Histopathological dimensions differ between aganglionic and ganglionic bowel wall in children with Hirschsprung’s disease

Hirschsprung’s disease (HSCR), with a reported incidence of 1:5000, should be treated by surgical resection of the aganglionic segment and establishment of bowel continuity by a proper anorectal reconstruction [1, 2]. For diagnosing HSCR histo- and immunopathologically stained rectal biopsy is mandatory, and for determining resection length, i.e. confirming the presence of ganglionic bowel wall, it is necessary to analyze intra-operative frozen biopsy samples of bowel wall [3‐5]. The intra-operative waiting time for the results of frozen biopsy means a prolonged anesthesia time for the child, which is a particular problem if repeated biopsies are required. A faster, more precise and immediate diagnostic method at disease level is warranted. Ultra-high frequency ultrasonography (UHFUS) using 50–70 MHz frequency enabling a resolution down to 30 μm, however at the cost of restricting imaging depth to only 5–10 mm, has shown promising results in precise diagnostics for small tissue structures [6, 7] as well as in a preliminary report on bowel wall in HSCR, where the histopathological layers (muscularis interna, externa and myenteric layer) have been reported to appear differently in aganglionic compared to ganglionic bowel wall [8]. For developing and validating UHFUS as a diagnostic method in HSCR, knowledge of disease-specific histopathological effects possibly caused by an absence of ganglia cells is needed. The long-term goal is to replace intraoperative biopsies or even primary diagnostic biopsy with UHFUS diagnostics. Therefore, it is worth exploring whether any differences in thickness of bowel wall layers in aganglionic versus ganglionic bowel wall can be found. Such findings could be useful in the creation of an algorithm for UHFUS diagnostics in HSCR.

The results of the study show that, in histopathological prepared bowel specimens, the thickness of the muscularis interna was thinner, the muscularis externa thicker and the ratio of the interna/externa thickness was greater in aganglionic bowel wall. In addition, the myenteric layer, where the myenteric nerve plexus is found in normal bowel wall, tended to be thicker in aganglionic bowel wall but the thickness varied greatly especially in aganglionosis. The thickness of the whole bowel wall did not differ between aganglionic and ganglionic bowel wall, and both ganglionic and aganglionic bowel wall thickness increased with the patient’s weight, i.e. the thickness of the bowel wall seemed to grow with the patient regardless of the presence or absence of ganglion cells.

The cardinal histological feature of bowel wall in HSCR patients is the absence of ganglion cells in the myenteric layer and in the submucosa [10, 11, 13]. However, secondary effects of aganglionosis, such as histopathological dimensional differences, become relevant when exploring new diagnostic opportunities such as UHFUS in HSCR diagnostics [8]. To study the muscle thicknesses of a developing organ, e.g. the gut, in the presence of a complex, highly variable condition such as HSCR, is challenging and dimensional differences in the bowel walls between aganglionic versus ganglionic bowel have, to date, to the authors’ best knowledge, only been described in murine models [14, 15]. In one of those, when only long segment HSCR was studied, both the muscle layers were thicker in aganglionic bowel [14]. In the other study both layers were thicker in ganglionic bowel [15]. The findings were explained speculatively as hypertonicity with an increased number of muscle cells in the muscularis interna in aganglionosis [16] resulting in the development of a muscle hypertrophy due to a high working force against a contracted aganglionic segment [15]. These speculations were contradictory to our results in which we showed that the muscularis interna was thinner and the externa thicker when ganglion cells were absent. Reasons for these differences can only be speculated upon. The most obvious differences between the former studies and ours are that the former studies were performed on animal models, they included total colonic aganglionosis and that they explored bowel exposed for bowel obstruction without decompressing treatment (wash outs). In addition none of them reported on the two bowel wall’s muscle layers separately. This was in contrast to our study which exclusively explored human bowel wall with only rectosigmoid aganglionosis in children who had their obstruction decompressed before surgical resection, and that our study explored the muscularis interna and externa as separate anatomical structures. Speculatively the regular wash outs for decompression might have impacted on the thickness of the bowel wall, and maybe also the interrelationship between the muscularis interna and externa.

In our study the myenteric layer tended to be more prominent in aganglionic bowel, but the results differed considerably between patients. Also, in our UHFUS preliminary report a more prominent and uneven myenteric layer was described in imaged aganglionosis [8]. An increased thickness of this layer might speculatively and, according to some literature, be caused by hypertrophy of nerve trunks and/or increased deposition of extracellular matrix replacing the absent ganglia cells in aganglionosis [17‐19]. Such transition could vary over time and individually, or the layer’s appearance might depend on the thickness of the surrounding tissue (muscularis interna and externa), which might explain its variation. More detailed studies of the appearance of the myenteric layer on UHFUS are currently being undertaken by our research group.

In our pilot assessments it seemed that the height of submucosal amplitudes differed between aganglionotic and ganglionotic bowel. However, after multiple measurements in our main calculations and analyses, we could not verify such a difference statistically. Instead the submucosa varied greatly within each patient, and especially in aganglionic specimens. This could, speculatively, depend on the fact that aganglionosis to a larger extent can be present with various amounts of fibrosis or collagen deposits in different parts, e.g. when replacing ganglia cells. This issue is not solved and constitutes a research question in our current project.

