Dual Registration of Abdominal Motion (DRAM) is one of our most recent developments designed to allow the analysis of motility in free-breathing data sets. Organs including the stomach and colon have dedicated contraction patterns that take place over minutes making breath-hold protocols unfeasible.
Small bowel respiration correction
In the examples on this page, a relatively low-resolution small bowel data is processed with DRAM. 1) Is raw free breathing data, 2) is DRAM registered displaying the removal of respiratory motion. 3) shows again raw free breathing data this time in the colon and 4) shows the respiratory motion correction following DRAM. The corrected data can then be analysed as per GIQuant to extract the parametric motility information or propagate annotations for ROI analysis.
Large bowel respiration correction
Respiratory motion occurs largely in the head-toe direction and causes a significant problem for manual investigation of motility especially in time series containing large numbers of images where the analysis becomes impractical. DRAM uses a pre-processing step based on robust principle component analysis (RPCA) to remove the comparatively 'low-rank' respiratory data from the the local deformation caused by bowel wall movement and then apply GIQuant to either propagate regions of interest or produce parametric motility maps.