At the Society of Nuclear Medicine and Molecular Imaging (SNMMI) Annual Meeting, the National Cancer Institute (NCI) presented a novel artificial intelligence (AI) method that can produce high-quality positron emission tomography (PET)/computed tomography (CT) images while decreasing radiation exposure to patients.
Patients with cancer often need imaging tests for diagnosis and treatment, such as repeat PET/CT scans. The CT portion of the test exposes patients to radiation, but the research team said that the CT portion of the test is redundant. The new method bypasses the need for CT-based attenuation correction, according to the presenters, which can allow more frequent PET imaging to monitor disease progression and treatment effectiveness without repeat radiation exposure from CT.
The researchers used an AI model to generate virtual attenuation–corrected PET scans in 305 PET/CT studies. Each study involved 3 separate scans: non-attenuation–corrected PET, attenuation-corrected PET, and low-dose CT.
Two nuclear medicine physicians reviewed 40 PET/CT studies. The studies were randomized, and the physicians were blinded to the type of image. The physicians noted the number and locations of PET-positive lesions and reviewed overall noise and image quality. The readers successfully detected lesions on the generated PET images with reasonable sensitivity.
“High-quality AI-generated images preserve vital information from raw PET images without the additional radiation exposure from CT scans,” said Kevin Ma, PhD, a postdoctoral researcher at the NCI in Bethesda, MD. “This opens opportunities for increasing the frequency and number of PET scans per patient per year, which could provide more accurate assessment for lesion detection, treatment efficacy, radiotracer effectivity, and other measures in research and patient care.”