Unraveling the site-specific features in small intestinal stromal tumors: a retrospective study

Abstract

Background

Gastrointestinal stromal tumors (GISTs) are a rare and less well-characterized disease. There is limited information on the clinical features of small intestinal GISTs at different sites.

Aims

To enhance the understanding of the clinical characteristics and disease behavior of small intestinal GISTs based on their sites.

Methods

We conducted a retrospective review of medical records for 317 patients diagnosed with primary small intestinal GISTs confirmed by surgical pathology, comparing their clinical features and tumor characteristics.

Results

According to this cohort’s data, duodenal GISTs presented with longer disease durations and higher prevalence of melena (44.6%), while jejunal GISTs manifested as abdominal masses (11.5%) and acute gastrointestinal bleeding (GIB) (13.3%), with the highest rate of emergency surgeries (16.8%). Ileal GISTs were associated with an older age of onset and a higher prevalence of hematochezia (19.6%), with one-third of cases discovered incidentally during gynecological procedures. Notably, the biological behavior of small intestinal GISTs varied significantly by sites. Tumors demonstrated different immunochemical markers and a progressive increase in diameter, mitotic activity, T and M stages, and risk classification from the duodenum to the jejunum and ileum. These findings warrant further validation in prospective multicenter studies.

Conclusions

Small intestinal GISTs might exhibit distinct clinical presentations and oncological features depending on their sites.

Visual abstract

Highlights

WHAT IS KNOWN: small intestinal GIST exhibits worse biological behaviors and prognosis compared to gastric GIST.

WHAT IS NEW HERE: distinct clinical and oncological features of small intestinal GISTs.

Introduction

Although gastrointestinal stromal tumors (GISTs) are rare disease, they are the most common type of gastrointestinal sarcoma, accounting for 1–2% of primary gastrointestinal malignancies [1]. GISTs are primarily characterized by mutations in the KIT or PDGFRA receptor tyrosine kinases, and mostly originate from precursors of the interstitial cells of Cajal [2], which play a crucial role in gastrointestinal motility. The incidence of GISTs varies widely across the world [2, 3, 4], and reported incidence also differ significantly across different regions in China [5, 6, 7]. Many studies consistently find that the stomach is the most common site for GISTs (55–65%), followed by the small intestine (25–30%) [1, 3].

Small intestinal GISTs are often insidious and lack specific or suggestive symptoms, with many patients (5–40%) diagnosed incidentally [3]. Gastroscopy, colonoscopy, and other screening strategies cannot access most of the small intestine, which complicates early diagnosis and presents significant challenges [8]. Furthermore, long-term missed or unclear diagnoses might significantly impact patient prognosis. Although the incidence of small intestinal GIST is lower than that of gastric GIST, studies indicate that small intestinal GIST exhibits a higher rate of lymph node metastasis, a greater degree of malignancy, more frequent recurrence and metastasis and worse prognosis compared to gastric GIST [8, 9, 10], which is influenced by tumor size and mitotic activity [10].

The small intestine comprises of segments (duodenum, jejunum, ileum) that exhibit distinct characteristics in histology, anatomy, and many other aspects. However, research on small intestinal GISTs is relatively limited, and there is a lack of attention on detailing the clinicopathological features and prognosis of small intestinal GISTs of different primary anatomic sites. Therefore, the purpose of this study is to thoroughly elaborate the clinical features and biological behavior of small intestinal GISTs at different sites, with the hope of aiding the diagnosis and treatment of this unusual and less characterized disease.

Methods

Patients

Through a search and retrieval of medical records from the Hospital Information System, we retrospectively reviewed information of primary small intestinal stromal tumor patients who were diagnosed and hospitalized at our medical center between May 2012 and December 2021. Patients meeting all the following inclusion criteria were enrolled in subsequent analysis: (1) Age ≥ 18 years; (2) Pathological results of surgery specimens confirmed the diagnosis of small intestinal stromal tumor based on cell morphology and immunohistochemical findings. Patients who met any of the following exclusion criteria were excluded: (1) A prior diagnosis of small bowel stromal tumor; (2) Incomplete data regarding important clinical information. Ultimately, a total of 317 patients were included in the study.

