Gastrointestinal stromal tumor
1Department of Pathology and Laboratory Medicine, School of Medicine, University of California, Irvine, California; 2Integrated Oncology, LabCorp,
Irvine, California, USA
Gastrointestinal stromal tumor
|
Review Article
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Abstract
Gastrointestinal stromal tumor has received a lot of attention over the last 10 years due to its
unique biologic behavior, clinicopathological features, molecular mechanisms, and treatment implications.
GIST is the most common mesenchymal neoplasm in the gastrointestinal tract and has emerged from a
poorly understood and treatment resistant neoplasm to a well-defined tumor entity since the discovery of
particular molecular abnormalities, KIT and PDGFRA gene mutations. The understanding of GIST biology
at the molecular level promised the development of novel treatment modalities. Diagnosis of GIST depends
on the integrity of histology, immunohistochemistry and molecular analysis. The risk assessment of the
tumor behavior relies heavily on pathological evaluation and significantly impacts clinical management. In
this review, historic review, epidemiology, pathogenesis and genetics, diagnosis, role of molecular analysis,
prognostic factor and treatment strategies have been discussed.
Key words
Gastrointestinal stromal tumor; GIST; KIT mutation; imatinib
Submitted Apr 06, 2012. Accepted for publication Apr 26, 2012.
DOI: 10.3978/j.issn.2078-6891.2012.031 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Introduction
Gastrointestinal stromal tumor (GIST) is the most common
(80%) mesenchymal tumor of the alimentary cannel (1-3). It
accounts for less than 1% of all gastrointestinal tumors and
about 5% all sarcomas (2-4). It represents a wide clinical
spectrum of tumors with different clinical presentations,
locations, histology and prognosis. GIST can occur
throughout the entire gastrointestinal (GI) tract and may
have extragastrointestinal involvement as well. The clinical
relevance of this tumor was generated by the discovery of
its molecular biology and, consequently, of a drug effective
in treating the tumor. The following review will discuss
the GISTs in all aspects including history, epidemiology,
clinical presentation, diagnosis, prognosis and treatment
and emphasize on those relevant to diagnosis. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Historic overview
Stromal tumors arising from the GI tract were initially classified
as smooth muscle neoplasms including leiomyomas (5),
leiomyoblastomas or sarcomas (6), following description by
Stout and colleagues in 1940 (7). These descriptions were
widely used until the 1970s when electron microscope found
little evidence of the smooth muscle origin of these tumors
(8,9). With the advent of immunohistochemistry during the
1980’s it was soon appreciated that a large number of these
tumors did not have immunophenotypic features of smooth
muscle, and conversely, expressed antigens related to neural
crest cells (10).
The term of “stromal tumors” was first described as
a separate entity by Mazur and Clark (11) in 1983 and
Schaldenbrand and Appleman in 1984 (12). However,
this term was not widely accepted. In 1989, a distinctive
subset of these stromal tumors revealing autonomic neural
features was recognized and named “plexosarcoma” (13)
and subsequently as gastrointestinal autonomic nerve
tumor (GANT) (14). There was considerable confusion
regarding the origin, differentiation and even clinical
behavior of these tumors. In 1994, it was discovered that
a significant proportion of GANTs were immunopositive
for CD34 (15,16), which was the first relatively specific
marker of GISTs during the mid-1990s. Based on the
CD34 immunopositivity the possibility that GIST might
be related to the interstitial cells of Cajal was raised by
investigators (17). Interstitial cells of Cajal, also known
as pacemaker cells for peristaltic contraction, are a group
of cells found in the muscularis propria and around the myenteric plexus along the GI tract and have the
immunophenotypic and ultrastructural characteristics of
both the neural and smooth muscle elements. Meantime,
additional studies found that interstitial cells of Cajal
express KIT and are developmentally dependent on stem
cell factor which is regulated through the KIT kinase (17,18).
However, the following critical issues were not resolved:
the exact origin of GIST, the best way to diagnose GIST,
and differentiation of benign from malignant GIST. As
the developments in studies of GISTs, describing gain-offunction
mutations and consequently, constitutive activation
of KIT receptors in several human tumor cell lines was
reported in the mid-1990s (19,20).
Finally in 1998, Hirota and colleagues (21) discovered
a specific mutation in the intracellular domain of
the c-KIT protooncogene in GISTs as well as a nearuniversal
expression of KIT protein in GISTs by
immunohistochemistry. In the same year, Kindblom and
colleagues (22) corroborated findings from Hirota and
colleagues by showing the immunoreactivity for KIT in 78 of
78 GISTs studied and GISTs shared striking ultrastructural
and immunophenotypic similarities with interstitial cells
of Cajal. Both studies supported the hypothesis that GIST
may indeed derive from stem cells that differentiated
toward interstitial Cajal phenotype and confirmed KIT as a
diagnostic tool for GIST (23). The KIT mutation implied a
gain-of function linked to the activation of the kinase even in
the absence of the binding of the ligand. The identification of
the KIT mutation was a major breakthrough in the biology of
GIST and overall, in cancer biology.
