Segmentation of Contrast Enhancing Gliomas and Whole Brain Multi-Parametric Mapping in a 6-minute Scan without Injecting Contrast

Multiple brain MRI sequences are routinely applied for tissue characterization and lesion detection, by acquiring multi-contrast anatomical images. We developed a single continuous data acquisition to generate multi-contrast images with high acquisition efficiency, which allows automatic brain tissue segmentation and lesion detection without contrast injection. Multi-compartment model-based dictionary searching method was developed to derive multi-parametric mapping, including T1, T2 and macromolecular proton fraction mapping. This short imaging acquisition (6 mins versus conventional ~30 mins scan time) without use of contrast injection and the efficient processing pipeline (automatic segmentation and quantitative mapping) can provide great potentials for imaging brain tumor patients.