Extrapolating the histopathological findings, presented here, to the developmental process of UHFUS in HSCR diagnostics, the capabilities of UHFUS need to be considered. The transducer of 50–70 MHz image has an optimal depth of only 10 mm and facilitates a resolution down to 30 µm [6, 7, 20]. This could be compared to the most frequently used ultrasound transducers with delivery frequencies of 2–18 MHz, capturing image depths of several cm [20] which is not suitable for a detailed exposure of the bowel wall. According to the histopathological morphometric results, the muscularis interna and externa thicknesses ranged from 0.15–1.60 mm on a depth of 2–3 mm from the serosa. Therefore, and in line with our UHFUS report [8], the UHFUS is quite capable of enabling imaging of the muscularis interna and externa. However, a more uncertain histopathological layer for UHFUS assessment might be the myenteric layer. According to our results, the myenteric layer ranged from 6 to 150 µm which means that some specimens had a myenteric layer thickness too thin for UHFUS assessment, although the myenteric layer is located at a reasonable depth of about 0.5 mm from the serosa. The submucosa was sometimes found at a depth of 3–5 mm, so if the signal becomes too weak at this depth, the 30–50 MHz UHFUS-transducer to probe deeper can be used instead, or the submucosa can be assessed from a mucosal approach. One concern that has been raised is if UHFUS in vivo examinations could be affected by colonic movements. However, in our clinical pilot studies with UHFUS during surgical procedures, the colon did not show any signs of movement at all.

Bowel wall thickness did not differ between aganglionic and ganglionic segments, and was shown to increase with the child’s weight, regardless of whether aganglionosis or ganglionosis was present. This finding impacts the development and validation of UHFUS in HSCR diagnostics, in terms of whether the same transducer frequency can be used for assessing both aganglionosis and ganglionosis. On the other hand in our study we only examined children weighing less than 10 kg, and greater weights might limit the tissue depth. Bowel wall thickness has been shown to be age dependent in healthy children examined with normo-frequent ultrasound [21, 22], but specific information on weight and bowel wall thickness are lacking, and especially in children with HSCR. In children with HSCR, it should be considered that ganglionic bowel wall could be secondarily affected by the disease, which is why our results on ganglionic bowel might not be able to be fully extrapolated to healthy children. This is supported by our finding that bowel wall thickness correlated to the time of the surgery (disease duration).

Another finding worthy of discussion, is that the thickness of the muscularis externa, in the majority of cases, was greater in ganglionosis. The absence of innervation is generally considered to diminish muscle mass, and to cause atrophy. In our results this was true for the muscularis interna, but not for the muscularis externa, at least not in most patients. Speculating, it could be that the interna and externa muscularis work as a unit: when one becomes thicker (stronger) the other gets thinner, i.e. compensating for each other. If so, the interrelationship between the interna and externa muscularis here described as the ratio, becomes important.

One strength of the study was the high quality of the tissue samples. They all contained well-preserved tissue structures of all histological layers, and together with stringent tissue handling and staining, precise differentiation of the various layers and accurate measurements were possible. A statistically relevant strength was that patients served as their own controls and therefore weight diversity did not matter for paired testing, and the paired testing increased the power of the otherwise low number of patients. Another strength was the digital imaging analysis, enabling exact consistent measurements including to three decimal places, and quality controls and re-analysis by a pathologist of the saved images. One obvious limitation was the studying of formalin-fixed specimens, when fresh bowel in vivo would be more relevant for clinical use. The histopathological preparation, including formalin treatment, has been reported to potentially alter the tissue, impacting particularly on the collagen deposits, which hypothetically could play a key role in HSCR [19] as the dimensions of the histopathological layers might be affected unequally in aganglionic versus ganglionic tissue. Such preparational effects have, to the authors’ best knowledge, never before been studied in human bowel wall but it is important to bear in mind that the histopathologic thickness may not fully correspond to that of in vivo tissue. Another limitation was the lack of control material which influences the capacity to show diagnostic efficacy compared to healthy specimens. Dimensional differences have, to our best knowledge, never been studied before in healthy human bowel but would be warranted in order to explore whether dimensional differences between the rectum and sigmoideum are evident. However, such studies on healthy children, in which the child constitutes its own control, would be difficult to undertake since healthy children very seldom require rectosigmoidal resections or sigmoid full wall biopsies. Another limitation was that the transition zone’s histopathological features with regard to histoanatomical thickness were not studied. Hypothetically the muscle layers in the transition zone could be either thinner or thicker as a result of physiological or anatomical reasons.

Despite several limitations, the histopathological results presented here will be useful and transferred to UHFUS imaging of bowel wall. In particular, the finding of the index indicating a lower thickness ratio muscularis interna/externa followed by a thinner muscularis interna will be useful. Importantly, the index here presented is only used to illustrate the interrelationship between the interna and externa muscularis in each patient. It could not be implemented clinically before several more examinations have been carried out, validating the results presented in this study. Furthermore, a very important aspect of exploring the role of UHFUS in HSCR, is that both the ultrasound examination and the expertise of the pathologist caring for the pathological material could be very in-person dependent. This means that generalization of results could be difficult, calling for a very careful interpretation.

This study demonstrates that thicknesses of the bowel wall’s muscle layers differ between aganglionic and gangionic bowel wall specimens of children with rectosigmoid HSCR. The finding of a lower thickness ratio of muscularis interna/externa and a thinner muscularis interna in aganglionosis, constitutes an index to be considered when validating UHFUS in HSCR diagnostics.