The identification of the biologic driver, activating
mutations in KIT provided a therapeutic target for the
treatment of GIST. One patient with metastatic GIST
refractory to multiple types of therapies was treated
with STI-571 (Imatinib mesylate- Gleevec; Novartis,
Basel, Switzerland), which is a small molecule tryosine
kinase inhibitor (TKI) with potent activity against the
transmembrane receptor KIT, ABL kinase and chimeric
BCR-ABL fusion oncoprotein product of chronic myeloid
leukemia. The treatment yielded an early, rapid, and
sustained response (24) with supportive preclinical data
(25,26). This case provided proof of principle that inhibition
of KIT by drug therapy was associated with improvement
in the disease and brought phenomenal growth in the
understanding of GIST biology and therapeutics. Imatinib
occupies the ATP binding pocket of KIT, thereby preventing
substrate phosphorylation, downstream signaling, and
thereby inhibiting cell proliferation and survival (23).
The remarkable therapeutic efficacy of imatinib in
patients with GIST along with accurate diagnoses using
CD117 expression (a marker of KIT receptor tryosine
kinase) resulted in subsequent approval of imatinib in this
indication by the US Food and Drug Administration in
February 2002 (27). In 2003, Heinrich and colleagues (28)
and Hirota and colleagues (29) all found platelet-derived
growth factor receptor alpha (PDGFRA) gene mutations
as an alternative pathogenesis in GISTs without KIT gene
mutation. In January 26, 2006, Sunitinib, a multitargeted
TKI with activity against KIT, PDGFR, vascular endothelial
growth factor (VEGF) receptor (VEGFR), and FLT-1/KDR,
also received FDA approval for the management of patients
who are refractory or intolerant to imatinib (30).
Overall, about 85% of GISTs are reported to have
activating mutation in KIT or PDGFRA (28,31,32). CD117
(c-Kit) immunohistochemistry has proven to be a reliable
and sensitive diagnostic tool (22,33,34). With the TKI
therapies against KIT and PDGFRA (imatinib and sunitinib),
inoperable or metastatic GISTs are now treatable, and a
number of additional alternative drugs are in clinical trials. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Epidemiology
Although the exact incidence of GISTs in the world is
hard to determine since the entity was not uniformly
defined until the late 1990s, a few estimates and studies
indicate the incidences of approximately 14.5 cases/
million/year in Sweden (35), 14.2 in Northern Italy (36),
13.7 in Taiwan (37), 12.7 in Holland (38), 11 in Iceland
(39) and 6.5 in Norway (49). In a recent report, about
5,000 new cases of GISTs were diagnosed annually (41)
and a incidence of 6.8/million from 1992 to 2000 (38) in
the United States. The overall incidence rates of GIST,
therefore, ranges between 6.5 and 14.5 per million per
year. In general, little information on the prevalence of
GIST was available. It is believed that the prevalence of
GIST is higher, as many patients live with the disease for
many years or develop small GISTs only detected at autopsy
or if a gastrectomy is performed for other causes (42). A study
performed in Germany on consecutive autopsies revealed small
(<10 mm) GISTs in 22.5% of individuals who were older
than 50 years (43). Rubin and colleagues used the SEER
(surveillance, epidemiology, and end results) cancer registry
in US for patients with GIST from 1993-2002 to determine
incidence, prevalence, and 3-year survival and found the
overall incidence, prevalence, and 3-year-servival rate were
3.2/million, 16.2/million, and 73%, respectively (44).
GIST mainly affects middle aged to elderly adults,
typically in their 60s (35,45) with no clear gender
predilection (46) although some studies demonstrated a
slight male predominance (39,47). GISTs are uncommonly
seen in patients younger than 40, however, cases in children
and young adults have been reported (46). The true incidence of GIST in children is unknown. An incidence
rate of 0.06/million/year was reported among young adults
(20-29 years of age) (37). Other large series studies showed
the percentage of patients with GIST below the age of 21 years
ranged from 0.5% to 2.7% (45,46,48). Data from the UK
National Registry revealed an annual incidence of 0.02 per
million children below the age of 14 years, which appears
to be the most accurate epidemiological data to date on
pediatric GIST (49). Pediatric GISTs are considered a rare
entity that can be quite different from its adult counterpart
and seen predominantly in the second decade (46,50,51)
with a predilection for female patients (46).
Sporadic GISTs are most common and familial GISTs with
germline mutation of the KIT gene are rare, but have been
well described (52-55). These patients usually have multiple
GISTs and cutaneous hyperpigmentation (53). In addition,
GIST rarely occurs in association with other syndromes
such as neurofibromatosis type I (56-59) or Carney’s triad,
a nonfamilial condition with gastric GIST, paraganglioma,
and pulmonary chondroma (60,61). The latter should
be distinguished from Carney-Stratakis syndrome, an
inherited tumor syndrome comprising gastric GIST and
paragangliomas (62.
GIST co-existing with other tumors has been reported
mainly as case report (63) and mostly with colorectal
carcinomas or adenomas, followed by gastric carcinomas
(64,65). p53, one of the most common involved genes in
colorectal carcinogenesis, has also been found to have a
prognostic significance in GISTs, and mutations in this
tumor suppressor gene are more often observed in the highrisk
GISTs (66). GIST colliding with other tumors, mostly
gastric adenocarcinomas, is rarely seen in literature (67-69).
