Intestinal “Piggybacking Lipoma”, A Unique Lipoma Composed of Lipoma and Overlying Epithelial Lesions: A Case-control Study and Review of Literature

Abstract

Background/Aim

Lipomas are rare but the most common benign mesenchymal lesions of the gastrointestinal (GI) tract, composed of mature adipose cells. The “piggybacking lipoma” is formed by lipomas with overlying polypoid epithelial lesions, such as sessile serrated lesion, tubular adenoma, or hyperplastic polyp, and the literature on these lesions is limited. In this study, we systematically investigated the clinical, endoscopic, and pathologic characteristics of these unique lipomas.

Patients and Methods

This is a single-institution retrospective study of gastrointestinal tract lipomas diagnosed from 2016-2021. Those with concurrent polypoid epithelial or mesenchymal lesions during the same endoscopic episode were included and reviewed in this study, and the lipomas were classified as “piggybacking lipoma” or “non-piggybacking lipoma” depending on whether the concurrent lesion was overlying the lipoma or was at a different location in the intestine. Demographic, clinical, and endoscopic data were obtained from electronic medical records.

Results

A total of 100 lipomas with concurrent epithelial or mesenchymal lesions were included in this study. Among them, 21 cases were classified as “piggybacking lipoma” and 79 were classified as “non-piggybacking lipoma”. Patients with piggybacking lipomas showed a female predilection, and were more likely to be symptomatic and less likely to exhibit classic endoscopic features of lipoma. Histologically, the piggybacking polyps showed overlying sessile serrated lesions (SSL) (76.2%) and tubular adenoma (TA) (19%), whereas the non-piggybacking group had differing characteristic lesions with TA (57.5%) and SSL (6.0%).

Conclusion

Piggybacking lipomas are rare lipomas with overlying polypoid epithelial lesions, most commonly SSL. They present different clinical, endoscopic, and pathologic features compared to non-piggybacking lipomas.

Lipoma is the most common mesenchymal polyp and the second most common benign colon tumor after adenomatous poly (1). Since they are often small and asymptomatic, they were mostly detected during autopsies in the past (2). However, in recent years with an increasing number of screening colonoscopies, there has been a rise in the incidence of these lesions (3). Those over 2 cm can present with anemia, abdominal pain, constipation, intestinal bleeding, and intussusception (4).

Endoscopically, lipomas can be sessile/flat or polypoid/ pedunculated. Histologically there is a well-circumscribed proliferation of benign mature adipocytes with variable amounts of fibrous stroma and an intact overlying colonic epithelium (5). Various morphological subtypes have been described in the literature based on the location of fat, architectural appearance, and associated histological features (2). Adipose tissue proliferation centered in the mucosa blending with the crypts is termed “intramucosal lipoma” and is associated with Cowden syndrome (6). Those with adipose tissue proliferation beneath the muscularis mucosae are termed “submucosal” and are the most common gastrointestinal (GI) tract mesenchymal tumors. They frequently occur in older females and the right colon (7). Other subtypes like angiolipoma and fibrolipoma have also been described (2).

The overlying mucosa can show ulceration, necrosis, and reactive changes mimicking a carcinoma (8). Additionally, there can be hyperplasia of the overlying mucosa. Rarely adenomatous polyps like tubular or tubulovillous adenomas can grow over the surface of a lipoma (9). Only isolated case reports have documented such polypoid epithelial lesions overlying a lipoma. In this study, we systematically investigated the clinic, endoscopic and pathologic characteristics of these unique lipomas.

Patients and Methods

Data collection. Lipomas of the gastrointestinal tract diagnosed at one academic institution from 2016-2021 were retrieved and reviewed. The study was approved by the university Institutional Review Board (#STUDY00007060) and performed in accordance with the Declaration of Helsinki. Those with concurrent polypoid epithelial or mesenchymal lesions during the same endoscopic episode were included in this study. The entire study cohort was divided into two groups. Cases with the concurrent epithelial/mesenchymal polyp and with lipoma occurred at the same site were classified as “piggybacking lipoma.” The other group was termed “non-piggybacking lipoma” which comprised cases in which the lipoma and concurrent epithelial/mesenchymal polyp occurred at different intestinal locations during the same procedure. Clinical, demographic, and endoscopic data were obtained from electronic medical records.

