Encoder-Weighted W-Net for Unsupervised Segmentation of Cervix Region in Colposcopy Images.

Cervical cancer can be prevented and treated better if it is diagnosed early. Colposcopy, a way of clinically looking at the cervix region, is an efficient method for cervical cancer screening and its early detection. The cervix region segmentation significantly affects the performance of computer-aided diagnostics using a colposcopy, particularly cervical intraepithelial neoplasia (CIN) classification. However, there are few studies of cervix segmentation in colposcopy, and no studies of fully unsupervised cervix region detection without image pre- and post-processing. In this study, we propose a deep learning-based unsupervised method to identify cervix regions without pre- and post-processing. A new loss function and a novel scheduling scheme for the baseline W-Net are proposed for fully unsupervised cervix region segmentation in colposcopy. The experimental results showed that the proposed method achieved the best performance in the cervix segmentation with a Dice coefficient of 0.71 with less computational cost. The proposed method produced cervix segmentation masks with more reduction in outliers and can be applied before CIN detection or other diagnoses to improve diagnostic performance. Our results demonstrate that the proposed method not only assists medical specialists in diagnosis in practical situations but also shows the potential of an unsupervised segmentation approach in colposcopy.

Quantitative Screening of Cervical Cancers for Low-Resource Settings: Pilot Study of Smartphone-Based Endoscopic Visual Inspection After Acetic Acid Using Machine Learning Techniques.

BACKGROUND: Approximately 90% of global cervical cancer (CC) is mostly found in low- and middle-income countries. In most cases, CC can be detected early through routine screening programs, including a cytology-based test. However, it is logistically difficult to offer this program in low-resource settings due to limited resources and infrastructure, and few trained experts. A visual inspection following the application of acetic acid (VIA) has been widely promoted and is routinely recommended as a viable form of CC screening in resource-constrained countries. Digital images of the cervix have been acquired during VIA procedure with better quality assurance and visualization, leading to higher diagnostic accuracy and reduction of the variability of detection rate. However, a colposcope is bulky, expensive, electricity-dependent, and needs routine maintenance, and to confirm the grade of abnormality through its images, a specialist must be present. Recently, smartphone-based imaging systems have made a significant impact on the practice of medicine by offering a cost-effective, rapid, and noninvasive method of evaluation. Furthermore, computer-aided analyses, including image processing-based methods and machine learning techniques, have also shown great potential for a high impact on medicinal evaluations. OBJECTIVE: In this study, we demonstrate a new quantitative CC screening technique and implement a machine learning algorithm for smartphone-based endoscopic VIA. We also evaluated the diagnostic performance and practicability of the approach based on the results compared to the gold standard and from physicians' interpretation. METHODS: A smartphone-based endoscope system was developed and applied to the VIA screening. A total of 20 patients were recruited for this study to evaluate the system. Overall, five were healthy, and 15 were patients who had shown a low to high grade of cervical intraepithelial neoplasia (CIN) from both colposcopy and cytology tests. Endoscopic VIA images were obtained before a loop electrosurgical excision procedure for patients with abnormal tissues, and their histology tissues were collected. Endoscopic VIA images were assessed by four expert physicians relative to the gold standard of histopathology. Also, VIA features were extracted from multiple steps of image processing techniques to find the differences between abnormal (CIN2+) and normal (≤CIN1). By using the extracted features, the performance of different machine learning classifiers, such as k-nearest neighbors (KNN), support vector machine, and decision tree (DT), were compared to find the best algorithm for VIA. After determining the best performing classifying model, it was used to evaluate the screening performance of VIA. RESULTS: An average accuracy of 78%, with a Cohen kappa of 0.571, was observed for the evaluation of the system by four physicians. Through image processing, 240 sliced images were obtained from the cervicogram at each clock position, and five features of VIA were extracted. Among the three models, KNN showed the best performance for finding VIA within holdout 10-fold cross-validation, with an accuracy of 78.3%, area under the curve of 0.807, a specificity of 80.3%, and a sensitivity of 75.0%, respectively. The trained model performed using an unprovided data set resulted in an accuracy of 80.8%, specificity of 84.1%, and sensitivity of 71.9%. Predictions were visualized with intuitive color labels, indicating the normal/abnormal tissue using a circular clock-type segmentation. Calculating the overlapped abnormal tissues between the gold standard and predicted value, the KNN model overperformed the average assessments of physicians for finding VIA. CONCLUSIONS: We explored the potential of the smartphone-based endoscopic VIA as an evaluation technique and used the cervicogram to evaluate normal/abnormal tissue using machine learning techniques. The results of this study demonstrate its potential as a screening tool in low-resource settings.

Fast learning method for convolutional neural networks using extreme learning machine and its application to lane detection.

