Date: November 29, 1999

To: Office of Rare Disease, NIH

From: Jules J. Berman, Ph.D., M.D., CDP, DCTD, NCI

RE: Workshop on GIST (Gastrointestinal Stromal Tumors) markers


GIST (Gastrointestinal Stromal Tumor) is a rare tumor of the GI tract. Overtly malignant tumors account for at most ten percent of GISTs. Malignant GISTs occur in the population at an incidence of approximately 3.3 per million, giving rise to approximately 800 new diagnoses and 600 deaths yearly in the United States (1). The remaining 90% of GISTs include a substantial number of tumors (10-20%) for which histologic criteria are unable to provide a clinically useful assessment of malignant potential. As a result, a large number of patients receive inappropriate medical attention due to over- or under-treatment of their tumors.

The important biological issue for diagnosticians and clinicians is that there are no reliable histologic criteria that can distinguish benign from malignant GISTs. Every new diagnosis of GIST provokes a clinical dilemma stemming from the unpredictable nature of this peculiar neoplasm. Even the name,"Gastrointestinal Stromal Tumor," conveys a noncommital evaluation of the neoplasm's biologic category.

In general, large tumors (exceeding 5 cm.), tumors with necrosis, and cellular tumors with many mitotic figures or with nuclear atypia, are usually considered to be malignant. Small tumors (< 5 cm) with fewer than 5 mitotic figures/50 high power fields are usually considered benign. However, the occasional histologically benign GIST will, despite surgical resection, give rise to fatal widespread metastases. Up to 20% of GISTs are given a clinically disappointing diagnosis of "GIST of uncertain malignant potential." Since pathologists currently rely on morphological determinations that are arbitrary or subjective, there is a pressing need for efforts that determine reliable and measurable indices of tumor biology.

The past few years have seen the first signficant advances in our understanding of the molecular pathogenesis of GIST. Gain-of-function mutations of the c-kit gene have been found in GISTs (2). Most GISTS (89%) express the KIT protein and 57% of GISTs have a specific exon 11 missense mutation (3). GISTs with the exon 11 mutation were more likely to be malignant, with more frequent recurrences (P=0.0005) and higher mortality (P=0.0001) than did GIST patients without the exon 11 mutation (3). A familial form of GIST, in which multiple tumors occur, is associated with a c-kit gene mutation found not only in tumor cells, but also in peripheral leukocytes (4). Mutations of the c-kit gene have only been found in GIST and in mastocytosis. The particular c-kit mutation found in mast cell neoplasms is distinct from the c-kit mutations found in GIST (3). Immunohistochemical and ultrastructual studies suggest that the putative precursor cell for GIST is the interstitial cell of Cajal (sometimes called the intestinal pacemaker cell). Recent evidence has shown that germ-line loss-of-function of the c-kit gene is associated with depletion of the interstitial cells of Cajal. All of these findings are very strong evidence that c-kit is a necessary gene for the growth of the putative precursor cells of GIST and that c-kit mutations are involved in most GISTs. However, the clinical utility of specific c-kit mutations as prognostic markers is still unevaluated.

In the past decade, there have been a flurry of small studies (usually under 60 cases in each study) attempting to establish reliable prognostic indicators for this tumor (5-8). Assessment of cytologic atypia and histologic pattern does not improve the ability to predict outcome in multivariate studies, and the gains in accuracy that result from consideration of tumor location are very modest. Tumor aneuploidy, as assessed by flow or image cytometry, is associated with poorer prognosis, but multivarate analysis suggests that it conveys no information beyond that provided by the mitotic index and a knowledge of whether metastases were present at surgery. Studies on the clinical utility of determining S-phase fraction have given conflicting results. Assessment of proliferation using immunohistochemical staining for proliferating cell nuclear antigen (PCNA) and Ki-67 (MIB-1) appears to be a little better than counting mitotic figures, but significantly more laborious. A recent study of potential significance is the observation that GISTs that show telomerase activity have a worse prognosis than those in which telomerase is not active (9). The relationship between telomerase activity and KIT mutation is not known.

Classical cytogenetic studies, together with comparative genomic hybridization results, suggest that GISTs can have a wide variety of cytogenetic abnormalities. The observation of cytogenetic abnormalities at the distal end of chromosome 1p may have special diagnostic importance, since loss of heterozygosity in this region has been associated with adverse outcome in several other tumors, including neuroblastoma, colorectal carcinoma, and breast carcinoma.

