The clinical data, histopathology, immunohistochemistry, treatment, and prognosis of two hospitalized patients and one consulted patient diagnosed with IGN were reviewed. In addition, one patient underwent molecular genetic analysis. This study conformed to the provisions of the institutional ethics committee and the Declaration of Helsinki (as revised in 2013). The patient’s parents shared all procedures, including treatment, and signed written informed consent. Written informed consent was obtained for the publication of any potentially identifiable images or data included in this article.

Specimens for all patients were obtained from endoscopic biopsy or surgical resection, fixed in 10% buffered formalin, dehydrated in graded concentrations of ethyl alcohol, and embedded in paraffin. Then, they underwent routine staining for hematoxylin and eosin (H&E).

Additional 4 μm sections were deparaffinized, rehydrated, and pretreated with 3% H₂O₂ to eliminate endogenous peroxidase activity. Moreover, they were treated with ethylenediaminetetraacetic acid (EDTA) (pH 9) or citrate buffer (pH 6) for heat-mediated antigen retrieval before commencing with the immunohistochemical (IHC) staining protocol. The primary antibodies used included S-100 (clone 4C4.9, ready-to-use solution), protein gene product 9.5 (PGP9.5) (polyclone, ready-to-use solution), synaptophysin (Syn) (clone MX038, ready-to-use solution), paired-like homeobox 2B (PHOX2B) (clone EP312, ready-to-use solution), neuron-specific nuclear protein (NeuN) (clone A60, ready-to-use solution), neurofilament (NF) (clone 2F11, ready-to-use solution), neuron-specific enolase (NSE) (clone 3-3-C, ready-to-use solution), SRY-related HMG-box 10 (SOX10) (clone EP268, ready-to-use solution), Vimentin (clone MX034, ready-to-use solution), Ki67 (clone MX006, ready-to-use solution), CD117 (clone YR145, ready-to-use solution), CD34 (clone QBEnd/10, ready-to-use solution), and CD163 (clone MX081, ready-to-use solution), which were purchased from http://www.maxim.com.cn (Fuzhou, China) and http://www.gzlbp.com (Guangzhou, China). The sections were incubated overnight at 4°C, followed by incubating with a general secondary antibody for 1 h at room temperature. Finally, the sections were developed with diaminobenzidine (DAB) and counterstained with hematoxylin. Omitting the first antibody was prepared as the negative control.

The EDTA-anticoagulated whole blood samples were collected from one patient and his parents. Trio-WES was performed by MyGenostics (a commercial genetic testing company) using the Illumina HiSeq X ten platform. DNA libraries were prepared with TruSeq DNA Library Preparation Kit following the manufacturer’s instructions. The raw reads were mapped to the reference human genome (GRCh37/hg19). Genome Analysis Toolkit (GATK)¹ was applied for variation calling to summarize single nucleotide variants (SNVs) and indels. The ANNOVAR software and Enliven^® Variants Annotation Interpretation System were employed for annotation and interpretation. Data were filtered in 1,000 Genome,² NHLBI Exome Sequencing Project (ESP6500),³ Genome Aggregation Database (gnomAD),⁴ dbSNP152,⁵ and Exome Aggregation Consortium (ExAC).⁶ Damage prediction of the genetic variants was conducted by Combined Annotation Dependent Depletion (CADD)⁷ for scoring and Mutation Significance Cutoff (MSC)⁸ for further comparing. The MSC server was applied to CADD, PolyPhen 2,⁹ and SIFT,¹⁰ with a 99% confidence interval and database source of HGMD and ClinVar.¹¹ Genomics England PanelApp,¹² a crowdsourcing tool, was utilized for analysis based on the variant-disease and gene-disease associations. Human Phenotype Ontology (HPO),¹³ Online Mendelian Inheritance in Man (OMIM),¹⁴ and HGMD database were used to match phenotype descriptions with variant and gene prioritization results. According to the American College of Medical Genetics and Genomics (ACMG) guidelines, genetic variants were classified as pathogenic, likely pathogenic, variants of uncertain significance (VUS), likely benign, and benign. The pathogenicity of filtered-out variants was predicted by MutationTaster¹⁵ as well. We conducted protein modeling of RTEL1 by SWISS-model¹⁶ with a UniProtKB code Q9NZ71, and the mutated structure was analyzed and visualized using PyMol.¹⁷ In addition, a total of 647 colorectal cancer samples and 51 normal controls from TCGA¹⁸ COADREAD were selected. The differential expression of RTEL1 mRNA was analyzed by the Wilcoxon rank-sum test and visualized by ggplot2 (3.3.3 version). The protein expression of RTEL1 in Caco-2 cells was displayed by The Human Protein Atlas (HPA), using an antibody coded HPA078328.

