Gastrointestinal stromal tumours (GIST, smooth muscle tumour of uncertain malignant potential, STUMP, gastrointestinal autonomic nerve tumour, GANT, gastrointestinal pacemaker cell tumour, GIPACT)

This is a heterogeneous group of tumours, previously thought to be tumours of smooth muscle differentiation (leiomyo(sarco)mas). About 95% of stromal tumours of the stomach are classifiable as GISTs21. They are now thought to arise from the interstitial cells of Cajal, which act as pacemakers for peristalsis. Gain-of function mutations of c-kit are thought to be important in the pathogenesis of GIST, most commonly mutations in the juxtamembrane domain encoded by exon 1110. Germline mutations of c-kit occur in familial GISTs. A subset of GISTs, negative for CD117 by immunohistochemistry, lack a KIT mutation; PDGFRA mutations may provide an alternative oncogenic mechanism in these cases20.

Epidemiology

GIST is the most common mesenchymal tumour of the gastrointestinal tract, with an incidence of 4 per million of the population per annum. The median age is 63 years: less than 3% of cases occur in those under 21 years of age and only 10% are under 40 years21. There is no sexual predominance. They may be multiple in Carney's triad (GIST / phaeochromocytoma / paraganglioma / pulmonary chondroma syndrome) and, rarely, in neurofibromatosis6,21,24,25.

They are found throughout the gastrointestinal tract, most often in the stomach:

 

Clinical presentation

The most common presentation is with GI bleeding, followed by pain and less commonly a large mass or an incidental finding at surgery, rarely obstruction or an acute abdomen due to tumour rupture21.

Macroscopic appearances

Gross features vary widely, variously being described as well demarcated, nodular, lobulated or bosselated21. Small tumours may form a plaque. Larger tumours may form polyps projecting into the lumen, often ulcerated, a spherical mass or a pedunculated mass within the peritoneal cavity. Cystic change is common. Colour varies from grey-white through yellow to tan-brown with haemorrhagic foci.

Histopathology

Various patterns have been described:

Immunohistochemistry

The published literature is variable and confusing (Reference 3 is based on a review of the literature3):

CD117

221/2443, 13/138, 24/269, 145/14517, 631/69321

CD34

168/2243; a lower rate in colorectal than gastric tumours7, 9/138, 23/269, 141/14517, 644/78821

muscle-specific actin

variable: 11/642

SMA

20-40%12, higher in small intestinal than gastric tumours7, 3/138, 9/269, 30/14517, 150/81121

desmin

variable7, 4/642 (non-neoplastic smooth muscle cells may become incorporated6), 0/138, 7/14517, 40/78821

h-caldesmon

10/138, 117/14517

Calponin

2/138

S-100

some focally positive7, 0/138, 5/14517, 7/79321

CD68

scattered positive cells2

PGP9.5

16/269

NSE

20/269

tau

20/269

bcl-2

20/269

VS38c

positive (1/1, plus unpublished data)15

collagen type IV

4/118

laminin

2/118

cytokeratin 18

2/118

Cam 5.2

3/4218

AE1/3

0/14517

EMA

0/14517

chromogranin

0/118

synaptophysin

0/118, 4/269

neurofilament

0/269

CD20

0/118

CD45 (LCA)

0/118

calretinin

1/3214

inhibin

0/3214

vimentin

95% positive7, 62/642, 13/138, 26/269, 145/14517

CD117 in GISTs is usually seen in the majority of cells: positivity of less than 10% of cells is an indication for cautious interpretation. Scanty cells positive for CD117 in a spindle cell tumour are likely to be mast cells6. The staining pattern is usually diffuse cytoplasmic and granular with membrane accentuation. There may be focal perinuclear staining10.

Molecular genetics

KIT mutations are more common in spindle cell tumours and PDGFRA mutations are more common in epithelioid tumours. Point mutations tend to occur in tumours with benign behaviour21.

 

 

epithelioid

spindle

mixed

progressive disease

 
 

KIT exon11 deletion +/- point mutation

12/72

56/72

4/72

11/72

 

KIT exon11 point mutation

4/36

30/36

2/36

0/36

KIT exon11 internal tandem duplications

0/9

8/9

1/9

1/9

PDGFRA exon 18

52/86

15/86

19/86

9/86

PDGFRA exon 12

4/9

3/9

2/9

0/9

         

 

Ultrastructure

Features of neural differentiation correlate with immunohistochemistry. Skenoid fibres (extracellular amorphous eosinophilic interwoven modified collagen) are a further feature of neural differentiation. GANT is only reliably distinguished from GIST on ultrastructural grounds.

