 0 and SSF = 4. First-order texture features, when combined with age at presentation, show good accuracy in predicting GBM survival.

MR texture analysis in differentiating renal cell carcinoma from lipid-poor angiomyolipoma and oncocytoma.

To assess the utility of magnetic resonance texture analysis (MRTA) in differentiating renal cell carcinoma (RCC) from lipid-poor angiomyolipoma (lpAML) and oncocytoma. After ethical approval, 42 patients with 54 masses (34 RCC, 14 lpAML and six oncocytomas) who underwent MRI on a 1.5 T scanner (Avanto, Siemens, Erlangen, Germany) between January 2011 and December 2012 were retrospectively included in the study. MRTA was performed on the TexRAD research software (Feedback Plc., Cambridge, UK) using free-hand polygonal region of interest (ROI) drawn on the maximum cross-sectional area of the tumor to generate six first-order statistical parameters. The Mann-Whitney U test was used to look for any statically significant difference. The receiver operating characteristic (ROC) curve analysis was done to select the parameter with the highest class separation capacity [area under the curve (AUC)] for each MRI sequence. Several texture parameters on MRI showed high-class separation capacity (AUC > 0.8) in differentiating RCC from lpAML and oncocytoma. The best performing parameter in differentiating RCC from lpAML was mean of positive pixels (MPP) at SSF 2 (AUC: 0.891) on DWI b500. In differentiating RCC from oncocytoma, the best parameter was mean at SSF 0 (AUC: 0.935) on DWI b1000. MRTA could potentially serve as a useful non-invasive tool for differentiating RCC from lpAML and oncocytoma. There is limited literature addressing the role of MRTA in differentiating RCC from lpAML and oncocytoma. Our study demonstrated several texture parameters which were useful in this regard.

High mean entropy calculated from cardiac MRI texture analysis is associated with antitachycardia pacing failure.

Antitachycardia pacing (ATP), which may avoid unnecessary implantable cardioverter-defibrillator (ICD) shocks, does not always terminate ventricular arrhythmias (VAs). Mean entropy calculated using cardiac magnetic resonance texture analysis (CMR-TA) has been shown to predict appropriate ICD therapy. We examined whether scar heterogeneity, quantified by mean entropy, is associated with ATP failure and explore potential mechanisms using computer modeling. A subanalysis of 114 patients undergoing CMR-TA where the primary endpoint was delivery of appropriate ICD therapy (ATP or shock therapy) was performed. Patients receiving appropriate ICD therapy (n=33) were dichotomized into "successful ATP" versus "shock therapy" groups. In silico computer modeling was used to explore underlying mechanisms. A total of 16 of 33 (48.5%) patients had successful ATP to terminate VA, and 17 of 33 (51.5%) patients required shock therapy. Mean entropy was significantly higher in the shock versus successful ATP group (6.1 ± 0.5vs 5.5 ± 0.7, P=.037). Analysis of patients receiving ATP (n=22) showed significantly higher mean entropy in the six of 22 patients that failed ATP (followed by rescue ICD shock) compared to 16 of 22 that had successful ATP (6.3 ± 0.7vs 5.5 ± 0.7, P=.048). Computer modeling suggested inability of the paced wavefront in ATP to successfully propagate from the electrode site through patchy fibrosis as a possible mechanism of failed ATP. Our findings suggest lower scar heterogeneity (mean entropy) is associated with successful ATP, whereas higher scar heterogeneity is associated with more aggressive VAs unresponsive to ATP requiring shock therapy that may be due to inability of the paced wavefront to propagate through scar and terminate the VA circuit.

Mean entropy predicts implantable cardioverter-defibrillator therapy using cardiac magnetic resonance texture analysis of scar heterogeneity

