Practical Pathology of Hirschsprung Disease

Practical Pathology of Hirschsprung Disease The only obligate diagnostic feature of the malformation, known as Hirschsprung disease (HSCR), is absence of ganglion cells in the myenteric and submucosal plexuses of the terminal rectum. Hypoganglionosis (paucity, but not complete loss of ganglion cells), segmental aganglionosis that does not involve the terminal rectum, and various forms of neuronal dysmorphogenesis in which ganglion cells are present, but dysplastic, should not be regarded as HSCR, even though the clinical presentation may be similar. It is very likely that the pathogenesis of these conditions is distinct from HSCR. The pathology of some of these other conditions is reviewed elsewhere.[1-3] Submucosal Biopsies Aganglionosis can be diagnosed by evaluation of either submucosal or myenteric ganglia. Since, in either instance, the pathologist must establish absence of ganglion cells, adequate sampling is essential and the recognition of ancillary histopathological findings that correlate with aganglionosis can be helpful. In most centers, initial diagnosis of HSCR is based on histopathological study of suction rectal biopsies. Diagnostic biopsies must be taken at least 1-1.5 cm proximal to the anorectal squamocolumnar junction, measure >2 mm in diameter, and include sufficient submucosa. When properly oriented and sectioned adequately (75-100 levels), H&E-stained, paraffin-embedded sections are generally sufficient to exclude the presence of submucosal ganglion cells and suggest the diagnosis of HSCR. The presence of multiple hypertrophic nerve fibers (>40 µm diameter) is observed in many, but not all cases, and helps establish the diagnosis. Similarly, the frequent, but not invariable, presence of abnormally thick and numerous AChE-stained nerve fibers in the muscularis mucosa +/- lamina propria can be used to confirm the diagnosis, particularly when neither ganglion cells nor hypertrophic nerves are observed. The hypertrophic nerves that exist in most patients with HSCR arise from extrinsic autonomic and sensory fibers, which enter along with vessels from perirectal region and project for a finite distance rostrally. It is the number and diameter of these fibers that increases in HSCR, giving rise to the hypertrophic nerves that are frequently, but not always, observed in the myenteric plexus, submucosa, and mucosa of HSCR patients. In short-segment HSCR, these nerves extend into the ganglionic bowel, and are considered by some to be a marker of the transitional zone (TZ) between normal bowel and aganglionic bowel. Because these fibers only project for a finite distance proximal to the rectum, hypertrophic innervation may not be observed in biopsies taken rostrally in long-segment disease. Furthermore, extrinsic nerve hypertrophy of the rectum may not be observed in patients with combined deficiency of intrinsic ganglion cells and other peripheral ganglia (more common with long-segment HSCR) or very premature infants with delayed extrinsic innervation. On the other hand, nerve thicknesses in the normal distal rectum increase with age, and it is not uncommon to find multiple nerves with diameters greater than 40 µm in biopsies (particularly deep biopsies) from children, as opposed to young infants. The typical submucosal biopsy is 3 mm across and the average distance between submucosal ganglia is probably about 500 microns, although this increases with age and in the distalmost rectum.[3] Most submucosal ganglia contains 2-7 ganglion cells, each of which measures 20-30 µm in diameter. With 3-5 µm-thick paraffin sections, ganglia are missed in many random sections. Therefore, it is important to evaluate alot of sections (>75). With an adequate, well-oriented, not crushed, H&E-stained biopsy one can be fairly confident about the diagnosis of HSCR simply by careful examination of 1

