Denecke K.,Innovation Center Computer Assisted Surgery |
Krieck M.,Niedersachsisches Landesgesundheitsamt |
Otrusina L.,Brno University of Technology |
Smrz P.,Brno University of Technology |
And 3 more authors.
Methods of Information in Medicine | Year: 2013
Objectives: Detecting hints to public health threats as early as possible is crucial to prevent harm from the population. However, many disease surveillance strategies rely upon data whose collection requires explicit reporting (data transmitted from hospitals, laboratories or physicians). Collecting reports takes time so that the reaction time grows. Moreover, context information on individual cases is often lost in the collection process. This paper describes a system that tries to address these limitations by processing social media for identifying information on public health threats. The primary objective is to study the usefulness of the approach for supporting the monitoring of a population's health status. Methods: The developed system works in three main steps: Data from Twitter, blogs, and forums as well as from TV and radio channels are continuously collected and filtered by means of keyword lists. Sentences of relevant texts are classified relevant or irrelevant using a binary classifier based on support vector machines. By means of statistical methods known from biosurveillance, the relevant sentences are further analyzed and signals are generated automatically when unexpected behavior is detected. From the generated signals a subset is selected for presentation to a user by matching with user queries or profiles. In a set of evaluation experiments, public health experts assessed the generated signals with respect to correctness and relevancy. In particular, it was assessed how many relevant and irrelevant signals are generated during a specific time period. Results: The experiments show that the system provides information on health events identified in social media. Signals are mainly generated from Twitter messages posted by news agencies. Personal tweets, i.e. tweets from persons observing some symptoms, only play a minor role for signal generation given a limited volume of relevant messages. Relevant signals referring to real world outbreaks were generated by the system and monitored by epidemiologists for example during the European football championship. But, the number of relevant signals among generated signals is still very small: The different experiments yielded a proportion between 5 and 20% of signals regarded as "relevant" by the users. Vaccination or education campaigns communicated via Twitter as well as use of medical terms in other contexts than for outbreak reporting led to the generation of irrelevant signals. Conclusions: The aggregation of information into signals results in a reduction of monitoring effort compared to other existing systems. Against expectations, only few messages are of personal nature, reporting on personal symptoms. Instead, media reports are distributed over social media channels. Despite the high percentage of irrelevant signals generated by the system, the users reported that the effort in monitoring aggregated information in form of signals is less demanding than monitoring huge social-media data streams manually. It remains for the future to develop strategies for reducing false alarms. © Schattauer 2013.
Ondruschka B.,University of Leipzig |
Sommer G.,Innovation Center Computer Assisted Surgery |
Schober K.,University of Leipzig |
Schober K.,German Airforce Institute of Aviation Medicine |
And 4 more authors.
Journal of Neurotrauma | Year: 2013
Postmortem analysis of relevant biomarkers might aid in characterizing causes of death and survival times in legal medicine. However, there are still no sufficiently established results of practical postmortem biochemical investigations in cases of traumatic brain injury (TBI). The two biomarkers-S100 protein subunit B (S100B) and neuronal specific enolase (NSE)-could be of special interest. Therefore, the aim of the present study was to investigate changes in their postmortem levels for further determination of brain damage in TBI. In 17 cases of TBI (average age, 58 years) and in 23 controls with different causes of death (average age, 59 years), serum and cerebrospinal fluid (CSF) samples were analyzed with a chemiluminescence immunoassay for marker expression. An increase in serum S100B, as well as a subsequent decrease after survival times>4 days, were detected in TBI cases (p<0.01). CSF NSE values >6,000 ng/mL and CSF S100B levels >10,000 ng/mL seem to indicate a TBI survival time of at least 15 min (p<0.01). It is of particular interest that CSF S100B levels (p<0.01) and serum S100B levels (p<0.05) as well as CSF NSE values (p<0.01) were significantly higher in TBI cases in comparison to the controls, especially when compared with fatal non-head injuries. In conclusion, the present findings emphasize that S100B and NSE are useful markers in postmortem biochemistry in cases of suspected TBI. Further, S100B may be helpful to estimate the survival time of fatal injuries in legal medicine. © 2013 Mary Ann Liebert, Inc.
Denecke K.,Bern University of Applied Sciences |
Deng Y.,Innovation Center Computer Assisted Surgery |
Declerck T.,German Research Center for Artificial Intelligence
CEUR Workshop Proceedings | Year: 2016
Important information is captured in medical documents. To make use of this information and intepret the semantics, technologies are required for extracting, analysing and interpreting it. As a result, rich semantics including relations among events, subjectivity or polarity of events, become available. The First Workshop on Extraction and Processing of Rich Semantics from Medical Texts, is devoted to the technologies for dealing with clinical documents for medical information gathering and application in knowledge based systems. New approaches for identifying and analysing rich semantics are presented. In this paper, we introduce the topic and summarize the workshop contributions.
