OBJECTIVEWe aimed to determine the corneal confocal microscopy (CCM) parameter that best identifies diabetic sensorimotor polyneuropathy (DSP) in type 1 diabetes and to describe its performance characteristics.RESEARCH DESIGN AND METHODSConcurrent with clinical and electrophysiological examination for classification of DSP, CCM was performed on 89 type 1 diabetic and 64 healthy subjects to determine corneal nerve fiber length (CNFL), density, tortuosity, and branch density. Area under the curve (AUC) and optimal thresholds for DSP identification in those with diabetes were determined by receiver operating characteristic (ROC) curve analysis.RESULTSDSP was present in 33 (37%) subjects. With the exception of tortuosity, CCM parameters were significantly lower in DSP case subjects. In ROC curve analysis, AUC was greatest for CNFL (0.88) compared with fiber density (0.84, P = 0.0001), branch density (0.73, P < 0.0001), and tortuosity (0.55, P < 0.0001). The threshold value that optimized sensitivity and specificity for ruling in DSP was a CNFL of ≤14.0 mm/mm2 (sensitivity 85%, specificity 84%), associated with positive and negative likelihood ratios of 5.3 and 0.18. An alternate approach that used separate threshold values maximized sensitivity (threshold value ≥15.8 mm/mm2, sensitivity 91%, negative likelihood ratio 0.16) and specificity (≤11.5 mm/mm2, specificity 93%, positive likelihood ratio 8.5).CONCLUSIONSAmong CCM parameters, CNFL best discriminated DSP cases from control subjects. A single threshold offers clinically acceptable operating characteristics, although a strategy that uses separate thresholds to respectively rule in and rule out DSP has excellent performance while minimizing unclassified subjects. We hypothesize that values between these thresholds indicate incipient nerve injury that represents those individuals at future neuropathy risk.

Development of corneal confocal microscopy may need to focus on the measurement of corneal nerve fibre length, as it appears to have superior reliability in comparison with other parameters, and as evidence exists for its potential as a clinical biomarker of early diabetic sensorimotor polyneuropathy.

No abstract

Little consideration has been given to the effect of different segmentation methods on the variability of data derived from microarray images. Previous work has suggested that the significant source of variability from microarray image analysis is from estimation of local background. In this study, we used Analysis of Variance (ANOVA) models to investigate the effect of methods of segmentation on the precision of measurements obtained from replicate microarray experiments. We used four different methods of spot segmentation (adaptive, fixed circle, histogram and GenePix) to analyse a total number of 156 172 spots from 12 microarray experiments. Using a two-way ANOVA model and the coefficient of repeatability, we show that the method of segmentation significantly affects the precision of the microarray data. The histogram method gave the lowest variability across replicate spots compared to other methods, and had the lowest pixel-to-pixel variability within spots. This effect on precision was independent of background subtraction. We show that these findings have direct, practical implications as the variability in precision between the four methods resulted in different numbers of genes being identified as differentially expressed. Segmentation method is an important source of variability in microarray data that directly affects precision and the identification of differentially expressed genes.