PET/CT relies on physiological imaging contrast agents to generate clear images, which presents two major challenges to its clinical application: the radiation suffered by patients and the cost of producing or purchasing contrast agents. In South Africa, the limited production capacity of contrast agents has led to a sharp increase in the cost of PET/CT scans.
We developed a groundbreaking, AI-based solution called Hyper Deep Progressive Reconstruction (HDPR), which debuted at Wits Donald Gordon Medical Centre in South Africa.
After repeated practices with a large amount of ultra-high-resolution and whole-body molecular imaging data, it can accurately assist PET/CT in detecting tiny lesions at a lower dose.
With this algorithm, users only need to use 1/3 of the conventional nuclide dose to achieve the same imaging quality.
Today, more than 70 hospitals and clinics around the world have adopted HDPR technology and enjoy substantial clinical and economic benefits.
Our scientists and partners from Fukushima Medical University and Fujita Medical University in Japan have jointly discovered that HDPR can significantly reduce the dosage of contrast agents, which helps to improve image contrast and lower noise.