Gastrointestinal Stromal Tumors (GIST): A Population-Based Study Using the SEER Database, including Management and Recent Advances in Targeted Therapy

Simple Summary Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms of the gastrointestinal (GI) tract. All GISTs are potentially malignant and, in general, the risk is proportional to the tumor size. Due to their location within the muscularis propria, GISTs require an endoscopic ultrasound for intramural sampling, although jumbo forceps sampling can be diagnostic in ulcerated lesions. GISTs usually affect people aged between 60 and 70 years, with no gender predominance. Symptomatic tumors mostly present with GI bleeding, anemia, early satiety, and abdominal fullness. The main treatment of GISTs is complete surgical resection, ideally with the preservation of tumor integrity to prevent intra-abdominal dissemination due to rupture and spillage. Advanced-stage tumors require targeted therapy with tyrosine kinase inhibitors and multidisciplinary oncologic care. Abstract Introduction: Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasm of the gastrointestinal (GI) system. Most GISTs originate from the interstitial cells of Cajal (ICC), the pacemaker cell situated between the circular and longitudinal layers of the muscularis propria along the GI tract. In this population-based study using the SEER database, we sought to identify demographic, clinical, and pathologic factors that affect the prognosis and survival of patients with this neoplasm. Molecular genetic advances, current management guidelines, and advances in targeted therapy are discussed. Methods: Demographic and clinical data from GIST patients were retrieved from the SEER research plus database for the period 2000–2018. Statistical analysis was performed with IBM SPSS® v20.2 software using the Chi-square test, paired t-test, multivariate analysis, and Kaplan–Meier functions. Results: A total of 10,833 patients with GIST were identified. Most patients were between 60–74 years of age: 40%, Caucasian: 68%, and the male to female ratio was 1.1:1. The most common primary tumor sites were stomach: 63%, small intestine: 30%, rectum: 3%, and esophagus: 0.7%. When reported, the grade of differentiation was well: 38%, moderately: 32%, undifferentiated: 19%, poorly: 12%. The size of most tumors ranged between 6–10 cm: 36% and they were treated by surgical intervention: 82% and/or chemotherapy/targeted therapy: 39%. The stage was localized: 66%, advanced: 19%, and regional: 15%. The 5-year survival was 74% (95% confidence interval (95% CI) = 72.6–74.7), and the 5-year cause-specific survival 82% (95% CI = 80.7–82.6). The 5-year cause-specific survival by treatment included surgery at 86% (95% CI = 85.4–87.3), chemotherapy/targeted therapy with or without surgery at 77% (95% CI = 75.7–78.9), and radiation at 75% (95% CI = 74.5–80). On multivariable analysis tumor size > 5 cm, poorly and undifferentiated grade, age > 60, and distant metastases at presentation were associated with worse overall survival. Conclusion: GISTs comprise 1–2% of malignancies of the GI tract, usually affect male Caucasians between the ages of 60 and 74 years, most tumors occur in the stomach and small intestine, and are usually >5 cm, but still localized, at the time of diagnosis. Most tumors receive multimodality surgical and chemotherapy/targeted therapy treatment, with a 5-year overall survival of 74% and cause-specific survival of 82%. GIST patients would benefit from enrollment in large clinical trials to establish better therapy guidelines for unresectable, treatment-refractory, and recurrent tumors.


