Gastrointestinal stromal tumors of the upper GI tract: population-based analysis of epidemiology, treatment and outcome based on data from the German Clinical Cancer Registry Group

Gastrointestinal stromal tumors (GIST) are rare mesenchymal tumors that originate from the interstitial cells of Cajal [1]. They represent less than 1% of all neoplasms in the gastrointestinal tract. A global epidemiologic analysis demonstrated that most studies report an incidence of 10–15 per million per year (Søreide et al. 2016). The stomach is the most common site of GIST (60%), followed by the small intestines (35%), and lastly by the colon and rectum (5%). Less than 1% of tumors are found in the esophagus and gastroesophageal junction (Miettinen and Lasota 2006; Joensuu et al. 2013).

The prognosis of GIST depends on a variety of factors. Since 2001, many classifications have been proposed for the survival stratification of patients with GIST (Miettinen et al. 2005; Miettinen and Lasota 2006; Joensuu 2008; Gold et al. 2009; Agaimy 2010). Currently, the Armed Forces Institute of Pathology (AFIP) classification by Miettinen et al. is widely used and recommended by the ESMO guidelines (Casali et al. 2022). This classification uses location, size, and mitotic activity to estimate the risk of recurrence and progression. According to AFIP classification, gastric GISTs have a lower risk of recurrence or progression for the same size and mitotic activity compared to GIST of the small bowel and colon/rectum. Furthermore, positive resection margin and intraoperative tumor rupture are important negative prognostic parameters since they are associated with a higher likelihood of recurrence and poorer overall survival (Rutkowski et al. 2007; Takahashi et al. 2007).

Due to the low incidence of GIST in the esophagus and GEJ, information on their clinicopathological characteristics and clinical outcome are very limited. Most of the available literature on GIST focuses on the features of gastric and proximal small intestinal GIST, as these are the more common locations for GIST (Miettinen and Lasota 2006; Joensuu et al. 2013). As such, GIST of the esophagus and GEJ are not specifically addressed by the current classifications and specific treatment guidelines are lacking. In this context, the usage of big data sets such as population-based registries might represent a valuable source of information. Especially when targeting a subpopulation that is difficult to assemble by a single or small number of institutions. Therefore, this study aimed to evaluate the clinical features and outcomes of patients with GIST of the entire upper gastrointestinal (GI) tract including the esophagus, the GEJ, and the stomach based on a large data set of population-based clinical cancer registries in Germany provided by the German Cancer Registry Group of the Society of German Tumor Centers – Network for Care, Quality and Research in Oncology (ADT).

Data for the current study is derived from the German Cancer Registry Group of the Society of German Tumor Centers (ADT). The ADT aims to combine data from different German Clinical Cancer Registries all over the country on a voluntary basis. Hence, data for this study originates from 17 regional clinical cancer registries in 11 federal states. This study was approved by the local ethics committee of the University of Lübeck, Germany (20–327) and the ADT.

A total of 93,069 patients with a diagnosis of esophageal or gastric cancer (C15 and C16) were recorded between 2000 and 2016. Using the histological ICD-O-3 code “8936”, we identified approximately 1500 patients with gastrointestinal stromal tumors (GIST). After thorough preparation of the data set and testing for plausibility (e.g. exclusion of duplicates, exclusion of patients with the missing year of birth or diagnosis, exclusion of patients if the year of diagnosis was not recorded to be between 2000–2016, or exclusion of patients if the date of death was before or equal to the date of diagnosis), 1361 patients were finally included into the analysis (Fritz, Percy et al. 2000).

The following information was retrieved from the registry: sex, age at diagnosis (years), lymph node metastases (N0, N +), T-stage (T1–T4), Tumor location (esophagus (C15.0–15.9), cardia / GEJ (C16.0), stomach (C16.1–16.9)), type of therapy (none, operation alone, neoadjuvant systemic/chemotherapy/radiochemotherapy plus operation with/without adjuvant therapy, operation plus adjuvant systemic/chemotherapy/radiochemotherapy, systemic/chemotherapy/radiochemotherapy alone), operation type according to OPS-codes (partial resection of the esophagus (5–423; 5–424), total resection of the esophagus (5–426), local excision of the stomach (5–433), partial gastric resection (5–434, 5–435), subtotal gastrectomy (5–436), total gastrectomy (5–437), total gastrectomy with esophageal resection (5–438)), follow-up time (months after diagnosis), status at last follow-up (dead, alive).

