Non Contrast
Non-contrast imaging, using techniques like non-contrast CT and MRI, aims to improve medical diagnosis and monitoring without the need for contrast agents, reducing risks and costs. Current research heavily utilizes deep learning, employing architectures such as U-Nets, convolutional neural networks, and transformers, to analyze non-contrast images for automated segmentation, lesion detection, and quantitative measurements of various organs and pathologies (e.g., coronary calcium, kidney volume, stroke lesion volume). These advancements offer the potential for faster, cheaper, and safer diagnostic tools across numerous medical specialties, improving patient care and streamlining clinical workflows. The focus is on improving the accuracy and generalizability of these AI-driven methods across diverse patient populations and imaging protocols.
Papers
Cluster-Induced Mask Transformers for Effective Opportunistic Gastric Cancer Screening on Non-contrast CT Scans
Mingze Yuan, Yingda Xia, Xin Chen, Jiawen Yao, Junli Wang, Mingyan Qiu, Hexin Dong, Jingren Zhou, Bin Dong, Le Lu, Li Zhang, Zaiyi Liu, Ling Zhang
Identification of Hemorrhage and Infarct Lesions on Brain CT Images using Deep Learning
Arunkumar Govindarajan, Arjun Agarwal, Subhankar Chattoraj, Dennis Robert, Satish Golla, Ujjwal Upadhyay, Swetha Tanamala, Aarthi Govindarajan