Three-dimensional Pouchography: A Proof-of-concept Study of a Breakthrough Technique for Visualising Ileoanal Pouch Anatomy and Morphology in Normal and Mechanical Pouch Complication Patients

Abstract

Background

Herein, we present a proof-of-concept study of three-dimensional [3D] pouchography using virtual and printed 3D models of ileal pouch-anal anastomosis [IPAA] in patients with normal pouches and in cases of mechanical pouch complications.

Methods

We performed a retrospective, descriptive case series of a convenience sample of 10 pouch patients with or without pouch dysfunction, who had CT scans appropriate for segmentation who were identified from our pouch registry. The steps involved in clinician-driven automated 3D reconstruction are presented.

Results

We included three normal patients who underwent CT imaging and were found to have no primary pouch pathology, and seven patients with known pouch pathology identifiable with 3D reconstruction [including pouch strictures, megapouch, pouch volvulus, and twisted pouches], underwent 3D virtual modelling; one normal and one twisted pouch were 3D-printed. We discovered that 3D pouchography reliably identified staple lines [pouch body, anorectal circular and transverse, and tip of J], the relationship between staple lines, and variations in pouch morphology and pouch pathology.

Conclusion

Three-dimensional reconstruction of IPAA morphology is highly feasible using readily available technology. In our practice, we have found 3D pouchography to be an extremely useful adjunct to diagnose various mechanical pouch complications and improve planning for pouch salvage strategies. Given its ease of use and helpfulness in understanding the pouch structure and function, we have started to routinely integrate 3D pouchography into our clinical pouch referral practice. Further study is needed to formally assess the value of this technique to aid in the diagnosis of pouch pathology.

1. Background

Three-dimensional [3D] segmentation of ileoanal pouches [IPAA] using computed tomography [CT] imaging has not yet been described.^1,2 Herein, we present the first description of radiographic 3D models of IPAA in humans.

Despite technological advances in IPAA construction, including laparoscopic, robotic, and transanal techniques,³ currently imaging relies on two-dimensional [2D] fluoroscopic and cross-sectional imaging. However, in 2D imaging, comprehensive visualisation of the pouch is not possible, because only parts of the pouch and pouch staple lines are seen on any given 2D image cross-section; herein lies the inherent value of applying 3D reconstruction methods to better understand normal and pathological pouch anatomy.

The complex 3D anatomy of IPAAs can lead to a variety of structural anomalies and ultimately pouch dysfunction. Given the ability of modern CT scans to visualise stainless steel staple lines and abnormal gastrointestinal [GI] anatomy, we aimed to use 3D technology to visualise healthy and dysfunctional pouches, to aid in the diagnosis of pathologies that result in pouch dysfunction.⁴

2. Materials and Methods

After institutional review board [IRB] approval, a retrospective descriptive case series was performed. A convenience sample of 10 IPAA patients with a diagnosis of ulcerative colitis or Crohn’s disease, who had CT scans appropriate for segmentation [ideally 1-mm resolution, pouch distended with stool or gas, without enteric contrast] were identified from the Cleveland Clinic pouch registry and included.

2.1. Segmentation and 3D printing process

For the automated clinician-generated image workflow, CT scans were imported into the TeraRecon Aquarius™ software package from our local picture archiving and communication system [PACS]. Briefly, the segmentation process [Table 1] involved the manipulation of system-generated 3D reconstructed images that included only bones, vasculature, surgical staples, and hardware. First, patient identifiers were removed. The bone tool was then used to automatically exclude bones and table. Next, the freehand region of interest [ROI] tool was used to exclude vasculature and anatomical features other than pouch staples. The pouch staples were then centred and zoomed, and the contrast was adjusted to increase the intensity of the staples. Additional freehand ROI cleanup was then performed on the contrast-adjusted images, as needed. The resultant 3D rotating models, which may be rotated along the vertical [anterior-posterior], lateral [left-to-right], and longitudinal [cranio-caudal or head-foot] axes, were then saved to the PACS system for viewing by clinicians caring for the patients. Representative images were then exported to still images and cine clips for offline annotation to highlight anatomical features [Figure 1, top panel].