Only one case of gastric GIST colliding with angiosarcoma
was reported (70). | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Pathogenesis and genetics
In 1995 Huizinga and colleagues reported a knockout
mice model of KIT failed to express in interstitial cells of
Cajal cells (17). This finding led to the hypothesis that
KIT was essential for the development of interstitial cells
of Cajal cells. In 1998, Hirota and colleagues published a
groundbreaking discovery of KIT mutations in GISTs (21)
and 95% GISTs are immunohistochemically positive for the
receptor tyrosine kinase KIT (also known as CD117) (21,22).
It is now established that KIT mutations, which cause the
constitutive activation of the kinase, are found in 70-80%
of GISTs. CD117 becomes a crucial diagnostic marker for
GIST, and mutant KIT provides an important therapeutic
target clinically in GIST treatment.
Initially, GISTs lacking any evidence of KIT mutation were
classified as “wild type” (WT). In 2003, novel mutations
in PDGFRA were found in WT GIST by Heinrich and
colleagues (28). Currently PDGFRA mutations account for
5-10% of known mutations in GIST. About 9-15% of all
GISTs do not exhibit mutations in either KIT or PDGFRA
and are now termed “wild type” (WT) (71).
KIT is a member of the type III transmembrane receptor
tyrosine kinase (RTK) family that includes PDGFRA and
PDGFRB, as well as macrophage colony-stimulating-factor
receptor (CSF1R) and Fl cytokine receptor (FLT1) (72).
Normally, binding of the KIT ligand, stem cell factor (SCF) to
KIT results in receptor dimerization and kinase activation (73). In
contrast, the presence of KIT receptor-activating mutations
will bypass the ligand binding requirement for activation
and therefore become oncogenic, which has been implicated
in the pathogenesis of several human tumors in addition to
GIST and chronic myelogenous leukemia (CML), including
seminomas (74), mastocytosis (19), acute myelogenous
leukemia (75) and, more recently, in melanomas (76).
KIT oncogenetic activation is the dominant pathogenetic
mechanism in GIST (77). Although familial GIST with
germline mutations have been reported (52,55), the
majority of KIT mutations in GIST are somatic. The most
common mutations in KIT are found in the juxtamembrane
domain that is encoded by the 5' end of exon 11 of the KIT
receptor (Figure 1). Mutations in exon 11 change the normal
juxtamembrane secondary structure and cause the active
conformation of the normal kinase activation loop (78). The
mutations vary from in-frame deletions of variable sizes, point
mutations to deletions preceded by substitutions (79). The
deletions are associated with a more aggressive behavior in
comparison to other exon 11 mutations (80-83). Particularly,
deletions involving codon 557 and/or codon 558 are
associated with malignant behavior (84,85). A less common
mutant spot is located at the 3' end of exon 11, which
includes mainly internal tandem duplications mutations
(ITDs) (86). These ITD-type mutations are considered
to have a more indolent clinical course and a predilection
in GISTs located in the stomach (86). The second most
common KIT mutation, between 10% and 15% of GISTs, is
a mutation in an extracellular domain encoded by exon 9 (87).
GISTs with KIT exon 9 mutations are characterized by small
bowel location and aggressive clinical behavior (86).
 Figure 1 Schematic distribution of KIT or PDGFRA receptor mutations, frequency of mutations and TKI (Abbreviations: Ex, Exon; S,
sensitive; R, resistant)
A minority of GISTs that lack KIT gene mutations have
high levels of phosphorylation of PDGFRA resulted from
an activation by mutations or small deletions (28). PDGFRA
is a close homologue of KIT (28). Mutations in PDGFRA
and KIT in GIST are mutually exclusive and about onethird
of GISTs without KIT mutations harbor a mutation of
PDGFRA, within exons 12, 14 or 18 (28,88,89). In GIST,
mutant forms of PDGFRA have constitutive kinase activity in the absence of their ligand-PDGFRA similar to those
for KIT mutations, and the activated downstream pathways
(28,29) are identical to those in KIT-mutant GISTs (28,90).
In spite of the similarities in molecular aspect, most GISTs
with mutated PDGFRA have distinct pathologic features,
including gastric location, epithelioid morphology, variable/
absent CD117 by immunohistochemistry and an indolent
clinical course (88,91,92).
Recent studies indicate that a small portion of GIST wildtype
for both KIT and PDGFRA genes may harbor mutations
of the BRAF gene (93) and KRAS and BRAF mutations
predict primary resistance to imatinib in GISTs (94).
Furthermore, GISTs demonstrate typical patterns of
chromosomal gains and losses, including losses at 1p, 14q,
15q, and 22q. Tumor site appears to be associated with
distinct chromosomal imbalances; for example, gastric
GISTs show predominantly losses 14q, whereas intestinal
GISTs more frequently exhibit losses of 15q (95). | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Clinical presentation
Most GISTs remain ‘silent’ until reaching a large size.
Symptoms vary according to location and size. Symptomatic
GIST patients generally present with nonspecific symptoms
including abdominal pain, fatigue, dyspepsia, nausea,
anorexia, weight loss, fever and obstruction. Patients
may present with chronic GI or overt bleeding due to
mucosal ulceration or tumor rupture with life-threatening
intraperitoneal hemorrhage. Some patients with large
GISTs may have externally palpable masses (96,97).
Aggressive GISTs have a defined pattern of metastasis to
the liver and throughout the abdomen or both (45). Lymph
node metastasis is not common. Spreading to the lung and
bone in advanced cases has been reported (98). Metastasis
often occurs 10-15 years after initial surgery (45).