Variants

Differential diagnosis

leiomyoma/leiomyosarcoma

may be CD34+ but reliably CD117negative12. GIST is usually negative for desmin but may be positive for SMA

Schwannoma

may be CD34+ but reliably CD117-

retroperitoneal dedifferentiated liposarcoma

 

vascular tumours

 

fibromatosis

 

solitary fibrous tumour

CD34 positive, consistently CD117 negative12

Kaposi' sarcoma

CD34 positive, consistently CD117 negative12

inflammatory myofibroblastic tumour v inflammatory GIST

 

fibroid inflammatory polyp v inflammatory GIST

 

spindle cell carcinoid

 

granular cell tumour of oesophagus

CD34 and CD117 negative, S-100 positive19

spindle cell melanoma

S-100 positivity usually diffuse, may be CD117 positive (9/2512)

clear cell sarcoma

may be CD117 positive (7/1512)

malignant fibrous histiocytoma

may be CD117 positive (1/2012)

spindle cell carcinoma

 

haemangiopericytoma

CD 34 positive, may be CD117 positive (2/1012)

dermatofibrosarcoma protuberans

CD 34 positive, may be CD117 positive (1/712)

follicular dendritic cell sarcoma1 v inflammatory GIST

this tumour is rare and would be CD21+, CD35+, CD68+, CD34- & CD117-

Although CD117 does not have a sensitivity of 100% for GIST, unless a pathologist is experienced, a diagnosis of GIST should not be made in the absence of immunoreactivity for CD1176.

Treatment

There are results from Phase I and II trials that indicate that the tyrosine kinase inhibitor STI-571 (GleevecTM; Novartis, Basel, Switzerland, Imatinib mesylate) may be highly efficacious in the treatment of GISTs5. Responsiveness to STI-571 may be a function of the exon in which the KIT mutation is found.6 In GISTs that are refractory to imatinib, the presence of an exon 9 c-kit mutation is associated with a greater response to SU11248, a multi-targeted tyrosine kinase inhibitor22.

Prognosis

10% to 30% of GISTs show malignant behavior (gastrointestinal stromal sarcoma)7. Local recurrence usually precedes metastasis. When metastasis occurs, it is usually to liver, peritoneum or lungs, rarely to lymph nodes or ovaries. The literature relating to prognosis is conflicting:

 

"Benign" / low risk

Borderline

Malignant

de Saint Aubain Somerhausen

spindle cell, no atypia & MI = 0-2/30 HPF

spindle cell, no atypia and MI = 3-4/30 HPF

spindle cell, no atypia and MI>4/30 HPF

spindle cell, mild atypia and MI = 2-3/30 HPF

spindle cell, frank atypia and MI > 2/30 HPF

epithelioid & MI=0/30HPF

epithelioid and MI = 1/30 HPF

epithelioid and MI > 1/30 HPF

Miettinnen

gastric and MI = 0-1/10 HPF

gastric and MI = 2-5/10 HPF

gastric and MI > 5/10 HPF

gastric and size < 5 cm

gastric and size = 5-10 cm

gastric and size > 10 cm

Kindblom

bland cytology, no necrosis and on mucosal infiltration & MI = 0/10 HPF

 

some but not all of the features of malignancy

frank atypia, necrosis, mucosal infiltration, haemorrhage & MI >= 1/10 HPF

GIST Workshop6

<2 cm and no mitoses

many tumours fall into this category with uncertain malignant potential

> 5cm and >50 mitoses/50 HPF

or small intestine, any size and >10 mitoses / 50 HPF

MI = mitotic index, HPF = high power field.

 

Seidal and Edvardsson derived a formula to predict behavior2:

Z=2.428. x value for Ki67 - 2.092 x value for CD34 - 0.944 x value for grade + 0.096 x size in cm-0.041x patient age - 2.377

where the Ki67 staining was scored as: 0; no positivity, 1; <1% of nuclei, 2; >1% & <10% nuclei, 3; >10% nuclei staining, CD34 scored as 0 or 1, grade 1; monomorphic, little cellular atypia, grade 2; more cellular, moderate cytological atypia, grade 3; highly cellular with severe atypia.

