Efficacy and safety of leadless pacemaker: A systematic review, pooled analysis and meta-analysis.

BACKGROUND: Leadless pacemakers have been designed as an alternative to transvenous systems which avoid some of the complications associated with transvenous devices. We aim to perform a systematic review of the literature to report the safety and efficacy findings of leadless pacemakers. METHODS: We searched MEDLINE and EMBASE to identify studies reporting the safety, efficacy and outcomes of patients implanted with a leadless pacemaker. The pooled rate of adverse events was determined and random-effects meta-analysis was performed to compare rates of adverse outcomes for leadless compared to transvenous pacemakers. RESULTS: A total of 18 studies were included with 2496 patients implanted with a leadless pacemaker and success rates range between 95.5 and 100%. The device or procedure related death rate was 0.3% while any complication and pericardial tamponade occurred in 3.1% and 1.4% of patients, respectively. Other complications such as pericardial effusion, device dislodgement, device revision, device malfunction, access site complications and infection occurred in less than 1% of patients. Meta-analysis of four studies suggests that there was no difference in hematoma (RR 0.67 95%CI 0.21-2.18, 3 studies), pericardial effusion (RR 0.59 95%CI 0.15-2.25, 3 studies), device dislocation (RR 0.33 95%CI 0.06-1.74, 3 studies), any complication (RR 0.44 95%CI 0.17-1.09, 4 studies) and death (RR 0.45 95%CI 0.15-1.35, 2 studies) comparing patients who received leadless and transvenous pacemakers. CONCLUSION: Leadless pacemakers are safe and effective for patients who have an indication for single chamber ventricular pacing and the findings appear to be comparable to transvenous pacemakers.

TSInsight: A Local-Global Attribution Framework for Interpretability in Time Series Data.

With the rise in the employment of deep learning methods in safety-critical scenarios, interpretability is more essential than ever before. Although many different directions regarding interpretability have been explored for visual modalities, time series data has been neglected, with only a handful of methods tested due to their poor intelligibility. We approach the problem of interpretability in a novel way by proposing TSInsight, where we attach an auto-encoder to the classifier with a sparsity-inducing norm on its output and fine-tune it based on the gradients from the classifier and a reconstruction penalty. TSInsight learns to preserve features that are important for prediction by the classifier and suppresses those that are irrelevant, i.e., serves as a feature attribution method to boost the interpretability. In contrast to most other attribution frameworks, TSInsight is capable of generating both instance-based and model-based explanations. We evaluated TSInsight along with nine other commonly used attribution methods on eight different time series datasets to validate its efficacy. The evaluation results show that TSInsight naturally achieves output space contraction; therefore, it is an effective tool for the interpretability of deep time series models.

Two-stage framework for optic disc localization and glaucoma classification in retinal fundus images using deep learning.

BACKGROUND: With the advancement of powerful image processing and machine learning techniques, Computer Aided Diagnosis has become ever more prevalent in all fields of medicine including ophthalmology. These methods continue to provide reliable and standardized large scale screening of various image modalities to assist clinicians in identifying diseases. Since optic disc is the most important part of retinal fundus image for glaucoma detection, this paper proposes a two-stage framework that first detects and localizes optic disc and then classifies it into healthy or glaucomatous. METHODS: The first stage is based on Regions with Convolutional Neural Network (RCNN) and is responsible for localizing and extracting optic disc from a retinal fundus image while the second stage uses Deep Convolutional Neural Network to classify the extracted disc into healthy or glaucomatous. Unfortunately, none of the publicly available retinal fundus image datasets provides any bounding box ground truth required for disc localization. Therefore, in addition to the proposed solution, we also developed a rule-based semi-automatic ground truth generation method that provides necessary annotations for training RCNN based model for automated disc localization. RESULTS: The proposed method is evaluated on seven publicly available datasets for disc localization and on ORIGA dataset, which is the largest publicly available dataset with healthy and glaucoma labels, for glaucoma classification. The results of automatic localization mark new state-of-the-art on six datasets with accuracy reaching 100% on four of them. For glaucoma classification we achieved Area Under the Receiver Operating Characteristic Curve equal to 0.874 which is 2.7% relative improvement over the state-of-the-art results previously obtained for classification on ORIGA dataset. CONCLUSION: Once trained on carefully annotated data, Deep Learning based methods for optic disc detection and localization are not only robust, accurate and fully automated but also eliminates the need for dataset-dependent heuristic algorithms. Our empirical evaluation of glaucoma classification on ORIGA reveals that reporting only Area Under the Curve, for datasets with class imbalance and without pre-defined train and test splits, does not portray true picture of the classifier's performance and calls for additional performance metrics to substantiate the results.

