Hot Takes — A Look at the Latest Trends in Dose Reduction
By Keith Loria
Vol. 20 No. 10 P. 14
Responsible use of radiation in medical imaging continues to be a focus in the United States and Europe, as countries aim to improve patient outcomes and population health. Christopher Martel, PhD, DABHP, a clinical research scientist for DoseWise at Philips, says nine years ago, the FDA launched its Initiative to Reduce Unnecessary Radiation Exposure from Medical Imaging, which strengthened the overall commitment in health care to quality, patient satisfaction, and safety.
“During this time, our industry has seen a significant focus on training and education, adoption of technologies and practices [such as the National Electrical Manufacturers Association’s] XR-29 standard, and active engagement by professional organizations, all with the common goal of ensuring patients receive the right exam, at the right time, with the right radiation dose,” he says. “This focus has resulted in better trained staff and patients, better devices, and a wealth of data. The forthcoming update to [National Council on Radiation Protection and Measurements] Report 160 should reveal the results of these efforts.”
The industry is also seeing a renewed emphasis on understanding and reducing staff dose in interventional procedures.
“Many of the attempts to reduce staff dose have been through introducing additional tools such as pads placed on the patient to reduce scatter radiation and lightweight aprons worn by the staff,” Martel says. “However, the addition of real-time dosimeters to monitor staff exposure, allowing real-time corrective actions, has complemented these efforts.”
Girish Muralidharan, vice president and general manager of performance intelligence analytics for GE Healthcare, explains, unlike dose reduction, dose optimization aims to maximize the benefit of imaging procedures while minimizing safety risks from ionizing radiation, which are very low or hypothetical for the vast majority of procedures.
“In other words, while a hospital dose optimization program may result in lowering some procedure doses, radiation exposure may increase in other exams to improve image quality and diagnostic accuracy,” he says. “A typical example would be increasing allowable dose levels for bariatric patients where image quality is suffering.”
Dose optimization efforts continue to be driven by professional societies, dose awareness campaigns such as Image Gently and Image Wisely, accreditation requirements, and regulations.
Most recently, The Joint Commission enacted a requirement for recording fluoroscopy dose indices, effective January 1, 2019, establishing thresholds for follow-up action and review of exams exceeding threshold limits.
Trends in dose optimization include expanded accreditation requirements requiring collection and assessment of dose indices; adoption of decision support systems to aid physicians in ordering appropriate imaging procedures; increased availability of imaging system dose reporting and prospective alerting; standardized, size-based protocols for CT imaging; development of imaging system features, from positioning and acquisition to image formation that reduces dose while improving image quality; and widespread adoption of software solutions to manage dose, assess quality, and identify opportunities for improvement.
Elliot K. Fishman, MD, a professor of radiology, surgery, oncology, and urology at Johns Hopkins Medicine in Baltimore, says people have figured out that trying to get the lowest dose possible was not the best for imaging, and a lot of the focus has turned to how to get the right protocol for the right patient.
“To me, the best dose reduction is to do the study right the first time. Doing it once and reading it correctly is best,” he says. “When you finally do the study, it’s important to take advantage of what you have on your scanner. The manufacturers have done a good job of building things into their scanners.”
Mahadevappa Mahesh, PhD, MS, FACR, chair of the ACR’s Commission on Medical Physics, says there is still a lot of work to be done in balancing dose and image quality but, in general, the common trend he sees in the clinic is people being more confident in using CT dose modulation because the technology has matured.
David H. Foos, chief technology officer and senior director of imaging systems research and development for Carestream Health, says many flat panel digital radiography detectors now have exceptional detective quantum efficiency (DQE) characteristics, a measure of a detector’s signal-to-noise ratio performance that is a quantitative measure of the detector’s intrinsic dose efficiency. In general, DR detectors that have better DQE performance can deliver images with good image quality at proportionally lower dose, as compared with detectors having lesser DQE.
“New image processing methods continue to emerge that help to reduce noise appearance while maintaining sharpness appearance. For a given DR detector, as dose is reduced, the appearance of noise increases,” he says. “Historically, image processing methods that reduced noise appearance in imagery resulted in some degree of blurring or contrast reduction—in other words, a tradeoff.”
Some of the newest noise reduction methods are leveraging AI techniques to deliver even greater degrees of noise reduction while doing a better job of preserving sharpness appearance.
“The ability to reduce noise appearance while simultaneously preserving image structures—the signal—could prove effective in leading to lower dose radiographic techniques,” Foos says.
CT manufacturers are heeding the dose optimization trend with new tools for their scanners.
“In CT, we are harnessing artificial intelligence to dramatically improve image quality at low dose,” Muralidharan says. “TrueFidelity, approved by the FDA in April 2019, is the next generation of image reconstruction, using deep learning technology to improve image quality and reader confidence at low dose. Deep learning image reconstruction preserves the look of higher dose images while keeping doses lower for patient safety.”
To spare sites the significant time and resource cost of optimizing CT protocols, the University of Wisconsin has created standardized CT protocols for the majority of GE models. The protocols include small, medium, and large adult versions, as well as five size categories for pediatric patients. Dose benchmarking and dose check settings are provided with the protocols.
Additionally, DoseWatch, GE Healthcare’s vendor-neutral solution for dose management, recently released peak skin dose and skin dose map features for fluoroscopic imaging. Muralidharan says peak skin dose is the most accurate metric available for predicting a skin reaction to X-ray radiation.
