Segyeong Joo

Computer-aided diagnosis of solid breast nodules: use of an artificial neural network based on multiple sonographic features

A computer-aided diagnosis (CAD) algorithm identifying breast nodule malignancy using multiple ultrasonography (US) features and artificial neural network (ANN) classifier was developed from a database of 584 histologically confirmed cases containing 300 benign and 284 malignant breast nodules. The features determining whether a breast nodule is benign or malignant were extracted from US images through digital image processing with a relatively simple segmentation algorithm applied to the manually preselected region of interest. An ANN then distinguished malignant nodules in US images based on five morphological features representing the shape, edge characteristics, and darkness of a nodule. The structure of ANN was selected using k-fold cross-validation method with k = 10. The ANN trained with randomly selected half of breast nodule images showed the normalized area under the receiver operating characteristic curve of 0.95. With the trained ANN, 53.3% of biopsies on benign nodules can be avoided with 99.3% sensitivity. Performance of the developed classifier was reexamined with new US mass images in the generalized patient population of total 266 (167 benign and 99 malignant) cases. The developed CAD algorithm has the potential to increase the specificity of US for characterization of breast lesions.

Computer-aided diagnosis of solid breast nodules on ultrasound with digital image processing and artificial neural network

A computer-aided diagnosis algorithm identifying breast nodule malignancy using multiple ultrasonography features and artificial neural network classifier was developed from a database of 584 histologically-confirmed cases containing 300 benign and 284 malignant breast nodules. The features were extracted from sonographic images through digital image processing. An artificial neural network then distinguished malignant nodules based on those features. The trained artificial neural network showed the normalized area under the receiver operating characteristic curve of 0.95.

A portable microfluidic flow cytometer based on simultaneous detection of impedance and fluorescence

A portable microfluidic flow cytometer with dual detection ability of impedance and fluorescence was developed for cell analysis and particle-based assays. In the proposed system, fluorescence from microparticles and cells is measured through excitation by a light emitting diode (LED) and detection by a solid-stated photomultiplier (SSPM). Simultaneous impedometric detection provides information on the existence and size of microparticles and cells through polyelectrolyte gel electrodes (PGEs) operated by custom designed circuits for signal detection, amplification, and conversion. Fluorescence and impedance signals were sampled at 1 kHz with 12 bit resolution. The resulting microfluidic cytometer is 15x10x10 cm(3) in width, depth, and height, with a weight of about 800 g. Such a miniaturized and battery powered system yielded a portable microfluidic cytometer with high performance. Various microbeads and human embryonic kidney 293 (HEK-293) cells were employed to evaluate the system. Impedance and fluorescence signals from each bead or cell made classification of micro particles or cells easy and fast.

β‐Amyloid is transmitted via neuronal connections along axonal membranes

β‐amyloid plaque is a critical pathological feature of Alzheimer disease. Pathologic studies suggest that neurodegeneration may occur in a retrograde fashion from axon terminals near β‐amyloid plaques, and that plaque may spread through brain regions. However, there is no direct experimental evidence to show transmission of β‐amyloid.

In vivo calibration of the subcutaneous amperometric glucose sensors using a non-enzyme electrode.

A new two-point calibration method for the subcutaneous amperometric continuous glucose sensor is reported. The proposed method is based on direct measurement of the background current (I(o)) using a non-enzyme electrode. For in vivo test, three electrodes were implanted in rabbits. Two of the three were identical needle-type enzyme electrodes with perfluorinated polymer outer layers (Pt/enzyme layer/Kel-F/PTFE/Kel-F/Nafion) that were placed in subcutaneous tissue and in a vessel (ear artery), respectively. And one non-enzyme electrode with exactly the same membrane composition as those of other two was in the subcutaneous layer to measure the background current. Implantation in the subcutaneous layer generated many crevices on the protecting layers of the electrodes. The signals from enzyme electrodes were effectively corrected by the measured background current from the non-enzyme electrode. In addition, a telemetric monitoring system was developed and evaluated for in vivo continuous glucose monitoring in order to alleviate the problems of motion artifact.