Correction to: Two-stage framework for optic disc localization and glaucoma classification in retinal fundus images using deep learning.

Following publication of the original article.

FuseAD: Unsupervised Anomaly Detection in Streaming Sensors Data by Fusing Statistical and Deep Learning Models.

The need for robust unsupervised anomaly detection in streaming data is increasing rapidly in the current era of smart devices, where enormous data are gathered from numerous sensors. These sensors record the internal state of a machine, the external environment, and the interaction of machines with other machines and humans. It is of prime importance to leverage this information in order to minimize downtime of machines, or even avoid downtime completely by constant monitoring. Since each device generates a different type of streaming data, it is normally the case that a specific kind of anomaly detection technique performs better than the others depending on the data type. For some types of data and use-cases, statistical anomaly detection techniques work better, whereas for others, deep learning-based techniques are preferred. In this paper, we present a novel anomaly detection technique, FuseAD, which takes advantage of both statistical and deep-learning-based approaches by fusing them together in a residual fashion. The obtained results show an increase in area under the curve (AUC) as compared to state-of-the-art anomaly detection methods when FuseAD is tested on a publicly available dataset (Yahoo Webscope benchmark). The obtained results advocate that this fusion-based technique can obtain the best of both worlds by combining their strengths and complementing their weaknesses. We also perform an ablation study to quantify the contribution of the individual components in FuseAD, i.e., the statistical ARIMA model as well as the deep-learning-based convolutional neural network (CNN) model.

Effect of hyperventilation on electroencephalographic activity.

OBJECTIVE: To observe the effect of voluntary hyperventilation on electroencephalographic activity during routine EEG recording on patients referred to a tertiary care hospital. METHODS: This was an observational study conducted at Neurophysiology Lab, Department of Neurology at Liaquat National Hospital, Karachi from May 2007 to September 2007. Data of 326 patients was collected prospectively and analyzed by SPSS version 10.0. At least 3 minutes voluntary hyperventilation was performed by the subjects. All those patients who were able to perform voluntary hyperventilation adequately were included in the study. RESULTS: Of 326 recordings, 256 (78.8%) were normal and 69 (21.2%) were abnormal. Focal epileptiform discharges were identified in 8.6% and generalized in 8.3% of subjects. Physiological slowing was found in 31 records. Out of 55 epileptics, 3 had discharges only during HV and 9 had increase in epileptiform discharges. CONCLUSION: Hyperventilation has significant effect on background rhythm during EEG recording. It is a useful activation method utilized to increase the yield of EEG.

Factors delaying hospital arrival of patients with acute stroke.

