Geometric distortion in clinical MRI systems - Part II: correction using a 3D phantom

In this paper, we present the correction of the geometric distortion measured in the clinical magnetic resonance imaging (MRI) systems reported in the preceding paper (Part 1) using a 3D method based on the phantom-mapped geometric distortion data. This method allows the correction to be made on phantom images acquired without or with the vendor correction applied. With the vendor's 2D correction applied, the method corrects for both the residual geometric distortion still present in the plane in which the correction method was applied (the axial plane) and the uncorrected geometric distortion along the axis non-nal to the plane. The evaluation of the effectiveness of the correction using this new method was carried out through analyzing the residual geometric distortion in the corrected phantom images. The results show that the new method can restore the distorted images in 3D nearly to perfection. For all the MRI systems investigated, the mean absolute deviations in the positions of the control points (along x-, y- and z-axes) measured on the corrected phantom images were all less than 0.2 mm. The maximum absolute deviations were all below similar to0.8 mm. As expected, the correction of the phantom images acquired with the vendor's correction applied in the axial plane performed equally well. Both the geometric distortion still present in the axial plane after applying the vendor's correction and the uncorrected distortion along the z-axis have all been restored. (C) 2004 Elsevier Inc. All rights reserved.

The relative benefits of endurance and strength training on the metabolic factors and muscle function of people with type 2 diabetes mellitus

Objective: To compare the effects of a 4-month strength training (ST) versus aerobic endurance training (ET) program on metabolic control, muscle strength, and cardiovascular endurance in subjects with type 2 diabetes mellitus (T2D). Design: Randomized controlled trial. Setting: Large public tertiary hospital. Participants: Twenty-two T21) participants (I I men, I I women; mean age +/- standard error, 56.2 +/- 1.1 y; diabetes duration, 8.8 +/- 3.5y) were randomized into a 4-month ST program and 17 T2D participants (9 men, 8 women; mean age, 57.9 +/- 1.4y; diabetes duration, 9.2 +/- 1.7y) into a 4-month ET program. Interventions: ST (up to 6 sets per muscle group per week) and ET (with an intensity of maximal oxygen consumption of 60% and a volume beginning at 15min and advancing to a maximum of 30min 3X/wk) for 4 months. Main Outcome Measures: Laboratory tests included determinations of blood glucose, glycosylated hemoglobin (Hb A(1c)), insulin, and lipid assays. Results: A significant decline in Hb A, was only observed in the ST group (8.3% +/- 1.7% to 7.1% +/- 0.2%, P=.001). Blood glucose (204 +/- 16mg/dL to 147 +/- 8mg/dL, P <.001) and insulin resistance (9.11 +/- 1.51 to 7.15 +/- 1.15, P=.04) improved significantly in the ST group, whereas no significant changes were observed in the ET group. Baseline levels of total cholesterol (207 +/- 8mg/dL to 184 +/- 7mg/dL, P <.001), low-density lipoprotein cholesterol (120 +/- 8mg/dL to 106 +/- 8mg/dL, P=.001), and triglyceride levels (229 +/- 25mg/dL to 150 +/- 15mg/dL, P=.001) were significantly reduced and high-density lipoprotein cholesterol (43 +/- 3mg/dL to 48 +/- 2mg/dL, P=.004) was significantly increased in the ST group; in contrast, no such changes were seen in the ET group. Conclusions: ST was more effective than ET in improving glycemic control. With the added advantage of an improved lipid profile, we conclude that ST may play an important role in the treatment of T2D.

An Incremental EM-based Learning Approach for On-Line Prediction of Hospital Resource Utilization

Objective: Inpatient length of stay (LOS) is an important measure of hospital activity, health care resource consumption, and patient acuity. This research work aims at developing an incremental expectation maximization (EM) based learning approach on mixture of experts (ME) system for on-line prediction of LOS. The use of a batchmode learning process in most existing artificial neural networks to predict LOS is unrealistic, as the data become available over time and their pattern change dynamically. In contrast, an on-line process is capable of providing an output whenever a new datum becomes available. This on-the-spot information is therefore more useful and practical for making decisions, especially when one deals with a tremendous amount of data. Methods and material: The proposed approach is illustrated using a real example of gastroenteritis LOS data. The data set was extracted from a retrospective cohort study on all infants born in 1995-1997 and their subsequent admissions for gastroenteritis. The total number of admissions in this data set was n = 692. Linked hospitalization records of the cohort were retrieved retrospectively to derive the outcome measure, patient demographics, and associated co-morbidities information. A comparative study of the incremental learning and the batch-mode learning algorithms is considered. The performances of the learning algorithms are compared based on the mean absolute difference (MAD) between the predictions and the actual LOS, and the proportion of predictions with MAD < 1 day (Prop(MAD < 1)). The significance of the comparison is assessed through a regression analysis. Results: The incremental learning algorithm provides better on-line prediction of LOS when the system has gained sufficient training from more examples (MAD = 1.77 days and Prop(MAD < 1) = 54.3%), compared to that using the batch-mode learning. The regression analysis indicates a significant decrease of MAD (p-value = 0.063) and a significant (p-value = 0.044) increase of Prop(MAD

Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America

Background There are substantial social inequalities in adult male mortality in many countries. Smoking is often more prevalent among men of lower social class, education, or income. The contribution of smoking to these social inequalities in mortality remains uncertain. Methods The contribution of smoking to adult mortality in a population can be estimated indirectly from disease-specific death rates in that population (using absolute lung cancer rates to indicate proportions due to smoking of mortality from certain other diseases). We applied these methods to 1996 death rates at ages 35-69 years in men in three different social strata in four countries, based on a total of 0.6 million deaths. The highest and lowest social strata were based on social class (professional vs unskilled manual) in England and Wales, neighbourhood income (top vs bottom quintile) in urban Canada, and completed years of education (more than vs less than 12 years) in the USA and Poland. Results In each country, there was about a two-fold difference between the highest and the lowest social strata in overall risks of dying among men aged 35-69 years (England and Wales 21% vs 43%, USA 20% vs 37%, Canada 21% vs 34%, Poland 26% vs 50%: four-country mean 22% vs 41%, four-country mean absolute difference 19%). More than half of this difference in mortality between the top and bottom social strata involved differences in risks of being killed at age 35-69 years by smoking (England and Wales 4% vs 19%, USA 4% vs 15%, Canada 6% vs 13%, Poland 5% vs 22%: four-country mean 5% vs 17%, four-country mean absolute difference 12%). Smoking-attributed mortality accounted for nearly half of total male mortality in the lowest social stratum of each country. Conclusion In these populations, most, but not all, of the substantial social inequalities in adult male mortality during the 1990s were due to the effects of smoking. Widespread cessation of smoking could eventually halve the absolute differences between these social strata in the risk of premature death.

Typing mineral deposits using their grades and tonnages in an artificial neural network

A test of the ability of a probabilistic neural network to classify deposits into types on the basis of deposit tonnage and average Cu, Mo, Ag, Au, Zn, and Pb grades is conducted. The purpose is to examine whether this type of system might serve as a basis for integrating geoscience information available in large mineral databases to classify sites by deposit type. Benefits of proper classification of many sites in large regions are relatively rapid identification of terranes permissive for deposit types and recognition of specific sites perhaps worthy of exploring further. Total tonnages and average grades of 1,137 well-explored deposits identified in published grade and tonnage models representing 13 deposit types were used to train and test the network. Tonnages were transformed by logarithms and grades by square roots to reduce effects of skewness. All values were scaled by subtracting the variable's mean and dividing by its standard deviation. Half of the deposits were selected randomly to be used in training the probabilistic neural network and the other half were used for independent testing. Tests were performed with a probabilistic neural network employing a Gaussian kernel and separate sigma weights for each class (type) and each variable (grade or tonnage). Deposit types were selected to challenge the neural network. For many types, tonnages or average grades are significantly different from other types, but individual deposits may plot in the grade and tonnage space of more than one type. Porphyry Cu, porphyry Cu-Au, and porphyry Cu-Mo types have similar tonnages and relatively small differences in grades. Redbed Cu deposits typically have tonnages that could be confused with porphyry Cu deposits, also contain Cu and, in some situations, Ag. Cyprus and kuroko massive sulfide types have about the same tonnages. Cu, Zn, Ag, and Au grades. Polymetallic vein, sedimentary exhalative Zn-Pb, and Zn-Pb skarn types contain many of the same metals. Sediment-hosted Au, Comstock Au-Ag, and low-sulfide Au-quartz vein types are principally Au deposits with differing amounts of Ag. Given the intent to test the neural network under the most difficult conditions, an overall 75% agreement between the experts and the neural network is considered excellent. Among the largestclassification errors are skarn Zn-Pb and Cyprus massive sulfide deposits classed by the neuralnetwork as kuroko massive sulfides—24 and 63% error respectively. Other large errors are the classification of 92% of porphyry Cu-Mo as porphyry Cu deposits. Most of the larger classification errors involve 25 or fewer training deposits, suggesting that some errors might be the result of small sample size. About 91% of the gold deposit types were classed properly and 98% of porphyry Cu deposits were classes as some type of porphyry Cu deposit. An experienced economic geologist would not make many of the classification errors that were made by the neural network because the geologic settings of deposits would be used to reduce errors. In a separate test, the probabilistic neural network correctly classed 93% of 336 deposits in eight deposit types when trained with presence or absence of 58 minerals and six generalized rock types. The overall success rate of the probabilistic neural network when trained on tonnage and average grades would probably be more than 90% with additional information on the presence of a few rock types.

Regression to the mean: what it is and how to deal with it

Background Regression to the mean (RTM) is a statistical phenomenon that can make natural variation in repeated data look like real change. It happens when unusually large or small measurements tend to be followed by measurements that are closer to the mean. Methods We give some examples of the phenomenon, and discuss methods to overcome it at the design and analysis stages of a study. Results The effect of RTM in a sample becomes more noticeable with increasing measurement error and when follow-up measurements are only examined on a sub-sample selected using a baseline value. Conclusions RTM is a ubiquitous phenomenon in repeated data and should always be considered as a possible cause of an observed change. Its effect can be alleviated through better study design and use of suitable statistical methods.