Risk stratification of ventricular arrhythmia remains complex in patients with ischemic and nonischemic cardiomyopathy. The purpose of this study was to determine whether scar heterogeneity, quantified by mean entropy, predicts appropriate implantable cardioverter-defibrillator (ICD) therapy. We hypothesized that higher mean entropy calculated from cardiac magnetic resonance texture analysis (CMR-TA) will predict appropriate ICD therapy. Consecutive patients underwent CMR imaging before ICD implantation. Short-axis left ventricular scar was manually segmented. CMR-TA was performed using a Laplacian filter to extract and augment image features to create a scar texture from which histogram analysis of pixel intensity was used to calculate mean entropy. The primary end point was appropriate ICD therapy. A total of 114 patients underwent CMR-TA (ischemic cardiomyopathy [ICM]: n = 70; nonischemic cardiomyopathy [NICM]: n = 44) with a median follow-up of 955 days (interquartile range 691-1185 days). Mean entropy was significantly higher in the ICM group (5.7 ± 0.7 vs 5.5 ± 0.7; P= .045). Overall, 33 patients received appropriate ICD therapy. Using optimized cutoff values from receiver operating characteristic curves, Kaplan-Meier survival analysis demonstrated time until first appropriate therapy was significantly shorter in the high mean entropy group (P = .003). Multivariable analysis showed that mean entropy was the sole predictor of appropriate ICD therapy (hazard ratio 1.882; 95% confidence interval 1.083-3.271; P = .025). In the ICM group, mean entropy remained an independent predictor of appropriate ICD therapy, whereas in the NICM group, precontrast T1 values were the sole predictor. Scar heterogeneity, quantified by mean entropy using CMR-TA, was an independent predictor of appropriate ICD therapy in the mixed cardiomyopathy cohort and ICM-only group, suggesting a potential role for CMR-TA in predicting ventricular arrhythmia and risk-stratifying patients for ICD implantation.

Magnetic resonance texture analysis utility in differentiating intraparenchymal neurosarcoidosis from primary central nervous system lymphoma: a preliminary analysis

Neurosarcoidosis and primary central nervous system lymphomas, although distinct disease entities, can both have overlapping neuroimaging findings. The purpose of our preliminary study was to assess if magnetic resonance texture analysis can differentiate parenchymal mass-like neurosarcoidosis granulomas from primary central nervous system lymphomas. A total of nine patients was evaluated, four with parenchymal neurosarcoidosis granulomas and five with primary central nervous system lymphomas. Magnetic resonance texture analysis was performed with commercial software using a filtration histogram technique. Texture features of different sizes and variations in signal intensity were extracted at six different spatial scale filters, followed by feature quantification using statistical and histogram parameters and 36 features were analysed for each sequence (T1-weighted, T2-weighted, fluid-attenuated inversion recovery, diffusion-weighted, apparent diffusion coefficient, T1-post contrast). The non-parametric Mann-Whitney test was used to evaluate the differences between different texture parameters. The differences in distribution of entropy on T2-weighted imaging, apparent diffusion coefficient and T1-weighted post-contrast images were statistically significant on all spatial scale filters. Magnetic resonance texture analysis using medium and coarse spatial scale filters was especially useful in discriminating neurosarcoidosis from primary central nervous system lymphomas for mean, mean positive pixels, kurtosis, and skewness on diffusion-weighted imaging ( P < 0.004-0.030). At spatial scale filter 5, entropy on T2-weighted imaging ( P = 0.001) was the most useful discriminator with a cut-off value of 6.12 ( P = 0.001, area under the curve (AUC)-1, sensitivity (Sn)-100%, specificity (Sp)-100%), followed by kurtosis and skewness on diffusion-weighted imaging with a cut-off value of -0.565 ( P = 0.011, AUC-0.97, Sn-100%, Sp-83%) and-0.365 ( P = 0.008, AUC-0.98, Sn-100%, Sp-100%) respectively. Filtration histogram-based magnetic resonance texture analysis appears to be a promising modality to distinguish parenchymal neurosarcoidosis granulomas from primary central nervous system lymphomas.

Filtration-histogram based magnetic resonance texture analysis (MRTA) for glioma IDH and 1p19q genotyping

Background: To determine if filtration-histogram based texture analysis (MRTA) of clinical MR imaging can non-invasively identify molecular subtypes of untreated gliomas.Methods: Post Gadolinium T1-weighted (T1+Gad) images, T2-weighted (T2) images and apparent diffusion coefficient (ADC) maps of 97 gliomas (54 = WHO II, 20 = WHO III, 23 = WHO IV) between 2010 and 2016 were studied. Whole-tumor segmentations were performed on a proprietary texture analysis research platform (TexRAD, Cambridge, UK) using the software’s freehand drawing tool. MRTA commences with a filtration step, followed by quantification of texture using histogram texture parameters. Results were correlated using non-parametric statistics with a logistic regression model generated.Results: T1+Gad performed best for IDH typing of glioblastoma (sensitivity 91.9%, specificity 100%, AUC 0.945) and ADC for non-Gadolinium-enhancing gliomas (sensitivity 85.7%, specificity 78.4%, AUC 0.877). T2 was moderately precise (sensitivity 83.1%, specificity 78.9%, AUC 0.821). Excellent results for IDH typing were achieved from a combination of the three sequences (sensitivity 90.5%, specificity 94.5%, AUC = 0.98). For discriminating 1p19q genotypes, ADC produced the best results using unfiltered textures (sensitivity 80.6%, specificity 89.3%, AUC 0.811).Conclusion: Preoperative glioma genotyping with MRTA appears valuable with potential for clinical translation. The optimal choice of texture parameters is influenced by sequence choice, tumour morphology and segmentation method.

