Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumor of the gastrointestinal tract. Soon after GIST was recognized as a tumor driven by a KIT or platelet-derived growth factor receptor mutation, it became the first solid tumor target for tyrosine kinase inhibitor therapies. More recently, alternative molecular mechanisms for GIST pathogenesis have been discovered. These are related to deficiencies in the succinate dehydrogenase complex, NF1-gene alterations in connection with neurofibromatosis type 1 tumor syndrome, and mutational activation of the BRAF oncogene in very rare cases.
Key points
- •
What is now known as gastrointestinal stromal tumor (GIST) used to be called as gastrointestinal (GI) smooth muscle tumor: leiomyoma if benign, leiomyosarcoma if malignant, and leiomyoblastoma if with epithelioid histology.
- •
GISTs typically occur in older adults, and the median patient age in the major series has varied between 60 and 65 years although are rarely seen in children and young adults.
- •
Most GISTs, approximately 85% to –90%, contain oncogenic KIT or platelet-derived growth factor receptor alpha mutations. However, loss of function of succinate dehydrogenase complex has been identified as alternative pathogenesis, especially in GISTs in young patients.
- •
Most common clinical symptoms of GIST are GI bleeding and gastric discomfort or ulcerlike symptoms.
- •
Wedge resection is the most common surgery for a small-sized to medium-sized gastric GIST, and sufficient margins can usually be obtained.
Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumor of the gastrointestinal (GI) tract. Soon after GIST was recognized as a tumor driven by a KIT or platelet-derived growth factor receptor alpha (PDGFRA) mutation, it became the first solid tumor target for tyrosine kinase inhibitor therapies. More recently, alternative molecular mechanisms for GIST pathogenesis have been discovered. These are related to deficiencies in the succinate dehydrogenase (SDH) complex, neurofibromatosis type 1 (NF1) gene alterations in connection with NF1 tumor syndrome, and mutational activation of the BRAF oncogene in very rare cases.
Clinically GISTs are diverse. They can involve almost any segment of the GI tract, from distal esophagus to anus, although the stomach is the most common site. From an oncologic perspective, GIST varies from a small, harmless tumor nodule to a metastasizing and life-threatening sarcoma. This article presents the clinical, pathologic, prognostic, and to some degree, oncological aspects of GISTs with attention to their clinicopathologic variants related to tumor site and pathogenesis.
History of GIST and terminology
What is now known as GIST, used to be called GI smooth muscle tumor: leiomyoma if benign, leiomyosarcoma if malignant, and leiomyoblastoma if with epithelioid histology. Tumors previously classified as GI autonomic nerve tumors have also turned out to be GISTs, as have many tumors historically classified as GI schwannomas or other nerve sheath tumors.
Electron microscopic studies from the late 1960s and on demonstrated that most of the “GI smooth muscle tumors” differed from typical smooth muscle tumors by their lack of smooth muscle–specific ultrastructure. Immunohistochemically they lacked smooth muscle antigens, especially desmin. As they also lacked Schwann cell features, GIST was then proposed as a histogenetically noncommittal term for these tumors. The discovery of KIT expression and gain-of-function KIT mutations in GIST in 1998 was the basis of the modern concept of GIST – a generally KIT positive and KIT mutation-driven mesenchymal neoplasm specific to the GI tract.
History of GIST and terminology
What is now known as GIST, used to be called GI smooth muscle tumor: leiomyoma if benign, leiomyosarcoma if malignant, and leiomyoblastoma if with epithelioid histology. Tumors previously classified as GI autonomic nerve tumors have also turned out to be GISTs, as have many tumors historically classified as GI schwannomas or other nerve sheath tumors.
Electron microscopic studies from the late 1960s and on demonstrated that most of the “GI smooth muscle tumors” differed from typical smooth muscle tumors by their lack of smooth muscle–specific ultrastructure. Immunohistochemically they lacked smooth muscle antigens, especially desmin. As they also lacked Schwann cell features, GIST was then proposed as a histogenetically noncommittal term for these tumors. The discovery of KIT expression and gain-of-function KIT mutations in GIST in 1998 was the basis of the modern concept of GIST – a generally KIT positive and KIT mutation-driven mesenchymal neoplasm specific to the GI tract.
Epidemiology of GIST
GIST, once considered an obscure tumor, is now known to occur with an incidence of at least 14 to 20 per million, by population-based studies from northern Europe. These estimates represent the minimum incidence, as subclinical GISTs are much more common. In an US study, 10% of well-studied resection specimens of gastroesophageal cancer harbored a small incidental GIST in the proximal stomach. An autopsy study from Germany also found a 25% incidence of small gastric GISTs.
