Photon-counting CT (PCCT) technology has potential for increased spatial resolution and decreased noise compared to traditional energy integrating CT in detection of liver lesions, according to research presented Sunday.
Wildman-Tobriner
"With any new technology we are looking for opportunities to see if it offers improvements in our work. We wanted to investigate liver lesions because this is such a common use of CT for surveillance and oncology screening," said Benjamin Wildman-Tobriner, MD, director, abdominal imaging fellowship program and associate professor of radiology at Duke University, Durham, NC.
Duke University's Center for Virtual Imaging, which is a National Center for Biomedical Imaging and Bioengineering supported by the National Institute of Biomedical Imaging and Bioengineering, facilitated this project with virtual imaging to test PCCT. Virtual imaging trials allow researchers to test a hypothesis with fewer resources than it would take to do a similar experiment in real life.
"It's a virtual rendition of a real patient based on thousands of data points. It's really amazing and an incredible resource with which to conduct experiments like this," Dr. Wildman-Tobriner said. "Time, ethics, and money are all limited when you deal with human subjects, but you can do it in a matter of months when you do it virtually."
Ongoing Research Could Support Clinical Benefits
The study produced more than 1,200 scans using 50 anthropomorphic virtual patients, or phantoms, with between one and six liver lesions of various sizes. All scans were then independently evaluated by three subspecialty-trained abdominal radiologists using a web-based platform for image viewing and grading.
Readers marked all identifiable liver lesions in each scan, assigned a confidence rating (0-100) for every detected lesion, and gave an overall image quality score (range 0-100). Reader detection of liver lesions was compared to the known ground truth of the inserted lesions, and performance metrics (sensitivity and precision) and image quality scores were evaluated across all imaging conditions.
Mean sensitivity across all readers for lesion detection was 91.8% for PCCT and 87.7% for traditional CT. For lesions smaller than 1 cm, sensitivity on PCCT was 84.9% compared to 73.5%. Improvements were most visible using low dose radiation and in larger patients.
"We found that there was a small, but real, increase in sensitivity for detection of small liver lesions. That could have been a predictable result because of the better resolution and reduced noise on a PCCT, but it was nice to confirm with data," Dr. Wildman-Tobriner said. "We are seeing the benefits of photon-counting play out nicely in real time."
While PCCT is not being widely used in abdominal radiology yet, Dr. Wildman-Tobriner said that the ability to detect smaller lesions could have a real impact on patient care, particularly among larger patients where improvements were seen to be strongest.
"Having an accurate map of where all of a patient's liver lesions are can drive treatment decisions and make a difference in what kind of treatment they will receive," he said.
It could also be an early sign of metastasis and help physicians stage disease. Wildman-Tobriner said additional studies need to be done to compare PCCT to Dual Energy CT (DECT) which is commonly used.
"This adds to the growing body of literature about the potential uses of photon-counting CT and that we need to continue to investigate where its impact may be the most profound, both in and out of the abdomen," Dr. Wildman-Tobriner concluded.
Access the presentation, "Benefits of Photon Counting CT for Liver Lesion Detection Compared to Conventional Scanners: A Pilot Reader Study Using Virtual Imaging Trials (S2-SSGI02-2)," on demand at Meeting.RSNA.org.