PubMed | Ludwig Maximilians University of Munich, Innovation Center Computer Assisted Surgery, University of Leipzig and University Medical Center Mainz
Type: Journal Article | Journal: Laryngo- rhino- otologie | Year: 2016
The follow-up for head and neck cancer (HNC) focussed on therapy control. Accessory long term functionality is important. Impairment of function is observed, but a comparable documentation is not established. Additional we frequently see psychooncological comorbidities, what complicates the assessment. This was the reason why Tschiesner et al. developed on the base of the ICF Core set for head and neck cancer a guideline for the Assessment of Function in HNC. In consequence of good results in other tumour entities we developed an electronic version (OncoFunction).In a proof of concept study all patients of our follow up consultation from 07/13 to 03/14 were included. OncoFunction was given to patients in a digital form using tablet computers. The results were visible to the physician in a concentrated form before consultation and were supplemented by a physician questionnaire. Furthermore we evaluated the usability in 202 patients.We had 682 patient contacts. 530 patient contacts (77, 7%) used the questionnaire. The physician questionnaire was answered in 470 times. Finally there are from 69.8% of the patient contacts full datasets available. Between users and non-users of the questionnaire we see no difference.The use of a computer-based screening and feedback system (OncoFunction) in clinical use is feasible and excellent assessed by patients. The patient data are visible in a compact form for the physician and problems can clear addressed to the patient. One more benefit is the standardized follow up documentation and the use of comparable data in research.
Denecke K.,Innovation Center Computer Assisted Surgery
Studies in Health Technology and Informatics | Year: 2014
Analysing medical social media data gains in importance given an increased availability of such data. In this paper, we analyse the language of medical blogs by means of a sublanguage analysis. More specifically, verb usage, semantic categories of used words as well as co-occurrence patterns are determined by means of natural language processing tools. The results show that in this text type, many concepts refer to the semantic categories Living Beings and Chemicals and Drugs. In contrast to clinical documents, the spectrum of verbs in blogs is very broad creating semantic relations of different types. From these language characteristics, we conclude for automatic processing tools for medical blogs that methods for reference resolution and for relation extraction where the relation type does not need to be specified in advance are required. © 2014 European Federation for Medical Informatics and IOS Press.
Gaebel J.,Innovation Center Computer Assisted Surgery |
Kolter T.,ID Information und Dokumentation im Gesundheitswesen |
Arlt F.,University of Leipzig |
Denecke K.,Innovation Center Computer Assisted Surgery
Studies in Health Technology and Informatics | Year: 2015
Clinical documentation is usually stored in unstructured format in electronic health records (EHR). Processing the information is inconvenient and time consuming and should be enhanced by computer systems. In this paper, a rule-based method is introduced that identifies adverse events documented in the EHR that occurred during treatment. For this purpose, clinical documents are transformed into a semantic structure from which adverse events are extracted. The method is evaluated in a user study with neurosurgeons. In comparison to a bag of word classification using support vector machines, our approach achieved comparably good results of 65% recall and 78% precision. In conclusion, the rule-based method generates promising results that can support physicians' decision making. Because of the structured format the data can be reused for other purposes as well. © 2015 IMIA and IOS Press.
PubMed | Innovation Center Computer Assisted Surgery and University of Leipzig
Type: Journal Article | Journal: Acta neurochirurgica | Year: 2016
Reliable intraoperative resection control during surgery of malignant brain tumours is associated with the longer overall survival of patients. B-mode ultrasound (BUS) is a familiar intraoperative imaging application in neurosurgical procedures and supplies excellent image quality. However, due to resection-induced artefacts, its ability to distinguish between tumour borders, oedema, surrounding tissue and tumour remnants is sometimes limited. In experienced hands, this bright rim effect could be reduced. However, it should be determined, if contrast-enhanced ultrasound can improve this situation by providing high-quality imaging during the resection. The aim of this clinical study was to examine contrast-enhanced and three-dimensional reconstructed ultrasound (3D CEUS) in brain tumour surgery regarding the uptake of contrast agent pre- and post-tumour resection, imaging quality and in comparison with postoperative magnetic resonance imaging in different tumour entities.Fifty patients, suffering from various brain tumours intra-axial and extra-axial, who had all undergone surgery with the support of neuronavigation in our neurosurgical department, were included in the study. Their median age was 56years (range, 28-79). Ultrasound imaging was performed before the Dura was opened and for resection control at the end of tumour resection as defined by the neurosurgeon. A high-end ultrasound (US) device (Toshiba Aplio XG) with linear and sector probes for B-mode and CEUS was used. Navigation and 3D reconstruction were performed with a LOCALITE SonoNavigator and the images were transferred digitally (DVI) to the navigation system. The contrast agent consists of echoic micro-bubbles showing tumour vascularisation. The ultrasound images were compared with the corresponding postoperative MR data in order to determine the accuracy and imaging quality of the tumours and tumour remnants after resection.Different types of tumours were investigated. High, dynamic contrast agent uptake was observed in 19 of 21 patients (90%) suffering from glioblastoma, while in 2 patients uptake was low and insufficient. In 52.4% of glioblastoma and grade III astrocytoma patients CEUS led to an improved delineation in comparison to BUS and showed a high-resolution imaging quality of the tumour margins and tumour boarders. Grade II and grade III astrocytoma (n=6) as well as metastasis (n=18) also showed high contrast agent uptake, which led in 50% to an improved imaging quality. In 5 of these 17 patients, intraoperative CEUS for resection control showed tumour remnants, leading to further tumour resection. Patients treated with CEUS showed no increased neurological deficits after tumour resection. No pharmacological side-effects occurred.Three-dimensional CEUS is a reliable intraoperative imaging modality and could improve imaging quality. Ninety percent of the high-grade gliomas (HGG, glioblastoma and astrocytoma grade III) showed high contrast uptake with an improved imaging quality in more than 50%. Gross total resection and incomplete resection of glioblastoma were adequately highlighted by 3D CEUS intraoperatively. The application of US contrast agent could be a helpful imaging tool, especially for resection control in glioblastoma surgery.