Deep learning has received significant attention recently as a promising solution to many problems in the area of artificial intelligence. Among several deep learning architectures, convolutional neural networks (CNNs) demonstrate superior performance when compared to other machine learning methods in the applications of object detection and recognition. We use a CNN for image enhancement and the detection of driving lanes on motorways. In general, the process of lane detection consists of edge extraction and line detection. A CNN can be used to enhance the input images before lane detection by excluding noise and obstacles that are irrelevant to the edge detection result. However, training conventional CNNs requires considerable computation and a big dataset. Therefore, we suggest a new learning algorithm for CNNs using an extreme learning machine (ELM). The ELM is a fast learning method used to calculate network weights between output and hidden layers in a single iteration and thus, can dramatically reduce learning time while producing accurate results with minimal training data. A conventional ELM can be applied to networks with a single hidden layer; as such, we propose a stacked ELM architecture in the CNN framework. Further, we modify the backpropagation algorithm to find the targets of hidden layers and effectively learn network weights while maintaining performance. Experimental results confirm that the proposed method is effective in reducing learning time and improving performance.

Metallothionein enhances angiogenesis and arteriogenesis by modulating smooth muscle cell and macrophage function.

OBJECTIVE: In a previous study we identified metallothionein (MT) as a candidate gene potentially influencing collaterogenesis. In this investigation, we determined the effect of MT on collaterogenesis and examined the mechanisms contributing to the effects we found. METHODS AND RESULTS: Collateral blood flow recovery was assessed using laser Doppler perfusion imaging, and angiogenesis was measured using a Matrigel plug assay. Smooth muscle cells were isolated from MT knockout (KO) mice for functional assays. Gene expression of matrix metalloproteinase-9, platelet-derived growth factor, vascular endothelial growth factor, and Fat cadherin in smooth muscle cells was measured by real-time polymerase chain reaction, and protein levels of vascular endothelial growth factor and matrix metalloproteinase-9 were determined using enzyme-linked immunosorbent assay and Western blot. CD11b(+) macrophages were tested for invasiveness using a real-time impedance assay. Both flow recovery and angiogenesis were impaired in MT KO mice. Proliferation, migration, and invasion were decreased in MT KO smooth muscle cells, and matrix metalloproteinase-9, platelet-derived growth factor, and vascular endothelial growth factor expression were also decreased, whereas FAT-1 cadherin expression was elevated. MT KO CD11b(+) cells were more invasive than wild-type cells. CONCLUSIONS: MT plays an important role in collateral flow recovery and angiogenesis, an activity that appears to be mediated, in part, by the effects of MT on the functionality of 3 cell types essential for these processes: endothelial cells, smooth muscle cells, and macrophages.

Patterns of glaucomatous visual field loss in sita fields automatically identified using independent component analysis.

PURPOSE: To determine if the patterns uncovered with variational Bayesian-independent component analysis-mixture model (VIM) applied to a large set of normal and glaucomatous fields obtained with the Swedish Interactive Thresholding Algorithm (SITA) are distinct, recognizable, and useful for modeling the severity of the field loss. METHODS: SITA fields were obtained with the Humphrey Visual Field Analyzer (Carl Zeiss Meditec, Inc, Dublin, California) on 1,146 normal eyes and 939 glaucoma eyes from subjects followed by the Diagnostic Innovations in Glaucoma Study and the African Descent and Glaucoma Evaluation Study. VIM modifies independent component analysis (ICA) to develop separate sets of ICA axes in the cluster of normal fields and the 2 clusters of abnormal fields. Of 360 models, the model with the best separation of normal and glaucomatous fields was chosen for creating the maximally independent axes. Grayscale displays of fields generated by VIM on each axis were compared. SITA fields most closely associated with each axis and displayed in grayscale were evaluated for consistency of pattern at all severities. RESULTS: The best VIM model had 3 clusters. Cluster 1 (1,193) was mostly normal (1,089, 95% specificity) and had 2 axes. Cluster 2 (596) contained mildly abnormal fields (513) and 2 axes; cluster 3 (323) held mostly moderately to severely abnormal fields (322) and 5 axes. Sensitivity for clusters 2 and 3 combined was 88.9%. The VIM-generated field patterns differed from each other and resembled glaucomatous defects (eg, nasal step, arcuate, temporal wedge). SITA fields assigned to an axis resembled each other and the VIM-generated patterns for that axis. Pattern severity increased in the positive direction of each axis by expansion or deepening of the axis pattern. CONCLUSIONS: VIM worked well on SITA fields, separating them into distinctly different yet recognizable patterns of glaucomatous field defects. The axis and pattern properties make VIM a good candidate as a preliminary process for detecting progression.

Interanimal variability in preexisting collaterals is a major factor determining outcome in experimental angiogenesis trials.