Statistical analysis. Univariate descriptive statistics are presented as means with ranges. Continuous variables were compared using the independent t-test. Categorical variables were analyzed using Fisher’s exact or chi-square test. We used a p<0.05 threshold for statistical significance. All statistical analyses were performed using GraphPad Prism (GraphPad Software, Boston, MA, USA).

Results

A total of 400 GI tract lipomas were diagnosed at our institution from 2016 to 2021. Among these, 100 (25%) lipomas from 98 patients were found to have concurrent polypoid epithelial or mesenchymal lesions. These lipomas were commonly seen in elderly people with a median age of 63.9 years. No sex predilection (51 female vs. 47 male) was observed. The lipomas most commonly occurred in the right colon (58%), followed by the left colon (36%) and small bowel (6%). They were predominantly located in the submucosa (96%), with rare cases in the mucosa (4%). Among the four intramucosal lipoma, one was found to be associated with Cowden’s syndrome.

Twenty-one out of these 100 cases (21%) presented with overlying epithelial lesions on the top and with submucosal or intramucosal lipoma on the bottom, and thereby were classified as “piggybacking lipomas”. The remaining 79 cases were defined as “non-piggybacking lipomas”. The age of patients did not differ significantly between these two groups (61.8 vs. 64.4, p=0.152). However, compared to non-piggybacking lipomas (34 women vs. 43 men), piggybacking lipomas were more common in women than in men (17 women vs. 3 men; p=0.003). Patients with piggybacking lipomas were more likely to present with clinical symptoms (42.9% vs. 15.9%; p=0.030). The most common symptom is abdominal pain, followed by diarrhea and anemia.

Endoscopically, colonic lipomas usually can be easily differentiated from other submucosal lesions by their unique characteristic features. They are yellow-colored, soft, smooth and with a “pillow sign”. In this study, 56.8% of non-piggybacking lipomas exhibited these endoscopic appearances and were diagnosed as “lipoma” by gastroenterologists. In contrast, these classic features were identified in only 4.8% of the piggybacking polyps (Table I).

Table I. Demographic, clinical, endoscopic, and histopathologic features of piggybacking and non-piggybacking lipomas.

There was no significant difference in the size of these polyps (10.5 mm vs. 10.7 mm, p=0.462). Most piggybacking lipomas were located in the ascending colon (66.7%), followed by descending colon (23.8%) and transverse colon (9.5%), which was not significantly different from the non-piggybacking lipomas (p=1.000) (Table I).

Histologically, concurrent epithelial or mesenchymal polyps/lesions in the non-piggybacking lipoma group consisted of tubular adenoma (57.5%), hyperplastic polyp (26.9%), sessile serrated lesion (6%), tubulovillous adenoma (1.5%), inflammatory polyp (3.0%), ganglioneuroma (1.5%) and mucosal Schwann cell hamartoma (0.7%). In contrast, the piggybacking lipomas showed overlying sessile serrated lesions in a majority (76.2%) of the cases. Tubular adenoma and hyperplastic polyps were identified in 19% and 4.8% of the cases, respectively (Figure 1). Overall, the difference in associated lesions between piggybacking and non-piggybacking lipomas was highly statistically significant (p<0.001). Among the four intramucosal lipomas, one was a piggybacking lipoma with an overlying tubular adenoma.

Figure 1. Histologic features of the “Piggybacking lipoma”. A) Lipoma with overlying sessile serrated lesion. B) Lipoma with overlying tubular adenoma. C) Lipoma with overlying hyperplastic polyp. D) Intramucosal lipoma with overlying tubular adenoma.

Ten of the twenty-one piggybacking lipomas had follow-up colonic biopsies (median 3 years, range=1-5 years). Of note, none of these cases showed either recurrent lipomas or piggybacking lipomas, although recurrent epithelial lesions were identified.

Discussion

Lipomas are a common benign lesion composed of well-defined proliferating mature adipocytes with varying amounts of fibrous stroma. Most lipomas are asymptomatic, however, large lipomas can cause GI obstruction and abdominal pain. In this study, we characterized an interesting and previously rarely reported phenomenon, “piggybacking lipoma”, which consists of a lipoma with an overlying polypoid epithelial lesion. To the best of our knowledge, this is the first study that systematically investigated the clinical, endoscopic, and pathologic features of these unique lipomas.