Introduction
Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal malignancy (sarcoma) of the gastrointestinal (GI) tract; however, they only comprise 1-2% of all GI malignancies. GISTs originate from pluripotential mesenchymal cells committed to become interstitial cells of Cajal (ICC), which are the pacemaker cells situated between the circular and longitudinal layers of the muscularis propria along the GI tract [1]. Most tumors affect the stomach and small intestine. The incidence of GIST is 10-15 cases per million worldwide, with~5000 cases per year in the United States [2,3]. Small incidental lesions (tumorlets) are commonly identified during abdominal surgery, radiologic or endoscopic studies, and at autopsy [4]. Diagnosis relies on a combination of clinical, imaging, histopathology, immunohistochemistry, and molecular studies [5]. GISTs were formally recognized as a specific type of neoplasm in 2000, and their incidence has been increasing steadily due to improvements in diagnostic technologies [6]. Advanced-stage disease is reportedly present in 47% of cases, usually with metastases to the liver and peritoneum [7]. Symptomatic tumors manifest by a mass-like effect, leading to early satiety or obstruction, or symptoms associated with ulceration/tumor rupture, such as pain, bleeding, anemia, and perforation [8]. Most GISTs arise within the muscularis propria and appear as submucosal or mural masses in endoscopy or imaging [9]. Computed tomography (CT) is the gold standard for evaluating abdominal masses, since it provides information about the size, location, and presence or absence of regional and distant spread. The use of oral and intravenous contrast improves the evaluation of the tumor margins [10]. This study is one of the largest and most up-to-date database studies aimed at investigating the demographic, clinical, and pathological factors affecting the prognosis and survival of GIST patients, with special insights into emerging therapies.
Patients without microscopic confirmation of the diagnosis were excluded from the study. The data were exported to IBM's Statistical Product and Service Solutions (SPSS©), version SPSS ® v20.2 (IBM Corp, Armonk, NY, USA). Demographic and clinical data, including age, race, sex, primary tumor site, histologic grading, tumor size, SEER-provided tumor pathological stage, treatment received (surgical, chemotherapy, and radiotherapy), overall survival, and survival by therapeutic modality were included. The SEER-provided chemotherapy treatment category refers predominantly to tyrosine kinase inhibitors targeting KIT and PDGFRA, since cytotoxic chemotherapy has been phased out as a systemic option. Endpoints included overall survival, mortality, and disease-specific survival. For categorical variables, the Chi-square test was used; for continuous variables, the paired t-test and analysis of variance (ANOVA) were used. To determine the independent factors affecting survival, multivariate analysis was used. p < 0.05 was considered statistically significant.

Results
Data from 10,833 patients were extracted from the SEER database for the period 2000-2018.

Overall and Location Based
Treatment data were available for most of the patients, and included surgery: 82%, chemotherapy with or without surgery: 39%, and radiation: 0.1% ( Figure 2). Surgery was the most common treatment in almost all locations, followed by chemotherapy. Overall, the highest percentage of surgical resection was in appendicular GIST: 92%, followed by hepatic flexure: 78% and ascending colon: 78%. The highest percentage of chemotherapy with or without surgery was in GISTs located in the rectum: 43%, rectosigmoid: 42, and esophagus: 42%. Neoadjuvant radiation was only administered for GISTs located in the stomach and rectum (Table 3).  Abbreviations: NOS, Not otherwise specified; AD, Anoderm; #, Due to the small sample size, the data from these locations should be interpreted cautiously.

Treatment According to Tumor Size, Stage, and Grade
Surgery was the primary treatment modality, irrespective of the tumor size and grade, followed by chemotherapy. Chemotherapy was the most common treatment modality in patients with distant spread (Table 4).
Five-year disease-specific survival by treatment: The 5-year disease-specific survival with surgical resection was highest when tumors were located in the transverse colon: 92%, followed by the rectum: 90%, anal region: 89%, stomach: 88%, and small intestine: 84%. When chemotherapy with or without surgery was used as the primary treatment modality, the highest 5-year disease-specific survival rates were observed for tumors located in the rectum: 89%, rectosigmoid: 86%, transverse colon: 84%, small intestine: 82%, and sigmoid colon: 80% (Table 7). Survival trends for 5-year disease-specific survival by treatment: The disease-specific survival greatly improved from 2000: 75% to 2013: 85% (p < 0.05). Surgery offered the best disease-specific survival in 2012: 93%; however. disease-specific survival has been increasing significantly with chemotherapy over the years, reaching its highest point in 2010: 83% (Table 8, Figure 3). Table 8. Five-year disease-specific survival trend data by treatment over the years for 10,833 patients with gastrointestinal stromal tumor from the Surveillance, Epidemiology, and End Results (SEER) database 2000-2018.