More than 80% of GISTs are primarily located in
GI tract and may occur throughout the GI tract with
extra-GI tract GISTs reported in omentum, mesentery,
retroperitoneum, gallbladder and urinary bladder (99-101).
The majority of GISTs (60%) are seen in the stomach,
usually in the fundus (35,39). The percentages of GISTs
found in other portions of GI tract are reported as 30% in
jejunum and ileum, 5% in duodenum, 4% in colorectum,
and rarely in the esophagus and appendix (45,46,48,65).
Reported tumor size in the stomach varies from a few
millimeters to >40 cm with a mean size of 6 cm in the
largest reported series (65). Apparently, the tumor size is
one of the factors contributing to the clinical symptoms. A
population-based study showed that the tumor size is 8.9 cm
in patients with clinical symptoms, which is about 70% of
GISTs studied, 2.7 cm in patients without clinical symptoms,
20%, and 3.4 cm in patients with GISTs detected at autopsy, 10% (35). Many smaller GISTs are detected incidentally
during endoscopy, surgery, or computed tomography (CT)
scans (35). | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Diagnosis
The diagnostic evaluation of GISTs is based on imaging
techniques (Figure 2), with a special role of endoscopic
examination because it is usually accessible when tumors
are in the stomach, esophagus and large intestine. In
addition, endoscopic ultrasonography (EUS) also plays an
important role in the diagnostic work-up of GISTs and is
accurate and efficient in the diagnosis of GISTs (102. In
general, externally bilging tumors are more common than
intraluminal masses (103). Punch-out ulcer is the classical
appearance of a submucosal tumor (104).
 Figure 2 Computed tomography scan revealed a partially
exophytic, dumbbell shaped solid mass (arrow) arising from the
posterior aspect of the gastric fundus along the greater curvature,
measuring approximately 6.7 cm × 4.5 cm
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Macroscopy
Gastric GISTs are greyish-white sub-mucosal tumors with
smooth contours and usually well-circumscribed and highly
vascular tumors. They typically have a tan-white or fleshy
pink cut surface often with hemorrhagic foci, central cystic
degeneration, or necrosis (Figure 3). The overlying mucosa
of large tumors is typically ulcerated (46).
 Figure 3 A gastric GIST with a nodulular surface and thin capusle. The cut surface reveals coarse granular and solid white tan suface with hemarrhage and cavities
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Histopathology
Microscopically, GISTs have a broad morphological
spectrum. Three main histological subtypes have been
best widely accepted and they are spindle cell type (most
common, 70%), epithelioid type (20-25%), and mixed
spindle cell and epithelioid type (99,105,106) (Figure 4).
In general, GISTs have a wide variation ranging from
hypocellular to highly cellular with higher mitotic rates.
Nuclear pleomorphism is relatively uncommon, and occurs
more frequently in epithelioid type.
 Figure 4 Common histologic al features of GISTs. A. Spindle cell GIST with short fascicles and whorls (×100); B. Spindle cell GIST with longer fascicles in bundles (×100); C. Spindle cell GIST with extensive perinuclear vacuolization (×100); D. Spindle cell GIST with
prominent nuclear palisading (×100); E. Epithelioid cells GIST with pleomorphic nuclei and vacuolated cytoplasm (×400); F. Epithelioid cell
GIST with rhabdoid features (×400)
Spindle cell type of GIST is composed of cells in short
fascicles and whorls. They have pale eosinophilic fibrillary
cytoplasm, ovoid nuclei, and ill-defined cell borders. Gastric
spindle cell GISTs often reveal extensive perinuclear
vacuolization, a diagnostic feature formerly used for tumors of
smooth muscle origin. The stroma sometimes demonstrates
myxoid change or, rarely osseous metaplasia. Distinctive
histological patterns among spindle cell GISTs including sclerosing type and palisading-vacuolated type (65). The
sclerosing spindle cell GISTs have slender spindle cells with
no nuclear atypia and low mitotic activity and are usually
paucicellular with extensive extracellular collagen. They
are often small and contain calcifications. The palisadingvacuolated
type is one of the most common gastric GISTs
and usually cellular with plump and uniformed spindle
cells. Nuclear palisading with perinuclear vacuolization is
characteristic. There is usually limited atypia with mitotic
activity rarely more than 10/50 high power fields (HPFs).
However, some examples show diffuse hypercellular pattern,
and others sarcomatoid features with significant nuclear
atypia and mitotic activity (65,99,106).
Epithelioid cell GISTs are characterized by round cells
arranged in nests or sheets and with eosinophilic to clear
cytoplasm. They also have spectrums from sclerosing and
paucicellular to sarcomatous and mitotically inactive to
mitotically highly active. However, the epithelioid GISTs
with atypia, even with pleomorphism are sometimes benign
(65,99,106).
