Figure 4

Ultra-short echo-time imaging after ferumoxytol administration. Example of how ultra-short echo-time imaging can allow for detection of tissue anatomy even in the presence of high tissue iron content. The top row again reflects the same phantom as in Figure 1, imaged with the VIBE-based sequence (left panel). The middle panel shows the same phantom but now imaged with an ultra-short TE-based sequence, with the resulting anatomical image fairly unaffected even by high amounts of ferumoxytol aggregates. The right panel shows the mu-map derived from the uTE-based MR data. Below each phantom dataset is an example of anatomical data that can be obtained in vivo with the same sequence. It can be observed that in the left panel, the mu-map is incorrect due to high liver iron content, causing the liver to be classified as lung when using the VIBE-based mu-map. However, due to the very short TE, the uTE sequence is able to detect signal in the liver in the same scan which in turn is reflected by high signal in the uTE-derived mu-map. However, it can be observed that the uTE-derived mu-map lacks detail elsewhere in the body (for example, the tissue of the left lung is incorrectly segmented) and therefore requires further optimization for this purpose. The mu-map quality in the liver is reflected by the SUV values measured from the corresponding PET datasets (data not shown) with the uTE-based mu-map causing a slight overestimation of the observed SUV compared to baseline (+21%); however, this difference is much smaller than the underestimation due to the iron-induced artifacts in the VIBE-based mu-maps (−53%).