Nina Stumpf

Semester Work
Supervisors: Dr. Sergio Sanabria, Prof. Dr. Orcun Goksel

Optimisation of clinical workflow of speed-of-sound measurements by cloud processing of ultrasound images synchronised with Bluetooth distance sensor readings

The Computer-assisted Applications in Medicine (CAiM) group has developed a hand-held device combining a conventional ultrasound (US) B-mode imaging system with a Plexiglas strip reflector in order to calculate the speed of sound (SoS) in human tissue. The current version contains a Sylvac S Cal EVO BT caliper for measuring the distance between the US transducer and the reflector. Currently, US images are recorded with the ultrasound system and transferred to the storage (PACS) system of the hospital. The images are then post-processed to identify the reflector, which allows calculating the time of wave propagation between transducer and reflector. The distance between US transducer and reflector is manually annotated and the speed of sound is calculated as SoS = 0.5*time/distance. Although this procedure provides good results, it poses usability challenges for the clinicians. This semester project aims at facilitating and automatising the data acquisition process. The first goal was to evaluate technical alternatives for the distance readings and to explore ways to automatically record distance readings during the measurement and combine them synchronously with sequences of US images to calculate the SoS. Furthermore, possibilities to outsource the computations to a cloud service were examined. Three measurement series were conducted: one on a SonixTOUCH research machine using a 5 MHz linear probe (both by Ultrasonix) and examining the Breast Elastography Phantom 059 (by CIRS), the others at the Universitätspital Zürich on a GE Logiq E9 using a 9 MHz linear probe, once in water and once examining the said breast phantom. During the measurements, the distance readings were recorded at a rate of minimum 10 Hz using the Sylvac BT Smart App and then continuously transferred through a Bluetooth link to a smartphone. A GUI has been developed to upload US images and distance readings to the server set up at SafeSwissCloud, where the SoS is calculated and the results are sent back in order to be displayed in the GUI.