$54
Oral Platform Abstracts
Tuesday, November 8, 2005
A three-layered neural network utilizing metabolite concentrations as input was created. The output was either neoplastic or nonneoplastic etiology initially, followed by further categorization o f nonneoplastic lesions. Thirty cases (119 tuberculomas, 7 cysticerci, 6 gliomas) were utilized for training the network. The remaining (17 tuberculomas, 6 cysticerci, 7 gliomas, 2 tumefactive demyelination, 2 pyogenic abscesses) were analyzed using the network. Results: The network provided 94.0% sensitivity and 96.7% specificity in differentiating neoplastic from non-neoplastic lesions. Tumefactive demyelination accounted for the false positive predictions. Further categorization among non-neoplastic lesions yielded variable results which were in part due to lack of specificity of ~H M R Spectroscopy itself. Conclusion: A neural network based on metabolite profiles allows differentiation of neoplastic from non-neoplastic ring enhancing lesions. Further categorization of non-neoplastic lesions was dependent on the degree of training of the network and the limits of specificity of spectroscopic analysis. OPL035 DaTSCAN SPECT is tligtfly sensitive but less specific for the diagnosis of Parkinson's disease Ruzicka, E 1, Zarubova, K 2, Kupka, K s, Krizova, H 3, Bojar, M 2, Roth, J~. ~Charles University, Movement Disorders Centre, Prague,
Czech Republic," 2Charles University, Dept. of Neurology, Prague, Czech Republic," 3Charles University, Dept. of Nuclear Medicine, Prague, Czech Republic Background: Ioflupane SPECT (DaTSCAN) demonstrates deficits of dopamine transporters reflecting striatal dopaminergic denervation in Parkinson's disease (PD). We aimed to verify its efficiency in differential diagnosis of PD. Methods: We perforated DaTSCAN examination in 75 patients (40 M, 35 F, mean age 55, range 21-81 yrs) with suspected PD, presenting with unusual combinations of signs or with unclear treatment response. After intravenous administration o f ioflupane, the distribution of radioactivity in the brain was scanned with subsequent visual and semiquantitative evaluation. The results of DaTSCAN were later confronted with the final clinical diagnosis. Results: With the time delay of 2 to 48 months, P D was confirmed in 43 and excluded in 32 out o f 75 patients. Of those, p rob able or possible multisystem atrophy (MSA) was diagnosed in 15, essential tremor in 7. By visual evaluation, DaTSCAN results were qualified as abnormal in 57 and normal in 18 patients. The results were abnormal in 42 out of 43 patients in whom the final diagnosis of PD was made. However, abnormal DaTSCAN results were also found in 15 out of 32 patients with different diagnoses. The comparison of DaTSCAN results with the final clinical diagnosis showed the sensitivity of 0.98 and specificity of 0.53 for PD. Conclusion: DaTSCAN proved high sensitivity for determination of nigrostriatal dopaminergic denervation. However, the specificity of DaTSCAN for the diagnosis of PD was rather low, with high risk of false positive results, especially in MSA patients where presumably both presynaptic and postsynaptic parts of dopaminergic system were involved. OPL036 Outsourcing Neurology Night-thne Cover Using Intercontinental Telemediehle Patterson, V 13, Forbes, R 2, Wootton, R 1. JCentrefor Online Health, Brisbane, Australia," 2Royal I,Tetoria Hospital, Belfast, UK Background: Compliance with the European Working Time Directive (EWTD) is an increasing problem for small specialties like neurology. Outsourcing night-time cover to a different time zone (where it is daytime) is used in radiology in the USA. Since realtime telemedicine
can be used effectively for acute neurological hospital admissions we have investigated whether it could be used intercontinentally to provide nighttime neurology cover in the U K from daytime Brisbane. Method anti Results: First we analyzed demand between 10pm and 8am Sunday to Thursday (equivalent to 8am to 6pm Monday to Friday in eastern Australia) by recording workload o f the on-call neurology registrar in N.Ireland over 92 days; there were only 16 consultations on 14 patients. Second we performed 18 realtime videoconsultations on neurological patients in N. Ireland from Brisbane Australia, recording any technical difficulties; there were minor connection problems in seven patients and reliable connection at 384 kb/s could not be established for one patient. Third we estimated the cost per patient of telemedicine and conventional care; this was approximately £ 1200 for conventional care and £ 256 for telemedicine. Conclusion: Unlike the conventional system the telemedicine system could be scaled up to cover the population o f the U K using a single overseas neurologist. Outsourcing neurology cover is therefore possible, is less expensive than the present system and is a solution to the constraints of the EWTD.
Tuesday 8 November 2005 Stroke
- Epidemiology
OPL037 Gulf Cooperative council (GCC) Stroke Awaxeness Study Kamran, S. Hamad General Hospital, Qatar Objective: To assess the knowledge of stroke in general public in the GCC states. Background: The Arabian Gulf is a rapidly developing part of the world with major changes in the life style that can increase the risk of stroke. To design effective stroke treatment and prevention strategy, an assessment o f public knowledge of stroke is required. Methods: A cross sectional conmmnity based survey was conducted at primary health care centers, in urban and semi urban areas, o f the Gulf Cooperative Council (GCC) countries (Qatar, Saudi Arabia, Kuwait, Bahrain, UAE, Oman) on the level of Stroke awareness in the general public. 3750 face to face interviews were conducted by health care workers. Results: 1089 (29.0%) recognized the term stroke, (34.7"/0) recognized stroke involves brain and (29.3%) considered age group 30-50 at highest risk for stroke. The commonest risk factors identified were hypertension (17.7%), and smoking (27.3%). People who did not recognize the term stroke had higher incidence of diabetes, hypertension, high cholesterol and had more than one risk factors (p < 0.05). "rite most frequently identified stroke symptoms were weakness (21.6"/0) and speech problems (20"/0). The blockage of blood vessels was recognized as the commonest cause of stroke (22"/0) followed by tension/worrying (19%). Doctors and nurses were regarded as the best source of stroke information (70%). Level of stroke knowledge was poorest among women (homemakers), labor and clerical group. In the univariate comparison, younger age (p < 0.001), higher level of education (p < 0.0001), and family history of stroke (p < 0.004), predicted better stroke knowledge. In multivariate regression analysis level of education, family history of stroke and hypertension were independent variables predicting stroke knowledge. Conclusion: Majority of the patients had not even herd the term stroke. Stroke knowledge was poorest among groups that were at the tdghest risk for stroke. Stroke education has to focus on the high risk groups particularly women (homemakers) and labor/clerical group. The health care workers at the P H C and hospital will need instructions on providing stroke information to the public. The level of knowledge of stroke risk factors and symptoms emphasizes the need for stroke education efforts in the conmmnity.