Data collection

General information of all patients was collected, including sex, age of onset, age at surgery, duration of disease, history of other malignancies and family history of malignancy. We also documented clinical manifestations such as gastrointestinal bleeding (GIB), abdominal pain, abdominal bloating, abdominal masses, weight loss, and complications like bowel obstruction, intestinal perforation, intratumoral bleeding, tumor rupture along with hemoglobin (HGB) levels and treatment options. Detailed characteristics of gastrointestinal bleeding were recorded, including specific symptoms (melena, hematochezia, hematemesis, syncope, or shock), and classified into categories: acute GIB, transient GIB, chronic persistent GIB, and chronic intermittent GIB. Acute GIB refers to a sudden onset of significant bleeding that occurs over a short period. Transient GIB is temporary bleeding that stops on its own and typically resolves without long-term issues. Chronic persistent GIB involves continuous, long-term bleeding that does not stop, resulting in ongoing blood loss over an extended period. Chronic intermittent GIB is long-term bleeding that occurs in episodes, with alternating periods of bleeding and no bleeding.

Additionally, we paid attention to the immunochemical markers (c-KIT, also named as CD117, DOG-1, CD34, Ki-67, SDHB, SMA, S-100, Desmin) and biological behaviors of tumors, including tumor location, diameter, mitotic count per 50 high-power fields (HPF), TNM staging, and modified NIH (National Institutes of Health) risk stratification criteria [11, 12].

Statistical analysis

Continuous variables conforming to the normal distribution were presented as the mean ± standard deviation ($\:\stackrel{-}{x}$±s), while continuous variables not conforming to the normal distribution were illustrated as median (25th percentile, 75th percentile). Categorical variables were represented as percentages of the corresponding research population. The difference between groups of continuous variables conforming to a normal distribution was examined by analysis of variance (ANOVA) and Bonferroni multiple test, the difference between the groups of continuous variables not conforming to a normal distribution was examined by Kruskal-Wallis rank-sum test whereby a p value of less than 0.05 through SPSS 26.0 analysis was considered statistically significant and marked with an asterisk (*). The difference between groups of categorical variables was examined by chi-square test and Fisher’s exact test. When the p-value for a categorical variable was less than α = 0.05 in comparisons across three groups, or less than α’ = α/3  = 0.0167 between two groups, it was considered statistically significant and marked with an asterisk (*). Otherwise, it was not designated as significant and termed as ns. The figures were generated by GraphPad Prism (version 9.5.0).

Results

Among the 317 patients who underwent surgical treatment at our hospital over the past nine years, 167 were male (52.7%) and 150 were female (47.3%). There was no significant difference in sex composition among small intestinal GISTs at different sites. However, duodenal and jejunal GISTs had a slightly higher proportion of males (54.5%, 58.4%, respectively), while ileal GISTs showed a slightly higher proportion of females (56.5%). The median age at onset was 55 (46, 63) years, the median age at surgery was 56 (47, 64) years, and the median duration of disease was 90 (30, 365) days, with significant difference among different groups. Small intestine stromal tumors also exhibited distinct clinical characteristics depending on their sites, including abdominal pain (p = 0.031), abdominal mass (p = 0.005), and gastrointestinal bleeding (p = 0.025). Certain clinical manifestations were observed to present ubiquitously across different groups, including abdominal bloating, weight loss, intratumoral hemorrhage, tumor rupture, bowel obstruction, and intestinal perforation (Table 1). In our cohort, 5 cases experienced tumor rupture, which all had a diameter of ≥ 8 cm. Among them, 4 cases were spontaneous ruptures and underwent emergency surgery, while the remaining one (measuring 18.5 cm) ruptured intraoperatively due to its excessively large size.

Table 1 Summary and comparison of the clinical manifestations and treatment choices of small intestinal stromal tumors at different sites

Specifically, duodenal GISTs typically exhibit a longer disease duration, with a median of 165.0 days, compared with a median of 60.0 days for jejunal and ileal GISTs. Abdominal masses were rarely found in patients with duodenal GISTs (0.9%), whereas jejunal GISTs were more commonly associated with abdominal masses (11.5%). Ileal GISTs were characterized by an older age of onset (57.8 ± 13.4) and a higher incidence of abdominal pain (30.4%) (Table 1; Fig. 1). Regarding gastrointestinal bleeding features, patients with duodenal GISTs frequently presented with melena (44.6%) and less often with hematochezia (3.6%). Jejunal GISTs not only had the highest risk of gastrointestinal bleeding (54.9%) but also the highest risk of acute gastrointestinal bleeding (13.3%). Although not statistically significant, jejunal GISTs had the highest proportion of patients (15.9%) presenting syncope or shock. While ileal GISTs showed the lowest proportion of gastrointestinal bleeding and the lowest rate of melena (25.0%), they had the highest frequency of hematochezia (19.6%). Additionally, hematemesis was observed in 4 duodenal GISTs (3.6%) and 3 jejunal GISTs (2.7%), while no cases were noted in ileal GISTs. Additionally, the proportions of patients experiencing chronic persistent GIB, chronic intermittent GIB, and transient GIB were similar across the three groups, ranging from 10 to 20%. Chronic intermittent GIB was the most common type of GIB in all three groups (Table 2; Fig. 1).