If Z<0.66, the course is likely to be benign, if Z> 0.66, a malignant course is likely.

 

Trupiano et al13 found that tumor size > 7 cm, high cellularity, mucosal invasion, high nuclear grade, mitotic count >= 5 mitoses / 50 HPF, mixed cell type, extensive myxoid change and absence of hyalinisation all predicted an aggressive clinical course, but without allowing a confident separation of benign from malignant. More importantly, they believe that they can recognise clinically benign tumours are having the following features:

All other appearances are classed as "not benign".

Hawagaza17 on multivariate analysis found age >60 years, male sex, high tumour grade and tumour size 5 to 10 cm (relative risk 4.0) or >10 cm (relative risk 10.8) to be predictive of a fatal outcome.

 

GISTs on the greater curvature of the stomach, even if large, may be more likely to be benign6. Oesophageal tumours have the most favorable prognosis and those of the small intestine the worst6. Inadequate resection may account for the recurrence rate in the rectum6. Certain cytogenetic gains and amplifications may be associated with malignant behaviour6. Mutations of c-kit are associated with malignant behaviour10.

 

The largest long-term study21 identified differences between histological subtypes:

Coagulative necrosis is associated with progressive disease (50% vs 9%)21 but liquefactive necrosis is not21.

Mucosal infiltration in a "lymphoma-like" pattern correlated with sarcomatous tumours and a poor prognosis21.

However, the suggested guidelines for assessing the malignant potential are based on tumour size and mitotic activity21:

 

Reference 21

Number of cases

Alive, no disease

Alive with disease

Dead of unrelated cause

Dead of disease

Postoperative death

Dead of unknown cause

Prognostic conclusion

 

tumour < 2 cm & mitoses <= 5 / 50 HPF

76

34 (45%)

0

42 (55%)

0

0

32 (42%)

benign

tumour 2-5 cm & mitoses <= 5 / 50 HPF

320

172 (54%)

1

134 (42%)

5 (2%)

8

120 (38%)

probably benign

tumour 5-10 cm & mitoses <= 5 / 50 HPF

229

136 (59%)

3

80 (35%)

5 (2%)

5

50 (22%)

tumour >10 cm & mitoses <= 5 / 50 HPF

140

74 (53%)

1

47 (34%)

16 (11%)

2

44 (31%)

low to moderate malignant potential

subtotal mitoses <= 5 / 50 HPF

765

416 (54%)

5

303 (40%)

26 (3%)

15

246 (32%)

 

tumour < 2 cm & mitoses > 5 / 50 HPF

6

3

0

3

0

0

2

uncertain or low malignant potential

tumour 2-5 cm & mitoses > 5 / 50 HPF

99

57 (58%)

1

26 (26%)

15 (15%)

0

32 (32%)

low to moderate malignant potential

tumour 5-10 cm & mitoses > 5 / 50 HPF

96

25 (26%)

2

18 (19%)

50 (52%)

1

49 (51%)

high malignant potential

 

tumour >10 cm & mitoses > 5 / 50 HPF

108

10 (9%)

1

5 (5%)

88 (81%)

4

78 (72%)

subtotal mitoses > 5 / 50 HPF

309

95 (31%)

4

52 (17%)

153 (50%)

5

161 (52%)

 

total

1074

511 (48%)

9

355 (33%)

179 (17%)

20

407 (38%)

 

Patients with tumours containing an exon 11 c-kit mutation have a longer event-free survival than those whose tumour carry other mutations23.

 

References

1 K-C Chang et al. Follicular dendritic cell sarcoma of the colon mimicking stromal tumour. Histopathology 2001; 38:25-29.

2 Seidal, T., Edvardsson, H. Expression of c-kit (CD117) and Ki67 provides information about the possible cell of origin and clinical course of gastrointestinal stromal tumours. Histopathology 1999;34:416-424.

3 JF Graadt van Roggen et al. The histopathological differential of diagnosis of gastrointestinal stromal tumours. J Clin Pathol 2001; 54: 96-102.

5 CDM Fletcher. KIT (CD117) immunostaining and treatment with STI-571. Advances in Anatomic Pathology 2001;8:304.

6 Berman, J., O'Leary, T. J. Gastrointestinal stromal tumor workshop. Human Pathol 2001; 32:578-582.

7 Al-Nafussi, A., Wong, N. A.Intra-abdominal spindle cell lesions: a review and practical aids to diagnosis. [review] Histopathology 2001;38:387-402.

