The Detection of Gastrointestinal Inflammation with Microultrasound and The Feasibility of Ultrasound Capsule Endoscopy

Abstract

Crohn’s Disease (CD) is a lifelong inflammatory disorder of the gastrointestinal tract. It may manifest anywhere along the gut including anatomically remote sections of the small bowel. The principal motivation behind the work described here is to meet the clinical need for improved means of diagnosis and management of inflammatory disorders of the bowel, such as CD. Potentially, improvements can be met by augmenting capsule endoscopes with microultrasound. Microultrasound (µUS) frequencies are > 20 MHz and have the potential to provide highly detailed transmural images of the bowel wall. This would provide capsule endoscopes with a means of detecting and displaying subsurface disease and, provide clinicians with the means of treating the disease to achieve complete or histological remission.

Investigating the feasibility of ultrasound capsule endoscopy (USCE) utilised explanted human tissue, in vivo pig trials and mouse experiments. Human tissue experiments were aimed at identifying a suitable µUS frequency for USCE and examining the relationship between acoustic and histologic bowel layers. Pig experiments aimed to demonstrate the technological feasibility of USCE. The primary endpoint was to determine if sufficient coupling occurred between capsule ultrasound transducer(s) and the mucosa to generate a transmural µUS bowel image. Mouse experiments were aimed at determining whether µUS could directly detect inflammation via the infiltration and accumulation of white blood cells. This also included what was the lowest grade of inflammation detectable.

Chapter 4 describes the results of scanning explanted human colon with different microultrasound frequencies. The purpose of this pilot study was twofold. One aim was to tentatively identify a suitable frequency for inclusion into USCE. The other aim was to examine the relationship between acoustic tissue layers and histological layers. Surgically acquired tissue was scanned at various µUS frequencies in three different orientations with respect to the µUS transducer. This was done to determine the effect layer interface and layer components on acoustic layer generation. A total of five cases were included in this study and tissue was collected exclusively from the colon. Neoplastic disease was the reason for all five procedures. Tissue was collected approximately 20 cm from the tumour and presumed healthy. Three cases (1-3) were transmurally scanned with the transducer facing the mucosa (i.e. mucosa to serosa image). Microultrasound was able to depict multiple acoustic layers. Case 1, in particular, was able to depict up to 7 acoustic layers with full tissue penetration at 34 MHz. Case 4 was scanned in a novel orientation where the histologic layers were scanned individually. This permitted each layer to generate a signal based on its composition without contribution from layer interfaces. Five acoustic layers were generated at 25 MHz and seven acoustic layers were detected at 37 MHz. Case 5 was scanned transmurally from serosa to mucosa at 37 MHz. Five distinct layers were detected. Results indicated that a frequency ≈35 MHz was suitable for inclusion. This was due to the high degree of superficial tissue detail and both frequencies demonstrated sufficient transmural penetration. The second aim, to further explore the relationship between acoustic and histological layers, demonstrated that scanning in various orientations represents a practical means of examining the ultrasound-histology relationship.

Chapter 5 describes the results obtained with a prototype tethered USCE device in vivo addressing the feasibility of USCE in the gastrointestinal tract. The primary clinical endpoint of this series of experiments was to determine whether sufficient coupling occurred between the capsule transducer(s) and the gut mucosa of pigs to generate an ultrasound image. A total of 18 in vivo experiments were carried out with SonoCap, a tethered USCE prototype. Experiments were conducted in the oesophagus (N=6) and small bowel (N=12) in nonrecovery anaesthetised pigs. Scans were done in both static and dynamic capsule modes. Images were generated of the oesophagus in static and dynamic modes indicating sufficient transducer to tissue coupling occurred. Images were also generated of the small bowel in static and dynamic modes indicating capsule transducer to mucosa contact. Similar results were also achieved using an alternatively designed USCE prototype from colleagues at Shenzhen University. Results showed coupling between the capsule’s acoustic window and transmural images of the oesophagus and small bowel. Evidence of USCE feasibility was observed in both prototype capsules indicating that further research and capsule development is warranted.

Chapter 6 describes the results of mouse gastrointestinal (GI) inflammation experiments. The purpose of these experiments was twofold. The first objective was to determine whether high resolution microultrasound could directly detect leukocyte infiltration during GI inflammation. The second objective was to determine the lowest grade of inflammation detectable. Stage 1 was able to demonstrate the ability of μUS to detect overt or severe (i.e. visually perceptible) inflammation at 37 MHz. Subsequent attempts at 37 MHz (Stages 2 and 3) to detect low grades (i.e. mild and moderate inflammation) were unsuccessful. This was despite histopathologic signs of inflammation upon microscopic examination. Further experimentation at 62 MHz was also unable to detect lowest grade of inflammation. Although sever inflammation was detected, low grade inflammation may have been beyond the capabilities of the two frequencies used in this series of experiments. Improved methods of co-registration of the two modalities, µUS scanning and histology, would assist with additional experiments of this nature.

The research described here examined a number of facets in the development of USCE in order to fulfil a clinical need for an improved means of diagnosis and management of CD and has succeeded in demonstrating areas of potential utility for further study.

Date of Award	2020
Original language	English
Awarding Institution	University of Dundee
Sponsors	Engineering and Physical Sciences Research Council
Supervisor	Inke Nathke (Supervisor), Bob Steele (Supervisor) & Sandy Cochran (Supervisor)