Immunohistochemically, the vast majority of GISTs
(95%) are strongly and diffusely positive for KIT (CD117),
which makes the KIT to be a very specific and sensitive
marker in the differentiating GIST from other mesenchyma
tumors in the GI tract (21,22,34,107). The stain appears
as cytoplasmic, membrane-associated or sometimes as
perinuclear dots (34). Although KIT positivity appears to have
significant therapeutic implications, the intensity, extent and
patters of KIT staining neither correlates with the type of KIT
mutation nor have therapeutic significance (34). It is important
to note that negative KIT does not exclude the patient
from being treated with TKI (imatinib or sunitinib) since
some wild-type GISTs for both KIT and PDGFRA genes
respond to treatment with TKI (42). In addition, CD34 is
another common marker for GISTs but it is not as sensitive
or specific. It is positive in about 80% of gastric GISTs,
50% of small intestine GISTs, and in 95% of esophageal
and colorectal GISTs (48,108) (Figure 5). Other markers
which can be expressed by GISTs include h-caldesmon,
SMA, S100, desmin, Vimentin, and cytokeratins 8 and 18
(100). Recently other CD markers for GISTs are reported
including CD10 (109), CD133, and CD44 (110).
 Figure 5 Immunohistochemical features of GIST. A. Spindle cell GIST with strong and diffuse cytoplasmic staining of CD117 (c-kit) (×400); B. Spindel cell GIST with strong and diffuse membrane staining of CD34 (×400); C. Epithelioid cell GIST with strong cytoplasmic staining
of CD117 (×100); D. Epithelioid cell GIST with patchy and heterogeneous staining of CD34 (×400); E. Epithelioid cell GIST with punctate
staining of h-Caldesmon (×100); F. Epithelioid cell GIST with patchy mambrane staining of h-Caldesmon (×400)
A small minority of GIST (<5%) are negative for KIT, or
minimally, if any, positive for KIT by immunohistochemistry.
These tumors appear to be either KIT wild-type or with
mutant PDGFRA, have a predilection to stomach or
omentum/peritoneum, and be usually epithelioid or mixed
subtype (91,111). For the special interest in this subgroup
of KIT-negative GISTs, several new antibodies for the
diagnosis of GIST have been discovered based on the
molecular studies. DOG1 (discovered on GIST1), known
also as TMEM16A and ANO1, a transmembrane protein, has been found specifically in GISTs and has emerged as a
promising biomarker for GISTs (112,113). Recent studies
have shown that antibodies against DOG1 have even
higher sensitivity and specificity than KIT (CD117) and
CD34 with 75% to 100% overall sensitivity (113-116).
DOG1 is highly expressed in KIT mutant GISTs and also
can detect up to one-third of KIT-negative GISTs, which
mostly have PDGFRA mutation (113,116). In addition to
GISTs, DOG1 is also positive in normal gastric epithelium,
some carcinomas, germ cell tumors, melanomas, and some
mesenchymal tumors (113,114), such as recently reported
chondroblastoma (117). Like KIT, DOG1 is also expressed
in interstitial cells of Cajal serving as an internal positive
control. However, DOG1 does not stain mast cells which
are usually positive for KIT (112,114).
Non-gastric gists and gists in specific populations
Non-gastric GISTs may vary in clinical presentation,
histopathology, molecular profile, prognostic significance
and management strategy compared with gastric GISTs.
Small intestinal GISTs including the duodenal GISTs are
more homogeneous histologically and have a significant
tumor-related mortality if the tumor is >5 cm (48). They
typically harbor KIT exon 11 mutations as seen in gastric
GISTs and a small portion of small intestinal GISTs contain
duplication of two codons in KIT exon 9 (86,118). Usually,
small intestinal GISTs do not harbor PDGFRA mutations.
The sigmoid colon is the most common segment involved
by GISTs (39) in the colon. Histopathologic profile of
colonic GISTs is similar to that of small intestinal GISTs.
Pediatric GISTs account for about 1-2% of GISTs.
They are often misdiagnosed as having another acute
or chronic abdominal condition and they are usually
symptomatic and mostly located in the stomach with
mainly epithelioid pattern (35,46,50,51). GIST occurs in
children and young adults as a component of two distinct
syndromes: Carney triad and Carney-Stratakis syndrome.
Carney triad is composed of co-occurrence of GIST,
pulmonary chondroma, and paraganglioma. Carney triad
can be diagnosed when any of the two tumors are present
in a patient. However, if only GIST and paraganglioma are
present, it is considered to be Carney-Stratakis syndrome.
GIST in patients with Carney triad tends to be multifocal
and have high local recurrence rate and/or metastatic rate.
However, the clinical course of GIST in Carney triad is
usually indolent (61). Although pediatric GISTs express
KIT protein, the majorities lack KIT or PDGFRA mutations
(46,50,51). In 2002, a germline-inactivating mutation in the
hereditary paraganglioma gene was found to be unique for
Carbey-Stratakis (119,120). This germline mutation results
in a cancer predisposition syndrome including GIST.
Patients with neurofibromatosis type 1 (NF1) have a high
risk for GIST. Some autopsy studies have demonstrated as
many as one of three NF1 patients to have GISTs (121).
NF-associated GIST typically occur in duodenum or small
intestine and often multifocal and small. They commonly
have low risk parameters and are clinically indolent (57,121).
In contrast to sporadic adults GISTs, NF1-associated GISTs
lack KIT and PDGFRA mutations (57,121,122).