each section. In this situation, the presence of certain ancillary features of HSCR can be quite helpful, but is not required. The problem is that very often the biopsies are suboptimal due to paucity of submucosa, crush, or both. It is in these cases, where acetylcholinesterase (AChE)-histochemistry or calretinin immunohistochemistry can be particularly valuable. AChE staining and interpretation Histochemical staining for AChE activity is a useful adjunct for the diagnosis of HSCR. The procedure is only performed with frozen sections and therefore requires a second suction biopsy, if paraffin sections are also going to be evaluated. The traditional protocol for AChE staining requires approximately 90 min,[3] but rapid procedures have been developed that require 5-10 minutes.[4, 5] AChE staining in the rectum of normal children includes staining of nerves in the submucosa and small fibers in the muscularis mucosae. Usually the latter are confined to the inner half of the muscularis mucosae.[6] If fibers are positively stained in the lamina propria, they are extremely thin and few. Most, but not all, of the rectal submucosal biopsies from HSCR patients contain more densely packed, large, AChE-positive fibers through the full thickness of the muscularis mucosae. In addition, prominent AChE-positive fibers are often present in the lamina propria. The latter finding should not be relied upon too heavily because abnormally hypertrophic nerves are only in the muscularis mucosae in biopsies from a significant subset of HSCR patients, particularly young infants. Suction Rectal Biopsies: Some Recommendations Routinely request biopsies from 3 different levels (e.g., 1-, 2-, and 3- cm above anorectal junction) Routinely cut at least 50 sections from each biopsy (e.g., 5 slides 5 ribbons/slide - 3 sections/ribbon) Cut and save two additional sections (mid-biopsy) for possible IHC (we routinely run calretinin IHC on every case) Do not hesitate to level block to exclude ganglion cells Unequivocal ganglion cells typically - have characteristic nuclear and cytoplasmic features - can be resolved at 10X - are visible in two or more adjacent levels In some parts of the world, diagnosis of HSCR is based solely on histochemically stained frozen sections of suction rectal biopsies; paraffin sections are not required.[3] The latter protocols generally stain separate sections from each biopsy for AChE and one or two markers of ganglion cell bodies, such as lactate dehydrogenase (LDH) and succinate dehydrogenase (SDH). A potential added value of this histochemical battery is that it may disclose subtle forms of enteric dysganglionosis (e.g., ganglion cell dysmaturity), apart from HSCR, which may explain the HSCR-like symptoms in a patient with ganglion cells.[3] Calretinin immunohistochemistry and interpretation Calretinin immunohistochemistry is an alternative ancillary diagnostic technique that is employed much like AChE histochemistry in the analysis of suction rectal biopsies. Calretinin is a calcium-binding protein that is expressed by a subset of submucosal and myenteric ganglion cells, some of which extend neurites into the mucosa.[7] In aganglionic suction rectal biopsies from an HSCR patient, calretinin-immunoreactive neurites in the lamina propria and muscularis mucosae are absent.[8] Be aware of the 2

fact that calretinin-positive neurites may persist in the extrinsic nerves that innervate the submucosa or myenteric plexus. The important diagnostic feature of aganglionosis is absent immunoreactive small nerves or individual neurites in the lamina propria and muscularis mucosae. In the context of appropriate positive and negative control sections, absent calretinin immunoreactivity has been shown to be equal, if not superior, to AChE histochemistry as a confirmative finding in HSCR.[9, 10] Moreover, calretinin immunostaining can be performed on a paraffin section from a biopsy used for H&E analysis, thereby eliminating any need for additional biopsies or frozen sections. Equivocal or misleading calretinin immunostaining results are rare, but do occur. Calretinin-immunoreactive nerves may be present in most proximal part of the aganglionic zone (typically less than 1 cm distal to ganglionic bowel). A biopsy from this region will be aganglionic or severely hypoganglionic, but show intact calretinin innnervation. In a suction rectal biopsy, this may indicate a low transition zone (see ultrashort-segment HSCR below). With regard to diagnosis of HSCR, experience is undoubtedly the most important variable that influences diagnostic accuracy. The pathologist must be secure with what constitutes adequate specimen, slides, orientation, etc. Above all, he/she should have the confidence