Gastrointestinal stromal tumors of the esophagus and gastroesophageal junction (GEJ) are very rare tumors and have an estimated incidence of 0.1 to 0.3 per million people (Miettinen and Lasota 2006; Joensuu et al. 2013). Therefore, little is known about their clinicopathological features and clinical outcome. With this work, we present the first population-based analysis for GIST of the entire upper gastrointestinal tract, including esophageal, gastroesophageal junction, and gastric GIST using large population-based clinical cancer registry data provided by the German Cancer Registry Group of the Society of German Tumor Centers—Network for Care, Quality and Research in Oncology (ADT).

Our results showed a steady increase in the reported diagnosis of gastroesophageal GIST for the period 2000–2016, reaching a peak of 5% of all upper GI tumors in our registry in 2015. In addition, we found that GIST of the esophagus and GEJ are rare tumors, compared to gastric GIST representing 0.04% and 0.08% of all upper GI tumors as well as 2.7% and 5.1% of all reported GIST in our cohort. These observations are consistent with other population-based studies (Miettinen et al. 2005; Tran et al. 2005; Briggler et al. 2018). The reasons for this internationally observed increase in the rate of new cases are manifold. Firstly, the diagnostic criteria have changed over time. For example, the immunohistochemical markers CD117 (KIT) and DOG1 became standard in routine diagnostics, accompanied by an increasing awareness of the disease among physicians. Epidemiologic cancer registries had long included only GIST coded as malignant; that is, only tumors with ICD-O codes 8931/3 or 8936/3 were included in incidence estimates. However, the latest update of the WHO classification for sarcomas now classifies all GISTs as "malignant," regardless of tumor size, location, or mitotic rate (Cancer 2020). Another possible reason might be an improved surveillance and endoscopy rate among the general population in germany in the last decades(Rey, Hoffman et al. 2014). However, information on surveillance or increased use of endoscopy can not be analysed using the present dataset.

Furthermore, our data revealed that the incidence of GIST was higher in male compared to female patients younger than 75 years, but was subsequently higher in fmela compared to male patients at an age > 75 years (Fig. 1). This observation might be biased by the current longer life expectancy of females in the general population worldwide (Kontis et al. 2017). In addition, we demonstrated in the univariable analysis that male patients have shorter median OS compared to female patients (165 vs 115 months, p = 0.003). As suggested by Jzerman et al. this might be due to the more aggressive tumor characteristics at baseline such as higher mitotic and increased tumor rupture in male patients(NS, van Werkhoven et al. 2022). Nevertheless, in our cohort, the available baseline characteristics like tumor size and presence of metastasis were not associated with sex differences in patients with gastroesophageal GISTs.

Moreover, GIST of the esophagus or GEJ were associated with shorter median overall survival when compared to gastric GIST (111 months vs. 130 months, p = 0. 016 respectively). More importantly, these tumors also demonstrated increased metastatic potential compared to gastric GIST. Specifically, distant metastasis was documented in 27% of patients with esophageal GIST compared to only 15.7% and 13.2% of patients with GEJ and gastric GIST (p = 0.047). Similar results were reported by Theiss et al. where the incidence of metastases in esophageal GIST ranged between 17.0% to 48.0%(Theiss and Contreras 2019). Although more than half of the tumors in the esophagus were small (≤ 2 cm), esophageal tumors represented the subgroup with the greatest metastatic potential. As previously suggested, this phenomenon could be explained by the lack of serosal covering of the esophagus which can propagate its metastatic potential (Tran et al. 2005; Lott et al. 2015; Briggler et al. 2018). Until now, tumor location of GIST in the esophagus and GEJ have not been specifically addressed by any of the GIST risk stratification classifications, including AFIP, and Joensuu criteria. Nevertheless, our data suggest that esophageal and GEJ GIST can be classified into a higher-risk category.

Furthermore, it is well known that gastroesophageal GIST occurs in up to > 80% of cases in adults older than 50 years (Mucciarini, Rossi et al. 2007, Joensuu et al. 2013). This was reflected in our data, in fact, 92.1% of all patients were ≥ 50 years of age and only 1.8% were younger than 40 years of age at the time of diagnosis.