“Availability of this new metric, previously requiring hours of physicist data collection and calculations, allows high-risk cases to be quickly and accurately identified,” he says. “The skin dose map, a visual presentation of dose across the patient’s skin, helps care providers understand where on the body a skin reaction may occur. In addition, impact of every exposure on skin dose can be displayed in movie format, allowing providers to better understand how specific imaging choices affect patient dose.”
Tim Nicholson, senior marketing manager for the CT business unit at Canon Medical Systems USA, says the trend the company sees, and what customers are asking for, is not only the use of low-dose techniques but also using low dose and improving image quality.
“For years, the trend was to use as low as possible and get the image quality you needed to make a diagnosis, but now you want that image to be improved, whether that’s facial recognition or low-contrast detectability,” he says.
With that in mind, Canon recently released its Advanced Intelligent Clear-IQ Engine, a deep convolutional neural network image reconstruction technology for CT that is capable of suppressing noise and enhancing signal with its deep learning algorithm for sharper images scanned on Canon’s premium imager, Aquilion ONE GENESIS.
“We’ve been able to train a machine to reconstruct CT images and identify signals in the images and identify noise in the images,” Nicholson says. “The machine could eliminate the noise, enhance the signal, and create some of the best image qualities we have seen on Canon CT scanners.”
Martel says the first step in building an effective radiation dose management strategy is collecting complete and accurate data to enable analysis and measurement of current performance. The Philips DoseWise Portal—part of the Philips PerformanceBridge portfolio—is a radiation exposure management solution that provides a streamlined, efficient way to review data on radiation exposure to both patients and staff, which can help facilitate regulatory compliance.
“The web-based, vendor-agnostic software solution automatically collects, measures, analyzes, and reports radiation exposure data on an individual and group basis and allows for real-time staff exposure measurements using Philips DoseAware,” Martel says. “The invisible nature of radiation makes the real-time assessment of radiation dose critical.”
DoseAware consists of wireless, real-time radiation badges, worn by all of the members of the team and quantitatively displaying the scattered X-ray dose being absorbed by each team member. The base station hangs adjacent to the imaging monitors and displays the amount of dose exposure. This provides instantaneous feedback to the operator to reduce X-ray dose, such as by using a shield, stepping back a few feet, or lowering the detector.
“They are all quantitatively demonstrated to the operators in the room during the procedure and serve as positive reinforcement,” Martel says. “The fact that the operators can immediately review the effect of their actions completes the feedback loop and helps to develop a healthy way of working.”
In addition to DoseWise Portal, Martel says many hospitals use Philips’ CT Iterative Model Reconstruction (IMR) to help manage dose and iPatient to help personalize dose management. IMR is the first knowledge-based solution that overcomes the traditional challenges of motion sensitivity, allowing it to be used with applications such as cardiac CT angiography and perfusion studies.
Radiation safety for patients and caregivers is essential, yet the complex multivendor imaging environment can make it challenging to gain control. Despite significant technology advancements, challenges remain. When imaging facilities lack an effective change management program, they may fail to effectively implement better imaging practices or dose reduction technology.
“If a facility does not have a shared vision, assigned accountability, and cross-discipline support—from technologists, medical physicists, and imaging physicians—progress will be frustratingly slow,” Muralidharan says. “Other barriers include lack of personnel resources, budget constraints, [lack of] executive support, or [lack of] a motivating influence such as state or Joint Commission requirements.”
One particular hurdle Muralidharan points to is a lack of adoption of cloud-based architectures for dose tracking, either due to institutional reluctance or regulatory barriers.
“Cloud-based systems have the potential to reduce infrastructure and maintenance costs, speed deployment of new software features, provide cross-facility and vendor benchmarking, and maintain a complete patient dose history regardless of the imaging provider,” he says.
Foos notes the main challenge in lowering dose is to ensure that the resultant image captured at a lower dose retains sufficient information content for the intended diagnostic task, in order for the radiologist to form a confident and accurate diagnostic impression.
“Another challenge is to ensure that incidental findings are not missed due to inadequate signal-to-noise ratio,” Foos says. “Additionally, reducing dose heightens the risk of increased frequency of need to repeat images because of poor image quality.”
While the industry has seen a maturing of the tools used in radiation dose tracking, the disconnects in the health care system continue to pose challenges in tracking patient radiation doses across multiple institutions. Martel says one institution that is uniquely positioned to effectively tackle this problem is the VA.
“Veterans who receive health care from VA medical facilities use a single medical record number. Now, the challenge is to create a network of medical centers connected to a common dose tracking system,” he says. “In April 2019, the VA Louis Stokes Medical Center in Cleveland received a grant to do just that; they selected the Philips DoseWise Portal as the single dose tracking system to link over 30 VA medical centers across the nation.”
Martel thinks the next breakthrough in dose reduction will likely come from vendors on postprocessing techniques that allow users to get as much diagnostic information as possible with as few photons as possible.
Muralidharan believes the future holds substantial opportunities for dose optimization, with nuclear medicine and molecular imaging likely to be one of the next focus areas for requirements on dose management. The doses in nuclear medicine and molecular imaging are often comparable with those from CT.
“Applications of deep learning in image generation, augmentation, and dose reduction are still in the early stages—there will undoubtedly be more to come across all modalities,” he says. “Dose management systems are growing in sophistication, providing previously unavailable insights and monitoring capabilities. It is expected that this technology will expand to further guide quality and performance improvements across medical imaging.”
— Keith Loria is a freelance writer based in Oakton, Virginia.