GISTs typically occur in older adults, and the median patient age in the major series has varied between 60 and 65 years. GISTs are relatively rare under the age of 40 years, and only less than 1% occurs below age 21. Some series have shown a mild male predominance. Over half of the GISTs occur in the stomach. Approximately 30% of GISTs are detected in the jejunum or ileum, 5% in the duodenum, 5% in the rectum, and less than 1% in the esophagus. Based on our review of Armed Forces Institute of Pathology (AFIP) cases, as many as 10% of all GISTs are detected as advanced, disseminated abdominal tumors whose exact origin is difficult to determine.
Despite occasional reports to the contrary, the authors do not believe that GISTs primarily occur in parenchymal organs outside the GI tract at sites such as the pancreas, liver, and gallbladder. At the 2 first mentioned organs, GISTs are metastatic or direct extensions from gastric or duodenal or other intestinal primary tumors. The authors are skeptical about primary GISTs in the gallbladder and note that the reported evidence for this diagnosis is tenuous and that molecular genetic documentation is absent. Furthermore, review of all gallbladder sarcomas in the AFIP failed to find any GISTs. Similarly, GISTs diagnosed in prostate biopsies are of rectal or other GI and not prostatic origin.
GIST is phenotypically related to GI Cajal cells
Almost all GISTs express the KIT receptor tyrosine kinase, similar to the GI Cajal cells that regulate the GI autonomic nerve system and peristalsis. These cells have a stem cell–like character, as demonstrated by their ability to transdifferentiate into smooth muscle. KIT-deficient mice lack GI Cajal cells and those with introduced KIT-activating mutations develop Cajal cell hyperplasia and GISTs, supporting the role of Cajal cells in GIST oncogenesis.
KIT and PDGFRA mutation as a driving force of GISTs
Most GISTs, approximately 85% to 90%, contain oncogenic KIT or PDGFRA mutations. KIT and PDGFRA are 2 highly homologous cell surface tyrosine kinase receptors for stem cell factor and platelet derived growth factors. Normally these kinases are activated (phosphorylated) up on dimerization induced by ligand binding. However, mutated receptors may self-phosphorylate in a ligand-independent manner, rendering the kinase constitutively activated, which has a critical role in cell proliferation and is considered the driving force of GIST pathogenesis. However, additional genetic changes are necessary for malignant progression as mutations are already detectable in the very small GISTs, most of which probably never grow to clinically detectable tumors.
How KIT mutations cause increased cell proliferation has been demonstrated in several ways. KIT mutations introduced in a lymphoblastoid cell line increased cellular proliferation. Families with germline KIT or PDGFRA mutation and transgenic mouse with similar “knock-in” KIT mutations have predisposition to GIST, often developing multiple GISTs. Also, it has been shown that KIT mutations are associated with constitutively phosphorylated KIT and that KIT tyrosine kinase inhibitors such as imatinib mesylate can in vitro and in vivo abolish the phosphorylated status and normalize the increased cell proliferation.
The clinical significance of mutation type analysis includes assessment of sensitivity to tyrosine kinase inhibitors (especially imatinib), and in some cases, mutation type, can also offer a prognostic clue. The rare presence of homozygous mutation is associated with an aggressive course of disease. The main KIT mutation types and their clinical significance are summarized in Table 1 .
| Gene | Domain | Exon | Type of Mutation | Frequency | Clinicopathologic Features |
|---|---|---|---|---|---|
| KIT | EC | 9 | Duplication | 10% | Associated with intestinal tumors |
| JM | 11 | Deletion Deletion/insertion Duplication Insertion Substitution | 70% | Deletions and deletion/insertions associated with malignant course especially in gastric tumors Duplication associated with gastric tumors and favorable course | |
| TK1 | 13 | Substitution | 1% | Slightly more frequent in intestinal tumors Associated with spindle cell morphology | |
| TK2 | 17 | Substitution | 1% | More frequent in intestinal tumors Associated with spindle cell morphology | |
| PDGFRA | JM | 12 | Deletion Deletion/insertion Duplication Substitution | 1% | Associated with gastric tumors, epithelioid cell morphology and more indolent course |
| TK1 | 14 | Substitution | <1% | ||
| TK2 | 18 | Deletion Deletion/insertion Duplication Substitution | 5% |
KIT exon 11 mutants
Approximately 90% of KIT mutations involve exon 11, the juxtamembrane domain. Their KIT activating potential is believed to be related to disruption of the alpha-helical structure of the juxtamembrane domain then allowing spontaneous dimerization and phosphorylation of KIT. Exon 11 KIT mutations include a spectrum of in-frame deletions of 3 to 21 or rarely more base pairs, single nucleotide substitutions, internal tandem duplications, and combinations of all, and rarely true insertions of nonduplicative genomic sequences.
