PTH-243 Intra-abdominal adhesions: identifying the “sticking point(s)” using novel cine-magnetic resonance image analysis. pilot results

Abstract

<h3>Introduction</h3> Peritoneal adhesion formation is a common complication of abdominal surgery and is associated with chronic abdominal pain, bowel obstruction, female infertility and reoperation complications. Non-invasive diagnosis has the potential to improve patient management by determining the cause of symptoms. Cine-MR (cMR) has been used for this purpose¹but in our experience has proven labour intensive and with high inter-operator variability. We previously presented a mathematical interpretation of abdominal movement in cMR to infer gross abnormalities.²We have since refined our approach to detect more subtle changes. Here we present the preliminary results of our new technique. <h3>Method</h3> Sagittal cMR images were obtained using a previously defined scanning protocol.²Images from 3 patients were analysed, 2 with abdominal wall adhesions and 1 without. The principle driving the methodology is the expectation that an adhesion’s resistance to visceral slide produces a quantifiable reduction in shear at that location. The method entails image processing of cMR images in which visceral slide is captured over several breathing cycles. Specifically this involves: Defining a line of demarcation between the sliding regions, typically the abdominal contents and the abdominal wall. Independent interrogation of the respective modes of motion in these separate regions using image registration software. Deduction of the relative motion at the boundary between the two regions Extraction of the magnitude of shear and its presentation in a ‘sheargram’. <h3>Results</h3> The images display the accumulated shear over approximately 3 respiratory cycles in patients with and without abdominal wall adhesions. The left image (with adhesions) shows a marked reduction in shear in the lower fifth of the abdominal wall correlating to a surgically verified adhesion. The reliability of these preliminary results is also supported by the more even distribution of shear observed in a patient without anterior abdominal wall adhesions (right image). <h3>Conclusion</h3> Our pilot results demonstrate the potential of this newly developed image analysis technique. In combination with other relevant diagnostic information it facilitates the efficient localisation of adhesions to the abdominal wall. A larger study is in progress to confirm these results. <h3>Disclosure of interest</h3> None Declared. <h3>References</h3> Lienemann, <i>et al</i>. <i>Radiology</i>2000;<b>217</b>(2):421–5 Fenner, <i>et al</i>. <i>Phys Medica</i>2014;<b>30</b>(4):437–447