Optimizing Medical Imaging for Children of All Sizes

A proposed framework balances quality and safety of computed tomography protocols across a range of body sizes in pediatric populations.

Pediatric patients come in many sizes, from newborns to teenagers, and the wide range poses an ongoing challenge for medical imaging. In a study published in August in the Journal of Medical Imaging, researchers set out to formulate a systematic, evidence-based method to minimize radiation dose while maximizing the diagnostic performance of computed tomography (CT) imaging for young patients. According to the authors, their model could be used to optimize individual scan parameters and provide for consistent diagnostic performance across the broad range of body sizes in children.

“This study provides a framework by which one can determine the exact technique and the amount of radiation dose that is appropriate as a function of patient size and age, such that the diagnostic quality from the images is ensured,” said first author Ehsan Samei, a professor of radiology at Duke University School of Medicine in North Carolina. “This is important, as it provides a scientific basis for proper CT imaging of children, ensuring the safety and quality of the examination.”

For the most part, past efforts to develop size-specific CT protocols for pediatric populations have relied on radiation dose alone. Ideally, these protocols should also take into account diagnostic accuracy and the trade-offs between these two factors for different pediatric patient sizes. “The imaging technique needs to be adjusted according to the patient size and age, but up to now, there has been little evidence on how this adjustment should take place,” Samei said.

To address this issue, Samei and his collaborators developed a framework for balancing quality and safety for varying pediatric patient sizes. The researchers devised their methodology based on data from their two prior studies on 30 pediatric patients who had undergone clinical chest CT imaging. In one of these studies, they determined doses as a function of patient size and scan parameters. In the other study, they characterized image quality in terms of detection accuracy and scan parameters. For each case, expert pediatric radiologists examined the images to detect lung nodules representing pulmonary metastatic diseases.

After analyzing the data, the researchers found that CT protocols could be improved by assessing the relationship between radiation dose and diagnostic accuracy across a range of pediatric sizes. Before optimization, protocols using a similar dose resulted in an accuracy ranging from 89 percent for the youngest patients to 67 percent for the oldest patients. When the researchers used optimized individual scan parameters, they achieved a consistent accuracy of 83 percent across patient age/size categories using a wide range of radiation doses.

“This work is among the first to combine realistic measures of task-based image quality and patient risk to help guide optimization of CT acquisition settings,” said Adam Alessio, a professor of radiology at the University of Washington in Seattle. “The primary take-away from this work is that as dose increases, there is a patient-specific and task-specific amount of dose where diagnostic performance plateaus and further increases in dose result in no real performance gains.”

Although additional studies are needed, the authors suggest that their framework could be applied to other imaging systems and clinical tasks. “We hope that the technology can be put to use by other medical centers, and further commercialized in near future,” Samei said. “It is ready for clinical translation, and we are in the process of precisely doing so in the follow-up projects underway.”

But other experts remain concerned about the cost-effectiveness and practicality of this approach. “This is a pretty complicated paper. I am not sure this can be incorporated into routine clinical practice,” said Eric J. Hall, a special research scientist at the Center for Radiological Research at Columbia University Medical Center in New York.

According to Keith Strauss, a clinical imaging physicist at Cincinnati Children’s Hospital Medical Center in Ohio, the new method is more rigorous than previous, simpler approaches. “However, these simpler approaches usually result in the correction of the majority of the problems,” he said. “Ideally, patient size, dose, and image quality should be rigorously addressed on all CT scanners, but this may only be possible in academic centers with extensive support in the clinic from medical physicists. There is always a point of diminishing returns. Just because we can do something in the clinic does not always mean we should do it.”

Janelle Weaver, Contributor | bit.ly/2zGZnCx | Mon, 10/02/2017