Deletions in the KIT juxtamembrane domain most frequently involve the 5′ portion of the exon 11 between codons 550 and 560. Tumors containing deletions in this area are clinically more aggressive than those with single nucleotide substitutions. In large site-specific series, this was especially true for gastric GISTs.
KIT exon 11 single nucleotide substitutions are generally limited to 4 codons: 557, 559, 560, and 576. Internal tandem duplications are essentially restricted to the 3′ portion of exon 11 and are associated with gastric tumors and favorable course. In general, most KIT exon 11 mutants are sensitive to the tyrosine kinase inhibitor imatinib mesylate, although some rare variants may show different level of resistance.
Other KIT mutants
KIT extracellular domain exon 9 mutations are rare and essentially restricted to intestinal GISTs based on Western studies. However, in a Japanese series these mutations have also been detected in some gastric GISTs, raising the possibility of population differences. Most mutations are identical 2 codon duplications introducing a tandem alanine-tyrosine pair (AY502-503). These KIT exon 9 mutant GISTs are notable for their poorer response to imatinib, so that dose escalation of imatinib from 400 mg/d to 800 mg/d or use of alternative tyrosine kinase inhibitor has been advocated. However, in a large preimatinib series, KIT exon 9 mutant GISTs did not seem to have an inherently worse prognosis than exon 11 mutant tumors.
Rarely, KIT tyrosine kinase 1 domain (exon 13) or tyrosine kinase 2 domain (exon 17) is mutated. Exon 13 mutations usually involve codon 642, whereas exon 17 mutations most often occur in codon 822. These mutants are variably sensitive to imatinib.
PDGRFA mutations in GISTs
PDGFRA mutations were discovered in 30% of KIT wild-type (WT) tumors. KIT and PDGFRA mutations were shown to be mutually exclusive in GISTs. PDGFRA mutants are essentially restricted to gastric GISTs, comprising approximately 10% of such cases overall. Most PDGFRA-mutant gastric GISTs represent clinically indolent tumors. Also, there is some predilection to epithelioid morphology, and some of these tumors show weaker KIT expression. Although they express PDGFRA, standard immunohistochemistry is not helpful for the detection of PDGFRA-mutant GISTs because PDGFRA is widely expressed in GISTs of any type and also in other tumors. PDGFRA mutations, similar to KIT mutations, include in-frame deletions, single nucleotide substitutions, and internal tandem duplications. Most of these mutations are D842V substitutions involving the PDGFRA tyrosine kinase 2 domain (exon 18), although rare mutations have been identified in juxtamembrane domain (exon 12) and tyrosine kinase 1 domain (exon 14). D842V mutants are notorious for their primary resistance to imatinib. Therefore for initial therapy, if oncologically indicated, an alternative, more potent, tyrosine kinase inhibitors should be selected.
Clinical manifestations of GIST
Most common clinical symptoms of GIST are GI bleeding and gastric discomfort or ulcerlike symptoms. The bleeding varies from chronic insidious bleeding often leading to anemia to acute life-threatening episodes of melena or hematemesis. Few GISTs manifest as other abdominal emergencies, such as intestinal obstruction or tumor rupture with hemoperitoneum. Nearly one-third of GISTs are incidentally detected during surgical or imaging procedures or endoscopic screening for gastric carcinoma. Some rectal GISTs are detected during prostate or gynecologic examination.
Most of the currently detected GISTs are localized tumors less than 5 cm, but retrospective studies had larger mean tumor sizes. In general, small intestinal GISTs are larger in average, and the percentage of metastasizing tumors is higher than among gastric GISTs. Peritoneal cavity and liver are the typical sites of metastases. Rarely, GISTs metastasize into bones. In our experience, bone metastases have a predilection to axial skeleton, especially the spine. Cutaneous and peripheral soft tissue metastases are rare. In contrast to other sarcomas, malignant GISTs very rarely, if ever, metastasize to lungs, even if they have extensive other metastases. Although some GISTs metastasize in 1 to 2 years or sooner, metastatic spread is possible after a very long delay. The longest interval from primary tumor to liver metastasis observed by us was 42 years; this indicates the need for long-term patient follow-up.
Identification of GISTs
Radiologists, endoscopists, and surgeons and pathologists can suspect a GIST whenever there is a rounded to oval, circumscribed mural or extramural nonmucosa-based mass of any size that involves or is closely associated with the stomach, intestinal segments, or lower esophagus. However, in some cases such lesions prove to be other mesenchymal tumors, unusual variants of carcinomas, neuroendocrine tumors, or even lymphomas. In most cases, examination of a biopsy easily resolves the differential diagnostic problem. A GIST should also be considered for any palpable abdominal mass.
