Understanding the learned behavior of customized convolutional neural networks toward malaria parasite detection in thin blood smear images.

Convolutional neural networks (CNNs) have become the architecture of choice for visual recognition tasks. However, these models are perceived as black boxes since there is a lack of understanding of the learned behavior from the underlying task of interest. This lack of transparency is a serious drawback, particularly in applications involving medical screening and diagnosis since poorly understood model behavior could adversely impact subsequent clinical decision-making. Recently, researchers have begun working on this issue and several methods have been proposed to visualize and understand the behavior of these models. We highlight the advantages offered through visualizing and understanding the weights, saliencies, class activation maps, and region of interest localizations in customized CNNs applied to the challenge of classifying parasitized and uninfected cells to aid in malaria screening. We provide an explanation for the models' classification decisions. We characterize, evaluate, and statistically validate the performance of different customized CNNs keeping every training subject's data separate from the validation set.

Variations in cyclic fatigue resistance among ProTaper Gold, ProTaper Next and ProTaper Universal instruments at different levels.

AIM: To compare the cyclic fatigue resistance of ProTaper Gold (PTG, Dentsply Tulsa Dental Specialities, Tulsa, OK, USA), ProTaper Next (PTN, Dentsply Tulsa Dental Specialities) and ProTaper Universal (PTU, Dentsply Tulsa Dental Specialities) instruments at different levels. METHODOLOGY: A total of 72 files were used to evaluate the cyclic fatigue of PTU (F2), PTN (X2) and PTG (F2) at 5 mm (n = 12) and 8 mm (n = 12) from the tip in 3-mm-radius steel canals with a 60° angle of curvature. The time to fracture was recorded. The Kolmogorov-Smirnov test was used to assess the normality of the samples distribution, and the statistical analysis was performed using the independent sample t-test (P < 0.01). RESULTS: Significant differences were found amongst the instruments 5 mm from the tip (P < 0.01). The PTG files had the highest CF resistance, and the PTN files displayed greater CF resistance than the PTU files. No significant differences were found between the PTG and PTN files 8 mm from the tip (P > 0.01). The PTG and PTN files demonstrated greater CF resistance than the PTU files (P < 0.01). CONCLUSIONS: The PTG instruments were the most resistant 5 and 8 mm from the tip; however, at 8 mm, there was no difference between the PTG and PTN instruments. The PTU files had the lowest CF resistance at all levels.

The effects of good glycaemic control on left ventricular and coronary endothelial functions in patients with poorly controlled Type 2 diabetes mellitus.

OBJECTIVE: Diabetics are at risk for developing overt heart failure and subclinical left ventricular (LV) dysfunction. Also, impaired coronary flow reserve (CFR) reflecting coronary microvascular dysfunction is common in diabetics. However, no substantial data regarding the effects of good glycaemic control on subclinical LV dysfunction and CFR are available. CONTEXT: To investigate whether good glycaemic control had favourable effects on subclinical LV dysfunction and CFR. DESIGN: Prospective, open-label, follow-up study. PATIENTS: Diabetics (n = 202) were classified based on baseline HbA1C levels: patients with good (group 1) (<7·0%) and poor glycaemic control (≥7·0%). MEASUREMENTS: All patients underwent echocardiographic examination at baseline evaluation, and it was repeated at months 6 and 12. Based on HbA1C levels obtained at month 6, the patients with poor glycaemic control were divided into two groups: achieved (group 2) and not achieved good glycaemic control (group 3). RESULTS: The groups were comparable with respect to diastolic function parameters including left atrium diameter, mitral E/A, Sm , Em /Am , E/E' and Tei index, and these parameters did not significantly change at follow-up in the groups. At baseline, CFR was slightly higher in group 1 than in group 2 and group 3, but it did not reach statistically significant level. At follow-up, CFR remained unchanged in group 1 (P = 0·58) and group 3 (P = 0·86), but increased in group 2 (P = 0·02: month 6 vs baseline and P = 0·004: month 12 vs baseline). CONCLUSIONS: Diabetics with poor and good glycaemic control were comparable with respect to echocardiographic parameters reflecting subclinical LV dysfunction, and good glycaemic control did not affect these parameters. However, good glycaemic control improved CFR.

