Presentations at the World Congress on Medical Physics and Biomedical Engineering in Prague, Czech Republic 3-8 june 2018.
Microwave technology for detecting abdominal bleeding
Stefan Candefjord1,2,3), Tomas Lukas Szulc1), Sigrún Helga Davíðsdóttir1), Mikael Persson1,2), Andreas Fhager1,2), Bengt Arne Sjöqvist1,2,3), Mikael Elam4,2), Marianne Oropeza-Moe5), Nils Petter Oveland6)
1)Electrical Engineering, Chalmers University of Technology, Gothenburg, Sweden
2)MedTech West, Sahlgrenska University Hospital, Gothenburg, Sweden
3)SAFER Vehicle and Traffic Safety Centre at Chalmers, Chalmers University of Technology, Gothenburg, Sweden
4)Clinical Neurophysiology, Sahlgrenska University Hospital, Gothenburg, Sweden
5)Department of Production Animal Clinical Sciences, Faculty of Vetbio, Norwegian University of Life Sciences, Sandnes, Norway
6)Department of Anesthesiology and Intensive Care, Stavanger University Hospital, Stavanger, Norway
Injury accounts for 10 % of global mortality, and tens of millions new victims each year face lifelong disabilities. A significant proportion of traumatic deaths are preventable if detected and treated earlier. Hemorrhage represents a substantial proportion of preventable deaths. Hemoperitoneum (abdominal bleeding) is one injury that is frequently lethal and challenging to detect in the prehospital setting. Ultrasound can be used for detection, but it requires a trained operator and is currently rarely used in prehospital care. In this study, we have evaluated the potential for microwave technology to detect hemoperitoneum. A porcine model of hemoperitoneum using anesthetized pigs was developed. A belt with eight microwave antennas was strapped around the pig’s abdomen. Measurements spanning a frequency interval of 0.1–2.0 GHz were performed on ten pigs, and hemoperitoneum of 500 mL and 1000 mL were induced using the pig’s own blood and compared to baseline (no bleeding). The blood accumulated predominantly around the midaxillary line on both sides of the body, as confirmed by ultrasound. Therefore, we compared the transmission coefficients for the two outer antennas of the belt, placed close to the midaxillary line. Visual inspection revealed that the bleeding dampened the magnitude of the signal, as was expected due to the high electrical conductivity of blood. An ANOVA test confirmed that the reductions of the magnitude, derived by calculating and comparing the area under the curves, for both the 500 mL and 1000 mL levels were statistically significant for both sides (p < 0.05). This shows that microwave technology has potential for detecting and monitoring hemoperitoneum. However, the baseline signals for the different pigs varied substantially, likely due to anatomical differences, which complicates detection. This warrants further studies to explore how the changes in magnitude due to hemoperitoneum can be effectively identified despite baseline variations.
Stroke and trauma diagnostics for ambulances and helicopters
Persson et al, Electrical engineering, Chalmers University of Technology, Gothenburg, Sweden
This paper reports on the development and testing of of novel, compact, portable diagnostics system for stroke and trauma for use in ambulances and helicopters. Portable diagnostics for the pre-hospital setting remains a major challenge for the 30 million individuals who each year sufferers a stroke or brain trauma. When an ambulance arrives at the scene of an accident or a suspected stroke the objective is to reduce the risk for death and provide optimal care of the patient. The patient is evaluated and decisions made, with limited information, on which hospital the patient needs to be transported to and on whether to initiate some form of treatment. Both these groups of patients, which are subject to these life and death decisions, would clearly benefit if medical personal had access to the additional support of diagnostic tools. No such tools are today available in clinical praxis but as the the first group in the world and have, with our clinical partners, started to evaluate the first generations in clinical trials. The latest from ongoing clinical and biomedical engineering research will be presented.
Numerical study of a Microwave Haemorrhagic Stroke Detector
Andreas Fhager, Stefan Candefjord, Mikael Persson
Electrical engineering, Chalmers University of Technology, Gothenburg, Sweden
Microwave systems constitute a novel technology that enables prehospital diagnostics of stroke and intracranial bleedings caused for example by trauma. The healthcare sector is in need of new innovations in both diagnostics and treatment of these diseases in order to further improve patient outcomes. Stroke and brain trauma are very common, they cause huge suffering among patients and are very costly for the society. Effective treatments exist that could benefit many patients. Treatment must however be preceded by diagnostics and a limitation of today’s healthcare is the delays in the diagnostics, which has to be made at hospitals with CT. As a result, many patients receive treatment either too late to be effective, or not at all. This is extremely unsatisfactory, since a treatment given in time often can be lifesaving and lead to improved recovery. Our research aims at developing a portable diagnostic system that can be used in ambulances and is capable of detecting intracranial bleedings. We have previously, with encouraging results, performed several smaller clinical studies to obtain proof-of-concept. Much investigation however remains to optimise the design of the system in order to maximise the performance. To assess the performance of a machine learning algorithm for detection a numerical study of the system has been conducted, simulating microwave measurements of stroke patients. The goal is to study the performance of the detection algorithm as a function of different design parameters of the system and the clinical studies, for example the number of patients included in the training data set. As a result of this numerical study we conclude that this approach is feasible, but that it may require a large patient group, up to one thousand stroke patients or more, for training the algorithm in order to reach high sensitivity and specificity levels.