The determination of three subcutaneous adipose tissue compartments in non-insulin-dependent diabetes mellitus women with artificial neural networks and factor analysis.

The optical device LIPOMETER allows for non-invasive, quick, precise and safe determination of subcutaneous fat distribution, so-called subcutaneous adipose tissue topography (SAT-Top). In this paper, we show how the high-dimensional SAT-Top information of women with type-2 diabetes mellitus (non-insulin-dependent diabetes mellitus (NIDDM)) and a healthy control group can be analysed and represented in low-dimensional plots by applying factor analysis and special artificial neural networks. Three top-down sorted subcutaneous adipose tissue compartments are determined (upper trunk, lower trunk, legs). NIDDM women provide significantly higher upper trunk obesity and significantly lower leg obesity ('apple' type), as compared with their healthy control group. Further, we show that the results of the applied networks are very similar to the results of factor analysis.     

Relationships between bioelectric impedance and subcutaneous adipose tissue thickness measured by LIPOMETER and skinfold calipers in children

The aim of this study was to compare the relationships between bioelectrical impedance and thicknesses of adipose tissue measured by traditional skinfold caliper (double thickness) or a LIPOMETER device (single non-compressed thickness) in 9- to 12-year-old boys (n=52) and girls (n=44). In total, nine skinfolds (triceps, subscapular, biceps, iliac crest, supraspinale, abdominal, front thigh, medial calf, mid-axilla) were measured. Measurement for the thickness of subcutaneous adipose tissue layers (SAT-layers) by LIPOMETER were performed at 15 body sites (neck, triceps, biceps, upper back, front chest, lateral chest, upper abdomen, lower abdomen, lower back, hip, front thigh, lateral thigh, rear thigh, inner thigh, calf). Body bioelectrical impedance was measured with a multiple-frequency impedance device Multiscan-5000 (Bodystat, UK). Impedance at 50 kHz highly correlated with body mass (r=–0.47 in boys, r=–0.46 in girls, r=–0.47 in total group). The relationship with body height was significant only in girls (r=–0.42). Skinfold thicknesses measured by caliper did not correlate significantly with body impedance at 50 kHz. SAT-layers measured by LIPOMETER at triceps, front thigh, lateral thigh and rear thigh sites in boys and at the lateral thigh site in girls correlated significantly with body impedance measured at 50 kHz. Stepwise multiple regression analysis indicated that the iliac crest and front thigh skinfold thicknesses measured by caliper characterized only 5.7–12.0% of the impedance at 50 kHz in the total group (n=96). From the measured 15 SAT-layers, the most significant was the lateral thigh layer which characterized 20.0%, 11.9% and 13.6% of the impedance at 50 kHz in boys, girls and the total group, respectively. It was concluded that the influence of subcutaneous adipose tissue on body impedance is relatively low in children. However, SAT-layers have a slightly higher influence on body impedance than skinfold thicknesses measured by caliper. The sum of skinfolds or SAT-layers did not correlate significantly with body impedance in any group.     

Measurement of subcutaneous adipose tissue topography (SAT-Top) by means of a new optical device,LIPOMETER, and the evaluation of standard factor coefficients in healthy subjects

The quantification of obesity in respect to subcutaneous adipose tissue and fat distribution is a matter of interest. We recently reported on a new optical device, LIPOMETER, and its ability to measure the thickness of subcutaneous adipose tissue and its advantages compared with other methods. To describe the subcutaneous adipose tissue distribution of the human body in a precise, reproducible, and comparable manner, 15 well-defined body sites distributed from neck to calf on the right body side were used. This set of sites defines subcutaneous adipose tissue topography (SAT-Top). To visualize SAT-Top for subjects or groups, special SAT-Top plots were used. Subcutaneous adipose tissue distribution can be recognized easily with these techniques. SAT-Top of 590 healthy men and women was measured. Factor analysis was used to extract the essential information from these 590*15 intercorrelated single measurements and to provide standard factor coefficients for later applications. As an example of how to use the results of factor analysis, the strong SAT-Top deviation of women with clinically proven type-2 diabetes mellitus (NIDDM) from healthy controls is described. Am. J. Hum. Biol. 12:231–239, 2000. © 2000 Wiley-Liss, Inc.     

