A recent study published in Oncotarget introduces a deep learning-based tool for PET/CT attenuation correction using AI technology. This tool aims to reduce the need for low-dose CT scans in oncology patients undergoing treatment follow-up. The research team from the National Cancer Institute developed a deep learning algorithm based on Pix-2-Pix GAN architecture to generate attenuation-corrected PET images from non-attenuation-corrected PET images. The study involved 302 prostate cancer patients and showed promising results in reducing the need for CT scans while maintaining image quality and quantitative markers. The AI-generated PET images demonstrated high correlation with original images, as indicated by various metrics such as NMSE, MAE, SSIM, PSNR, ICC, and RC. The study highlights the potential of AI technology in improving imaging processes and reducing the burden of radiation exposure in cancer patients. The research was published in Oncotarget and can be accessed for further information.
Source link
Source link: https://medicalxpress.com/news/2024-05-deep-based-body-psma-petct.amp
GIPHY App Key not set. Please check settings