Out-of-bore application – Qualitative validation

Qualitative validation was performed, in Volunteers E-H (Table 1), by comparing synthetic MR images to optically-tracked USI data. At least 4 short sequences (< 1 min) of simultaneous OCM and optically-tracked USI were performed for each volunteer. Although the stream of acquired MRI data had stopped, the creation of synthetic MRI continued based solely on OCM signals and previously learned correlations. Time-matched sagittal synthetic MRI and acquired USI slices were obtained, but these slices did not match spatially thus complicating the task of comparing them. Only for the left lobe were we able to acquire sagittal US liver images throughout the breathing cycle. Because of the upward curvature of the ribcage toward the solar plexus, the left lobe was seen even at expiration. The chosen MRI sagittal plane, through the main bulk of the organ, was typically more to the right than the USI plane through the left lobe. Furthermore, MRI and USI gave drastically different tissue contrasts; after all, MRI and USI being so notoriously difficult to register was part of the rationale for the present hardware-based solution to begin with. For all these reasons, it was not possible, from our data, to track the same landmarks in both modalities, for quantitative validation purposes. On the other hand, qualitative comparisons of the timing and amplitude of displacements was possible:

Between 1 and 3 clearly-visible structures, e.g. blood vessels and/or the lung-liver boundary, were manually selected in a master frame and automatically tracked in time with template matching. Optical tracking of the USI transducer allowed the in-plane motion of the hand-held probe to be accounted for. Results from all landmarks were averaged for each subject.