to distinguish diagnostic from equivocal findings and clearly communicate the results to the clinician. In some instances, repeat biopsy is indicated, particularly if the specimen is inadequate or equivocal results are obtained. Diagnosis of HSCR commits the patient to a surgical procedure and may lead to bowel resection and/or an ostomy without any intraoperative confirmation of aganglionosis. Therefore, accurate pre-operative diagnosis is paramount. Suction Rectal Biopsies Worrisome Features and Possible Solutions Transitional or squamous mucosa - rebiopsy at multiple levels No hypertrophic nerves - level the block - AChE or calretinin Small sample of submucosal tissue - level the block - AChE or calretinin Equivocal ganglion cell(s) - level the block - destain and immunostain - rebiopsy - AChE or calretinin Intraoperative Seromuscular Biopsies Many surgical approaches to HSCR are currently employed with a trend toward one-step procedures that are transanal. In other instances, diagnosis based on suction rectal biopsy is followed by a two-stage procedure that begins with placement of an ostomy proximal to the aganglionic segment. Intraoperative seromuscular biopsies are important to determine that ganglion cells are present at the level where the ostomy or anastomosis will be placed. Generally this is accomplished by sequential seromuscular or full-thickness biopsies, from distal-to-proximal, until ganglionic bowel is sampled. 3

Seromuscular biopsies should be a minimum of 4 mm in length and extend for a depth of 2-3 mm, so as to include the longitudinal and most of the circular layers of the muscularis propria. Proper orientation of the biopsy for frozen sections greatly facilitates sampling and identification of ganglion cells. The goal is to cut perpendicular to the serosal surface, thereby visualizing the both muscle layers and their interface in the histological sections. With a well-oriented biopsy, two-to-five sections are generally sufficient to confirm/exclude aganglionosis. Recognition of ganglion cells in usually not difficult, although inflammation sometimes obscures their cytological features. Identification of ganglion cells in a seromuscular or full-thickness biopsy from a portion of the bowel circumference should not be taken as evidence that the entire circumference is ganglionic. To the contrary, studies of the distal transitional zone (TZ) show that the interface between ganglionic and aganglionic bowel is irregular, and ganglion cells are often present 2-3 cm more distally along portions (mesenteric, antimesenteric or in between) of the circumference.[11, 12] In addition, the distributions of submucosal and myenteric ganglion cells are usually correlated, but may differ by a centimeter or so in their proximal-to-distal extent. If an anorectal pull-through is performed in this interface, the bowel immediately proximal to the anastomosis will be partially aganglionic, and the patient may continue to have obstructive symptoms post-operatively. Some have recommended resection 3 or more centimeters proximal to a ganglion-cell containing biopsy to avoid this type of TZ pull-through.[13] It is unclear whether this practice is effective because no randomized controlled studies have been performed. A retrospective comparison of patients with >5 cm versus < 5 cm of resected ganglionic bowel failed to reveal any significant difference in the rate of post-operative enterocolitis.[14] Despite the fact that well designed controlled studies have not been performed, TZ pullthrough is regarded in the surgical literature as an established reason for post-operative obstructive physiology in HSCR patients.[15, 16] As a consequence, surgeons and pathologists work together intraoperatively to minimize the likelihood that neuroanatomically abnormal bowel remains proximal to an ostomy or anorectal pullthrough. In addition to distal partial aganglionosis, hypoganglionosis and hypertrophic nerves are features of the TZ, which extend for a variable distance (usually less than 5 cm) proximal to the fully aganglionic bowel. Hypoganglionic bowel consists of widely spaced myenteric ganglia with limited neuropil and reduced numbers of ganglion cell bodies. Hypertrophic nerves are identical to the large submucosal and myenteric nerves those found in aganglionic