Impact of visceral adiposity measured by abdominal computed tomography on pulmonary function

Although an inverse relationship between abdominal adiposity and pulmonary function has been suggested, direct measurement of abdominal adipose tissue has rarely been attempted. Our object is to determine the impact of abdominal adiposity on pulmonary function by directly measuring abdominal adipose tissue with abdominal computed tomography (CT). In this cross-sectional study, we included never-smokers between the ages of 18 and 85 yr, who had undergone spirometry and abdominal adipose tissue analysis with CT scans during November 1, 2005 to October 31, 2009 as part of the comprehensive health examination. Among a total of 3,469 participants, 890 (25.7%) were male. The mean body mass index and waist circumference among males and females were 24.6 kg/m(2) and 87.8 cm and 23.0 kg/m(2) and 83.0 cm, respectively. Although total adipose tissue (TAT) of the abdomen in males (269.1 cm(2)) was similar to that in females (273.6 cm(2)), the ratio of visceral adipose tissue (VAT)/subcutaneous adipose tissue (SAT) was different; 0.99 in males and 0.50 in females. In males, TAT, SAT, and VAT were inversely associated with the absolute value of forced vital capacity (FVC), and TAT and VAT were inversely associated with forced expiratory volume in one second (FEV(1)). However, in females, TAT and VAT, but not SAT, were inversely associated with absolute FVC and FEV(1) values. In conclusion, the amount of abdominal adipose tissue directly measured using CT is inversely associated with lung function.     

Estimating percentage total body fat and determining subcutaneous adipose tissue distribution with a new noninvasive optical device LIPOMETER

A newly developed optical device was applied to measure the subcutaneous adipose tissue (SAT) thickness of 20 healthy women and 18 healthy men at specified body sites. These measurements were used to derive equations to estimate percentage total body fat (TBF%). Total body electrical conductivity (TOBEC) was employed as a reference method; caliper techniques and measurements of absorbances of infrared light in fat versus lean tissue were also compared. The LIPOMETER results show good agreement with TOBEC data (r = 0.96). The technique allows the precise determination of the distribution of SAT thickness at specified body sites. The method also permits the construction of profiles of SAT thicknesses, e.g., the profiles are significantly different between women and men. Based on the normal profiles of healthy subjects, patients with proven type-2 diabetes mellitus were also evaluated. The patients showed significantly different profiles. By linear discriminant analysis, classification functions were extracted with good predictive accuracy classification of subjects according to the presence or absence of type-2 diabetes mellitus. The data suggest that measurement of SAT thickness might aid in the diagnosis and/or classification of metabolic disorders. Am. J. Hum. Biol. 12:221–230, 2000. © 2000 Wiley-Liss, Inc.     

Epicardial adipose tissue and pericoronary fat thickness measured with 64-multidetector computed tomography: potential predictors of the severity of coronary artery disease

OBJECTIVE:

The aim of the present study was to investigate the relationship between pericoronary fat and the severity and extent of atherosclerosis, quantified using 64-multidetector computed tomography, in patients with suspected coronary artery disease.

METHODS:

The study population consisted of 131 patients who were clinically referred for noninvasive multislice computed tomography coronary angiography for the evaluation of coronary artery disease. Patients were classified as follows: no atherosclerosis, Group 1; nonobstructive atherosclerosis (luminal narrowing <50% in diameter), Group 2; and obstructive atherosclerosis (luminal narrowing ≥50%) in a single vessel or obstructive atherosclerosis in the left main coronary artery and/or multiple vessels, Group 3. Epicardial adipose tissue was defined as the adipose tissue between the surface of the heart and the visceral layer of the pericardium (visceral epicardium). Epicardial adipose tissue thickness (mm) was determined in the right ventricular anterior free wall. The mean thickness of the pericoronary fat surrounding the three coronary arteries was used for the analyses.