5-Year Disease-Specific Survival, % (Confidence Interval)
Year Overall Disease-Specific Survival Surgery Chemotherapy ± Surgery Adjuvant Radiation

Discussion
The histopathologic diagnosis of GIST relies on the expression of c-KIT and DOG1 (Discovered on GIST1) by immunohistochemistry [1]. Originally, all GISTs were believed to be derived from the interstitial cells of Cajal (ICC), the pacemaker cells of the GI tract; however, different molecular GIST subgroups have been identified more recently, leading to the realization that each subset may arise from other specialized stromal cells. Tumors originating from ICCs with KIT mutations are considered the "classic GIST" and were initially identified in 1998. They comprise 75% of GISTs, most commonly affect the stomach, and are responsive to imatinib therapy [11]. GIST with PDGFRA mutations make up the second-most-common group and comprise~10% of all cases. PDGFRA Exon 12 and exon 18 mutations are more common in the duodenum and gastric antrum, respectively, and do not always respond to imatinib therapy [12]. The discovery that patients with germline PDGFRA mutations V561D and P653L develop multiple GISTs and inflammatory fibrous polyps in the stomach and small bowel led to the identification of a new PDGFRA+/CD34+/KIT+ interstitial stromal cell, the telocyte, which has been proposed as the cell of origin of PDGFRA-mutated GISTs with distinctive epithelioid histology and aggressive behavior [13][14][15].
GISTs without KIT or PDGFRA alterations are collectively dubbed "wild-type GISTs", and make up the remaining 10-15% of all cases. This group is heterogeneous, not typically responsive to imatinib or other typical tyrosine kinase inhibitors (TKI), and includes SDHdeficient tumors (14%), tumors with mutations in NF1 (type 1 neurofibromatosis), BRAF V600E, and NTRK (0.5% each), and other rare mutations. SDH-deficient GISTs have loss-offunction alterations in one of the four SDH genes that encode for the subunits that constitute the SDH complex (mitochondrial succinate dehydrogenase complex), presenting in younger patients, and are usually only identified in the stomach. Loss-of-function alterations in SDHA/B/C/D are either due to sporadic epigenetic silencing (promoter hypermethylation) or germline alterations (Carney triad syndrome, Carney-Stratakis syndrome) at a ratio of 1:3 [16,17]. Animal models of BRAF V600E-driven GISTs develop tumors beyond the myenteric plexuses, which have been hypothesized to be derived from smooth muscle cells [18].
To our best knowledge, this study includes the largest cohort of GIST patients ever published. We found a 5-year overall cumulative survival of 74% (95% CI = 72.6-74.7), which is lower than that of the Dutch GIST Registry Study that reported a 5-year survival of 85% for adults aged 18-40 years and 76% for adults aged > 40 years [19]. Our study includes patients diagnosed between 2000 and 2018, while the Dutch study included patients diagnosed between 2009 and 2019 only. The improved survival in the Dutch cohort is explained by the more widespread use of TKIs over time, availability of universal health care in the Netherlands, different life expectancies of Dutch and American men and women, and earlier detection in the Dutch study: advanced disease 24% versus 34%.
In our study, age > 60 years and male gender were associated with worse overall survival, consistent with a previous SEER database study covering data from 1998 to 2011 [20]. In our study, non-gastric/non-small intestinal location (esophagus, and ascending and sigmoid colon) was associated with worse overall survival. Equivalent results were reported in another SEER database study with data from 1998-2011, [21] in which there were no differences in OS and DSS for small bowel, rectal, and gastric GISTs, but colonic and extra-visceral GISTs had significantly worse outcomes. A previous study from Iceland also showed lower metastatic rates in gastric than non-gastric GISTs [22]. A SEER database study from 1990-2009 showed that intestinal location had a higher risk of recurrence after surgery, independent of tumor size and mitotic rate, [23]; however, the interpretation of this data is difficult as GISTs were not defined as an entity until 2000. Together, all of these studies support better outcomes for gastric GISTs and worse outcomes for colonic GISTs. Tumors in very uncommon locations, such as the esophagus, rectum, and anus, and extra-gastrointestinal tumors are presumed to be at higher risk than gastric GISTs; however, reliable information is lacking due to their rarity. Like other sarcomas, large tumors >5 cm, distant metastases, and poor differentiation or undifferentiated, were associated with worse survival outcomes. Regarding treatment, surgery offered the best 5-year disease-specific survival, followed by systemic therapy with TKIs.