bowel, which have ultrastructural and immunohistochemical features of non-enteric peripheral nerves. The transition between TZ and normoganglionic bowel is gradual and impossible to recognize without sophisticated morphometric studies. However, resection of the overtly abnormal mid and distal portions of the TZ is a reasonable goal that can be accomplished by intraoperative frozen section examination of the entire circumference of the proximal resection margin. A well-oriented H&E-stained frozen section of the proximal margin is suitable to exclude hypoganglionosis and/or hypertrophic innervation. Recognition of these features requires experience, but is not difficult. Practice with the permanent sections of transitional zones is helpful. 4

Intra-Operative Consultation: 3 Steps Frozen sections of seromuscular or full-thickness biopsies are used to find ganglionic bowel - for each biopsy 3-5 slides with 2 sections per slide is usually adequate Advise surgeon to perform resection at least 3 cm proximal to lowest point where ganglion cells are identified Frozen section of full-circumference of proximal resection margin should be examined to exclude - large gaps between ganglia (partial aganglionoisis of 1/8 th or more of the circumference) - hypoganglionosis (widely spaced small ganglia with only 1 or 2 ganglion cell bodies and minimal neuropil - hypertrophic myenteric or submucosal nerves (more than 2 large nerves in the same quadrant with sharp round contours, no ganglion cells, and conspicuous perineureum) APPENDICEAL AGANGLIONOSIS TOTAL COLONIC AGANGLIONOSIS In addition to hypoganglionosis, submucosal hyperganglionosis (intestinal neuronal dysplasia, type B; IND) has been reported in the proximal gut of HSCR patients. IND is a controversial form of submucosal hyperganglionosis that has been defined entirely based on histochemically stained biopsies.[17] Similar changes have been observed as isolated neuropathy in some children with HSCR-like symptoms, as well as in contexts of other primary disorders of dysmotility, including the transitional zone of HSCR. The significance of Hirschsprung-associated IND is hotly debated in the literature; some authors have advocated screening for IND with frozen sections at the time of surgery so as to extend the resection proximal to the affected area. However, no compelling data exists to suggest that IND-like changes in HSCR either predict a poor outcome or should be managed any differently from isolated HSCR. Analysis of HSCR Resections The pathologist s goals in analyzing resected aganglionic gut from an HSCR patient are to confirm the diagnosis of HSCR, map the transitional zone, and assess the integrity of the nervous system at the proximal end of the resection. Confirmation that the distal gut is aganglionic is straightforward. A map of the transition zone can be completed by sampling multiple areas along the length of the resected segment to document the presence/absence of submucosal and myenteric ganglia. Some pathologists prepare rolls from longitudinal strips of the entire length of the resection.[3] I prefer to use transverse sections because the circumferential interface between aganglionic and ganglionic gut is often irregular, which will not be apparent in any single longitudinal strip.[11, 12] Full-circumference sections are needed, particularly if the aganglionic zone appears extend within two cm of the proximal resection margin. Construction of a decent map may require repeated sampling sessions. A variety of other poorly understood histopathological findings are observed in aganglionic bowel and or the transitional zone. Some of these are summarize in Table 1. While most of these findings have no established clinical and/or genetic significance, their potential correlations with specific genetic defects and / or post-operative complications have not been adequately studied. 5

