Ultrasound for evaluating liver steatosis.

By Kate Madden Yee, AuntMinnie.com staff writer

May 31, 2019 -- Ultrasound is a reliable alternative to MRI for assessing liver steatosis in the clinical setting, according to a study published in the July issue of Clinical Radiology. The findings offer clinicians a cost-effective, accessible option for evaluating steatosis.

While MRI has become an accepted modality for detecting liver steatosis, it comes with a high cost and is generally less available than other modalities, according to a team led by Dr. Marie-Luise Kromrey of University Medicine Greifswald in Germany. "Ultrasonography is commonly used to detect liver steatosis and has the advantage of being cost-effective, simple, and widely available," the team wrote.

Fat storage in the liver can be an indicator of metabolic syndrome, a condition characterized by insulin resistance and a precursor to type 2 diabetes. Since liver steatosis is also a risk factor for a variety of other diseases, having an effective way to assess it is important, Kromrey and colleagues noted.

Yet although ultrasound is regularly used to detect liver steatosis, its diagnostic accuracy and reliability for assessing the severity of fatty liver have been unclear. So Kromrey and colleagues conducted a study to compare the modality's performance to that of MRI.

The study included 2,783 patients who underwent 1.5-tesla MRI scans of the liver; from these MRI exams, the group calculated proton-density fat fraction and transverse relaxation rate to estimate liver steatosis and iron overload. Patients also underwent B-mode ultrasound. Kromrey's team then assessed the sensitivity and specificity of ultrasound to identify different degrees of steatosis and amounts of liver iron (Clin Radiol, July 2019, Vol. 74:7, pp. 539-546).

MRI showed liver steatosis in 40% of participants (mild, 68.9%; moderate, 26.7%; severe, 4.4%), while ultrasound found liver steatosis in 37.8%, which corresponded to a sensitivity of 74.5% and a specificity of 86.6%.

The group also found that ultrasound sensitivity increased with the amount of liver fat present (65.1% for low fat content, 95% for moderate fat content, and 96% for high fat content). Liver iron did not affect ultrasound's ability to detect liver steatosis, Kromrey and colleagues noted.

"The present results show excellent sensitivity and specificity of ultrasound for the estimation of fatty liver disease in patients with moderate and high liver fat content," the group wrote.

However, since ultrasound didn't perform as well in patients with low liver fat content, additional evaluation methods may still need to be used, according to the team.

"The weakness of ultrasound in assessing small amounts of liver fat should be considered and compensated by additional liver enzyme quantification or MRI," the authors concluded.

Top 4 Priorities for AI Research in Medical Imaging

The top 4 priorities for AI research in medical imaging

By Erik L. Ridley, AuntMinnie staff writer

May 29, 2019 -- Bringing radiology artificial intelligence (AI) technology to routine clinical practice will require four major priorities: structured use cases, data sharing methods, validation and monitoring tools, and new standards and data elements, according to a report published online May 28 in the Journal of the American College of Radiology.

"An active AI ecosystem in which radiologists, their professional societies, researchers, developers, and government regulatory bodies can collaborate, contribute, and promote AI in clinical practice will be key to translating foundational AI research to clinical practice," wrote a team of authors led by Dr. Bibb Allen Jr. of the American College of Radiology (ACR) Data Science Institute.

Following up on an initial medical imaging artificial intelligence roadmap published April 16 in Radiology, which covered the challenges, opportunities, and priorities for foundational research in AI for medical imaging, Allen and colleagues turned their attention to the key priorities for translational research. Both articles were produced as a summary of last year's U.S. National Institute of Biomedical Imaging and Bioengineering (NIBIB) workshop on medical imaging, which was co-sponsored by the RSNA, the ACR, and the Academy for Radiology & Biomedical Imaging Research.

In their latest report, the authors highlighted four key translational research priorities:

• Create structured use cases to define and highlight the clinical challenges that AI could potentially solve.
• Create methods to encourage data sharing to support the training and testing of AI algorithms. This would promote generalizability of these algorithms to widespread clinical practice and mitigate unintended bias.
• Establish tools for validating and monitoring the performance of AI algorithms in clinical practice, to facilitate regulatory approval.
• Develop standards and common data elements to facilitate seamless integration of AI tools into existing clinical workflows.

In defining and prioritizing AI use cases, the medical imaging community should describe exactly what's important to radiology and what data scientists -- including researchers and developers -- can do to improve patient care, according to the authors.

"Those descriptions should go beyond narratives and flowcharts," they wrote. "Human language should be converted to machine-readable language using standardized data elements with specific instructions for standard inputs, relevant clinical guidelines that should be applied, and standard outputs so that inferences can be ingested by downstream HIT resources."

Standardized inputs would enable algorithms to run on the modality, on a local server, or in the cloud. Meanwhile, application programming interfaces (APIs) could be developed based on these standardized outputs to integrate AI into any system or electronic resource, according to the researchers.

Furthermore, structured use cases should include specifications for data that should be collected to inform the developer how the algorithm performs in actual clinical use, according to the researchers.

"Understanding performance variances that occur in different patient populations, across different equipment manufacturers, or using different acquisition protocols can then be used to refine the algorithm, modify the use case specifications, or inform regulatory agencies," they wrote.