Magnetic Resonance Texture Analysis in Identifying Complete Pathological Response to Neoadjuvant Treatment in Locally Advanced Rectal Cancer.

A certain proportion of patients with locally advanced rectal cancer experience complete response after undergoing neoadjuvant chemoradiotherapy. These patients might be suitable for a conservative "watch and wait" approach, avoiding high-morbidity surgery. Texture analysis is a new modality that can assess heterogeneity in medical images by statistically analyzing gray-level intensities on a pixel-by-pixel basis. This study hypothesizes that texture analysis of magnetic resonance images can identify patients with a complete response. This study aims to determine whether texture analysis of magnetic resonance images as a quantitative imaging biomarker can accurately identify patients with complete response. This is a retrospective diagnostic accuracy study. This study was conducted at Colchester General Hospital, January 2003 to 2014. All patients diagnosed with locally advanced rectal cancer who underwent long-course chemoradiotherapy had a posttreatment magnetic resonance scan and underwent surgery are included. Texture analysis was extracted from T2-weighted magnetic resonance images of the rectal cancer. Textural features that are able to identify complete responders were identified by a Mann-Whitney U test. Their diagnostic accuracy in identifying complete responders was determined by the area under the receiver operator characteristics curve. Cutoff values were determined by the Youden index. Pathology was the standard of reference. One hundred fourteen patients with first posttreatment MRI scans (6.2 weeks after completion of neoadjuvant treatment) were included. Sixty-eight patients had a second posttreatment scan (10.4 weeks). With no filtration, mean (p = 0.033), SD (p = 0.048), entropy (p = 0.007), and skewness (p = 0.000) from first posttreatment scans, and SD (p = 0.042), entropy (p = 0.014), mean of positive pixels (p = 0.032), and skewness (p = 0.000) from second posttreatment scans were all able to identify complete response. Area under the curve ranged from 0.750 to 0.88. Texture analysis of MRI is a new modality; therefore, further studies are necessary to standardize the methodology of extraction of texture features, timing of scans, and acquisition parameters. Texture analysis of MRI is a potentially significant imaging biomarker that can accurately identify patients who have experienced complete response and might be suitable for a nonsurgical approach. (Cinicaltrials.gov:NCT02439086). See Video Abstract at http://links.lww.com/DCR/A760.

Quantitative assessment of myocardial scar heterogeneity using cardiovascular magnetic resonance texture analysis to risk stratify patients post-myocardial infarction

To determine whether heterogeneity of cardiac scar, as assessed by cardiovascular magnetic resonance (CMR) texture analysis, may provide insight into better risk stratification for patients with previous myocardial infarction (MI). Patients with previous MI (n=76) were followed for a median of 371.5 days after late gadolinium enhancement (LGE) CMR. The primary endpoint was a composite of ventricular tachycardia, ventricular fibrillation, or unexplained syncope. Areas of LGE were identified and manually segmented on a short-axis projection. The characteristics of the scar heterogeneity were evaluated via CMR texture analysis. This is a filtration-histogram technique, where images are filtered using the Laplacian of a Gaussian filter to extract features different sizes (2-6 mm in radius) corresponding to fine, medium, and coarse texture scales followed by a quantification step using histogram analysis (skewness and kurtosis). Patients suffering arrhythmic events during the follow-up period demonstrated significantly higher kurtosis (coarse-scale, p=0.005) and lower skewness (fine-scale, p=0.046) compared to those suffering no arrhythmic events. Furthermore, Kaplan-Meier analysis showed significantly higher coarse kurtosis (p=0.004), and lower fine skewness (p=0.035) were able to predict increased incidence of ventricular arrhythmic events. In this pilot study, indices of texture analysis reflecting textural heterogeneity were significantly associated with a greater incidence of arrhythmic events. Further work is required to delineate the role of texture analysis techniques in risk stratification post-MI.

Multicentre magnetic resonance texture analysis trial using reticulated foam test objects

Texture analysis in magnetic resonance imaging has the ability to provide useful diagnostic information with respect to the discrimination of disease states of a single tissue or the separation of different tissues. However, for widespread use it is necessary to determine how texture measurements carried out in one center relate to those carried out in another. To this end, a multicentre trial has been performed where reticulated foam test objects have been scanned in six European centers according to a fixed protocol. It has been concluded that texture measurements are not transportable between centers. Principal component models calculated from the texture parameters collected in one center do not fit the data collected in another. Further trials are to investigate whether the reticulated foam test objects may be used to normalize tissue texture data collected in different centers.