RESULTS:

The average thickness over all three regions was 13.2 ± 2.1 mm. The pericoronary fat thickness was significantly increased in Group 3 compared with Groups 2 and 1. The epicardial adipose tissue thickness was significantly increased in Group 3 compared with Groups 2 and 1. A receiver operating characteristic curve for obstructive coronary artery disease was assessed to verify the optimum cut-off point for pericoronary fat thickness, which was 13.8 mm. A receiver operating characteristic curve for obstructive coronary artery disease was also assessed to verify the optimum cut-off point for epicardial adipose tissue, which was 6.8 cm.

CONCLUSION:

We showed that the epicardial adipose tissue and pericoronary fat thickness scores were higher in patients with obstructive coronary artery diseases.     

Blood supply of subcutaneous tissue in the leg and its clinical application

The vascular anatomy of the subcutaneous or adipose layer of the leg is described. The major arteries of the distal thigh and leg were injected with either a colored latex or an India ink and gelatin mixture to demonstrate the principal sources of blood supply to the skin and the superficial tissues of the leg. Microdissection and the techinique of Spalteholz for dehydrating and clearing the tissues (rendering them transparent) facilitated the study of the direct and indirect arterial branches which form the three principal networks: the deep fascia, the subcutaneous or adipose layer, and the skin or the dermoepidermic layer. The results are of practical importance since the fascio-subcutaneous tissue layer can be used as a flap for covering soft tissue defects of the leg. © 1995 WiIey-Liss, Inc.     

Structural remodeling of the lumbar multifidus,thoracolumbar fascia and lateral abdominal wall perimuscular connective tissues: A search for its potential determinants

Recently remodeling of lumbar soft tissues has received increased research attention. However, the major determinants that influence remodeling need to be elucidated in order to understand the impact of different rehabilitation modalities on tissue remodeling. The main aim of this study was to explore the between-subject variance of different measures of lumbar soft tissues quantified with rehabilitative ultrasound imaging (RUSI). RUSI measures (n = 8) were collected from 30 subjects without and 34 patients with LBP: (1) lumbar multifidus (LM) echogenicity (fatty infiltration/fibrosis) at three vertebral levels (L3/L4, L4/L5 and L5/S1) (n = 3); (2) posterior layer thickness of the thoracolumbar fascia (n = 1); and (3) thickness of the fasciae surrounding the external oblique (EO), internal oblique (IO), and transversus abdominis (TrA) (n = 4). Forward stepwise multivariate regression modeling was conducted with these RUSI measures as dependent variables, using the following independent variables as potential determinants: age, sex, the presence of LBP, body size/composition characteristics (height, weight, trunk length, subcutaneous tissue thickness over the abdominal, and LM muscles), trunk muscle function (or activation) as determined with the percent thickness change of LM, EO, IO, and TrA muscles during a standardized effort (RUSI measures), and physical activity level during sport and leisure activities as estimated with a self-report questionnaire. Two or three statistically significant predictors (or determinants) were selected in the regression model of each RUSI measure (n = 8 models), accounting for 26–64% of their total variance. The subcutaneous tissue thickness on the back accounted for 15–30% variance of LM echogenicity measures and thoracolumbar fascia thickness while the subcutaneous tissue thickness over the abdominals accounted for up to 42% variance of the fascia separating the subcutaneous adipose tissues and the EO muscle. The thickness of IO at rest accounted for 13–21% variance of all investigated abdominal fasciae except the fascia separating the subcutaneous adipose tissue and EO. Pain status accounted for 13–18% variance of the anterior and posterior fasciae of the TrA. Age accounted for 11–14% variance of LM echogenicity at all investigated vertebral levels while sex accounted for 15–21% variance of LM echogenicity at L3/L4 and fascia separating subcutaneous adipose tissue and EO muscle. The function (or activation) of EO and LM at L3/L4 accounted for 8–11% variance of the thoracolumbar fascia and fascia separating TrA and intra-abdominal content (TrA posterior fascia), respectively. Finally, the physical activity level during sport activities accounted for 7% variance of the fascia separating the subcutaneous adipose tissues and the EO muscle. These findings suggest that determinants other than body size characteristics may impact the remodeling of lumbar soft tissues, more importantly the subcutaneous adipose tissue deposits (thickness RUSI measures), which are associated with ectopic fat deposition in the LM and in the fasciae that are more closely positioned to the surface. While age, sex, and pain status explain some variability, modifiable factors such as physical activity level as well as trunk muscle thickness and function were involved. Overall, these results suggest that rehabilitation can potentially impact tissue remodeling, particularly in terms of intramuscular and perimuscular adipose tissues.     