Voruganti A.K.R.,Innovation Center Computer Assisted Surgery |
Mayoral R.,Innovation Center Computer Assisted Surgery |
Vazquez A.,Innovation Center Computer Assisted Surgery |
Burgert O.,Innovation Center Computer Assisted Surgery
International Journal of Computer Assisted Radiology and Surgery | Year: 2010
Purpose Continuous video is used with increasing frequency in the operating room for minimally invasive laparoscopic and endoscopic procedures. Video data communication inthe OR requires device interoperability, efficient data transfer methods, and specialized IT infrastructure. Methods A framework for digital video communication based on a two channel client-server architecture was developed and tested. One channel is used for stream handling and the second channel is used for data streaming.Avideo stream description (VSD) specification is defined to negotiate video stream characteristics and ensure semantic interoperability. Quality assessment of the streamed data employs an imagebased structural quality measure called the Structural Similarity (SSIM) Index. By introducing the stream description and a quality metric, the stream parameters can be modified as needed. The video communication framework ensures interoperability by defining interfaces for each of the streaming architecture modules. Results To prove the framework's feasibility, two prototype applications were developed and performance testswere performed on a dedicated OR network. The results showed acceptable network performance for streaming video in the OR network under clinically realistic conditions. Conclusion An OR video communications framework was developed that uses existing OR network infrastructure as an economical alternative to dedicated integrated OR solutions. This framework provides functional and semantic interoperability among imaging modalities for continuous video data communication. © CARS 2010.
PubMed | Innovation Center Computer Assisted Surgery, University of Leipzig and TU Munich
Type: Comparative Study | Journal: Laryngo- rhino- otologie | Year: 2016
Microsurgical preparation is limited by geometric and mechanical constraints. In preparation for clinical use, this study investigates performance, ease of handling and precision of a novel manipulator concept for microsurgery.A group of 15 ENT experienced doctors, as well as a group of 17 medical students with low/non surgical experience participated in the study. Each of the subjects carried out 4 trials of simulated surgeries on a phantom with built-in force sensors. The task was to apply a defined force between 1.5 and 2 N using a Fisch micro perforator, 16 cm length, 0.4 mm (Storz) targeting holes with a diameter of 0.5 mm. For comparison, the Fisch micro perforator was moved manually or with the manipulator.Assessing the total number of errors proved a significantly lower error number (p<0.0001) and an improvement of the accuracy of 76% with the manipulator. The time requirement for the procedure with the micro manipulator is on average 2-3 times higher than with manual control (p<0.0001). But it is notable that this time requirement significantly decreases with training (p<0.0001).The study shows a significant reduction in the number of errors by using a new manipulator concept compared to the non-augmented human hand in an experimental setup. We observed a significant learning effect when subjects applied the micro manipulator, resulting in reduction of the time requirement while maintaining a constant number of errors.
PubMed | Innovation Center Computer Assisted Surgery
Type: Journal Article | Journal: Yearbook of medical informatics | Year: 2015
Social media, web and mobile technologies are increasingly used in healthcare and directly support patientcentered care. Patients benefit from disease self-management tools, contact to others, and closer monitoring. Researchers study drug efficiency, or recruit patients for clinical studies via these technologies. However, low communication barriers in socialmedia, limited privacy and security issues lead to problems from an ethical perspective. This paper summarizes the ethical issues to be considered when social media is exploited in healthcare contexts.Starting from our experiences in social-media research, we collected ethical issues for selected social-media use cases in the context of patient-centered care. Results were enriched by collecting and analyzing relevant literature and were discussed and interpreted by members of the IMIA Social Media Working Group.Most relevant issues in social-media applications are confidence and privacy that need to be carefully preserved. The patient-physician relationship can suffer from the new information gain on both sides since private information of both healthcare provider and consumer may be accessible through the Internet. Physicians need to ensure they keep the borders between private and professional intact. Beyond, preserving patient anonymity when citing Internet content is crucial for research studies.Exploiting medical social-media in healthcare applications requires a careful reflection of roles and responsibilities. Availability of data and information can be useful in many settings, but the abuse of data needs to be prevented. Preserving privacy and confidentiality of online users is a main issue, as well as providing means for patients or Internet users to express concerns on data usage.