Loren Cobb

In vivo measurement of lumbar spine configuration is useful for constructing quantitative biomechanical models. Positional magnetic resonance imaging (MRI) accommodates a larger range of movement in most joints than conventional MRI and does not require a supine position. However, this is achieved at the expense of image resolution and contrast. As a result, quantitative research using positional MRI has required long reconstruction times and is sensitive to incorrectly identifying the vertebral boundary due to low contrast between bone and surrounding tissue in the images. We present a semi-automated method used to obtain digitized reconstructions of lumbar vertebrae in any posture of interest. This method combines a high-resolution reference scan with a low-resolution postural scan to provide a detailed and accurate representation of the vertebrae in the posture of interest. Compared to a criterion standard, translational reconstruction error ranged from 0.7 to 1.6 mm and rotational reconstruction error ranged from 0.3 to 2.6°. Intraclass correlation coefficients indicated high interrater reliability for measurements within the imaging plane (ICC 0.97–0.99). Computational efficiency indicates that this method may be used to compile data sets large enough to account for population variance, and potentially expand the use of positional MRI as a quantitative biomechanics research tool.

Convergence of the ensemble Kalman filter in the limit for large ensembles to the Kalman filter is proved. In each step of the filter, convergence of the ensemble sample covariance follows from a weak law of large numbers for exchangeable random variables, the continuous mapping theorem gives convergence in probability of the ensemble members, and L ^p bounds on the ensemble then give L ^p convergence.

The effects of inhalation of radon/radon decay products at different total doses, dose rates and 'unattached' fractions were investigated in a life span study in rats. 1574 rats inhaled radon/radon decay products in a purpose-built recirculating exposure system that provided stable/reproducible exposure conditions. 501 were maintained as controls. Lung tumour incidences were significantly elevated in most exposed groups. The study power was insufficient to resolve the shape of the dose and dose rate response curves, but combination of this data with that from other studies demonstrated that for high cumulative exposures, the lifetime excess absolute risk increases with increasing exposure durations and for low cumulative exposures the opposite trend occurs. Exposure did not increase leukaemia incidences. A small number of non-lung tumour types including mammary fibroadenoma showed elevated incidences in some exposed groups, however not consistently across all exposure groups and showed no dose or dose rate relationship. Radon/radon decay product exposure caused excess lung tumours in rats along with limited non-lung effects. The results are consistent with the findings that at low cumulative exposures decreasing exposure concentrations or protracting the time over which the dose is delivered, reduces lung tumour risk. At higher levels, decreasing exposure concentrations or protracting exposure time increases lung tumour risk.

To assess objectively the sensitivity and specificity of low-field-strength (0.064 T) magnetic resonance (MR) imaging, a prospective blind study of 280 examinations was performed to compare low-field-strength MR imaging with computed tomography (CT) and with high-field-strength (1.5-T) MR imaging of the cranium. The sensitivity (defined as the true-positive rate) with high-field MR imaging was superior to that with low-field MR imaging and CT in helping detect overall abnormalities. Sensitivities were generally similar over a broad range of specific cranial central nervous system diseases. Low-field and high-field MR imaging were equivalent in the blind diagnoses of neoplasms and white matter disease, whereas low-field MR and CT were equivalent in the blind diagnoses of contusion, subdural and epidural hematoma, sinus disease, normality, and abnormality. The specificities with low-field MR imaging and CT were substantially better than those with high-field MR imaging.

Forty normal kidneys were selected from a series of 146 renal samples obtained at forensic autopsies of American Indians and whites. Glomeruli were counted and the percentage of globally sclerosed, obsolescent glomeruli was calculated for each case. The mean for sclerotic glomeruli was less than 3% in any age group younger than 56 years. This finding is in agreement with the results of similar studies, but substantially lower than the figurs quoted in textbooks. Analysis of the combination of our data with studies performedby other authors shows that the 90th percentile of glomerular sclerosis can be calculated for any age by subtracting 10 from half the patient''s age. No ethnic-or sex-related differences were seen in any age group.

This study reports the results of a critical comparison of five statistical methods for estimating the density of viable cells in a limiting dilution assay (LDA). Artificial data were generated using Monte Carlo simulation. The performance of each statistical method was examined with respect to the accuracy of its estimator and, most importantly, the accuracy of its associated estimated standard error (SE). The regression method was found to perform at a level that is unacceptable for scientific research, due primarily to gross underestimation of the SE. The maximum likelihood method exhibited the best overall performance. A corrected version of Taswell's weighted-mean method, which provides the best performance among all noniterative methods examined, is also presented.

Six different statistical methods for comparing limiting dilution assays were evaluated, using both real data and a power analysis of simulated data. Simulated data consisted of a series of 12 dilutions for two treatment groups with 24 cultures per dilution and 1,000 independent replications of each experiment. Data within each replication were generated by Monte Carlo simulation, based on a probability model of the experiment. Analyses of the simulated data revealed that the type I error rates for the six methods differed substantially, with only likelihood ratio and Taswell's weighted mean methods approximating the nominal 5% significance level. Of the six methods, likelihood ratio and Taswell's minimum Chi-square exhibited the best power (least probability of type II errors). Taswell's weighted mean test yielded acceptable type I and type II error rates, whereas the regression method was judged unacceptable for scientific work.