I was very impressed with the IEEE EMBC 2008 as it was really obvious that the conference organizers were able to share cutting edge science while maintaining an engineering atmosphere. My presentation went smoothly! The over-all theme of the conference was to push the limits of current bio-medical technologies to make a better patient experience focusing on better diagnostics, ease of use and instrumentation intelligence. Vancouver ended up being a pretty nice city as well, after it stopped raining that is.
Most of the diagnostic improvements were in the form of better signal processing and data analysis. The key motivation was to get more use from already-available instruments without substantial additional investment. The star of this show was definitely breast cancer detection where the leading edge technology focused on using both elastography and magnetic resonance imaging. The key idea is that an ultrasound actuator can send elastic waves through the breast tissue that ultimately displace (cyclically) charged particles and this displacement can be detected on MR in 3D. Since this displacement depends on tissue properties, this is effectively a measurement of both electrical and mechanical tissue impedance where changes in this impedance can highlight potential tumor areas.
The focus of the ease of use designs was overrun by wireless applications. Everything from neural recordings, to patient vital sign measurement, to device configuration was being moved to a wireless network without any serious security considerations. There were various stimulation devices proposed, both for neural and cardiac applications, that used standard bluetooth or zigbee type protocols with no encryption or authentication whatsoever. The typical response I received from the presenters was that they were developing a proof-of-concept and would add the security features at a later point, however, my experience with network engineering leads me to believe that it is best to security considerations from the start. It is easy to tell when an authentication system was cobbled on as an afterthought and such a system either decreases performance or has some flaws at the interface. It would take a pretty sick person to launch a denial of service attack on a pacemaker, but I don’t doubt that such people exist.
Device intelligence was mainly about making devices that require less intervention and can be operated by unskilled technicians and patients. Patient operation stood out as there seems to be a substantial positive psychological effect when they are allowed to have some input on their therapies. The increased device intelligence would then attempt to control the patient input to prevent any harmful reactions. Furthermore, it was good to see at least a few medical devices that would initiate some sequence of events based on monitored patient activity as compared to strictly open-loop programs.
del.icio.us |
digg