TABLE 1: MISCELLANEOUS HISTOPATHOLOGICAL FINDINGS IN HSCR Histopathological Finding Location References Eosinophilic neural infiltrates Aganglionic and [18] transitional zones; transmural including nerve plexuses Submucosal arterial fibromuscular Transitional zone [19] dysplasia Loss of c-kit immunoreactive Conflicting data [20-22] interstitial cells of Cajal Peripheral nerve pattern of laminin Aganglionic and [23] expression transitional zones Loss of nnos innervation in muscularis propria Aganglionic zone [24-26] Skip Lesions and Zonal Aganglionosis It is important to be aware of two rare forms of congenital aganglionosis that deviate from the classic pattern, in which the distal rectum and uninterrupted contiguous bowel are devoid of nerve cell bodies. In the intestinal tracts of persons with zonal ( segmental ) aganglionosis, ganglion cells are present in the distal rectum, but are absent from a proximal segment of gut. In contrast, skip areas are ganglion cellcontaining segments of large intestine, flanked proximally and distally by aganglionic gut. Zonal aganglionosis is considered to be an acquired lesion (disruption) that results when ganglion cells (or their precursors) in a fully colonized segment of gut die due to ischemic, viral, immunologic, or other types of injury. The aganglionic segment can occur in small or large intestine. In some cases, a specific etiology is suggested by history (e.g., necrotizing enterocolitis) or other pathological findings (e.g., viral cytopathy). Alternatively, it has been suggested that zonal aganglionosis might result from failure of cranial and sacral neural crest-derived ganglion cells to converge in the gut wall. At the time this hypothesis was introduced, it was less certain that sacral crest cells naturally adopt an enteric neural fate. Given recent evidence that a subset of colonic neurons derive normally from the sacral crest, the proposal is more tenable. Some colleagues and I reviewed the subject of skip areas in 1995 and found that only eleven cases had been reported. Since then, I have been informed of several other cases that were encountered by colleagues, and I suspect the entity may be more common than the literature might suggest. With one exception, skip areas are located in the large intestine and bracketed by aganglionic areas that invariably include the distal rectum and appendix plus variable lengths of contiguous large and small intestine. Pathologists and surgeons must be cognizant of skip areas, and not use biopsies of the appendix as a means to diagnose total colonic aganglionosis since relatively large skip areas can be present in which ganglion cells exist. In at least two patients, the skip area was retained a functional segment of colon between small intestine and anus. Ultrashort-Segment Hirschsprung Disease Ultrashort-segment HSCR is a controversial entity that has been defined differently by various investigators. Initially, the term ultrashort-segment HSCR was reserved to 6

describe patients with clinical and radiological findings similar to those of HSCR, but with ganglion cells in their rectal biopsies. Others (myself included) reserve the diagnosis for those distal rectal biopsies that contain ganglion cells, but demonstrate an HSCR-like AChE-staining pattern, hypertrophic nerves, absent mucosal calretinin immunoreactive nerves, or some combination thereof. In any case, the underlying assumption is that an extremely short aganglionic zone in the distal rectum, inferior to the submucosal biopsy site, is responsible for obstructive symptoms. Another possibility is that such patients have functionally significant distal hypoganglionosis, which is extremely difficult to document because ganglion cells are normally sparse or absent in the terminal 1.5-2 cm of rectum. Diagnosis of ultrashort-segment HSCR requires considerable experience. I will not make this diagnosis without a series of suction rectal biopsies from different levels along the anorectal canal (e.g., 1-, 2-, and 3-cm), hypertrophic nerves in all three levels, and either calretinin immunohistochemistry or AChE histochemistry. If a diagnosis of ultrashort-segment HSCR is suspected, the surgeon (and family) need to be advised that intraoperative confirmation may not be possible and the rectal resection specimen may not yield a definitive diagnosis because the distal 1-2 cm of rectum incompletely resected during a pull through and the distal margin is traumatized by the resection procedure. Despite this disclaimer, in my experience, nerve hypertrophy and partial aganglionosis/hypoganglionosis are pretty obvious in these resections, over a distance of 2-5 centimeters from the distal margin. 7