Radhi et al. first described a lipoma with changes in the overlying mucosa reminiscent of a hyperplastic polyp (10). Subsequently, a few similar cases of colonic lipomas with overlying hyperplastic, sessile serrated, tubular or tubulovillous polyps and even adenocarcinoma were also reported by other authors (4,5,9,11-16) (Table II). In this study, 25% of all lipoma cases diagnosed at our institution presented with concurrent polypoid epithelial or mesenchymal lesions. They were classified into piggybacking (5%) and non-piggybacking lipomas (20%). Piggybacking lipomas were more commonly seen in older females and frequently occurred in the ascending colon. Both submucosal and intramucosal lipomas were found in the piggybacking lesions. Interestingly, patients with piggybacking lipomas were more likely to present with abdominal pain, change in bowel habits, anemia, GI bleeding and positive fecal occult blood test, as seen in previously published case reports (Table II).

Table II. Literature review and case report summary of demographic, clinical presentation, histopathologic features, and treatment of GI lipomas with overlying epithelial lesions.

Smooth yellow appearances and a “pillow sign” (a surface indentation observed upon pushing the lesion with a biopsy forceps) are classic characteristics for a lipoma during colonoscopy (17). Most times, gastroenterologists can recognize them by these features. However, only 4.8% of the piggybacking lipomas exhibited these features, probably due to interference from the overlying epithelial lesions.

While tubular adenomas were the most common concurrent lesions in non-piggybacking lipomas, sessile serrated lesions (SSLs) were the most common overlying lesions associated with piggybacking lipomas, which was consistent with the findings in a prior study (2). Among all concurrent epithelial lesions, SSL-type was found predominantly in female patients (79.2%) in this study, in line with prior reports (18).

Tubular adenomas and serrated polyps with dysplasia are notorious for associating with colorectal carcinomas (CRC), but with different pathways of pathogenesis (19,20). The serrated pathway is characterized by BRAF mutations, CpG island methylation and microsatellite instability. Hyperplastic polyps (HP) were not considered premalignant in the past. However, recent studies have demonstrated that a subset can progress to traditional serrated adenoma and carry a shallow risk of progression to CRC. There is also some genetic overlap between HP and sessile serrated polyps, and they could very well fall within a spectrum of the same lesions with similar pathogenesis (20). Lipoma with overlying SLL and malignant transformation was described in one case report (16) (Table II).

Sessile serrated lesions are less common than tubular adenomas overall but are significantly more frequent in association with piggybacking lipomas, indicating that this is not a coincidence or a random occurrence. The underlying mechanism of pathogenesis of piggybacking lipoma remains elusive. Capra et al. speculated that chronic mucosal irritation or trauma from the passage of formed feces could result in reactive epithelial hyperplasia and, eventually, adenomatous transformation of the overlying mucosa (15). The chronic inflammatory irritation could stimulate the expression of aromatase enzyme, which promotes local synthesis of estrogens in the adipose tissue (21,22). Estrogen has been reported to lower the risk of colorectal adenomatous polyps and colorectal cancers (23,24). The protective roles of estrogen are mainly through the nuclear estrogen receptor beta (ER-β), activating apoptotic signaling, increasing DNA repair, and inhibiting expression of oncogenes (24-26). Interestingly, expression of ER-β is significantly downregulated in sessile serrated lesions (27). This might represent an underlying mechanism by which SSLs are mostly commonly identified in piggybacking lipomas in females.

So far, it is not clear whether a higher risk of malignant transformation is associated with piggybacking epithelial lesions, compared to the non-piggybacking ones. Only one case with lipoma and overlying adenocarcinoma has been found in the published literature (Table II). In our limited follow-up data, none of the piggybacking lipoma cases showed recurrent lipomas or piggybacking lipomas. However, a careful and complete pathological evaluation of these lesions may be warranted.

Conclusion

Colonic lipomas are rare, benign, and usually asymptomatic. Some of them can form piggybacking lesions with overlying epithelial lesions, including sessile serrated lesion, tubular adenoma, or hyperplastic polyp. They present different clinic, endoscopic and pathologic features, compared to non-piggybacking lipomas. Considering the overlying premalignant nature of adenomas, further study exploring the pathogenesis and molecular profile of these piggyback lesions is warranted.

Conflicts of Interest

The Authors have no potential conflicts (financial, professional, or personal) of interest to disclose in relation to this study.

Authors’ Contributions

MR: acquisition of data; drafting of the manuscript. ME: study concept and design, drafting of the manuscript. XL: study concept and design, drafting of the manuscript. YH: study concept and design, acquisition of data; drafting of the manuscript, administrative support, study supervision.