Quantifying the 'appleness' or 'pearness' of the human body by subcutaneous adipose tissue distribution

To quantify subcutaneous adipose tissue topography (SAT-Top) describing individual SAT distribution for a subject or even a group we measured subcutaneous adipose tissue thickness at 15 specified body sites of 303 healthy women aged 20-69 yrs and 20 women with proven non-insulin-dependent diabetes mellitus (NIDDM) by the optical device 'LIPOMETER'. The type of upper-body- and lower-body-fat pattern, (apples or pears), was determined by factor analysis of the data. Upper body sites were highly loaded in factor 1, whereas factor 2 included highly loaded body sites from the lower extremities. For an individual, factor 1 scores > factor 2 scores, was described as an 'apple'-type, while factor 2 > factor 1 was described as a 'pear'-type. We found about 80% 'pears' and 20% 'apples' in 20-29 year olds and 20% 'pears' and 80% 'apples' in 60-69 year old women. Women with NIDDM tended to be 'super-apples'. SAT-Top provides a useful differentiation between apples and pears and we recommend this approach as a screening method.     

Subcutaneous Fat in the Upper Eyelids of Asians: Application to Blepharoplasty

The Asian eyelid has distinctive anatomical properties. Although the upper eyelid is reported in the literature to have no subcutaneous fat layer, we have often noticed adipose tissue below the upper eyelid skin during ophthalmic plastic surgeries. The aim of the present study was to examine the anatomical and histological features of subcutaneous adipose tissue in Asian upper eyelids using biopsies and microscopic morphological analysis. A total of 30 Chinese patients (60 upper eyelids) undergoing upper blepharoplasty were evaluated. The subcutaneous fat in the upper eyelids was surgically harvested and examined histologically. This study yielded two main findings. First, upper eyelid subcutaneous fat was found in 28 of the 30 patients. This fat layer was the inferior extension of the eyebrow subcutaneous fat layer, which descended and covered the superior half of the upper eyelid. Second, the lobules of subcutaneous fat in the upper eyelid were surrounded by dense fibrous tissues, and the fat cells were smaller than abdominal adipocytes but similar in size to orbital fat cells. Subcutaneous fat tissue is prominent in the upper eyelids of most Asians and is one important contributor to the bulky appearance of those eyelids. Clin. Anat. 33:338–342, 2020. © 2019 Wiley Periodicals, Inc.     

Some aspects of temperature regulation in newborn harp seal pups.