Fig. 1

Clinical features of patients with small intestinal stromal tumors at different sites. A violin plot showed the distribution of age of onset and age at surgery; B the distribution of duration of disease and duration of gastrointestinal bleeding (GIB); C the percent of abdominal mass in different groups; D the percent of abdominal pain; E the percent of melena; F the percent of hematochezia; G the percent of gastrointestinal bleeding; H the percent of acute gastrointestinal bleeding; I the percent of gynecological surgery; J the percent of emergency surgery

Table 2 Comparison of the different features of gastrointestinal bleeding in small intestinal stromal tumors at different sites

Furthermore, from the duodenum to the jejunum and the ileum, GISTs located further along the small intestine tended to exhibit different immunohistochemical characteristics and worse biological behavior. Significant differences were observed in the positivity of DOG-1 (p < 0.001), the Ki-67 index (p = 0.013), tumor diameter (p < 0.001), mitotic count (p = 0.002), T stage (p < 0.001), M stage (p = 0.034), and risk category (p < 0.001) among small intestinal GISTs located at different sites (Table 3; Fig. 2).

Table 3 Summary and comparison of the immunohistochemical features and biological behaviors of small intestinal stromal tumors at different sites

Fig. 2

Biological behaviors of small intestinal stromal tumors at different sites. A the median, interquartile range and range of diameter of tumors in different groups; B the median, interquartile range and range of the mitotic count in 50 high power fields (HPF); C the percent of distant metastasis (M) category; D the percent of primary tumor (T) category; E the percent of modified NIH risk stratification

The positivity rates of c-KIT and DOG-1 were both greater than 90% in all small intestinal GISTs, with duodenal GISTs showing the highest positivity rate of 100% (Table 3). Small intestinal GISTs exhibited similar expression patterns of CD34, with 12–15% negative, 15–21% partially positive, 17–19% vascular positive, and 48–52% positive. The Ki-67 index was highest in ileal GISTs and lowest in duodenal GISTs. Additionally, no significant differences were observed in the expression of SDHB, SMA, S-100, and Desmin (Table 3).

The median diameter of duodenal GISTs was 3.8 cm (2.6, 5.4), smaller than that of jejunal GISTs at 5.5 cm (3.1, 8.0), and ileal GISTs at 7.5 cm (5.2, 10.0). Ileal GISTs had the highest mitotic count, with a median of 5.0 (2.0, 10.0), compared to 2.0 (1.0, 4.5) for duodenal GISTs and 3.0 (1.5, 7.5) for jejunal GISTs. The most common T stage in duodenal GISTs was T2, while in ileal GISTs it was T3. Among all patients, lymph node metastasis was observed in only one case of duodenal GIST, while distant metastasis was most frequent in ileal GISTs (duodenum: 2.7%, jejunum: 7.1%, ileum: 12.0%). Consistent with the above-mentioned results, the predominant risk category was low risk for duodenal GISTs but high risk for ileal GISTs (Table 3; Fig. 2).

In terms of treatment, the proportion of emergency surgeries was highest in patients with jejunal GISTs (16.8%), which was associated with the highest risk of acute gastrointestinal bleeding. Notably, nearly one-third of ileal GISTs were discovered incidentally during gynecological surgeries, which is rare phenomenon for duodenal GISTs (0.9%) and jejunal GISTs (3.5%). The proportion of patients receiving preoperative neoadjuvant therapy showed no significant differences among the three groups. However, the proportion of patients who received postoperative targeted treatment was significantly higher in ileal GIST patients (64%), followed by jejunal GIST patients (47%) and then duodenal GIST patients (30%) (Table 1).

Discussion

In this study, we correlated clinical symptoms with tumor location and risk factors. The results suggest that GISTs can present with varying symptoms based on different primary sites, emphasizing the need for careful attention to these clues during clinical inquiries and physical examinations. Notably, although rarely detailed in prior studies, small intestinal GISTs from different primary sites exhibit markedly different tumor characteristics, which are closely related to risk stratification and prognosis. This indicates that a more detailed risk stratification of small intestinal GIST patients is likely to be meaningful for treatment and management.

