Autonomous and Ultra Rapid Cavitation Control Feedback Loop for Focused Ultrasound Surgery


"Coding transparency in the same LabVIEW environment across PC, real-time, and FPGA platforms was critical to getting a working proof of concept in just six months."

- David Lines, Diagnostic Sonar

The Challenge:
Monitoring cavitation activity fast enough to use it in a feedback loop to provide adaptive control of bubble creation and activity, for advanced and safe operation. This could provide a mechanism for efficacious and targeted drug delivery.

The Solution:
Demonstrating that we can control cavitation via an ultra rapid feedback loop between an acoustic detector and the source of the focused ultrasound. The FlexRIO system is capable of analysing key cavitation emissions and outputting modulated intensity signals to the source, thereby providing controlled cavitation activity.

David Lines - Diagnostic Sonar
Bjoern Gerold -
Paul Prentice - University of Dundee, Insitutute of Medical Science and Technology

Figure 1. Cavitation Control Block Diagram


Focused ultrasound surgery is an emerging therapeutic modality, setting new clinical standards for treating a range of conditions. The technique involves noninvasively applying high-intensity focused ultrasound, with current clinical practise mediated by heating effects within tissue located at the focus. Cavitation bubble activity is generally avoided, as the phenomenon is not well understood and can cause considerable damage to healthy tissue.

There is, however, the possibility that cavitation could offer an alternative form of therapy, whereby bubbles act to mechanically disrupt tissue structures, rendering the site highly susceptible to enhanced therapy. Solid tumours, for example, exert a pressure gradient such that delivery of extraneous drugs is currently very challenging, requiring systemic administration of high doses of an inherently toxic substance. Our vision is to selectively and controllably ‘fizz’ tumour tissue with focused ultrasound for targeted drug delivery, substantially reducing the quantity of drug required and concentrating the effect to the disease site. In this role, cavitation would act to both permeate the tumour and actively pump drugs to the volume of tissue where they are needed. As focused ultrasound is typically administered from outside the body, there is no requirement for surgical incision, drastically reducing patient recovery time and risk of infection.  Demonstrating real-time monitoring, control, and manipulation of cavitation activity in focused ultrasound is critical for realising this potential.

The Team

This project was jointly carried out by the Cavitation Research Group at the Institute for Medical Science and Technology (IMSaT), University of Dundee, and Diagnostic Sonar Ltd (DSL), Livingston, UK.

IMSaT was founded in 2006 as an interdisciplinary institute positioned at the interface of physics, engineering, and clinical and life sciences. Medical ultrasound is a major focus of the institute’s work.

DSL, founded in 1975, successfully introduced the first real-time medical ultrasound scanner manufactured in the UK. Over the following four decades, the company has expanded into other areas, including industrial nondestructive evaluation and medical phantom supply. DSL has used NI hardware and software to develop this technology for many years, mainly using FlexRIO modules and LabVIEW. DSL finished products, including the NI PXI device, are shown below in Figures 2 and 3.


Figure 2. DSL PXI Front End



Figure 3. DSL PXI Front End


The work was supported by the Scottish Universities Physics Alliance and the European Research Council.

The Experiment

The Cavitation Research Group at Dundee recently developed a technique for studying acoustic cavitation at unprecedented spatial and temporal resolution via high-speed microphotography. This technique facilitated the development of custom-fabricated acoustic detectors, sensitive over key frequency bandwidths, to detect signature signals emitted by cavitation activity. For this project, we sought to demonstrate that a feedback loop between sensor and transducer could deliver a controllable level of cavitation activity. As cavitation develops at a timescale defined by the frequency of the focused ultrasound (MHz regime for therapeutic ultrasound), it requires that the decision-making algorithm be undertaken at submillisecond intervals.

System Criteria

 The essential criteria for such a system would be:

 1) The capability of deciphering cavitation signals to assess current level of activity

2) Comparing the current level to a user-defined level (according to application)

3) Generating an output signal to the source transducer, modulating the intensity of the focused ultrasound to deliver sustained cavitation at the user-defined level

An overriding precondition is that, if at any time cavitation activity proliferates to a volume deemed beyond a controllable level, the focused ultrasound exposure terminates. This represents the fail-safe setting for the intended therapeutic application.

