Multivariate prediction of multiple sclerosis using robust quantitative MR-based image metrics.

OBJECTIVES: The current work investigates the performance of different multivariate supervised machine learning models to predict the presence or absence of multiple sclerosis (MS) based on features derived from quantitative MRI acquisitions. The performance of these models was evaluated for images which are significantly degraded due to subject motion, a problem which is often observed in clinical routine diagnostics. Finally, the difference between a true multivariate analysis and the corresponding univariate analysis based on single parameters alone was addressed. MATERIALS AND METHODS: 52 MS patients and 45 healthy controls where scanned on a 3T system. The datasets showed variable degrees of motion-associated artefacts. For each dataset, the average of T1, T2*, total and myelin bound water content was determined in white and grey matter. Based on these parameters, different multivariate models were trained and their cross-validated performance to predict the presence of MS was evaluated. Furthermore, the univariate distributions of each quantitative parameter were employed to define optimised cut-offs that differentiate MS patients from healthy controls. RESULTS: For data not affected by motion, 83.7% of all subjects were correctly classified using a crossvalidated multivariate model. Inclusion of data with significant artefacts reduces the rate of correct classification to 74.5%. T1 in grey and myelin water content in white matter where the most discriminating variables in the multivariate analysis. In contrast, the total water content in white matter and the ratio of white and grey matter total water content each resulted in 77% correct classifications in a univariate regression analysis. CONCLUSION: The results demonstrate that even simple quantitative MRI-based measures allow for an automated prediction of the presence/absence of multiple sclerosis with good specificity. Importantly, even highly degraded datasets due to motion-artefacts could be correctly classified, especially when pooling features derived from grey and white matter. Finally, the advantage of a multivariate over a univariate analysis of quantitative MR data was shown.

Positive and negative plant interactions contribute to a north-south-patterned association between two desert shrub species.

Abiotic factors are often thought to be the predominant forces shaping desert plant communities. But both positive and negative interactions between plants are frequently observed in deserts, and it is an open question whether they can strongly affect the spatial structure of a desert community. The goal of this study was to answer this question for a plant community in the North American Mojave Desert. Two semi-shrub species, Ambrosia dumosa and Acamptopappus sphaerocephalus, were the focus of this study. At the study site, seedlings emerged predominantly on the northern side of shrubs, indicating positive effects of canopy shading on emergence, but survival of Ambrosia seedlings was much higher in open areas than at the edge of conspecific shrubs. Negative intraspecific interactions also affected Ambrosia shrubs, which did not increase in size over a 4-year period unless the nearest conspecific neighbor had been removed. These negative intraspecific interactions among different life stages of Ambrosia appear to contribute to spatial segregation observed among shrubs of this species. In contrast, Acamptopappus shrubs and their seedlings were aggregated with Ambrosia shrubs, and occurred more often on the northern side of Ambrosia than expected by chance. Removal of Ambrosia neighbors positively affected growth of Acamptopappus, but only when the neighbor was removed on the northern side. For Acamptopappus, an Ambrosia neighbor on the southern side may have some positive effects, which appear to neutralize the negative effects found for northern neighbors. These positive effects were likely at least partly due to shading. Removal of Ambrosia neighbors negatively affected predawn xylem pressure potentials of Acamptopappus, but this effect was only found during one growing season and was briefly reversed during the next. In summary, negative intraspecific interactions appear to cause spatial segregation of Ambrosia shrubs, while a combination of positive and negative interactions apparently contribute to the directional association between Ambrosia and Acamptopappus. Thus plant interactions in this desert appear to shape community structure in at least two dimensions by influencing the distances and in which directions to their neighbors plants can grow and survive.

Multicentre absolute myelin water content mapping: Development of a whole brain atlas and application to low-grade multiple sclerosis.