There was no significant difference in sex distribution among GISTs patients [2, 3]. While some studies have shown that GISTs were slightly more common in males than in females [1, 4], and the male sex was an independent risk factor for higher GIB risk [13] and worse overall survival [4, 7]. Our cohort shows slight sex differences across small intestinal GISTs at different sites, with a slightly higher proportion of males in duodenal and jejunal GISTs, and a slightly higher proportion of females in ileal GISTs. Besides, patients with small intestinal GISTs are generally younger, compared to the median diagnosis age in the mid-60s for most GIST studies [3, 14].

Approximately 20–30% of patients in our small intestinal GIST cohort were asymptomatic, which is consistent with previous studies on GISTs that reported a rate of 18.7% [3]. There are suggestive but nonspecific symptoms associated with different small bowel GISTs. The varying clinical manifestations of GISTs at different sites in the small intestine can be partially explained by their anatomical locations and physiological functions. Abdominal masses are less frequently observed in duodenal GISTs, likely due to the duodenum’s deep and predominantly retroperitoneal anatomical position, as well as the smaller tumor diameter. Regarding jejunal GISTs, the frequent presence of abdominal masses may be due to their location in the proximal gastrointestinal tract, where food is not fully digested and is prone to obstruction. Additionally, since the jejunum is neither retroperitoneal like the duodenum nor located in the pelvic cavity like the ileum, masses are easier to palpate.

Consistent with previous research findings, we also found that gastrointestinal bleeding and abdominal pain are the most common clinical manifestations of GISTs [15]. Research has demonstrated that small intestinal GISTs increased the odds of GIB compared to gastric GISTs(13, 16). Additionally, factors such as male and tumor sizes might elevate the risk of GIB [13]. Our findings further indicate that the characteristics of GIB vary depending on tumor location. The duodenum, being the most proximal part of the small intestine, is more likely to present gastrointestinal bleeding as melena rather than hematochezia. In contrast, bleeding occurring further down the small intestine often presents as hematochezia. While chronic intermittent GIB was the most common form across the three groups of small intestinal GISTs, jejunal GISTs were also prone to acute GIB. A study in South Korea also found that ileal tumors primarily present with abdominal pain (72.9%), while jejunal tumors mainly present with gastrointestinal bleeding (36.4%), with approximately half of these small intestinal tumors being GISTs [17]. The jejunum and ileum are supplied by the superior mesenteric artery, forming 15–18 arcades. The jejunum has fewer, larger arcades with longer vasa recta, while the ileum has more, smaller arcades with shorter and narrower vasa recta [18]. The richer blood supply in jejunum compared with ileum might partially contribute to the higher rate of bleeding in jejunum stromal tumors.

Ileal GISTs have fewer cases of melena compared to jejunal GISTs, but more cases of abdominal pain. They are more likely to present as pelvic masses and be misdiagnosed as gynecological diseases. The DOG-1 positivity rate is slightly lower, and the risk category is higher. Therefore, when a pelvic mass is accompanied by gastrointestinal symptoms or gastrointestinal bleeding, the possibility of ileal GISTs should be considered. Emergency surgery following massive bleeding in small intestinal GISTs has also been reported in many case reports [19, 20, 21]. Therefore, when encountering patients with acute or chronic GIB, GIST is a rare cause that cannot be ignored. Early diagnosis and proper treatment mainly determine the outcomes of patients presenting with GIST-related emergencies.

But the prognostic value of GIB in small intestinal GISTs remains controversial [16]. Meta-analyses have shown that GIST patients with GIB have poorer overall survival (OS) compared to those without GIB, but there is no significant difference in recurrence-free survival (RFS) [16]. Another study found that the 5-year OS rate of small intestinal GISTs without GIB was higher than that of small intestinal GISTs with GIB (81.3% vs. 66.2%, p = 0.519) through applying propensity score matching (PSM) to reduce confounding factors [22]. In contrast, small intestinal GIST patients with GIB were found to exhibit better RFS and OS rates compared to those without GIB in another PSM study [13]. Although some inconsistent results exist, gastrointestinal bleeding holds significant implications for the prognosis of GISTs and warrants close attention. Additionally, the mechanisms by which GIB affects the prognosis of GIST patients remain unclear and need to be elucidated.