Despite numerous animal trials reporting that cell therapy promotes collateral flow, clinical trials have not convincingly shown benefit. Patient-related risk factors are often used to explain these discrepancies. However, during the course of our own angiogenesis studies using mice, we noted large anatomical variability in collateral vessels. The purpose of the present investigation was to define how important this factor might be in determining intervention outcomes. Hindlimb ischemia was induced in BALB/c mice by ligating the superficial femoral artery. After 24 h, animals were treated by injecting the adductor muscle with either control media or cultured mesenchymal stem cells (MSCs). Blood flow recovery was measured using laser-Doppler [laser-Doppler perfusion imaging (LDPI) ratio]. In a second experiment, mice were stratified 24 h after arterial ligation before treatment by using a simple clinical score of the ligated leg: 1, able to flex, mild discoloration; 2, no flexion, mild discoloration; 3, severe discoloration; and 4, any necrosis. Without stratification, blood flow recovery significantly increased in the MSC-treated group (P < 0.05, n = 6 MSC group, n = 7 media group). In the experiment employing stratification, all differences between the groups disappeared (n = 11 MSC group, n = 10 media group; P = 0.3). Furthermore, we found a striking inverse correlation between clinical score on day 1 and the LDPI ratio on day 28 (P < 0.0001; n = 79). Anatomical confirmation of the disparity in preexisting collaterals was found in two different mouse strains using microscopic computed tomography. In conclusion, there is substantial interanimal variability in preexisting collateral flow, and this variability can importantly influence outcome. To overcome this, either animals must be stratified before treatment, the number of animals must be increased substantially, or, preferably, both.

C2-ceramide induces vasodilation in phenylephrine-induced pre-contracted rat thoracic aorta: role of RhoA/Rho-kinase and intracellular Ca2+ concentration.

It is known that ceramide may play an important regulatory role in vascular tone although its effect on vascular tone and the mechanisms involved are controversial. The present study was designed to investigate the effects of ceramide and its key initial regulators, TNF-alpha and neutral sphingomyelinase (SMase), on vascular tone of isolated rat thoracic aortic rings and elucidate the mechanisms involved in the changes in vascular tone induced by ceramide. Contractile responses and Fura-2 Ca2+ signals were measured in rat thoracic aortic rings or strips. 10(-5) M C2-ceramide, 0.1 U/ml neutral sphingomyelinase (SMase), and 5x10(-7) g/ml TNF-alpha had no effect on resting tone in rat thoracic aortic rings. However, in phenylephrine-induced pre-contracted rings, treatment with ceramide, SMase, and TNF-alpha evoked a gradual but sustained vasodilation. Vasodilation effect in response to 10(-5) M C2-ceramide was not significantly changed by the absence or presence of endothelium, a cyclooxygenase pathway inhibitor (10(-6) M indomethacin), or PKC inhibitors (10(-5) M H-7 & 5x10(-7) M calphostin-C). Pretreatment with 1 microM Y-27632, a RhoA/Rho-kinase inhibitor, significantly inhibited the phenylephrine-induced contraction itself as well as the C2-ceramide-induced vasodilation. Pre-treatment with 10(-5) M C2-ceramide had no effect on phasic rise in [Ca2+]i and tension evoked by stimulation with 10(-8) M phenylephrine, but post-treatment of C2-ceramide significantly reduced the phenylephrine-induced secondary tonic [Ca2+]i and tension plateau. Our results indicate that C2-ceramide induces vasodilation in phenylephrine-induced pre-contracted rat thoracic aorta. Furthermore, inhibition of phenylephrine-induced activation of RhoA/Rho-kinase pathway as well as phenylephrine-induced elevations in [Ca2+]i are clearly a key factors in C2-ceramide-induced vasodilation.

Rapamycin attenuates atherosclerotic plaque progression in apolipoprotein E knockout mice: inhibitory effect on monocyte chemotaxis.

Rapamycin has been shown to reduce neointimal thickening in the setting of balloon angioplasty and chronic graft vessel disease. This study was designed to test the effect of oral rapamycin on atherosclerotic plaque progression and the possible mechanism involved. Apolipoprotein E (apoE) knockout mice were fed either a diet supplemented with cholesterol or with cholesterol and rapamycin. At 4 and 8 weeks, quantitative analyses of plaque area and macrophage numbers were determined. Plasma cholesterol, triglyceride, and whole-blood rapamycin levels were measured. Rapamycin could be detected in the blood of mice (117+/-7 pg/mL). In mice fed with rapamycin, atherosclerotic lesions covered 22% of the aortic arch as compared with 41% in cholesterol-fed mice. The macrophage count was significantly lower in the rapamycin-fed mice as compared with cholesterol-fed mice. Rapamycin, in a dose-dependent manner, inhibited monocyte chemotaxis elicited by stromal cell-derived factor-1. Lesions in the cholesterol-fed mice had complex atherosclerotic plaque with acellular core, cholesterol clefts, and an abundant collection of monocytes/macrophages. Lesions in the rapamycin-fed mice were mainly composed of monocytes/macrophages. Oral rapamycin is effective in slowing the progression of atherosclerosis. Along with its multitude actions, attenuation of monocyte chemotaxis may be one more way by which rapamycin attenuates plaque progression.