Familial GISTs were reported and account for a very
small portion of GISTs (<0.1%). They have typically
activated germline KIT or PDGFRA mutations with an
autosomal dominant inheritance and high penetrance
(52,55,123,124). They occur usually in middle age of life
and typical multifocal or diffuse in the GI tract. Most of
these GISTs have a benign course. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Differential diagnosis
Although GISTs are the most common mesenchymal tumor
of the GI tract, a variety of other tumors should be included
in the differential diagnosis. Accurate recognition of GIST
is obviously important as the treatment differs according
to the tumor type. The main differential diagnoses include
smooth muscle tumors, schwannoma, desmoid fibromatosis,
inflammatory myofibroblastic tumor, inflammatory fibroid
polyp, solitary fibrous tumor, synovial sarcoma, follicular
dendritic cell sarcoma, glomus tumor, and melanoma.
Kirsch and colleagues have published extensive review of
diagnostic challenges and practical approach to differential
diagnosis of GISTs (125).
Anatomic location may be helpful in differential
diagnosis. Intramural leiomyomas most commonly
locate in the esophagus and are rare in the stomach and
small intestine (126). Morphologically, leiomyomas have
brightly eosinophilic cytoplasm with distinct cell borders
whereas GISTs usually reveal syncytial cell morphology.
Immunohistochemically, GISTs and leiomyomas share some
markers, such as SMA and h-caldesmon, but spindle cell
GISTs are rarely positive for desmin which is more specific
for leiomyomas. Rare epithelioid GISTs that lack KIT
expression do stain positive for desmin (116). Leiomyomas
are negative for CD117.
Although gastric schwannomas are not commonly seen,
they can be morphologically very similar to certain spindle
cell GISTs. Distinct peripheral cuffing of lymphocytes and
strong reactivity with S-100 and GFAP readily differentiate
them from GIST in addition to the negativities of CD117
and CD34 (127).
Mesenteric fibrous lesions can be very challenging in
terms of diagnosis of itself and confusion with GIST due
to the location and gross appearance. Microscopically,
intraabdominal desmoid fibromatosis usually display long
sweeping fascicles of spindle cells embedded within a
collagen matrix with an infiltrating patter at peripheral
of the tumor. Immunohistochemical stain of betacatenin
is positive in about 75% of cases (128-130).
Inflammatory myofibroblastic tumors are commonly seen
in pediatric or young adult patients and recognized as a
mesenteric mass. Microscopically, this tumor has cellular
fascicular fibroblastic/myofibroblastic proliferation with
a prominent mixed inflammatory components including
significant number of plasma cells. About 50% of tumors
express ALK-1 (131), which is essentially negative in
GIST. Inflammatory fibroid polyp is a polypoid lesion of
mucosa with collagenous or myxoid stroma admixed with
fibroblasts. It can be CD34 positive but should be negative
for CD117 and DOG1 (113,114,132). Interestingly, same
PDGFRA mutations as seen in GISTs are also discovered in
inflammatory fibroid polyps (133).
Histologically, epithelioid GISTs need to be distinguished
from other epithelial or epithelioid tumors including carcinoma,
melanoma, glomus tumor, germ cell tumor and clear cell
sarcoma. Immunohistochemical studies play a major rule on
the differential diagnosis and the evaluation of appropriate
immunophenotypic markers in context with morphology in
most cases allows an accurate classification (Table 1).
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Role of molecular analysis
Mutational analysis of the KIT gene including exons 11,
9, 13, and 17, and PDGFRA gene including exons 12,
14, and 18 can be helpful in confirming the diagnosis of
GISTs if immunohistochemical studies fail to support
the diagnosis (particularly in CD117/DOG1-negative
spindle cell suspect cases). Corless and colleagues (134)
summarized the mutations of GISTs and classified GISTs
based on the molecular findings (Table 2). Furthermore,
mutational analysi s probably has more clini cal
significance in therapeutic aspect as it has predictive value
for sensitivity to molecular-targeted therapy (including
dosage) and prognostic value. It is strongly recommended
that it should be included in the diagnostic work-up
of all GISTs (135). The correlation between KIT and
PDGFRA mutational status and the response to tyrosine
kinase inhibitors and their role in primary and secondary
resistance has been widely investigated (31,136). Tumors
harboring KIT exon 11 mutations have a better outcome
under imatinib treatment than tumors harboring
different mutation, whereas tumors with PDGFRA
exon 18 mutations (D842V) have primary resistance to
imatinib both in vivo and in vitro (27,71,137). Therefore,
GIST mutational analysis is strongly recommended in current NCCN (National Comprehensive Cancer
Network) clinical practice guidelines (Figure 6) and in
ESMO (European Society for Medical Oncology) clinical
recommendations (138,139).
 Figure 6 NCCN Guidelines Version 1.2012, Gastrointestinal Stromal Tumors (GIST) (Abbreviations: H&P, history & physical examination; Mets, metastatic disease; IM, imatinib; Preop, preoperative; DX, diagnosis; SU; sunitinib; mo, month; y, year)
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Prognostic factors, grade and stage
The risk of relapse of GISTs is estimated based on mitotic
rate, tumor size, tumor site, surgical margins and the
status of tumor rupture. Tumor size and mitotic count are considered to be the most useful and best studied
prognostic factors by the 2002 Consensus risk classification
(Table 3) (99). It is believed that indicating a risk level of
GIST (low, intermediate, or high) is more appropriate than
definitively labeling the tumor as benign or malignant. This
risk classification was based on the cumulative experience of
the authors in the committee. The most important cut-offs
as indicators of aggressive clinical behavior were tumor size
of 5 cm and 5 mitoses/50 HPF. This consensus guideline
indicated that all GISTs may have malignant potential (99).