Targeted Treatments and Emerging Therapies
Complete surgical excision without lymph node dissection is the gold standard for localized tumors, similar to other visceral sarcomas [24][25][26]. Adjuvant imatinib for 3 years is indicated when a high risk of relapse exists; this benefit is unclear for wild-type GIST [27]. There is an ongoing phase-3 randomized trial comparing 3 vs. 5 years of adjuvant imatinib (Table 10. Newer TKIs are continuously emerging and being used for advanced, unresectable, relapsed, or metastatic GISTs based on the high prevalence of mutations leading to constitutively activated tyrosine kinase receptors (e.g., c-KIT 75%, PDGFRA 10-15%). Imatinib works best for c-KIT mutations in exon 11 (standard-dose imatinib), exon 9 (high-dose imatinib), and non-D842V PDGFRA mutations [28,29]. The KIT mutation A502_Y503dup comprises 90% of exon-9 mutations and is associated with locally aggressive tumors, spindle cell morphology, and decreased progression-free survival compared with exon 11 mutations, but better relapse-free survival after curative resection [30,31]. KIT exon-13 mutations are rare (1-2%) and most commonly occur as a secondary mutation after the initiation of therapy [32]. Esophageal and gastric KIT exon 13 mutant GISTs are rare, but tend to be larger and more aggressive on average. This has led them to be slightly overrepresented in clinical trials [32]. Exon-9, -13, and -14 c-KIT mutations can be treated with sunitinib. Exon-17 c-KIT mutations, which are usually the result of clonal evolution, may respond to ponatinib [28,29]. Avapritinib is recommended to treat symptomatic or progressive PDGFRA exon-18 D842V-mutated GISTs as an upfront therapy or whenever the mutation is discovered.
In practice, imatinib at a standard dose of 400 mg per day, sometimes followed by a trial of higher dose of 800 mg per day if disease progression occurs on the standard dose, sunitinib, regorafenib, and the recently approved ripretinib can be used for empiric treatment as first, second, third, and fourth lines for advanced GISTs, respectively [29,33,34]. These TKIs are not effective for c-KIT and PDGFRA wild-type GISTs (10-15% of total cases) [29,33,34]. Testing for alterations in BRAF, NTRK, SDH, and NF1 should be pursued to detect targetable mutations treatable with BRAF inhibitors or BRAF-MEK inhibitor combinations and NTRK inhibitors, such as Larotrectinib [28]. There are 76 active studies registered on clinicaltrials.gov (accessed on 12 February 2022) enrolling patients with GISTs.
GIST-driver mutations' clinical characteristics and therapies are summarized in Table 11. A summary of active phase-3/4 treatment trials is provided in Table 10.

Limitations
The limitations of this study are mainly related to the quality and completeness of the information entered and retrievable from the database. The data regarding specific TKI treatment could not be extracted, data on genetic mutations and the mitotic index were not available, syndromic and sporadic cases could not be segregated from each other, and the type of surgical resection or margin status were not available, limiting the quality of the analysis.

Conclusions
Our study showed that tumor size > 5 cm and high histologic grade had the highest impact on survival. Additional negative risk factors were location other than the stomach and small bowel, age > 60 years, and distant spread. Female gender and primary surgery were associated with improved overall survival. Numerous existing and emerging targeted therapies are now available for high-risk or advanced disease. Optimal targeted therapy should be guided by genomic profiling, and this information should be collected in national databases to improve the existing treatment guidelines for unresectable or recurrent disease. Institutional Review Board Statement: All data are publicly available, and no IRB was required.
Informed Consent Statement: Patient consent was waived due to this article using data from the SEER database, which are publicly available deidentified patient data from the National Cancer Institute (NCI), USA.
Data Availability Statement: All data are publicly available. The data of this manuscript were presented at the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES), Las Vegas, Nevada, 31 August-3 September 2021.