The tumor location was also found to be closely related to immunohistochemical characteristics and biological behavior of the tumors. Consistent with previous literature, we also found that c-KIT and DOG-1 were the most commonly expressed markers, followed by SDHB, CD34, and SMA, with S-100 and Desmin being the least commonly expressed [2, 23]. Additionally, we observed distinct patterns of immunohistochemical marker expression across small intestinal GISTs, which may offer valuable insights into their underlying pathogenesis. Furthermore, from the duodenum to the jejunum and then to the ileum, tumor diameter increases, mitotic counts rise, and both T and M staging are higher, along with more advanced risk classifications and a higher rate of postoperative targeted therapy. A previous study [24] found that the proportion of high-risk ileal GISTs reached 66.0%, significantly higher than in duodenal GISTs (36.8%) and jejunal GISTs (43.9%). Similarly, the study noted that the mean diameter of ileal GISTs (9.77 cm) was significantly larger than that of duodenal (7.41 cm) and jejunal GISTs (8.14 cm). Although similar trends were observed, no significant differences in risk categories or tumor size were found between jejunal and duodenal GISTs in that study [24].

The proportion of patients in all three groups of small intestinal GISTs who received preoperative neoadjuvant therapy was low. However, postoperative targeted treatment for ileal GISTs was significantly more frequent than for jejunal GISTs, which was in turn significantly higher than that for duodenal GISTs. Whether a more detailed differentiation of small intestinal GISTs is necessary during risk stratification still requires further evidence, which is important for the decision for subsequent treatment.

Results also indicate that patients with duodenal GISTs often experience a long disease course, yet their tumor staging, such as T stage, tends to be low. This may be due to the accessibility of the duodenum via gastroscopy, facilitating earlier detection. Although the statistical significance is lacking, it is noteworthy that the duration of bleeding in duodenal GISTs is among the longest. This suggests that the earlier identification of duodenal lesions through gastroscopy does not fully explain the phenomenon. Another possibility is that duodenal GISTs can cause bleeding even when quite small (especially intermittent bleeding) or lead to nonspecific symptoms like abdominal distension at early stages. In these cases, symptomatic treatments or management for gastric or duodenal ulcers may result in improvement, or the bleeding itself is intermittent or transient, causing patients not seeking further diagnosis or additional investigations, thus missing the detection of duodenal stromal tumors. Besides, ileal GISTs tend to occur in older patients but have a shorter disease course and higher T stage. This may be because ileal GISTs are more common in elderly individuals and tend to progress more rapidly; or because their insidious onset, where symptoms are absent for a long time, leading to a late diagnosis.

Previous studies have not thoroughly explored the clinical differences among small intestinal stromal tumors at various sites, and the highlight of this article is to address that gap with a relatively large sample size. Our findings contribute to a more detailed risk stratification for GIST patients, thereby guiding subsequent monitoring and treatment. However, this study has certain limitations. It is a retrospective study conducted at a single medical center. Determining the disease duration of GIST patients and collect their clinical symptoms and signs is challenging for retrospective studies, as it can only be assessed based on limited available data. Therefore, these findings necessitate subsequent verification through prospective multicenter studies to address the inherent limitations of the study design. Furthermore, to clarify the diagnosis of GIST, we only included newly diagnosed patients who underwent surgery and obtained pathological results, which may have introduced selection bias and might contribute to the relatively low rate of lymph node metastasis compared to previous studies. Moreover, we did not collect long-term follow-up data or prognostic results for these patients in account of the study design and the higher rate of loss of follow-up. Studies have indicated that after propensity score matching, there is no significant difference in RFS and OS between patients with duodenal GISTs and those with jejunal or ileal GISTs [25]. However, some studies have also shown that patients with duodenal GISTs have a poorer prognosis compared to those with other small intestinal GISTs [26]. Therefore, further investigation and the establishment of predictive models for the prognosis of different GIST patients are essential. Furthermore, we did not collect germline mutation information for small intestinal GISTs in this study. The frequency of driving mutations varies among GISTs in different anatomical sites, with KIT exon 11 mutations accounting for half of small intestinal GISTs, followed by a relatively high proportion of KIT exon 9 mutations (20 ~ 25%)(2). Future research should include this information to provide more guidance for selecting targeted treatment options.

With a deeper understanding and increased awareness, we may achieve earlier diagnoses, thereby reducing the incidence of ileal GISTs discovered incidentally during gynecological surgeries, decreasing the risk of emergency surgery for jejunal GISTs, and lowering the TNM stage of these tumors. Such vigilance is crucial for improving patient prognosis.

Conclusion

Small bowel stromal tumors at different sites exhibited distinct clinical manifestations. Adequate attention should be paid to the clinical symptoms of abdominal pain, melena, hematochezia, weight loss, and abdominal mass, which can help to diagnose GIST early and avoid events such as GIB and emergency surgery.