Based on long-term follow-up of more than 1,600 GISTs
(1,055 gastric, 629 small intestinal, 144 duodenal, and 111
rectal), Miettinen and colleagues proposed risk classification
incorporates primary tumor site in addition to the mitotic
count and tumor size (Table 4) (140). It demonstrates the
fact that gastric GISTs have a better prognosis than small
intestine or rectal GISTs. The more recently updated
consensus NCCN guidelines from 2007 (141) includes
anatomic site as an additional parameter in risk assessment
for GIST. Based on those guidelines, GISTs that are smaller
than 2 cm are considered to be essentially benign. Recently,
Gold and colleagues proposed a nomogram for estimating
the risk of tumor progression (142), in which each GIST
was assigned points on a scale based on tumor site, size, and
mitotic index. The total points of a tumor should determine
the 2- and 5-year recurrence free survival probabilities.
From a clinical point of view, additional prognostic factors
including non-radical resection and tumor rupture, whether
spontaneous or at the time of surgical resection, are both
associated with adverse outcome independent of any
other prognostic factors (143). Furthermore, Takahashi
and colleagues suggested the inclusion of a “clinically
malignancy group” to include patients with peritoneal
dissemination, metastasis, and invasion into adjacent organs
or tumor rupture (144). In 2008, a proposal by Joensuu
based on the NIH system included the presence of tumor
rupture as a high risk factor irrespective of size and mitotic
count (145). The Joensuu’s revised NIH risk system is shown in Table 5.
In the TNM staging (AJCC, 7th edition, 2010) (146),
grading of GISTs is based on mitotic rate. Mitotic rate
less than 5/50 HPFs is considered to be low (grade 1) and
greater than 5/50 HPFs is considered to be high (grade 2).
Please note that the staging criteria are different for gastric
GISTs and small intestinal GISTs to emphasize the more
aggressive clinical course of small intestinal GISTs even with
similar tumor parameters (147). The seventh edition of the
international union against cancer (UICC) published at the
beginning of 2010 included for the first time a classification
and staging system for GIST (148). This represents a
significant step towards a more standardized surgical and
oncological treatment for patients with GIST and, more
importantly, may facilitate the establishment of a uniformed
follow-up system based on tumor stage (Table 6) (149).
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Treatment
Treatment of localized disease
Surgery
The only potentially curative treatment of GISTs, still,
is complete surgical resection if it is a locally resectable
or marginally resectable tumor (141,150). GISTs rarely
metastasize to lymph node (142,151) and therefore
regional lymph node dissection is generally not needed. In
addition, organ-sparing resection (segmental resection) is
also appropriate oncologically. However, about 40-90% of
surgically treated patients experience disease recurrence
(152). A recent study of 127 patients with localized GISTs
who underwent complete resection demonstrated a 5-year
recurrence-free survival (RFS) rate of 63% (153). This study
concludes tumor size 10 cm, mitotic rate 5/50HPFs, and tumor location in the small intestine were all independently
associated with an increased risk of recurrence. In addition,
intraperitoneal rupture or bleeding is also associated with
a high risk of postoperative recurrence of nearly 100%
(143,154,155).
Adjuvant therapy
Understanding the molecular changes of GISTs along with
target treatments resulted in a considerable transformation
in the management of GISTs. The remarkable efficacy
of imatinib in treating metastatic GISTs has prompted
interest in developing an adjuvant after complete resection
of GISTs. Resent phase III randomized trial involved 778
patients with localized GISTs who underwent complete
surgical resection followed by 1 year of imatinib (400 mg/day)
and revealed that adjuvant imatinib significantly improved
the 1-year RFS rate (98%) compared with the placebo (83%)
(P<0.0001) (156). Based on the results of this trial, FDA
approved imatinib as adjuvant therapy for GISTs (157). The
most recent management guidelines in US (NCCN) (138)
and Europe (ESMO) (139) recommended adjuvant imatinib
for at least 1 year following complete surgical resection in
patients with intermediate- to high-risk GIST. However,
the optimal duration of adjuvant therapy has not been
established yet.
Treatment of localized unresectable or metastatic gists
Although surgical intervention was applied to patients
with metastases prior to the imatinib era, it was unlikely
to completely resect the tumor and consequently with
earlier recurrence than localized disease (45). Nunoby and
colleagues (158) in Japan studied the outcome of surgical
resection in 18 patients with liver metastases of GISTs and
showed 83% complete resection of liver metastases with
64% 3-year postoperative overall survival (OS) rate and
34% 5-year postoperative OS rate. However, the recurrence
rate in the remnant liver and in other organs reached 94%
in this study. Surgical treatment alone for metastatic GISTs,
therefore, is only palliative (158).
The application of imatinib for patients with advanced
and non-resectable GISTs was first evaluated in the
palliative setting in 2000 (24). A recent large clinical study
of imatinib for unresectable or metastatic GISTs revealed
up to 57 months of median OS rate (159), which is almost
a threefold increase in OS from about 20 months (45) prior
to the application of imatinib. Based on the clinical practice
guidelines (NCCN & ESMO), treatment with imatinib
(400 mg/day) now is the standard of care for patients with
locally advanced, recurrent, or metastatic disease (138,139).