Impaired heart rate recovery in patients with endemic fluorosis.

The objective of the present study was to determine the heart rate recovery index (HRRI), a marker of autonomic nervous system function in patients with endemic fluorosis. Forty patients with endemic fluorosis (16 men/24 women) and 40 age-, sex-, and body mass index-matched healthy controls (16 men/24 women) with normal fluoride intake were enrolled in this study. HRRI was calculated by subtracting the heart rate values at the first, second, and third minutes of the recovery phase from the peak heart rate (HRRI 1, HRRI 2, HRRI 3). Urine fluoride levels of fluorosis patients were significantly (P < 0.001) higher than control subjects as expected. HRRI 2 was significantly lower in fluorosis patients than in the controls. The incidence of abnormal HRRI 1 was significantly higher in fluorosis patients than in the controls (P < 0.05). We observed that HRRI, a marker of autonomic nervous system function, is impaired in patients with chronic fluorosis.

Weight of apically extruded debris following use of two canal instrumentation techniques and two designs of irrigation needles.

AIM: To compare the weight of apically extruded debris associated with canal shaping using K-files, and the Revo-S system using two different needle tips, open-ended and side-vented. METHODOLOGY: Forty freshly extracted mandibular premolar teeth were divided into four groups as follows; Group 1a: Hand instrumentation of root canal with open-ended needle used to provide irrigation, Group 1b: Hand instrumentation with a two side-port needle, Group 2a: Revo-S preparation with open-ended needle, Group 2b: Revo-S preparation with a two side-port needle. One Eppendorf tube for each tooth was weighed with a 0.0001 electronic weighing machine before instrumentation procedures. During instrumentation, irrigation was performed after every instrument with 1 mL distilled water. On completion of the preparation, teeth were removed from the Eppendorf tubes. All of the tubes were incubated at 37 °C for 15 days, to evaporate the irrigant. After the incubation period, tubes were weighed again. The difference between pre-and post-debris weights was calculated, and statistical evaluation was performed using analysis of variance at 0.05 level of significance. RESULTS: No statistically significant difference was observed between K-files and the RevoS system. A significant difference was observed between needles (P < 0.05). Open-ended needles were associated with significantly more debris than side-vented needles (P < 0.05). CONCLUSIONS: Both hand and rotary canal instrumentation resulted in extrusion of debris beyond the apical foramen. No significant difference was observed between K-files and Revo-S system. Open-ended needles were associated with significantly more extruded debris than two side-port needles.

Increased plasma neopterin and hs-CRP levels in patients with endemic fluorosis.

Although fluoride induced inflammatory reactions have been shown in animals and in vitro humans, there are few studies about fluoride induced inflammatory reactions in human beings at clinical setting. We aimed to measure the plasma neopterin, a marker of activation of the monocyte/macrophage system, and high sensitivity C-reactive protein (hs-CRP) levels in patients with endemic fluorosis to investigate the possible role of inflammatory processes (monocyte/macrophage activity) in the underlying pathophysiology of fluoride toxicity at clinical level. Plasma neopterin and hs-CRP levels were determined in endemic fluorosis patients and control subjects. Plasma neopterin levels were significantly higher among patients with endemic fluorosis when compared with control group (2.40 ± 0.66 vs. 1.63 ± 0.27 ng/mL respectively; p < 0.001) and plasma hs-CRP levels were also significantly higher among patients with endemic fluorosis when compared with control group (2.41 ± 1.23 vs. 1.93 ± 0.64 mg/L respectively; p < 0.001). Plasma neopterin levels were positively correlated with urine fluoride levels (r = 0.67, p < 0.001) and serum hs-CRP levels were positively correlated with urine fluoride levels (r = 0.36, p < 0.001). We have found that plasma neopterin and hs-CRP levels are increased in patients with endemic fluorosis. We have concluded that inflammation play an important role in the pathophysiology of fluoride toxicity in patients with endemic fluorosis.

Impact of chronic fluorosis on left ventricular diastolic and global functions.

