Classification of papulo-squamous skin diseases using image analysis.

INTRODUCTION: Papulo-squamous skin diseases are variable but are very close in their clinical features. They present with the same lesions, erythematous scaly lesions. Clinical evaluation of skin lesions is based on common sense and experience of the dermatologist to differentiate features of each disease. AIM: To evaluate a computer-based image analysis system as a helping tool for classification of commonly encountered diseases. MATERIALS AND METHODS: The study included 50 selected images from each of psoriasis, lichen planus, atopic dermatitis, seborrheic dermatitis, pityrasis rosea, and pitryasis rubra pilaris with a total of 300 images. The study comprised three main processes peformed on the 300 included images: segmentation, feature extraction followed by classification. RESULTS: Rough sets recorded the highest percentage of accuracy and sensitivity of segmentation for the six groups of diseases compared with the other three used techniques (topological derivative, K-means clustering, and watershed). Rule-based classifier using the concept of rough sets recorded the best percentage of classification (96.7%) for the six groups of diseases compared with the other six techniques of classification used: K-means clustering, fuzzy c-means clustering, classification and regression tree, rule-based classifier with discretization, and K-nearest neighbor technique. CONCLUSION: Rough sets approach proves its superiority for both the segmentation and the classification processes of papulo-squamous skin diseases compared with the other used segmentation and classification techniques.

Association between dialysis dose improvement and nutritional status among hemodialysis patients.

BACKGROUND: Several studies suggest an association between improved survival and better nutritional status. It has been suggested that there is a correlation between dose of dialysis and nutritional status. However, in spite of the current practice, there are conflicting reports regarding the relationship between dose of dialysis or malnutrition, and biochemical outcome. In this article, we will discuss the impact of dose of dialysis on nutritional status and biochemical outcome in hemodialysis patients. We will also mention the interrelationships of dialysis dose, malnutrition, and biochemical outcome with respect to these patients. METHODS: Data were processed on 134 dialysis patients (mean age 48.21 +/- 13.38, 69 male, 65 female) on 3-times-per-week dialysis regimens. The overall study period was 3 months from June 1, 2005 to August 31, 2005. The patients were divided into two groups: the baseline group and the intervention group. The data of the baseline group were collected in June, 2005 and the data of the intervention group were collected in August, 2005 after applying the intervention or a protocol for dialysis adequacy improvement. RESULTS: The statistical analysis demonstrated that there was a significant improvement in mean URR and Kt/V from the baseline to the intervention group. The intervention group had a considerably higher rate than the baseline group for all nutritional and biochemical outcome parameters. The study showed a strong positive correlation between nPCR and Kt/V (p = 0.0001) and also a strong positive correlation between serum albumin and Kt/V (p = 0.00001). No correlations were found between Kt/V and biochemical outcomes such as hemoglobin (p = 0.4922), calcium (p = 0.650), phosphate (p = 0.508), and phosphatase (p = 0.091). CONCLUSION: All the available evidence in hemodialysis patients confirms the close association between dialysis dose and biochemical outcome. A body of evidence also highlights the existence of relationship between malnutrition and outcome among these patients. Dose of dialysis and nutrition are considered to be interrelated.