1. Ultrasound assessment of bowel motility & vascularity as predictors of necrotising enterocolitis Caron Parsons

2. What is NEC?  Necrosis of wall of any part of the GI tract  Most common acquired disorder of premature infant  7 – 10% infants < 1500g  No change in mortality in last 30 years  Serious long term sequelae

3. Pathophysiology

4. Data from NDAU

5. Mortality Data From NDAU

6. Clinical Presentation Vomiting Abdominal distension Temp instability Lethargy

7. How Do We Diagnose NEC?

8. How do we diagnose NEC?  X-ray  Bowel wall thickening/loop separation  Pneumatosis – gas in bowel wall  Portal venous gas  Free gas  Non-specific  Necessary to repeat

9. X-rays ?NEC

10. Definitive NEC

11. Are there other diagnostic methods?

12. Ultrasound of NEC Bowel wall thickening Focal fluid collections Pneumatosis/ Portal venous gas Changes in Doppler signal Bowel motility

13. What have other groups done? Motility Definite diagnosis of NEC Aperistalsis Counted number of contractions 30 secs – 1 min No data presented on reliability of data acquisition Reliability of analysis performed in one study No longitudinal studies & no evaluation of infants prior to the development of NEC Muchantef, K., et al., Sonographic and radiographic imaging features of the neonate with necrotizing enterocolitis: correlating findings with outcomes. Pediatr Radiol, 2013

14. What have other groups done? Vascularity Square regions of interest Counted number of dots of colour Doppler signal/cm 2 No data on reliability of data acquisition or analysis. Faingold, R., et al., Necrotizing enterocolitis: assessment of bowel viability with color doppler US. Radiology, 2005. 235(2): p. 587-94.

15. Faingold R et al. Radiology 2005;235:587-594 ©2005 by Radiological Society of North America

16. Faingold R et al. Radiology 2005;235:587-594 ©2005 by Radiological Society of North America

17. Feasibility Study – Aims To study appearances of bowel in healthy infants on NICU/SCBU To evaluate bowel motility on ultrasound To evaluate reliability of acquiring and analysing 3D power Doppler datasets Overall to explore possible methods of quantification

18. Feasibility Study – Longer Term Aims Establish normal parameters of motility and vascularity Enable sample size calculation for longitudinal study

19. Feasibility Study – Methodology Recruit 30 healthy infants (GA <40/40) Perform single ultrasound Motility – 2D cine loops Vascularity – 3D volumes with Doppler

20. Inclusion Criteria All infants admitted to NICU or SCBU at UHCW with gestation of 40+0 weeks or less.

21. Exclusion Criteria Gestation more than 40+0 weeks Current diagnosis/ treatment of NEC Too unstable to tolerate ultrasound as dictated by the clinical team Undergoing high-frequency oscillation ventilation Known diagnosis of congenital bowel disorder Previous bowel surgery/abdominal wall defect

22. Recruitment

23. Break  There was a change to the protocol during recruitment  Therefore  Evaluate the protocol and set-up of the study  Start evaluating the image analysis methodology

24. Motility – Original Methodology Acquire cine-loops in a four quadrant approach Two observers Analyse as per previous groups Presence/absence of contractions Number of contractions/minute Computerised vectorial analysis - ShIRT

25. Motility Methodology  This part of the study did not change  2D cine loops (10 secs)  Evaluation  Basic  ShIRT – by colleagues in Sheffield

26. 2D Cine-Loop

27. Initial ShIRT Evaluation

28. Vascularity – Original Methodology Acquire 3D power Doppler volumes 2 observers 2 sets of data per observer Off-line analysis in VOCAL software Select ROI Software calculates vascularity indices

30. Vascularity – Changes To Methodology  Suggestion of 2D acquisition over time rather than 3D volume  This was then trialed on the latter 6 of the recruited patients  Availability of proprietary software for 2D Doppler data  However…..  Issues with proprietary software  Cost  Understanding of physics/ mathematics behind calculations  Difficulty with obtaining trial versions

31. 2D Doppler Cine-Loop

32. Baby with NEC

33. Initial Evaluation with Q-Flow

34. Vascularity – Where next?  We designed our own.  Using MatLab software and a very clever physicist

35. Data Analysis  The idea is to use random ROI placement within the Power Doppler data clips to sample the bowel and calculate a global measure of vascularity for each subject  Two measures  %Colour – proportion of colour pixels to grey-scale pixels in ROI  Luminance – intensity of each colour pixel  Summated for ROI

36. MatLab Software

37. Luminance: Total v Sampled

38. Why sample?  In the future we will be acquiring repeated samples on a daily basis  Bowel is mobile  The width of the Doppler box cannot be standardised

39. What Next?  Motility – Waiting for results from Sheffield  Vascularity  Statistician looking at random circle ROI data  Intraclass correlations for reliability of  Data acquisition  Data analysis  Design software for 3D Doppler volumes  Next stage of feasibility study  Pre and post feeding in healthy premature infants

40. Summary NEC is difficult to diagnose in the early stages The role of ultrasound is increasing in NEC Motility and vascularity have postulated roles in the pathophysiology of NEC There is potential to characterize motility and measure vascularity with novel software

41. Study Group Dr. E. Helm, Consultant Radiologist, UHCW Prof. C. Hutchinson, Prof of Imaging, University of Warwick Dr. V. Sherwood, Clinical Scientist, UHCW Dr. P. Satodia, Consultant Neonatologist, UHCW Dr. R. DeBoer, Consultant Neonatologist, UHCW Dr. H.Parsons, Research Fellow in Statistics, University of Warwick Dr. J. Fenner, Medical Physics Group, University of Sheffield