Sixty three patients with endemic fluorosis (36 males/27 females; mean age 33.9+/-8.6years) and 45 age, sex and body mass index (BMI) matched healthy controls (30 males/15 females; mean age 32.7+/-8.8years) were included in this study. Basic echocardiographic measurements, left ventricular diastolic parameters and left ventricular myocardial performance index (MPI) were measured. The left ventricular MPI was calculated as (isovolumic contraction time+isovolumic relaxation time)/aortic ejection time by Doppler. The urine fluoride levels of fluorosis patients were significantly higher than control subjects as expected (1.9+/-0.1mg/l vs 0.4+/-0.1mg/l respectively; P<0.001). Isovolumic relaxation time (IVRT) and deceleration time (DT) were significantly higher in fluorosis patients than in controls (for IVRT 106.9+/-15.6ms vs 96.7+/-12.2ms; P<0.001 and for DT 211.7+/-30.7ms vs 188.0+/-30.0ms; P<0.001, respectively). MPI was significantly higher in fluorosis patients than in controls (0.62+/-0.15ms vs 0.49+/-0.10ms; P<0.001, respectively). We have shown that chronic fluorosis patients had left ventricular diastolic and global dysfunctions.

Aortic elasticity is impaired in patients with endemic fluorosis.

Sixty-three patients with endemic fluorosis (36 males/27 females; mean age 33.9 +/- 8.6 years) and 45 age-, sex-, and body mass index-matched healthy controls (30 males/15 females; mean age 32.7 +/- 8.8 years) were included in this study. Aortic stiffness indices, aortic strain (AS), aortic distensibility (AD), and aortic strain index (ASI) were calculated from the aortic diameters measured by echocardiography and blood pressure obtained by sphygmomanometry. The urine fluoride levels of fluorosis patients were significantly higher than control subjects as expected (1.9 +/- 0.1 mg/l vs. 0.4 +/- 0.1 mg/l, respectively; P < 0.001). AS and AD were significantly lower in fluorosis patients than in the controls (for AS 5.3 +/- 3.6 vs. 8.0 +/- 3.4%; P < 0.001 and for AD 0.2 +/- 0.1 vs. 0.3 +/- 0.1 cm(2) dyn(-1) 10(-3); P < 0.001, respectively). In contrast, significantly higher ASI was observed in fluorosis patients than in the controls (3.4 +/- 0.6 vs. 3.0 +/- 0.4; P < 0.001, respectively). The results of our study demonstrate that elastic properties of ascending aorta are impaired in patients with endemic fluorosis.

Cell spreading analysis with directed edge profile-guided level set active contours.

Cell adhesion and spreading within the extracellular matrix (ECM) plays an important role in cell motility, cell growth and tissue organization. Measuring cell spreading dynamics enables the investigation of cell mechanosensitivity to external mechanical stimuli, such as substrate rigidity. A common approach to measure cell spreading dynamics is to take time lapse images and quantify cell size and perimeter as a function of time. In our experiments, differences in cell characteristics between different treatments are subtle and require accurate measurements of cell parameters across a large population of cells to ensure an adequate sample size for statistical hypothesis testing. This paper presents a new approach to estimate accurate cell boundaries with complex shapes by applying a modified geodesic active contour level set method that directly utilizes the halo effect typically seen in phase contrast microscopy. Contour evolution is guided by edge profiles in a perpendicular direction to ensure convergence to the correct cell boundary. The proposed approach is tested on bovine aortic endothelial cell images under different treatments, and demonstrates accurate segmentation for a wide range of cell sizes and shapes compared to manual ground truth.

CELL SEGMENTATION USING HESSIAN-BASED DETECTION AND CONTOUR EVOLUTION WITH DIRECTIONAL DERIVATIVES.

The large amount of data produced by biological live cell imaging studies of cell behavior requires accurate automated cell segmentation algorithms for rapid, unbiased and reproducible scientific analysis. This paper presents a new approach to obtain precise boundaries of cells with complex shapes using ridge measures for initial detection and a modified geodesic active contour for curve evolution that exploits the halo effect present in phase-contrast microscopy. The level set contour evolution is controlled by a novel spatially adaptive stopping function based on the intensity profile perpendicular to the evolving front. The proposed approach is tested on human cancer cell images from LSDCAS and achieves high accuracy even in complex environments.