Pusan National University researchers develop breakthrough deep learning model that enhances handheld 3D medical imaging
Peer-Reviewed Publication
Updates every hour. Last Updated: 30-Dec-2025 09:11 ET (30-Dec-2025 14:11 GMT/UTC)
Conventional handheld photoacoustic and ultrasound Imaging (PAUS), while offering flexibility, offers only a narrow view of the target region, providing limited information on its structure. Alternative methods require external sensors and bulky equipment, which can also make measurements inaccurate. In a new study, researchers developed a MoGLo-Net, a deep-learning model for 3D reconstruction of 2D PAUS images. This method does not require any external sensors and can make treatments more accessible, safer, and effective.
Immune checkpoint blockade (ICB) therapies have revolutionized cancer treatment, yet many tumors remain resistant. Now, researchers at Korea University identifies the protein kinase WEE1 as a key driver of this resistance. They discovered that, outside its traditional role in the cell nucleus as tumor suppressor, cytoplasmic WEE1 fosters tumor growth and immune evasion by enhancing AKT hyperactivation. Targeting WEE1 with clinically available inhibitors may re-sensitize tumors to immunotherapy, offering new hope for treatment-resistant cancer patients.
New research explains how low levels of the electrolyte sodium in the blood can disrupt the timing of the heartbeat in patients taking widely used rhythm-control medications such as flecainide, which is commonly prescribed for atrial fibrillation and other fast or irregular heart rhythms.