The radiologic and gross appearances of GISTs, especially the gastric ones, can be highly variable including tumors with intraluminal, intramural, external components and with pedunculated extramural and cystic appearances. Any larger GIST in the intestines typically forms an externally extending mass that is often centrally cystic and may fistulate into the lumen ( Fig. 1 ). Some small intestinal GISTs form dumbbell-shaped masses with intramural and external components.
GIST is the most common mesenchymal tumor in all segments of the GI tract with 2 exceptions: most esophageal mesenchymal tumors are true leiomyomas and not GISTs and small mucosal leiomyomas are more common in the colon and rectum than are GISTs. In the stomach, GIST is by far the most common mesenchymal tumor, as there are only 4 leiomyomas or schwannomas reported for every 100 GISTs.
Sampling of a GIST for preoperative diagnosis via endoscopic mucosal biopsy is successful in only 20% to 30% cases, including those that involve the mucosa or superficial submucosa. Endoscopic ultrasound-associated fine-needle aspiration biopsy (EUS-FNA) is more promising. In a recent study, the diagnostic yield of EUS-FNA was 76%, whereas even better results were reported with Tru-cut histologic biopsies (97% yield of diagnostic material). The larger GISTs especially are amenable to CT-guided needle core biopsy.
Pathologic diagnosis of GIST is based on identification of a mesenchymal neoplasm with spindle cell or epithelioid histology that is generally positive for KIT (CD117 leukocyte antigen). Common histologic features in GISTs include spindle cells with sclerosing matrix, perinuclear vacuolization and nuclear palisading, epithelioid cytology, and sarcomatoid, mitotically active morphology in gastric GISTs ( Fig. 2 ). Small intestinal GISTs are characterized by extracellular collagen globules and a Verocay body or neuropil-like material, reflecting complex entangled cell processes ( Fig. 3 ).
Approximately 97% of GISTs are immunohistochemically positive for KIT, at least focally, but the pattern can vary from membranous and apparent cytoplasmic to a perinuclear dotlike pattern. Tumors with epithelioid cytology can be only focally positive, or rarely entirely negative, which can be the cases especially with some PDGFRA-mutant GISTs. The KIT-negative GISTs are usually positive for anoctamin-1 (Ano-1), a calcium-activated chloride channel protein also expressed in Cajal cells. Ano-1 is also known under the aliases DOG1 (discovered on GIST-1) and ORAOV2 (overexpressed in oral carcinoma). Approximately 97% of GISTs are Ano-1 positive including KIT-negative tumors, but some GISTs, especially small intestinal ones, are negative. Together KIT and Ano-1 capture nearly 100% of GISTs as their “shadow areas” tend to be different. Other potential but less-specific and sensitive markers in the overall detection of GISTs are protein kinase C theta and CD34. The latter is positive in 70% of GISTs and is nearly consistently detected in gastric spindle cell GISTs.
For comprehensive detection of GISTs, one should consider that most GI mesenchymal tumors are GISTs, and that most intraperitoneal mesenchymal tumors are also GISTs. Even in the retroperitoneal space, a GIST is more common than a leiomyosarcoma, in our experience. GIST should also be considered in the differential diagnosis of mesenchymal or epithelioid neoplasms involving the liver, pancreas, and pelvic cavity.
GIST prognosis at different sites
The best universally applicable prognostic parameters are tumor size (maximum tumor diameter in cm) and mitotic rate per 50 high-power fields (HPFs) (corresponding to 5 mm 2 ). A prognostic chart ( Table 2 ) has been devised by analysis of large series of gastric and small intestinal GISTs, based on AFIP cases with long-term follow-up. This chart shows that gastric and small intestinal GISTs less than or equal to 5 cm with mitotic count less than or equal to 5/50 HPFs have a very good prognosis with only 3% to 5% of metastatic risk. Also, the chart shows a marked prognostic difference between gastric and small intestinal GISTs. The latter shows significant to high metastatic rates in many categories that in the gastric location have much lower rate of metastases. Other intestinal GISTs have a prognosis approximately similar to small intestinal GISTs, although less data exist, because of their rarity. Prognostic nomograms creating these parameters as continuous variables have also been devised. Based on a relatively small sample size, such a nomogram was found to offer a more accurate prognostication. Addition of genetic parameters may further improve prognostication. The number and type of genomic losses and gains detected by comparative genomic hybridization and genome complexity index are examples of this development. In the latter, expression of aurora kinase was found an adverse prognostic factor.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