A 6-year-old boy with anal masses and bloody stools for 2 years was admitted to our hospital. The patient was born to healthy non-consanguineous parents as the second child in the family, and the sibling was healthy. No family members had similar symptoms. Physical examination on admission showed normal growth and development, with a height of 122 cm and a weight of 25.5 kg. There was no tenderness and no palpable mass in the abdomen. By finger examination, many spherical masses with diameters of 0.2–0.5 cm were palpable in the rectum, protruding from the surface of the rectal mucosa. Except for a reduction in hemoglobin (101 g/L) and red blood cell count (3.96 × 10¹²/L), other routine blood test indicators were in the normal range. Liver and kidney function and coagulation parameters were unremarkable. Colonoscopy revealed multiple polyps with paving stone-like changes in the descending colon (Figure 1A) of 10 cm from the anal orifice, extending outward to the dentate line and partly located in the rectal fold. The polyps were removed by electrocoagulation and cauterization. Gland hyperplasia, cavity expansion, interstitial blood vessel hyperplasia, necrosis, neutrophil infiltration, and hemosiderin deposits were observed in the lesions under microscopy (Figure 1B). The preliminary impression was juvenile polyps. In addition, there was a proliferation of the nerve plexus in the lamina propria and submucosa, consisting of nerve fibers, Schwann cells, and scattered ganglion cells (Figure 1C). IHC staining showed that the neoplastic hyperplasia was positive for S-100 (Figure 1D), PGP9.5, and Syn, but negative for PHOX2B. The Ki67 index was 5% approximately. Based on the findings, a diagnosis of IGN (ganglioneuromatous polyposis subtype) was made. To investigate systemic disorders correlating with IGN, an additional examination was performed. Cutaneous cafe’-au-lait spots and neurofibromas, thyroid and adrenal lumps, or trichilemmomas and “cobblestone” tongue lesions were all absent, and serum calcitonin and urine catecholamine levels were normal. However, due to the unavailability of genetic testing in this patient, NF1, MEN2B, and Cowden syndrome cannot be ruled out. The mutations in NF1, RET, and PTEN are pending to be detected. During the 3-year follow-up, the patient had multiple recurrences but with no malignant transformation.

A boy aged 1 year and 4 months with recurrent diarrhea for 2 months was referred to our hospital. Prior to admission, the patient had been hospitalized in the intensive care unit (ICU) of another hospital for 77 days due to severe pneumonia. After anti-infective therapy and invasive ventilator-assisted ventilation, the patient’s pneumonia improved. Subsequently, the patient repeatedly developed rashes and loose stools, which resembled a dilute water sample, more than 20 times a day. The boy was delivered by a full-term cesarean section and he weighed 4.1 kg at birth and was assessed as having macrosomia. The parents both claimed to be in good health, though his mother had a miscarriage in the third month of her first pregnancy for unknown reasons. No family members had similar symptoms. Compared with his peers, the patient’s growth and development were normal, with a height of 79.5 cm and a weight of 8.9 kg on physical examination at admission. Routine blood tests revealed elevated C-reactive protein (51.8 mg/L), neutrophil percentage (66.9%), and white blood cell count (8.4 × 10⁹/L), and decreased lymphocyte percentage (27.1%). Albumin (22.90 g/L), alanine aminotransferase (7.0 U/L), and lactate dehydrogenase (361 U/L) were at low levels. Routine stool examination displayed 3–5 white blood cells per high-power field (HPF), without pathogens. Metagenome sequencing screened out human cytomegalovirus (HHV-5) with high confidence. However, no nucleotide sequences of other pathogens were detected, such as Clostridium difficile. Abdominal computed tomography (CT) showed that the liver was slightly enlarged, with a reduced density. The rectum, sigmoid colon, and partial small intestine were dilated and effused, and the colon wall was slightly thickened. In addition, enlargement of multiple mesenteric lymph nodes and minimal ascites was noted. Although colonoscopy during the patient’s previous hospitalization showed erosive colitis with superficial ulcers and granulation tissue formation, our surgeon performed another colonoscopy to confirm the intestinal manifestation, when the colonoscope explored the transverse colon, a change to laparotomy was required due to difficulty in entering the region. In the surgical view, there was a moderate amount of clear ascites in the abdominal cavity. Starting from 15 cm of the flexor ligament, the small intestine exhibited multiple segmental stenoses and dilatations, with poor peristaltic function. The entire small intestinal mucosa presented pseudomembranous changes, raising suspicion of pseudomembranous enteritis. After clearing the necrotic mucosa, the entire thickness of the intestinal sample was taken and sent for pathological examination, followed by enterotomy. Histologically, significant inflammatory exudation and necrosis of the mucosal layer of the intestinal wall with ulcer formation, as well as eosinophils and histiocyte-like cells, were detected (Figure 1E). The nerve plexus in the submucosa and myenteria was significantly proliferated and enlarged, including ganglion cells, nerve fibers, and Schwann matrix (Figures 1F,G). Histiocyte-like cells in the necrotic region were immunoreactive for CD163, and neural elements were positive for Syn (Figure 1H) and PGP9.5. Hence, he was diagnosed with IGN (diffuse ganglioneuromatosis). The patient received symptomatic and supportive treatment of anti-infection and albumin supplementation. Because of severe and uncontrollable inflammation, the patient’s condition gradually deteriorated, and he unfortunately died 3 months after surgery.

