Single-breath-hold 3D abdominal metabolic MRI enables label-free diagnosis of liver cancer

Abstract

Abstract Chemical exchange saturation transfer (CEST) MRI could detect proteins/peptides, creatine, glucose, and glycogen by labeling their exchangeable amide, amine, and hydroxyl groups respectively, via frequency-selective RF pulses. Without the need for contrast agents or specialized hardware, CEST can be conveniently integrated into existing clinical MR protocols. However, its abdominal application is limited by long scan time (> 5 min) and susceptibility to respiratory motion (60–70% successful scan rate). We develop an ultra-fast 3D CEST MRI approach using spatial-spectral encoding (SSE), enabling a full spectral scan of whole-liver 3D images within a single breath-hold. SSE-CEST employs an efficient z-ω encoding pattern by applying a saturation gradient, followed by a data-driven spatial spectral reconstruction based on the low-rankness of CEST spectra. SSE-CEST is comprehensively evaluated in glycogen phantoms, ex vivo porcine liver, healthy volunteers and patients. Single breath-hold SSE-CEST largely improves successful rate, with a correlation of 0.95 between two repeated scans. SSE-CEST enables the detection of multi-metabolite changes in the liver and pancreas after an overnight fasting, and the dynamic mapping of hepatic glucose metabolism during an oral glucose test. For liver cancer patients, SSE could differentiate active lesions from post-treatment necrosis, featuring superior in-slice spatial resolution and motion-stabilized images. SSE-CEST MRI potentially could facilitate the diagnosis and patient management for liver and other abdominal diseases.