8 Miettinen, M., Paal, E., Lasota, J., Sobin, L. H. Gastrointestinal glomus tumors: a clinicopathologic, immunohistochemical, and molecular genetic study of 32 cases. Am J Surg Pathol 2002;26:301-311

9 Chambonniere, M. L., Mosnier-Damet, M., Mosnier, J. F. Expression of microtubule-associated protein tau by gastrointestinal stromal tumors. Human Pathol 2001;32:1166-1173.

10 Gibson, P. C., Cooper, K. CD117 (KIT): a diverse protein with selective applications in surgical pathology. Adv Anat Pathol 2002;9:65-69.

11 Miettinen, M., Sobin, L. H. Gastrointestinal stromal tumors in the appendix: a clinicopathologic and immunohistochemical study of four cases. Am J Surg Pathol 2001;25:1433-7.

12 Sarlomo-Rikala, M., Kovatich, A. J., Barusevicius, A., Miettinen, M. CD117: a sensitive marker for gastrointestinal stromal tumors that is more specific than CD34. Mod Pathol 1998;11:728-34.

13 Trupiano, J.K., Stewart, R.E., Misick, C., Appelman, H.D. and Goldblum, J.R. Gastric stromal tumors: a clinicopathologic study of 77 cases with correlation of features with nonaggressive and aggressive clinical behaviors. Am J Surg Pathol 2002;26:705-14.

14 Shah, V.I., Freites, O.N., Maxwell, P. and McCluggage, W.G. Inhibin is more specific than calretinin as an immunohistochemical marker for differentiating sarcomatoid granulosa cell tumour of the ovary from other spindle cell neoplasms. J Clin Pathol 2003;56:221-4.

15 Suarez-Vilela, D. and F. M. Izquierdo-Garcia (2003). "Cytotoxic T-lymphocyte-rich, gastrointestinal stromal tumour." Histopathology 43(4): 398-400.

16 Shek, T. W., I. S. Luk, et al. (1996). "Inflammatory cell-rich gastrointestinal autonomic nerve tumor. An expansion of its histologic spectrum." Am J Surg Pathol 20(3): 325-31.

17 Hasegawa, T., Y. Matsuno, et al. (2002). "Gastrointestinal stromal tumor: consistent CD117 immunostaining for diagnosis, and prognostic classification based on tumor size and MIB-1 grade." Hum Pathol 33(6): 669-76.

18 Nga, M. E., A. S. Wong, et al. (2002). "Cytokeratin expression in gastrointestinal stromal tumours: a word of caution." Histopathology 40(5): 480-1.

19 Prematilleke, I. V., V. Sujendran, et al. (2004). "Granular cell tumour of the oesophagus mimicking a gastrointestinal stromal tumour on frozen section." Histopathology 44(5): 502-3.

20 Heinrich, M. C., C. L. Corless, et al. (2003). "PDGFRA activating mutations in gastrointestinal stromal tumors." Science 299(5607): 708-10.

21 Miettinen, M., L. H. Sobin, et al. (2005). "Gastrointestinal Stromal Tumors of the Stomach: A Clinicopathologic, Immunohistochemical, and Molecular Genetic Study of 1765 Cases With Long-term Follow-up." Am J Surg Pathol 29(1): 52-68.

22 Maki RG, Fletcher CDM, Heinrich MC et al. Resutls from a continuation trial of SU11248 in patients with imatinib-resistant gastrointestinal stromal tumour. J Clin Oncol 2005;23(165):9011 (Abstract).

23 Hahn HP, Fletcher CDM. The role of cytogenetics and molecular genetics in soft tissue tumour diagnosis - a realistic approach. Current Diagnsotic Pathology 2005;11:361-370.

24 1.Perry CG, Young WF, Jr., McWhinney SR, et al. Functioning Paraganglioma and Gastrointestinal Stromal Tumor of the Jejunum in Three Women: Syndrome or Coincidence. Am J Surg Pathol 2006; 30:42-49

25 Miettinen M, Fetsch JF, Sobin LH, et al. Gastrointestinal Stromal Tumors in Patients With Neurofibromatosis 1: A Clinicopathologic and Molecular Genetic Study of 45 Cases. Am J Surg Pathol 2006; 30:90-96

This page last revised 14.1.2005.

©SMUHT/PW Bishop