The second workflow involved additional manual segmentation of the above models by a specialist [NM] to augment the automatically segmented 3D models for educational purposes, using the Materialize Mimics™ software platform. Voxels were added to fill in gaps in the staple lines, to make the staple lines continuous and suitable for 3D printing, and a semi-automated ROI capture tool was used to add to the bowel wall. These virtual 3D models were then exported to 3D portable document format [PDF] files for annotation [Figure 1, middle panel] and 3D printing of deformable models [Figure 1, bottom panel] with a translucent bowel wall and opaque staple lines using a Stratasys J850 Digital Anatomy™ 3D printer.

3. Results

3.1. Defining normal pouch 3D anatomy

We observed that 3D pouch reconstruction reliably demonstrated multiple features [Table 2] based on staple lines alone, including the pouch body staple lines, tip of the J staple line, and anorectal transverse and circular staple lines [Supplement 1: 3D Video Clip, Normal IPAA].

Pouch 3D reconstruction can also reveal the relationship between these structures; for example, if the tip of the J was constructed flush with the pouch, it appeared contiguous with the pouch body staple lines, whereas if the tip of the J is 1 or 2 cm long, the transverse staple line appears separate and to be ‘floating’ next to the pouch body. The location of the tip of the J can be cross-referenced using the operative note. Aberrant location may result from twisted pouches or occult leaks from the tip of the J.⁵ Although not seen without segmentation of the bowel wall, the location of the pouch inlet may be inferred to be opposite to the location of the tip of J.

We were also able to determine the relationship and variations between the anorectal transverse and circular staple lines, such that if the spike was placed through the centre of the transverse staple line [circle-slash manner configuration], the resultant dog-ears were visible, whereas if the spike was placed through one corner, a tennis racquet configuration was observed.⁶

Finally, the overall shape and orientation of the pouch were determined. In this case series we have observed several variations, including straight tubular, C-shaped, and reverse C-shaped, and spiralled. In terms of orientation, several normal variations were observed. Ideally, the anterior pouch-body linear staple line should be the anterior midline. However, as observed during pouch construction, the pouch may shift or be displaced to one side or the other, typically to the left with the mesentery to the right, making it difficult to obtain an exact anterior-posterior orientation. Thus, we observed cases in which the pouch was rotated to the left, but not more than 90°, in normal pouches.

3.2. Diagnosing abnormal pouch 3D anatomy in patients with pouch dysfunction

To date, we have used 3D pouch segmentation to gain insight into various mechanical pathologies.

3.3. Strictures

Pouch septum is a rare complication that occurs when the stapler linear is inserted from above and does not reach the spur leading to an undivided segment at the pouch apex, and may lead to chronic pouch outlet obstruction and consequently megapouch [dilated pouch with widening of the staple lines, and top of the pouch emanating out of the pelvis into the abdominal cavity]. In this case, enterolith formation was also observed [Figure 2, upper left panel].⁷ Pouch strictures [Figure 2, upper right panel] related to Crohn’s-like disease of the pouch are not uncommon in pouch referral practices, and may result in proximal pouch body dilation and a dilated, heart-shaped, proximal pouch.

3.4. Acute pouch volvulus

In the case of pouch volvulus, which occurs when a correctly oriented pouch acutely volvulises around the superior mesenteric artery mesenteric [SMA] axis in an organo-axial manner, may occur due to a lack of adhesions. Spiralling of the pouch body staple lines may be observed [Figure 2, lower left panel; Supplement 2: 3D Video Clip, Acute Volvulus].^8,9 Chronic pouch volvulus, in which part of the pouch pathologically adheres in an spiraled manner in the pelvis, may result in pouch dysfunction, and endoscopically and radiographically, spiralling of the pouch body staple lines mat be observed [Figure 2, lower right panel; Supplement 3: 3D Video Clip, Adhesive Volvulus].