Case 3, a consultation case, involved an 8-year-old boy who had a history of bloody stools for 3 months. No family members had similar symptoms. The patient had been hospitalized in a local hospital. Routine stool examination revealed occult blood positivity, though routine blood examination and coagulation function were normal. Electronic colonoscopy revealed multiple colorectal polyps. Two sigmoid colon and rectal polyps were removed, with a pathological diagnosis of juvenile polyps. Subsequently, intestinal endoscopic submucosal dissection (ESD) was performed, and 17 sigmoid colon polyps were excised. Then, the patient’s pathological slides were transferred to our Department of Pathology for consultation. Histologically, there were many acute and chronic inflammatory cell infiltrations in the intestinal mucosa, and hyperplastic nerve plexuses were seen in the lamina propria and submucosa, in which several ganglion cells were surrounded by nerve fibers and Schwann cells. The cells were positive for Vimentin, NF, NSE, S-100, SOX10, and Syn but negative for NeuN, CD117, and CD34. Therefore, the result of the consultation was IGN (ganglioneuromatous polyposis subtype). Because it was a consultation case, no more clinical examination information was obtained. The patient’s parents denied any family history of MEN2B, NF1, or Cowden syndrome, and declined genetic testing to detect any germline mutations in RET, NF1, and PTEN. The patient’s subsequent follow-up indicated recurrence. During the 2-year follow-up, there were still more than ten broad-based polypoid lesions in the descending colon, sigmoid colon, and rectum, which all indicated IGN.

The main manifestations of the 3 patients are provided in Table 1.

In Patient 2, a germline heterozygous variation, c.296C > T, was found in RTEL1 with the transcript of NM_032957 by WES. The variant was inherited from the healthy father, leading to a change of proline to leucine at the amino acid position 99 (p.Pro99Leu) in the domain of helicase ATP-binding (Figure 2A), which was not reported in the HGMD. The amino acid residue of the base substitution (p.Pro99Leu) was not conserved across various species (Figure 2B). The CADD score of c.296C > T was 12.030, predicted as an uncertain impact by MSC. Comprehensively, c.296C > T (absent from controls PM2 + insufficient evidence in silico BP4) was evaluated to be VUS according to the guideline of the ACMG, which need further study to confirm the association between the variant and phenotype. The nomenclature of variants was based on the recommendations of the Human Genome Variation Society (HGVS).¹⁹ Wild-type and mutated RTEL1 were modeled by PyMol (see text footnote 17)that showed no effect of Pro99Leu residue change on polar contact with the amino acids Gly97 and Ala101 (Figure 2C). In the TCGA database, the differential expression of RTEL1 mRNA in colorectal cancer was higher than normal control with statistical significance (p < 0.001) (Figure 2D). The protein expression of RTEL1 in Caco-2 cells was exhibited by data from HPA (Figure 2E). There were another 3 pathogenic or likely pathogenic variants identified. The heterozygous variants in F11 (c.1489C > T/p.Arg497Xaa) and NBAS (c.1514delC/p.Pro505Hisfs*15) were inherited from the healthy mother, while the FECH (c.315-48T > C/splicing) was inherited from father and mother. The genetic variants with potential significance are presented in Table 2.