INTRODUCTION: To determine the proportion of patients with acute stroke presenting late to hospital and to identify the factors that delay hospital arrival of patients with acute stroke. METHODS: A cross sectional study was carried out between Sept 2006 to Feb 2007 in the department of Neurology, Liaquat National Hospital Karachi. All patients of both genders, ago >18 years with symptoms of stroke and neuro- imaging (CT scan/MRI brain) findings consistent with stroke were included. RESULTS: Atotal of 165 patients attending the Emergency department were included. There were 86 (52%) males and 79 (47.9%) females. The mean age was 60.04+/-13.98 years, (males 58.2 years and females 61.9 years). The median delay from onset of symptoms to hospital arrival was six hours. Only 28.5% of the patients came within three hours while 71.5% after three hours. Attendants of 47 patients had a low throat perception, 53 (32%) of the patients did not know a single symptom of stroke and 63% (104) patients first contacted their General Practitioner who referred them to hospital. Similarly 60.6% of patients wore first taken to a local hospital not equipped to handle major emergencies. CONCLUSION: Time elapsed from onset of symptoms to hospital arrival is influenced by lack of knowledge of stroke symptoms, contact with a local doctor, low threat perception and non availability of ambulance services.

A wearable sensor based multi-criteria-decision-system for real-time seizure detection.

This paper presents a wireless, low power and low cost two part wearable for real-time epileptic seizure detection. Using parameters of Electro-cardiograph (ECG), Electro-dermal Activity (EDA), body motion and breathing rate (BR), a novel multi-criteria-decision-system (MCDS) is proposed that reduces false alarms and true negatives. The combination of a chest and hand worn wearable continuously senses these parameters transmitting the data to a smart phone application via BLE 4.0 where long-short-term-memory (LSTM) based anomaly detection algorithms and logistic classifiers decide on the occurrence of the seizure in real time. A 96% precision and 90% recall is achieved through testing on synthetic data.

Calcific Metamorphosis: A Review.

Dental trauma to the permanent dentition can lead to clinical complications and its management may considerably challenge a practitioner. The incidence of pulp canal obliteration following dental trauma has been reported to be approximately 4 - 24%. Attempting to locate canals following calcific metamorphosis and negotiating it to full working length may lead to iatrogenic errors such as fractured instrument and perforation. This review article describes the possible etiology of Calcific Metamorphosis, its clinical and radiographic features as well as its management.

Management of pulp canal obliteration using the Modified-Tip instrument technique.

The incidence of pulp canal obliteration following dental trauma has been reported to be approximately 4 - 24% and its management can be quite challenging for the practitioner. Locating the canal and negotiating it to full working length may lead to iatrogenic errors such as fractured instrument and perforation. This case report deals with such a case using the modified tip instrument technique to gain access and negotiate the canal to the apex followed by conventional root canal preparation and obturation.

Bactericidal efficacy of photodynamic therapy against Enterococcus faecalis in infected root canals: a systematic literature review.

OBJECTIVE: The aim was to review the bactericidal efficacy of photodynamic therapy (PDT) against Enterococcus faecalis (E. faecalis) in infected root canals. METHODS: To address the focused question "Does PDT exhibit bactericidal effects against E. faecalis in infected root canals?" PubMed/Medline and Google-Scholar databases were searched from 1985 up to August 2013 using various combinations of the following key words: "antibacterial; "bactericidal; "endodontic; "root canal" and "photodynamic therapy". Original studies, experimental studies and articles published only in English language were included. Letters to the editor, historic reviews and unpublished data were excluded. The pattern of the present review was customized to primarily summarize the pertinent information. RESULTS: Seventeen studies (16 ex vivo and one in vivo) were included. In these studies, numbers of teeth used ranged between 30 and 220 teeth. In these studies, wavelengths of diode laser used, diameter of fiber and power output ranged between 625 and 805 nm, 200 µm and 0.4 cm, and 40 mW and 5 W respectively. Twelve studies reported PDT to be effective in eliminating E. faecalis from infected root canals. Four studies reported conventional irrigation and instrumentation to be more efficient in killing E. faecalis than PDT. One study reported PDT and conventional endodontic regimes to be equally effective in eliminating E. faecalis from infected root canals. In most studies, toluidine blue and/or methylene blue were used as photosensitizers. CONCLUSION: Efficacy of PDT in eliminating E. faecalis from infected root canals remains questionable.