Search

In conventional retinal region detection methods for optical coherence tomography (OCT) images, many parameters need to be set manually, which is often detrimental to their generalizability. We present a scheme to detect retinal regions based on fully convolutional networks (FCN) for automatic diagnosis of abnormal maculae in OCT images. The FCN model is trained on 900 labeled age-related macular degeneration (AMD), diabetic macular edema (DME) and normal (NOR) OCT images. Its segmentation accuracy is validated and its effectiveness in recognizing abnormal maculae in OCT images is tested and compared with traditional methods, by using the spatial pyramid matching based on sparse coding (ScSPM) classifier and Inception V3 classifier on two datasets: Duke dataset and our clinic dataset. In our clinic dataset, we randomly selected half of the B-scans of each class (300 AMD, 300 DME, and 300 NOR) for training classifier and the rest (300 AMD, 300 DME, and 300 NOR) for testing with 10 repetitions. Average accuracy, sensitivity, and specificity of 98.69%, 98.03%, and 99.01% are obtained by using ScSPM classifier, and those of 99.69%, 99.53%, and 99.77% are obtained by using Inception V3 classifier. These two classification algorithms achieve 100% classification accuracy when directly applied to Duke dataset, where all the 45 OCT volumes are used as test set. Finally, FCN model with or without flattening and cropping and its influence on classification performance are discussed.

Objective: To determine the surgery-related and hardware-related complications of deep-brain stimulation (DBS) at a single centre. Methods: 262 consecutive patients (472 electrodes) operated for DBS in our department from February 1996 to March 2003 were retrospectively analysed to document acute adverse events (30 days postoperatively). The data of 180 of these patients were additionally revised to assess long-term complications (352 electrodes, mean follow-up 36.3 (SD 20.8) months). Results: The frequency of minor intraoperative complications was 4.2% (11/262 patients). Transient (0.2%) or permanent (0.4%) neurological deficits, and in one case asymptomatic intracranial haemorrhage (0.2%), were registered as acute severe adverse events caused by surgery. Among minor acute complications were subcutaneous bleeding along the extension wire (1.2%) and haematoma at the pulse generator implantation site (1.2%). Skin infection caused by the implanted material was registered in 15 of 262 patients (5.7%). The infection rate during the first observation period was 1.5% (4/262 patients) and the late infection rate was 6.1% (11/180 patients). Partial or complete removal of the stimulation system was necessitated in 12 of 262 (4.6%) patients because of skin infection. During the long-term observation period, hardware-related problems were registered in 25 of 180 (13.9%) patients. Conclusions: Stereotactic implantation of electrodes for DBS, if performed with multiplanar three-dimensional imaging and advanced treatment planning software, is a safe procedure with no mortality and low morbidity. The main causes for the patients’ prolonged hospital stay and repeated surgery were wound infections and hardware-related complications.

Background: Secondary prevention of stroke has been shown to dramatically reduce recurrence and has been described as suboptimal. Objective: To analyse patients’ awareness and knowledge about cerebrovascular risk factors (CVRF) and their influence on CVRF control. Methods: Patients (n = 164) who were attending a stroke outpatient clinic for the first time after hospital discharge (3 months) for a first stroke were asked to answer a short questionnaire including questions on awareness and knowledge of CVRF, visits to a CVRF specialist, number of visits to a general practitioner, adherence to drug treatments, cigarette smoking and cessation. Results: CVRF were spontaneously mentioned as relevant for their stroke by only13% of patients. A specialist was visited by only one-third of the patients and a general practitioner was not visited at all by 27% of the patients since their stroke. Awareness was inversely correlated with older age and good recovery. More than half of the patients had high blood pressure (≥140 mmHg for systolic and ≥90 mmHg for diastolic values) at the time of follow-up. These high values were correlated with poor awareness. Appropriate secondary stroke prevention measures were not received by one-fourth of the patients; this was also correlated with poor awareness. Conclusions: CVRF control is not optimal and is at least partially related to patients’ awareness and knowledge and suboptimal medical follow-up. Older patients and patients with excellent recovery are at particular risk for poor awareness and CVRF control.