Approximately 13.7% of patients had documented metastasis at diagnosis. Incidence of metastasis was associated with advanced age, tumor size, and with tumors located in the esophagus. While tumors ≤ 2 cm were only in 12.9% of the patients associated with documented distant metastases, those larger than 10 cm developed metastasis in 35.7% of the patients. Similar results were found in the literature (DeMatteo et al. 2000; Trupiano et al. 2002). Interestingly, tumor size showed no significant impact on survival in our cohort and the respective survival rates did not show any major differences in the first five years. This is surprising as tumor size has been shown in several studies to play an important prognostic role concerning malignancy and the tendency to distant metastasis of the GIST (DeMatteo et al. 2000; Trupiano et al. 2002). However, it should be noted that the mentioned studies were performed before or in the initial period of imatinib. As previously described, the prognosis of GIST has significantly improved since the introduction of imatinib. (Perez et al. 2006; Cavnar et al. 2021). Thus, tumor biology cannot be determined on the basis of tumor size alone, but other factors such as mitotic activity or mutation analysis should be taken into consideration. Especially, since activating mutations in KIT proto-oncogene receptor tyrosine kinase or in platelet-derived growth factor alpha are considered major drivers of GIST and their detection predicts treatment responsiveness or resistance (Joensuu et al. 2017). For example, the presence of PDGFRA mutation predicts resistance to Imatinib or the presence of KIT exon 9 mutation stratifies the patients who will benefit from higher doses of Imatinib(Debiec-Rychter et al. 2006). Unfortunately, mitotic activity and mutation analysis are not documented in the data set used for this analysis, hence we cannot provide an informative analysis of this aspect of tumor biology.

The organs affected by distant metastasis were predominantly the liver and the peritoneum (68.4% and 16.4%, respectively), Interestingly, we noticed a change in metastatic patterns according to the tumor size. While the percentage of liver metastases decreased with an increase in tumor size, the percentage of peritoneal metastasis increased. Similarly, Yang et al. reported that the incidence of liver metastases was associated with increasing tumor size (Yang et al. 2019).

In addition, we demonstrate that patients with gastroesophageal GIST have a 5- and 10-year overall survival of 78.9% and 51.8%. Interestingly, even the presence of metastasis at presentation was associated with relatively long 5- and 10-year overall survival rates of 69% and 39.4% and a median OS of 100 months, which is two to fivefold higher than in previously published results by Ma et al. which included patients with liver, bone, and lung metastatic GIST. Moreover, longer overall survival was associated with female gender, age < 75 years, gastric GIST, and absence of distant metastases.

Although this study represents a large population-based registry study, several limitations must be addressed for the proper interpretation of results. First, the German Clinical Cancer Registry Group collects data from several regional clinical cancer registries on a voluntary basis, implicating data entry by many different people in variable quality. In addition, several variables were not included in the predefined dataset, such as the mitosis rate and mutation analysis (like KIT and PDGFRA mutation), therefore the effect of these variables on patients’ survival could not be assessedand our analysis could not be adjusted for potential confounding factors. Another limitation, is that a multivariable analysis was not conducted due to variability in completeness of some variables and the consequently low number of patients among the different subgroups. Some variables, such as disease-free survival, were not analyzed due to a large number of missing information. Information regarding response to medical treatment was not available in our dataset. On the other hand, due to the large number of patients in our cohort, we were able to provide a representative overall picture of GIST arising in the upper GI tract and to describe the daily clinical practice in Germany. To our knowledge, only a few publications to date exist with a comparable caseload, emphasizing the urgent need for more research and the creation of GIST-specific registries.

This is the largest population-based study on the incidence and characteristics of GIST of the entire upper gastrointestinal tract (esophagus, GEJ, and stomach). Our results demonstrate that the proportion of female patients in the GIST population increased with increasing age and that male patients are twice as likely to develop tumors in the esophagus and the GEJ compared to females. Moreover, tumors in GEJ and esophagus are associated with shorter overall survival and increased metastatic potential compared to those in the stomach, which justifies in our opinion classification of this specific group of patients into a high-risk category.