Multiple phase III clinical trials have confirmed the
effectiveness of imatinib with standard-dose (400 mg/day)
or high-dose (800 mg/day) (159,160). Furthermore, the
efficacy of imatinib certainly also depends on the mutant
profile of GISTs. KIT exon 11 mutations show the greatest
benefit from imatinib treatment (400 mg/day) (Figure 1)
(135,161). KIT exon 11 codon 557/558 deletion/insertion
mutations have a more aggressive clinical behavior (162).
KIT exon 9 mutant GIST requires a higher imatinib dosage
to reach a better response (135,163). In addition, sunitinib,
another TKI, is beneficial for exon 9 mutated-GIST (30).
Although wild-type patients are not likely to benefit
from imatinib (161), some in vivo and in vivo studies on
sunitinib (164), nilotinib, and dasatinib (165) are promising.
Regarding PDGFRA-mutated GISTs, PDGFRA exon 18
mutations have better response to imatinib therapy but not
with PDFGRA exon 18 D842V-mutation (71).
According to the NCCN guidelines, patients with
progressive disease after imatinib treatment are allowed achieved in those cases (166-168). However, the timing
of the surgical intervention is very important and was
recommended as the time at which patients reached
maximum benefit from imatinib but before tumor
progression occurs (139,169). In addition, neoadjuvant
therapy with TKI should be considered to facilitate
complete resection and allow for a less morbid operation,
especially in duodenal GIST which can be sometimes hardly
resected completely (170,171). With a short neoadjuvant
imatinib therapy, tumor blood flow was decreased and
apoptosis was increased within 3-7 days of starting therapy
compared with pre-imatinib tumor tissue, although minimal
size reduction was observed (171).
Assessment of treatment response
According to the NCCN guidelines, imaging study of contrastenhanced
CT scan is the technique of choice to detect
recurrence or progression of GISTs (138,139,172). In rectal
GIST, MRI should be used or additional PET or PET-CT/
MRI may be useful for early detection of tumor response
to neoadjuvant therapy (172). Choi and colleagues (173)
proposed modified response evaluation criteria which
is considered to predict response more accurately than
previously proposed Response Evaluation Criteria in Solid
Tumor (RECIST) (174) and has a better correlation with
time to progression (175).
Resistant disease and alterative treatments
Although TKIs, especially imatinib, have resulted in
disease-free survival for patients following surgical resection
of their primary tumors and increased response rates and
survival for patients with metastatic disease, some patients
will eventually develop resistance to imatinib (176). Several
potential mechanisms of resistance were proposed and
include specific types of mutations (KIT exon 9, KIT wildtype
or PDGFRA exon 18) (31,135), acquisition of secondary
mutations within the KIT gene, KIT gene amplification, loss
of the wild-type allele, or inadequate imatinib plasma levels
(176-179). Sunitinib is the only second-line TKI approved
for use after imatinib failure due to its inhibitory function
on multi-kinases receptors (136). It has also been shown to
be effective against secondary mutations in vitro and in vivo
studies (136,161). However, as with imatinib, resistance
has recently been documented in patients with prolonged
exposure to sunitinib (180,181). In addition, it has been
shown that sunitinib can cause serious, life-threatening
adverse effects, including hypertension, cardiotoxicity, and
hypothyroidism (30,182,183). According to the NCCN and
ESMO guidelines, sunitinib is recommended as a secondline
therapy in patients who experience disease progression
after high-dose imatinib or who have life-threatening side
effects. If further progression occurs with sunitinib, patients
should be considered for clinical trials of new agents or new
combinations or discontinuation of anti-cancer therapy.
The role of newer generation KIT and PDGFRA kinase
inhibitors, e.g., nilotinib, remains to be determined in GIST
patients with multiple resistants after imatinib and sunitinib
therapies. Nilotinib has demonstrated activity against
imatinib- and sunitinib resistant GISTs (184) and displays,
by an ongoing pilot study (185), substantial clinical benefit
and is safe in the first-line treatment of advanced GIST.
Other agents, such as dasitinib (186), sorafenib (187), and
masitinib (188), target multiple oncogenic receptor tyrosine
kinases that have been implicated in the development and
growth of GIST. These newer agents and a wide number
of others (189) are currently under clinical trials for the
management of advanced and resistant GISTs and likely to
change the treatment of this disease soon. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Conclusions
GISTs have received much attention for many reasons.
The rapid expansion of molecular and clinicopathological
knowledge of GIST has given this disease a promising
future. The molecular targets for therapeutic interventions
are not only of importance for the treatment of GIST
patients, but also in the development of novel drugs and
new strategies in basic cancer therapy. Pathologists need
to know their role as the diagnostic information they
provided impacts on the choice of treatment as well as on
estimation of its efficacy. Molecular testing of GISTs should
be performed for treatment selection and assessment of
disease progression. The cause of GIST is still unknown;
therefore, little has been done preventively. However, with
gradual understanding the molecular mechanisms of GIST,
the etiology will be elucidated eventually. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Acknowledgements
Disclosure: The authors declare no conflict of interest.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
References
Cite this article as: Zhao X, Yue C. Gastrointestinal stromal
tumor. J Gastrointest Oncol 2012;3(3):189-208. DOI: 10.3978/
j.issn.2078-6891.2012.031
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