References 1. Kapur, R.P., Motor disorders, in Pathology of Pediatric Gastrointestinal and Liver Disease, P. Russo, E. Ruchelli, and D.A. Piccoli, Editors. 2004, Springer-Verlag: New York. p. 128-56. 2. Holschneider, A.M. and P. Puri, Hirschsprung's disease and allied disorders. 2nd ed. 2000, Amsterdam: Harwood Academic Publishers. 503. 3. Meier-Ruge, W. and E. Bruder, Pathology of chronic constipation in pediatric and adult coloproctology. Pathobiology, 2005. 72: p. 1-106. 4. Kobayashi, H., et al., A rapid technique of acetylcholinesterase staining. Arch Pathol Lab Med, 1994. 118: p. 1127-9. 5. Martucciello, G., et al., A new rapid acetylcholinesterase histochemical method for the intraoperative diagnosis of Hirschsprung's disease and intestinal neuronal dysplasia. Eur J Pediatr Surg, 2001. 11: p. 300-4. 6. Moore, S.W. and G. Johnson, Acetylcholinesterase in Hirschsprung's disease. Pediatr Surg Int, 2005. 21: p. 255-63. 7. Furness, J.B., The Enteric Nervous System. 2006, Malden: Blackwell Publishing. 274. 8. Barshack, I., et al., The loss of calretinin expression indicates aganglionosis in Hirschsprung's disease. J Clin Pathol, 2004. 57(7): p. 712-6. 9. Kapur, R.P., et al., Calretinin immunohistochemistry versus acetylcholinesterase histochemistry in the evaluation of suction rectal biopsies for Hirschsprung Disease. Pediatr Dev Pathol, 2009. 12(1): p. 6-15. 10. Guinard-Samuel, V., et al., Calretinin immunohistochemistry: a simple and efficient tool to diagnose Hirschsprung disease. Mod Pathol, 2009. 22(10): p. 1379-84. 11. White, F.V. and J.C. Langer, Circumferential distribution of ganglion cells in the transition zone of children with Hirschsprung disease. Pediatr Dev Pathol, 2000. 3: p. 216-22. 12. Gherardi, G.J., Pathology of the ganglionic-aganglionic junction in congenital megacolon. Arch Pathol, 1960. 69: p. 520-8. 13. Kapur, R.P., Practical pathology and genetics of Hirschsprung's disease. Semin Pediatr Surg, 2009. 18(4): p. 212-23. 14. Haricharan, R.N., et al., Older age at diagnosis of Hirschsprung disease decreases risk of postoperative enterocolitis, but resection of additional ganglionated bowel does not. J Pediatr Surg, 2008. 43(6): p. 1115-23. 15. Langer, J.C., Persistent obstructive symptoms after surgery for Hirschsprung's disease: development of a diagnostic and therapeutic algorithm. J Pediatr Surg, 2004. 39(10): p. 1458-62. 16. Levitt, M.A., B. Dickie, and A. Pena, Evaluation and treatment of the patient with Hirschsprung disease who is not doing well after a pull-through procedure. Semin Pediatr Surg. 19(2): p. 146-53. 17. Kapur, R.P., Neuronal dysplasia: a controversial pathological correlate of intestinal pseudoobstruction. Am J Med Genet, 2003. 122A: p. 287-93. 18. Lowichik, A. and A.G. Weinberg, Eosinophilic infiltration of the enteric neural plexuses in Hirschsprung's disease. Ped Pathol Lab Med, 1997. 17: p. 885-91. 19. Taguchi, T., et al., Abnormally shaped arteries in the intestine of children with Hirschsprung's disease: etiological considerations related to ischemic theory. J Pediatr Gastroenterol Nutr, 1994. 18: p. 200-4. 20. Vanderwinden, J.-M., et al., Interstitial cells of Cajal in human colon and in Hirschsprung's disease. Gastroenterology, 1996. 111: p. 901-10. 21. Taguchi, T., et al., An abnormal distribution of C-kit positive cells in the normoganglionic segment can predict a poor clinical outcome in patients with Hirschsprung's disease. Eur J Pediatr Surg, 2005. 15: p. 153-8. 22. Newman, C.J., et al., Interstitial cells of Cajal are normally distributed in both ganglionated and aganglionic bowel in Hirschsprung's disease. Pediatr Surg Int, 2003. 19: p. 662-8. 23. Alpy, F., et al., The expression pattern of laminin isoforms in Hirschsprung disease reveals a distal peripheral nerve differentiation. Hum Pathol, 2005. 36: p. 1055-65. 8

24. Bealer, J.F., et al., Nitric oxide synthase is deficient in the aganglionic colon of patients with Hirschsprung's disease. Pediatrics, 1994. 93(4): p. 647-51. 25. Kobayashi, H., H. Hirakawa, and P. Puri, NADPH-diaphorase histochemistry: a reliable test for the intraoperative diagnosis of Hirschsprung's disease. J Pediatr Surg, 1996. 31: p. 1552-3. 26. O'Kelly, T.J., et al., Abnormalities of nitric-oxide-producing neurons in Hirschsprung's disease: morphology and implications. J Pediatr Surg, 1994. 29: p. 294-300. 9