The potential role of resistin in atherogenesis.

Resistin, an adipocyte-derived cytokine linked to insulin resistance and obesity, has recently been shown to activate endothelial cells (ECs). Using microarrays, we found that along with numerous other pro-atherosclerotic genes, resistin expression levels are elevated in the aortas of C57BL/6J apoE-/- mice; these findings led us to further explore the relation between resistin and atherosclerosis. Using TaqMan PCR and immunohistochemistry, we found that ApoE-/- mice had significantly higher resistin mRNA and protein levels in their aortas, and elevated serum resistin levels, compared to C57BL/6J wild-type mice. Incubation of murine aortic ECs with recombinant resistin increased monocyte chemoattractant protein (MCP)-1 and soluble vascular cell adhesion molecule (sVCAM)-1 protein levels in the conditioned medium. Furthermore, human carotid endarterectomy samples stained positive for resistin protein, while internal mammary artery did not show strong staining. Patients diagnosed with premature coronary artery disease (PCAD) were found to have higher serum levels of resistin than normal controls. In summary, resistin protein is present in both murine and human atherosclerotic lesions, and mRNA levels progressively increase in the aortas of mice developing atherosclerosis. Resistin induces increases in MCP-1 and sVCAM-1 expression in murine vascular endothelial cells, suggesting a possible mechanism by which resistin might contribute to atherogenesis. Finally, PCAD patients exhibited increased serum levels of resistin when compared to controls. These findings suggest a possible role of resistin in cardiovascular disease.

Overlapping vs. one long stenting in long coronary lesions.

Intervention of long coronary lesions remains problematic, and optimal treatment strategy is yet to be determined. Despite advancement of stent technology, data are few regarding the efficacy of overlapping stents vs. a single long stent in long coronary lesions. This study was performed to evaluate the results of those strategies for long coronary lesions and to determine the predictors of in-stent restenosis (ISR). Sixty-four lesions (> 20 mm) in 64 patients were treated with either one long stent (group 1, n = 32) or two overlapping stents (group 2, n = 32). Overlapping stents were used at tortuous or calcified lesions and at lesions with diameter discrepancy or significant dissection. Immediate results, follow-up clinical and angiographic outcomes, and predictors of ISR were evaluated. Procedures were successful in all patients in both groups. Clinical and angiographic follow-ups were performed in 54 (84%) cases and 50 (78%) cases, respectively. During the follow-up, major adverse cardiac event occurred in 36% of group 1 and 29% of group 2 (P = 0.56). Six-month ISR rates were 39% in group 1 and 41% in group 2 (P = 0.91). Age (>/= 65 years old) was an independent risk factor of ISR (54% vs. 23%; OR = 4.4; P = 0.04), and distal reference diameter (RD) of less than 2.5 mm tended to predict ISR in multivariate analysis (60% vs. 25%; OR = 3.5; P = 0.06). In conclusion, stent overlapping can be used with outcome similar to that of one long stent in long coronary lesions. The optimal result may be obtained by considering the patient's age and the distal vessel diameter of the lesion.

Percutaneous intervention of spontaneous renal artery dissection complicated with renal infarction: a case report and literature review.

Spontaneous renal artery dissection (SRAD) is a rare condition that occurs before renal infarction. Using percutaneous intervention to treat SRAD remains controversial because it is not clear whether it is feasible or effective. We describe a 48-year-old male patient with SRAD complicated with renal infarction who was successfully treated with percutnaeous angioplasty and renal artery stenting.

Spontaneous disruption of mycotic aneurysm involving innominate artery.

We report a case of ruptured mycotic aneurysm involving innominate artery requiring an urgent surgical treatment. A 62-yr-old woman presented with fever and dyspnea. Previously, she was diagnosed with colon cancer and received right hemicolectomy and one cycle of adjuvant chemotherapy. On echocardiogram, pericardial effusion was noted and emergency pericardiocentesis was performed. CT scan revealed aortic aneurysm involving ascending aorta and innominate artery, and thrombi surrounding those structures. Patch repair of the defect in the ascending aorta and ringed Goretex graft to bypass the innominate and ascending aorta were performed. We believe that this is the first case of ruptured mycotic aneurysm involving innominate artery.