The current study investigates the whole brain myelin water content distribution applying a new approach that allows for the simultaneous mapping of total and relative myelin water content, T 1 and T 2* with full brain coverage and high resolution (1 × 1 × 2 mm(3)). The data was collected at two different sites in healthy controls to validate the independence of a specific setup. In addition, a group of patients with known white matter affections was investigated to compare two measures of myelin, i.e. relative and absolute myelin water content. Based on the first dataset, a quantitative myelin water content atlas was created which served as a control set for the other two datasets. Both control groups measured at different institutions yielded consistent results. However, distinct regions of reduced myelin water content were observed for the patient dataset, both on an individual basis and in a group-wise comparison. The comparison between the absolute and relative measurement of myelin water content in MS patients showed that the relative measurement, which is employed by many researchers, overestimates both disease volume and the corresponding reduction of myelin water content in white matter lesions. However, for normal appearing white matter, no difference between both approaches was detected. The results obtained in the current study demonstrate that absolute myelin water content can reliably be determined in a multicentre environment using standard MR sequences. The optimised protocol allows for a measurement of four quantitative parameters with full brain coverage in only 10 min. This might expedite a more widespread future use of quantitative MRI methods for clinical research and diagnosis.

Rapid myelin water content mapping on clinical MR systems.

We present an algorithm for the fast mapping of myelin water content using standard multiecho gradient echo acquisitions of the human brain. The method extents a previously published approach for the simultaneous measurement of brain T(1), T(2)(*) and total water content. Employing the multiexponential T(2)(*) decay signal of myelinated tissue, myelin water content was measured based on the quantification of two water pools ("myelin water" and "rest") with different relaxation times. As the existing protocol was focussed on the fast mapping of quantitative MR parameters with whole brain coverage in clinically relevant measurement times, the sampling density of the T(2)(*) curve was compromised to 10 echo times with a TE(max) of approx. 40ms. Therefore, pool amplitudes were determined using a quadratic optimisation approach. The optimisation was constrained by including à priori knowledge about brain water pools. All constraints were optimised in a simulation study to minimise systematic error sources given the incomplete knowledge about the real pool-specific relaxation properties. Based on the simulation results, whole brain in vivo myelin water content maps were acquired in 10 healthy controls and one subject with multiple sclerosis. The in vivo results obtained were consistent with previous reports which demonstrates that a simultaneous whole brain mapping of T(1), T(2)(*), total and myelin water content is feasible on almost any modern MR scanner in less than 10 minutes.

Adhesion, spreading, and aggregation of platelets in flowing blood and the reliability of the retention test Homburg.

Adhesion and aggregation are important parameters characterizing the function of intact platelets in flowing blood and in contact with a more or less thrombogenic surface. In the retention test Homburg (RTH), platelets are exposed to a standardized textured surface (Sysmex retention tubes) under defined conditions of flow. Platelet counts are performed before and after the Sysmex retention tube passage. The difference between these values indicates the percentage of retained platelets (retention index). Decreased retention in the RTH indicates a loss of function or defective platelet function; increase is associated with an increased activation of platelets, for example, in patients with vascular diseases. For further evaluation of the retention phenomenon the filters were fixed after cell passage and examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). SEM and TEM micrographs show activated platelets adhering and spreading on the filter surface, comparable with platelets on a disturbed endothelium. Also, we examined the influence of different G forces and centrifugation times on the retention behavior of the platelets in citrated platelet-rich plasma (PRP) and whole blood (WB). G forces influenced the retention index in PRP and WB significantly and in a different way. Finally, we used a platelet standard, as customary in the quality check, to determine the serial as well as the day-to-day precision.

Role of the retention test Homburg in evaluating platelet hyperactivity and in monitoring therapy with antiplatelet drugs.

Decreased retention in the retention test Homburg (RTH) indicates a loss of platelet function; increase is associated with an increased activation of platelets, for example, in patients with vascular diseases. Compared with other materials (e.g., collagen, glass pearls, etc.) the filter surface in the retention tubes is nonthrombogenic. Therefore, the RTH seems to be well suited for measuring an in vivo over-reactivity of platelets. In a pilot study using the RTH we evaluated the postoperative over-reactivity of platelets in 14 patients and observed a significant heterogeneity of the platelet population concerning size and stickiness. Reliable platelet function tests are also necessary for "drug monitoring," since they deliver important clinical laboratory parameters for efficient control of a therapy with antiplatelet drugs. Therefore, we evaluated in vitro how, after administration of platelet aggregation-inhibiting medications (such as acetylsalicylic acid, Prostaglandin and ReoPro in various concentrations, the adenosine diphosphate (ADP), collagen, ristocetin, or suprarenin increased retention can be reduced. The reaction of the platelets in platelet-rich plasma of different patients or donors to the addition of ADP is variable. The platelet function inhibitor effect is dose dependent. In a clinical pilot study, a significant platelet-inhibiting effect of clopidogrel using the RTH has been shown.