SERS decoding of micro gold shells moving in microfluidic systems

In this study, in situ surface‐enhanced Raman scattering (SERS) decoding was demonstrated in microfluidic chips using novel thin micro gold shells modified with Raman tags. The micro gold shells were fabricated using electroless gold plating on PMMA beads with diameter of 15 μm. These shells were sophisticatedly optimized to produce the maximum SERS intensity, which minimized the exposure time for quick and safe decoding. The shell surfaces produced well‐defined SERS spectra even at an extremely short exposure time, 1 ms, for a single micro gold shell combined with Raman tags such as 2‐naphthalenethiol and benzenethiol. The consecutive SERS spectra from a variety of combinations of Raman tags were successfully acquired from the micro gold shells moving in 25 μm deep and 75 μm wide channels on a glass microfluidic chip. The proposed functionalized micro gold shells exhibited the potential of an on‐chip microfluidic SERS decoding strategy for micro suspension array.

Prediction of spontaneous ventricular tachyarrhythmia by an artificial neural network using parameters gleaned from short-term heart rate variability

Highlights? Prediction of ventricular tachyarrhythmia with artificial neural network. ? Parameters of heart rate variability analysis were used as features. ? Sensitivities of the classifier were around 80%. Reducing casualties due to sudden cardiac death and predicting ventricular tachyarrhythmia (VTA), ventricular tachycardia (VT) or ventricular fibrillation (VF), is a key issue in health maintenance. In this paper, we propose a classifier that can predict VTA events using artificial neural networks (ANNs) trained with parameters from heart rate variability (HRV) analysis. The Spontaneous Ventricular Tachyarrhythmia Database (Medtronic Version 1.0), comprising 106 pre-VT records, 26 pre-VF records, and 126 control data, was used. Each data set was subjected to preprocessing and parameter extraction. After correcting the ectopic beats, data in the 5min window prior to the 10s duration of each event was cropped for parameter extraction. Extraction of the time domain and non-linear parameters was performed subsequently. Two-thirds of the database of extracted parameters was used to train the ANNs, and the remainder was used to verify the performance. Three ANNs were developed to classify each of the VT, VF, and VT+VF signals, and the sensitivities of the ANNs were 82.9% (71.4% specificity), 88.9% (92.9% specificity), and 77.3% (73.8% specificity), respectively. The normalized areas (Azs) under the receiver operating characteristic (ROC) curve of each ANNs were 0.75, 0.93, and 0.76, respectively.

A novel high-resolution anorectal manometry parameter based on a three-dimensional integrated pressurized volume of a spatiotemporal plot, for predicting balloon expulsion in asymptomatic normal individuals

Anorectal manometry with simulated evacuation (SE) has limited applicability in predicting balloon expulsion (BE) test results. The newly developed high‐resolution anorectal manometry (HRAM) technique can yield spatiotemporal plots with three‐dimensional pressurization. We aimed to define new parameters based on three‐dimensional integrated pressurized volume (IPV) for predicting the BE test results in asymptomatic normal individuals.

A comparison of fabrication methods for Iridium Oxide reference electrodes

Several methods for the manufacturing of Iridium Oxide (IrOx) electrodes have been discussed in the literature. Two commonly used fabrication methods are Sputtered Iridium Oxide Films (SIROF) and Activated Iridium Oxide Films (AIROF). Most of the studies for in vivo electrodes have reported optimizations to these methods in the context of stimulation of and recording from neural tissue. In this work we characterize three fabrication methods of IrOx films for use as reference electrodes during in vivo neurochemical recordings, and we conclude that AIROF electrodes are preferable as reference electrodes due to their superior open circuit potential (OCP) stability.

Early prediction of ventricular tachyarrhythmias based on heart rate variability analysis

Ventricular tachyarrhythmias (VTAs) are fatal events and it is obvious that early prediction of VTAs could help in reducing mortality rate due to sudden cardiac death (SCD). Heart rate variability (HRV) reflects all symptoms associated with autonomic nervous system (ANS) as well as heart disease. Thus, HRV has frequently been used in various studies. We collected 220 recordings (VTAs - ventricular tachycardia (VT) and ventricular fibrillation(VF): 110, Control data: 110) from 81 adult patients in Intensive care unit (lCU), Asan Medical Centar (AMC) and proposed three classifiers for prediction of VT As events using eleven HRV parameters. Our group already developed a predictor for VTAs using ventricular arrhythmias dataset in Physionet before 10 seconds ahead of the events. In this study, we tried to predict VTAs earlier than an hour using parameters from HRV analysis and artificial neural network (ANN) models. The ANN model for prediction of VTAs showed a significantly high accuracy as 86.11 % (189/220) and Area under the curve (AVC) of receiver operating characteristic (ROC) was 0.88.