The FlexRIO Solution

One advantage of FlexRIO architecture is that we can use LabVIEW to program both the host PC and the FPGA firmware. We use the latter to comprise the high-speed deterministic code for the feedback loop and to provide immediate shutdown response when the system detects unsafe cavitation. The host PC software offers system setup as well as the monitoring function.

The NI 5752 digitizer, optimised for ultrasound applications, can route the receive signal stream direct into the PXIe-7966R FlexRIO FPGA, where we can implement the fast Fourier transform in firmware using LabVIEW FPGA. The host PC loads the FPGA with the spectral profiles for rapid pattern matching. Decision output controls whether the therapy power is maintained, increased, decreased for stable operation, or shut down when safety is compromised.

With the FPGA sampling clock rate (50 MHz), the system offers sufficiently fast data processing for submillisecond feedback loop closure. The same card also controls the output, which responds to a cavitation-signal-detected energy and frequency analysis. A simple digital–to-analogue converter generates signals at a corrected voltage amplitude from the digitizer outputs (NI 5752), which pass through a radio frequency power amplifier (E&I 2100L) to the source transducer, shown in Figure 4.

Because the LabVIEW code is modular, we easily extended it to monitor the spectral response from multiple receive sensors in parallel, offering the additional benefit of triangulation to localize the cavitation source. LabVIEW code running on the host PC monitors this, as well as sending the multisensory parameters to the FPGA, in case cavitation activity outside the target volume occurs, requiring immediate shutdown.

Figure 4. Schematic Representation of the Feedback Loop System Configuration for Modulating Focused Ultrasound Cavitation Activity 


Experiments undertaken at the cavitation laboratory at IMSaT confirmed system functionality via high-speed camera observations. We observed cavitation bubble clouds with frame rates of up to 500x103 frames per second. We seeded cavitation inside an ultrasound coupling gel using a laser-nucleation technique. The high-speed camera recordings lasted for several milliseconds in order to observe the cavitation response to the system over several sets of 10+ iterations.

The system identifies:

  • Cavitation quantity: Bubble cloud spatial distribution and quantity. We could set a threshold number of clouds, above which the system stops the cavitation activity. For more information, view the Quantity video.

  • Cavitation quality: Cloud size, based on cavitation signal frequency content. The system modulates the acoustic intensity, keeping each bubble cloud below a certain size while sustaining activity at the required level. For more information, view the Quality video.

In both instances, the system could autonomously control the intensity of the driving voltage to the transducer, following pre established logic rules. The system regulates the acoustic intensity at every refresh cycle of ~100 µs to achieve the desired cavitation activity level.

By using LabVIEW throughout this project, our team could write code at an application level on a PC for initial simulation with previously acquired data, without using hardware. Once the system was viable, we simply moved the LabVIEW code over to the FlexRIO hardware using LabVIEW FPGA. Our researchers used cutting-edge NI FPGA equipment by continuing to develop using LabVIEW, without the need to learn another programming language. Coding transparency in the same LabVIEW environment across PC, real-time, and FPGA platforms was critical to getting a working proof of concept in just six months.

Future Work

This project demonstrates that cavitation control at high-speed refresh rates is feasible. The NI system configuration is also potentially capable of responding to several detectors in parallel, meaning that the system can react to the location of bubble clouds via triangulation with computational FPGA parallelism characteristics. Using the modular NI system combined with DSL components, we could simultaneously actively probe the cavitation activity and tissue imaging via a diagnostic ultrasound probe.

Other Potential Applications

Cavitation is known to play a pivotal role in industrial applications, including sonochemistry and precision acoustic cleaning. We anticipate that a system capable of delivering a controlled level of activity will also be of commercial interest to these sectors.


Author Information:
David Lines
Diagnostic Sonar

Bookmark and Share

Explore the NI Developer Community

Discover and collaborate on the latest example code and tutorials with a worldwide community of engineers and scientists.

‌Check‌ out‌ the‌ NI‌ Community

Who is National Instruments?

National Instruments provides a graphical system design platform for test, control, and embedded design applications that is transforming the way engineers and scientists design, prototype, and deploy systems.

‌Learn‌ more‌ about‌ NI