Harp seals are born on the drifting ice of the North Atlantic Ocean during arctic winter when temperatures of -20 degrees C, occasionally in combination with wind of 10 m/s, might prevail for days. At birth the pups lack subcutaneous blubber and the wet infantile fur has a conductance value of 30.0 W . m-2 . degrees C-1, as compared with only 2.0 W . m-2 . degrees C-1 when dry. While still wet immediately after birth the pups are nevertheless able to retain body core temperature by shivering. This activity leads to reduction of muscle fat and glycogen stores. Nonshivering thermogenesis commences in thermogenic adipose tissue by virtue of loosely coupled mitochondria. Thermogenic adipose tissue is found at birth both as a subcutaneous layer along the back and as internal deposits around venous plexuses in the neck, on the pericardium, on the kidneys, and the abdominal walls. After about 3 days of suckling the subcutaneous adipose tissue loses its thermogenic function being gradually transformed into blubber, whereas the internal deposits persist at least until the pups venture into water at the age of 3-4 wk.     

ROC and CART analysis of subcutaneous adipose tissue topography (SAT-Top) in type-2 diabetic women and healthy females

Women suffering from type-2 diabetes mellitus (NIDDM) show a more android fat pattern than healthy females, but to date no exact determination of their fat distribution differences exists. Measurements at 15 specified body sites with an optical device, the LIPOMETER, provide a subcutaneous adipose tissue topography (SAT-Top) of the individual. SAT-Top of 20 female NIDDM patients and 122 healthy controls was measured. ROC curve analysis was applied to evaluate the discriminative power of each body site and to provide cutoff values. Then a classification tree by the CART algorithm was established, showing SAT-Top differences between the two groups. Best discriminating results were achieved by the neck site (ROC area index = 0.76, sensitivity = 61.3%, specificity = 77.8%), the four sites of the thigh (area indices from 0.71 to 0.76), and a linear combination of all body sites stemming from a previous factor analysis, which provides condensed information of the extremities SAT-Top (area index = 0.80, sensitivity = 80.4%, specificity = 64.6%). The results could be improved by a summary measure of “android fat pattern” (area index = 0.89, sensitivity = 73.6%, specificity =88.3%) and a proportional measure of SAT-distribution, the relative neck (area index = 0.84, sensitivity = 83.0%, specificity = 70.5%). Overall, 136 (95.8%) of the 142 subjects were correctly classified by the classification tree (sensitivity = 75%, specificity = 99.2%). Both methods show the expected increased upper trunk obesity and decreased lower body obesity of NIDDM women compared with healthy females. Am. J. Hum. Biol. 12:388–394, 2000. © 2000 Wiley-Liss, Inc.     

Combination of Radiological and Gray Level Co-occurrence Matrix Textural Features Used to Distinguish Solitary Pulmonary Nodules by Computed Tomography

The objective of this study was to investigate the method of the combination of radiological and textural features for the differentiation of malignant from benign solitary pulmonary nodules by computed tomography. Features including 13 gray level co-occurrence matrix textural features and 12 radiological features were extracted from 2,117 CT slices, which came from 202 (116 malignant and 86 benign) patients. Lasso-type regularization to a nonlinear regression model was applied to select predictive features and a BP artificial neural network was used to build the diagnostic model. Eight radiological and two textural features were obtained after the Lasso-type regularization procedure. Twelve radiological features alone could reach an area under the ROC curve (AUC) of 0.84 in differentiating between malignant and benign lesions. The 10 selected characters improved the AUC to 0.91. The evaluation results showed that the method of selecting radiological and textural features appears to yield more effective in the distinction of malignant from benign solitary pulmonary nodules by computed tomography.     

Critical visceral adipose tissue thresholds aassociated with type 2 diabetes mellitus in chinese population

