PTB >>Treating Retinal Disease with FPGA Controlled Lasers

More than 50 percent of Americans diagnosed with diabetes are at risk of developing diabetic retinopathy, a retinal disease that can lead to blindness. The condition is a result of diabetes affecting the circulatory system of the retina and causing abnormal new blood vessel growth. It has become the leading cause of new blindness among U.S. adults.

Laser Photocoagulators
Since the 1970s, ophthalmologists have used a standard treatment called laser photocoagulation to treat diabetic retinopathy. The method, which has changed little since its inception, involves the controlled destruction of the peripheral retina using targeted laser pulses. This can free up oxygen to the rest of the eye and slow down or even prevent the blindness caused by the disease. While this type of treatment has proven effective at reducing the chances of vision loss by as much as 50 percent, it can require as many as 2,000 burns and ophthalmologists can deliver only one burn at a time.

A full course of treatment typically requires two to four sessions, each lasting 12 to 15 minutes in which physicians administer the laser burns using a joystick and foot pedal to aim and fire. The procedure is so painful that some patients do not return for follow-up visits, even though they are fully aware that the untreated condition can eventually lead to blindness. OptiMedica Corporation recognized the need for improved safety, precision, comfort, and speed of the photocoagulation procedure. Using a technique that was originally developed at Stanford University, they created the PASCAL Photocoagulator, which semi-automates a series of patterned laser shots and cuts the entire procedure to less than five minutes. The system provides significantly improved performance for the physician administering the treatment, as well as an enhanced therapeutic experience for the patient.

The PASCAL Photocoagulator uses a pattern scan laser to treat the condition using a single shot or predetermined pattern array. By automatically delivering multiple, shorter pulses rather than a series of manually placed lesions, the PASCAL system greatly reduces the risk of inadvertent application of the treatment to the fovea, which can result in localized vision loss. Physicians using conventional methods of photocoagulattion must manually attempt to avoid this critical region of central vision. Furthermore, the PASCAL system significantly reduces the overall procedure time, as well as the amount of discomfort felt by the patient. This is achieved by reducing the laser pulse time from 100 milliseconds to just 10 milliseconds and then automating multiple spots with each depression of the foot pedal.

FPGAs
At the highest level, FPGAs are reprogrammable silicon chips. Using prebuilt logic blocks and programmable routing resources, users can configure these chips to implement digital computing tasks in software and compile them down to a configuration file or bitstream that contains information on how the components should be wired together. Traditionally, using FPGA technology required expertise in hardware description languages (HDLs) like VHDL or Verilog. Now with higher-level design tools such as LabVIEW that abstract the complexities of digital design, engineers and scientists of any discipline can take advantage of FPGA technology. OptiMedica engineers, whose expertise is in ophthalmology and photocoagulation, used graphical block diagrams in LabVIEW FPGA to successfully embed their laser treatment algorithms in hardware.

For precise laser control OptiMedica uses a PCI-7833R intelligent DAQ device.

The trap leakage versus temperature behavior of the sensor must first be learned in order to determine the target maximum operating temperature for the planned maximum exposure length. Measuring the traps’ leakage characteristics is a straightforward but time-consuming process. First, a set of reference dark exposures is taken at different temperatures using a common exposure time. In this work, the standard exposure time was five minutes. Once this reference data is collected, the sensor is flooded using NIR light and a series of dark exposures are taken matching the length of the reference dark. After each dark exposure is captured, it is compared to the reference dark, and the difference is recorded. It is important to accurately measure and remove the offset from each dark frame. A good way to do that is by over scanning the horizontal shift register during readout and carefully measuring the zero signal level in the over scan region. Once the data is collected at several different operating temperatures, it can be plotted as shown in Figure 2.

For the PASCAL user interface, an LCD touch screen display is used to select from predetermined pattern types and administer up to 25 spots at a time.

In previous methods of retinal laser photocoagulation, the physician used a mechanical joystick and foot pedal to deliver single 100- millisecond laser pulses to abnormal blood vessels of the peripheral retina. The PASCAL Photo coagulator essentially automates and im proves this process by taking advantage of FPGA technology in National Instruments intelligent data acquisition (DAQ) devices. In the NI PCI-7833R, built-in digital-to-analog converters are used to generate the control signal for the laser position and built-in analog-to-digital converters ensure that the actual laser position matches the desired position. Other inputs and outputs, such as the laser power and the foot pedal, are read and written using isolated digital inputs and outputs. With this approach, OptiMedica engineers can achieve a closed-loop control system with a response time of four microseconds.

As with all medical devices, PASCAL incorporates redundancy with additional sensors for an increased level of safety. In addition to feedback built into the galvanometer, the laser beam is split and directed to an internal photo detector, which is constantly read by another analog input channel on the PCI-7833R. This photo detector gives PASCAL the ability to constantly test and monitor the laser power coming from the output beam. If the power is too low or too high, a second galvanometer deflects the beam away from the final lens and prevents faulty lasers from entering into the patient’s eye. PASCAL also uses this internal photo detector to ensure that the power is off when necessary, such as during the start-up sequence or when the system is in standby mode.

Using an FPGA-based data acquisition board, PASCAL can make hardware-timed decisions in microseconds. Although the laser control logic and safety logic are embedded within the same FPGA chip, dedicated blocks of I/O and silicon operate independently. These dedicated blocks of I/O give PASCAL the ability to fully take advantage of the parallelism that FPGAs offer. Engineers at OptiMedica chose to use NI LabVIEW for all software development of the host and FPGA applications. The graphical nature of LabVIEW makes it simpler to program and visualize tasks running in parallel on the block diagram. LabVIEW FPGA code can be migrated between PCI, PXI, and PCI Express off-the-shelf hardware devices. Each part of the LabVIEW FPGA application functions simultaneously with all other pieces of the application, allowing redundant logic to become more than just another set of commands that procedurally execute. PASCAL can continuously test and monitor itself and immediately identify possible component failures before undesirable results have occurred. Reading the photo detector, for example, will happen every five microseconds, while the voltage output loop to the galvanometer happens every 50 microseconds. Because each of the safety routines is independent of all other tasks, each part of the safety code runs quickly and reliably.

User Interface
For the user interface, OptiMedica engineers chose an off-the-shelf touch panel monitor running on a Windows system. From the main control panel, an LCD display with a touch screen control is used to select from predetermined pattern types and administer up to 25 spots at a time. All other equipment is very similar to conventional, single-shot photocoagulation, with a mechanical joystick and foot pedal, so physician training is minimal. Once the pattern is selected, a separate red beam is used for aiming and visualizing placement before delivery. The touch screen interface is also used for selecting various parameters, such as aim beam intensity, treatment laser power, exposure time, system status, and shutdown. Because each pulse size and pattern is predictable, PASCAL photocoagulation is a more precise way to evenly distribute patterned spots, which also increases the reproducibility of results for each treatment.

In the past, FPGA technology was available only to engineers with a deep understanding of digital hardware design. The rise of high-level design tools, however, has opened up this technology to a new class of users — domain experts with no embedded design experience. The result was that OptiMedica was able to quickly and efficiently prototype and deploy their final system using off-the-shelf hardware.

This article was written by Chris Delvizis, R Series Intelligent Data Acqusition Product Marketing Engineer, and PJ Tanzillo, Biomedical Segment Lead, National Instruments (Austin, TX). For more information, contact Mr. Delvizis at chris.delvizis@ni.com, Mr. Tanzillo at pj.tanzillo@ni.com, or visit http://info. hotims.com/22916-200.