3.5. Twisted pouch syndrome

We found that 3D pouch reconstruction is particularly useful for aiding in the diagnosis of twisted pouch syndrome [TPS]. TPS can be difficult to diagnose preoperatively, and the need to visualise the entire pouch body staple line was the inspiration for the development of 3D pouch reconstruction.¹⁰ Patients with this mechanical problem typically present with a triad of erratic bowel habits, often severe abdominopelvic pain, and obstructive symptoms [obstructive defaecation and/or small bowel obstructions]. The diagnosis is most often made at the time of exploratory laparotomy and planned pouch revision for pouch dysfunction. TPS typically results from spiralling of the distal pouch body during firing of the circular stapler, adhesions from occult tip of J leaks, or unintentional pouch construction in a 180° manner with the mesentery anterior or a full 360° from a twisted mesentery. 3D reconstructions [Figure 3] of the latter two may reveal what appears to be a normal pouch with straight pouch body linear staple lines; however, in the case of unintentional 180°, one may observe the spur with a lack of staples to be anterior instead of posterior, and in the case of unintentional 360°, the distal portion of the SMA spiralled 360°. After the redo IPAA, a normal pouch configuration was observed [Figure 4].

4. Limitations

Currently, the quality of 3D pouch reconstruction is limited by the quality of the CT scan used to derive the 3D models; cuts larger than 1 mm may have missing staples, and a lack of pouch distension hampers the interpretation of morphology. To overcome these limitations, we have developed a standard CT pouchography protocol consisting of a CT pelvis, or pelvic CT angiogram if vessel visualisation is needed, with 0.75-mm slices without enteric contrast and CO₂ insufflation of the pouch, which provides optimal imaging for 3D reconstruction. Another limitation of the current technique is the inability to automatically segment the bowel wall because its appearance is similar to that of the surrounding soft tissues. However, given the rapid developments in the field of artificial intelligence machine vision algorithms, we believe that in the near future, the manual segmentation process of the bowel wall described herein will be fully automated in commercially available software platforms, thus enabling small-bowel GI fly-throughs akin to virtual colonography.

5. Conclusion

Three-dimensional reconstruction of ileoanal pouch staple line morphology is highly feasible using readily available technology. In our practice, it is an extremely useful adjunct to diagnose mechanical pouch complications and improve surgical planning for pouch salvage surgery. Given its ease and helpfulness in understanding the pouch structure and function, we have started to routinely integrate this into our Cleveland Clinic pouch centre clinical practice. Presently we are implementing a blinded study to assess the value of this technique in the diagnosis of mechanical pouch complications.

Supplementary Data

Supplementary data are available at ECCO-JCC online.

Funding

There were no funding sources for this study.

Conflict of Interest

SDH: consulting fees, Takeda; funding: Crohn’s & Colitis Foundation; American Society of Colon & Rectal Surgeons. BLC: consultant for Takeda and Target RWE, speaker for AbbVie and Takeda. FR: Adiso, Adnovate, Agomab, Allergan, AbbVie, Arena, Astra Zeneca, Boehringer-Ingelheim, Celgene/BMS, Celltrion, CDISC, Celsius, Cowen, Ferring, Galapagos, Galmed, Genentech, Gilead, Gossamer, Granite, Guidepoint, Helmsley, Horizon Therapeutics, Image Analysis, Index Pharma, Landos, Jannsen, Koutif, Mestag, Metacrine, Mopac, Morphic, Organovo, Origo, Palisade, Pfizer, Pliant, Prometheus Biosciences, Receptos, RedX, Roche, Samsung, Sanofi, Surmodics, Surrozen, Takeda, Techlab, Teva, Theravance, Thetis, UCB, Ysios, 89Bio. The remaining authors have no disclosures.

Acknowledgements

The authors would like to acknowledge the assistance of Stephen Dombrowski PhD and Ryan Klatte, Cleveland Clinic, Lerner Research Institute, with arranging 3D manual segmentation; we are also grateful to our medical illustrators Amanda Mendelson BS, BFA and Joseph Pangrace BFA, CMI, Cleveland Clinic, for the contributing to the understanding of twisted pouch syndrome. This study was made possible with philantropic support from Mr David Horing.

Author Contributions

All authors made substantial contributions to the concept, design, and drafting of the work, and the final approval of the version to be published, and agreed to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved. An early version of this work was presented at Digestive Disease Week 2023 as an abstract.

Data Availability

The data for this work are not publicly available but may be shared upon reasonable request to the Principal Investigator..

References