Objective: To compare intervertebral location L2-L3 with L4-L5 as landmarks for measuring abdominal fat distribution and to determine critical levels of visceral adipose tissue (VAT) at those planes, exceeding which may lead to the development of type 2 diabetes. Methods: Abdominal fat distribution was measured using computed tomography (CT) in 29 diabetics (19 male, 10 female) and 30 non-diabetics (18 male, 12 female). CT images obtained at two intervertebral locations L2-L3 and L4-L5 were used to measure the areas of total fat, VAT and subcutaneous adipose tissue (SCAT) using slice thickness of 5mm and an attenuation range from -190 to -30 Hounsfield units (HU). Data were analyzed using logistic regression and Receiver-operating characteristic (ROC) analysis. Results: At L2-L3, diabetes and obesity were correctly classified at 91.53% and 83.05% respectively, while at L4-L5, the same were correctly classified at 84.75% and 88.14% respectively. VAT compared to SCAT, had significantly higher correctly classified percent values for predicting diabetes in both measurement sites. At L2-L3, VAT≥177.29cm2 or VAT≥51.52% of the total fat area had the highest correctly classified value for predicting diabetes in men, while VAT≥132.27cm2 or VAT≥45.7% of the total fat area had the highest correctly classified value for predicting diabetes in women. At L4-L5, VAT≥130.82cm2 or VAT≥45.54% of the total fat area had the highest correctly classified value for predicting diabetes in men, while VAT≥118.56cm2 or VAT≥32.24% of the total fat area had the highest correctly classified value for predicting diabetes in women. Conclusion: L2-L3 plane is a better landmark for measuring abdominal fat distribution for predicting diabetes, while abdominal fat distribution measured at L4-L5 has better association with obesity. Regardless of the measurement site, VAT compared to SCAT, has significantly stronger association with diabetes.     

人类脂肪来源成体干细胞体外定向成软骨的诱导实验

目的 评价人类脂肪来源成体干细胞(hADASc)体外培养转基因定向成软骨诱导的可行性.方法 hADASc体外培养后行PGL3.转化生长因子β1(TGF-β1)基因转染并鉴定,检测转染细胞Lueiferase活性相对光读数(RLU),再行反转录聚合酶链式反应(RT-PCR)检测和抗Ⅱ型胶原免疫组织化学染色(实验组:转染的hADASc,同比对照组:加入软骨细胞定向诱导培养基的未转染hADASc;阴性对照组:未转染的hADASc),自动图像分析系统分析免疫染色图片并计算阳性颗粒(PU)值.结果 人类皮下脂肪体外分离培养出具有干细胞特征的有核组织细胞.hADASc转染后,实验组测得RLU值(9 212.583±315.240)高于阴性对照组(317.000±20.710,P＜0.01).RT-PCR结果显示转染的hADASc表达TGF-β1;免疫染色结果显示,实验组与同比对照组呈阳性反应,且两组PU值(13.864±2.416,13.637±2.548)均与阴性对照组的PU值差异有统计学意义(6.013±0.827,P＜0.05).结论 人类皮下脂肪体外分离培养出具有干细胞特征的有核组织细胞.PGL3-TGF-β1基因转染的hADASc可表达TGF-β1.内源性TGF-β1可诱导hADASc分泌Ⅱ型胶原蛋白,与外源性TGF-β1作用相类似.     

The LPD-II: a modified locked percutaneous device that permits safe subcutaneous access

A locked percutaneous device, the LPD-I (previously described as LPD), is effective in overcoming the problems of skin downgrowth and local tissue infection; however, it can only be implanted at a site providing adequate subcutaneous adipose tissue to attach to the subcutaneous connector of the LPD-I. A modified device, the LPD-II, has been developed that has a thin dome and skin connector but does not have a subcutaneous tissue connector. In addition, a newly designed structure called the skin stop collar (SSC) has been developed. It is positioned just beneath the mesh collar described in the LPD-I to further improve the function of the LPD-II. Six rabbits were implanted with one LPD-II without the SSC (group 1) and five rabbits were implanted with one LPD-II with the SSC (group 2). For more than 6 months, two of the implants in group 1 rabbits were successful. Four of the implants in group 2 rabbits were successful for more than 1 year. One of the animals in group 2 died of causes unrelated to the device. We conclude the following: the mesh collar skin connector can function well as a locked percutaneous device without the subcutaneous tissue connector; the LPD-II can be implanted in any site and does not require the presence of subcutaneous adipose tissue; and the SSC may increase the success rate of LPD-II implantation provided the mesh collar is made of soft material.     

