Using NI FlexRIO for Photoacoustic Quantitative Ultrasound

 Read in  |   Print

"The development work was fast and easy because the NI products we chose gave us quick prototyping for cutting-edge medical research. With NI products, we bridged academic proof-of-concept studies and clinical trials."

- Pasi Karppinen, University of Helsinki

The Challenge:
Developing a real-time system to tune photoacoustic measurements to clinically assess osteoporosis.

The Solution:
Using NI FlexRIO and NI LabVIEW to connect the ultrasonic signal to a time-delayed laser diode array for optimized photoacoustic excitation of ultrasound in the patient.

Author(s):
Pasi Karppinen - University of Helsinki
Jari Tuovinen - University of Helsinki
Timo Karppinen - University of Helsinki
Edward Hæggström - University of Helsinki
Petro Moilanen - University of Jyväskylä
Jussi Timonen - University of Jyväskylä
Zuomin Zhao - University of Oulu
Risto Myllylä - University of Oulu

Introduction

Multimode skeletal quantitative ultrasound, based on ultrasonic guided waves, assesses bone properties such as mineral density, elastic stiffness, porosity, and cortical thickness. These properties correlate with the bone’s ability to resist fractures. Therefore, multimode QUS provides a means to predict the propensity for fractures and identify individuals with an increased risk. The main technical challenge with this method is stimulating and detecting ultrasonic signals through the overlying soft tissue. In particular, certain clinically useful ultrasonic wave modes do not penetrate soft tissue layers. We solved this problem with photoacoustics (PA). In this approach, a laser beam excites and detects broadband ultrasound at the surface of the radius bone, through the soft tissue. PA avoids the contact problems of traditional ultrasound excitation transducers, as well as adverse effects caused by the soft tissue.

Ultrasonic guided waves, used in our photoacoustic quantitative ultrasound (PAQUS) method, propagate in the cortical bone. These waves are the most sensitive to certain bone parameters of clinical interest. For each wave mode, a particular ultrasonic frequency band yields the best sensitivity. These characteristic frequency bands vary between individuals because soft tissue and bones differ in size and shape. This natural variation in anatomy causes significant measurement (and clinical assessment) bias, unless the frequency band is appropriately tuned to each individual. Such bias reduces the clinical applicability and relevance of the test. We solved this problem by implementing tuned excitation. However, the current tuning is too slow for practical in-vivo clinical measurements.

Clinical PAQUS method application requires rapid automated excitation tuning. To this end, we implemented a source featuring a phase-delayed array of laser diodes. Next, we figured out how to time the array element excitation rapidly enough to take full advantage of the tuned excitation. The excitation timing controls how energy transfers to the desired wave mode propagating in the bone in the specified frequency band. Thus, to implement rapid automatic tuning, we designed a feedback system that quickly finds appropriate, individualized timing with a frequency-domain-based feedback algorithm.

To build the feedback system, we needed a user-configurable field-programmable gate array (FPGA) and a digitizer with the capacity to collect at least 40 MS/s. The only available off-the-shelf solution was NI FlexRIO and its digitizer module, making it the obvious choice. In addition to the necessary sample rate, the digitizer’s 14-bit resolution gives us accurate measurements with a large dynamic range, which is important because signals that travel in bone may be substantially weaker than signals that travel in soft tissue.

Instrument Capabilities

The PAQUS device assesses the radius bone of the forearm. Ultrasound signals are excited by four 905 nm laser diodes, which are pulsed 1,000 times per second with 500 ns long pulses. An NI PXIe-7961R FPGA module independently controls each laser diode. NI 5732 digitizer digital outputs route the trigger signals to the laser diode array. Trigger time resolution is 150 ns. Optical fibers guide light pulses generated by the diodes to a linear array. The near-infrared light is collimated and directed onto the forearm. The ultrasound signals excited in the radial bone are detected by two piezoelectric contact transducers with a center frequency of 200 kHz.

The equipment is enclosed in a PXI chassis that provides a sturdy, portable platform for clinical use. Custom-made equipment preamplifies the signals, and an NI 5732 digitizes them. An NI PXIe-7961R module analyzes the received sound signals in real time and determines the delay of each laser diode. The implemented feedback algorithm optimizes the excitation to each individual. The analysis includes signal fast Fourier transform (FFT) and determination of the full width at half maximum (FWHM) of the received ultrasound signal FFT spectrum. The algorithm minimizes the derivate of the FFT-FWHM to optimize the excitation, with respect to individual parameters. When the derivate of the FFT-FWHM is minimized, the frequency band is as narrow as possible, and the ultrasound mode dispersion is strongest, so tester sensitivity is highest with regard to assessing radial bone cortical thickness.

Results

The feedback system we implemented helped us assess predictors of osteoporosis on test subjects. We learned that we can use the feedback algorithm for rapid patient positioning—to deduce whether the radial bone is parallel to the laser diode array. This is important because it improves individual measurement speed by a factor of up to five (from 5 minutes to 1 minute) and offers faster screening. It can also help eliminate operator bias with regard to cortical thickness estimates, which is important for proper diagnosis.

NI FlexRIO offered off-the-shelf hardware with a user-configurable FPGA and a fast digitizer. We took advantage of LabVIEW system design sofware for swift and easy programming, using its extensive libraries and complete drivers. This environment also provided the necessary tools to develop a solid user interface for the medical technician. The LabVIEW FPGA Module gave us a convenient, high-level language for FPGA programming, so we could quickly develop a fast, practical medical diagnostics device. By using NI hardware and software, we developed the product quickly. In just two years, we developed a device from scratch that is suited for clinical trials.

Conclusion

We developed a device to perform large-scale clinical trials for osteoporosis screening. The development work was fast and easy because the NI products we chose gave us quick prototyping for cutting-edge medical research. With NI products, we bridged academic proof-of-concept studies and clinical trials.

Author Information:
Pasi Karppinen
University of Helsinki

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