In summary, GIST is an uncommon tumor, and classic morphologic features fail to reliably distinguish benign from malignant forms. There have recently been several important advances leading to better understanding of the pathogenesis of this tumor. In addition, there are a variety of prognostic tests poised for further development and evaluation, either as single tests or in combination with other tests. It seems appropriate, at this time, for the leaders in this small field to assemble to assess the state of science in GIST diagnosis, and to discuss and suggest designs for those studies that will provide indicators for the diagnosis and treatment of GIST.

Reference List:

1. Fishman, AP. Gastrointestinal tract. In: J. Rosai (ed.), Ackerman’s Surgical Pathology, Ed. 8, pp. 691-693, St. Louis: Mosby, 1996.

2. Hirota S, Isozaki K, Moriyama Y, et al.Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors.Science 279:577-80, 1998.

3. Taniguchi M, Nishida T, Hirota S, et al. Effect of c-kit mutation on prognosis of gastrointestinal stromal tumors.Cancer Res 59:4297-4300, 1999.

4. Mishida T, Hirota S, Taniguchi M, et al. Famililial gastrointestinal stromal tumors with germ line mutation of the KIT gene. Nat. Genet 19:323-324, 1998

5. Cooper PN, Quirke P, Hardy GJ, Dixon MF. A flow cytometric, clinical, and histological study of stromal neoplasms of the gastrointestinal tract. Am.J.Surg.Pathol. 1992;16:163-70.

6. el-Naggar AK, Ro JY, McLemore D, Garnsey L, Ordonez N, Mackay B. Gastrointestinal stromal tumors: DNA flow-cytometric study of 58 patients with at least five years of follow-up. Mod.Pathol. 1989;2:511-5.

7. Suzuki H, Sugihira N. Prognostic value of DNA ploidy in primary gastric leiomyosarcoma. Br.J.Surg. 1993;80:1549-50.

8. Lee JS, Nascimento AG, Farnell MB, Carney JA, Harmsen WS, Ilstrup DM. Epithelioid gastric stromal tumors (leiomyoblastomas): a study of fifty-five cases. Surgery 1995;118:653-60; discussion 660-1.

9. Sakurai S, Fukayama M, Kaizaki Y, Saito K, Kanazawa K, Kitamura M, Iwasaki Y, Hishima T, Hayashi Y, Koike M. Telomerase activity in gastrointestinal stromal tumors.Cancer1998;83(10):2060-6.

10. Cunningham RE, Federspiel BH, McCarthy WF, Sobin LH, O'Leary TJ. Predicting prognosis of gastrointestinal smooth muscle tumors. Role of clinical and histologic evaluation, flow cytometry, and image cytometry. Am.J.Surg.Pathol. 1993;17:588-94.

11. Ernst SIB, Hubbs AE, Przygodzki RM, Emory TS, Sobin LH, O'Leary TJ. KIT Mutation portends poor prognosis in gastrointestinal stromal/smooth muscle tumors.Am.J.Surg.Pathol1998;in press.


This workshop will be cosponsored by NCI and AFIP (Armed Forces Institute of Pathology). The AFIP has a large collection of GIST tumors and actively supports work in this area. Dr. Timothy O’Leary, Chief of the AFIP’s Laboratory of Cellular Pathology and a major contributor to the GIST field (10,11), will be co-organizer of the workshop. The workshop wll bring together GIST researchers; pathologists who diagnose GIST and who have an interest in developing and applying new prognostic markers; oncologists who face the problem of deciding which patients with GIST need aggressive treatment; and statisticians who have reviewed the existing literature and who may provide advice regarding the design of new translational efforts in the GIST field.

The purpose of the workshop is to:

1) Review the current state of the science of GIST

2) Discuss the best practices for the diagnosis and treatment of GIST

3) Discuss views of the kinds of studies that are most needed to advance the field of GIST diagnosis and treatment.

3) Provide an opportunity for investigators in this small field to form collaborative efforts.

We expect that a workshop summary will be published in an oncology journal.

Tentative Agenda:

1. State of the Science (Formal Presentations)

a. Basic Science

b. Diagnosis

c. Prognosis

d. Treatment

f. Paradigms for new test development

2. Participant presentations:

Every interested participant will have up to 10 minutes to describe new

studies from their own laboratory.

3. Group discussion: Current status of methodology for GIST diagnosis, prognosis and treatment.

4 Group discussion: Future Goals for developing and evaluating new prognostic tests.