Changes in anthropometric indices of body composition with age in a healthy elderly population

Changes in body composition are known to occur with old age, but there is limited knowledge of the normal values rates of change in anthropometric indices of body composition or the distribution of these changes in the elderly. In the present study, 98 elderly white men and 122 elderly white women were followed prospectively for 6 years. These participants were divided into 5-years cohorts on the basis of age at the start of the study in 1980. Separate linear regressions of triceps skinfold thickness, midarm circumference, midarm muscle area, and weight divided by stature squared on age were computed for each participant. The mean annual changes or slopes for these indices were small for each sex. A loss of subcutaneous adipose tissue was indicated with age, but the effects of compression or measurement error may hamper the detection of changes in indices of body composition in the elderly.     

Spatial frequency domain spectroscopy of two layer media

Monitoring of tissue blood volume and oxygen saturation using biomedical optics techniques has the potential to inform the assessment of tissue health, healing, and dysfunction. These quantities are typically estimated from the contribution of oxyhemoglobin and deoxyhemoglobin to the absorption spectrum of the dermis. However, estimation of blood related absorption in superficial tissue such as the skin can be confounded by the strong absorption of melanin in the epidermis. Furthermore, epidermal thickness and pigmentation varies with anatomic location, race, gender, and degree of disease progression. This study describes a technique for decoupling the effect of melanin absorption in the epidermis from blood absorption in the dermis for a large range of skin types and thicknesses. An artificial neural network was used to map input optical properties to spatial frequency domain diffuse reflectance of two layer media. Then, iterative fitting was used to determine the optical properties from simulated spatial frequency domain diffuse reflectance. Additionally, an artificial neural network was trained to directly map spatial frequency domain reflectance to sets of optical properties of a two layer medium, thus bypassing the need for iteration. In both cases, the optical thickness of the epidermis and absorption and reduced scattering coefficients of the dermis were determined independently. The accuracy and efficiency of the iterative fitting approach was compared with the direct neural network inversion.     

Differential diagnosis of CT focal liver lesions using texture features, feature selection and ensemble driven classifiers

OBJECTIVES: The aim of the present study is to define an optimally performing computer-aided diagnosis (CAD) architecture for the classification of liver tissue from non-enhanced computed tomography (CT) images into normal liver (C1), hepatic cyst (C2), hemangioma (C3), and hepatocellular carcinoma (C4). To this end, various CAD architectures, based on texture features and ensembles of classifiers (ECs), are comparatively assessed. MATERIALS AND METHODS: Number of regions of interests (ROIs) corresponding to C1-C4 have been defined by experienced radiologists in non-enhanced liver CT images. For each ROI, five distinct sets of texture features were extracted using first order statistics, spatial gray level dependence matrix, gray level difference method, Laws' texture energy measures, and fractal dimension measurements. Two different ECs were constructed and compared. The first one consists of five multilayer perceptron neural networks (NNs), each using as input one of the computed texture feature sets or its reduced version after genetic algorithm-based feature selection. The second EC comprised five different primary classifiers, namely one multilayer perceptron NN, one probabilistic NN, and three k-nearest neighbor classifiers, each fed with the combination of the five texture feature sets or their reduced versions. The final decision of each EC was extracted by using appropriate voting schemes, while bootstrap re-sampling was utilized in order to estimate the generalization ability of the CAD architectures based on the available relatively small-sized data set. RESULTS: The best mean classification accuracy (84.96%) is achieved by the second EC using a fused feature set, and the weighted voting scheme. The fused feature set was obtained after appropriate feature selection applied to specific subsets of the original feature set. CONCLUSIONS: The comparative assessment of the various CAD architectures shows that combining three types of classifiers with a voting scheme, fed with identical feature sets obtained after appropriate feature selection and fusion, may result in an accurate system able to assist differential diagnosis of focal liver lesions from non-enhanced CT images.