Diagnosis and Risk Assessment

Identifying the type of cell from which a cancer has developed is a specialized art. The doctors who make these judgments are pathologists, and their job is absolutely critical! If the oncologist does not know what the type of cancer is, he or she does not know how to treat it. The pathology and diagnosis of GIST are only recently beginning to be understood, as we discuss below. Given the critical job of the pathologist, newly diagnosed GIST patients should consider a second opinion of their pathology reports (unless the pathology was initially done by a pathologist who was very experienced with GIST). While many pathologists may be able to accurately diagnose GIST now, predicting the behavior of GIST is very difficult. Experience is critical here. A team of doctors and pathologists who are all very experienced with GIST will usually be better able to develop and implement a specific treatment plan based on each patients unique situation.

Gastrointestinal Stromal Tumor (GIST) Workshop
The Novartis International website

A biopsy is usually used to aid in the diagnosis of cancer and GIST. Tissue samples are prepared from the biopsy sample. One method that the pathologist uses to classify tissues is called immunohistochemistry. Using this technique, the pathologist applies various "antibodies" to the tissues. These antibodies are each designed to react with specific features (proteins) on the cell surface. The most important antibody that is applied when GIST is suspected is the KIT antibody. When these antibodies bind to their specific target on the cell surface they produce a "stain" or change in color of the sample. So when the KIT antibody is applied, if the cell surface has the KIT protein present on the surface, the sample will "stain positive". This is what is known as "KIT positive" and means that this cell has KIT receptors on its surface.

With very rare exceptions ¹, GIST tumors will stain positive for KIT. This means that the cell is manufacturing or using the KIT protein. In the case of KIT, this protein is a receptor. (See KIT Receptor Image)

Using immunohistochemical staining, GIST cells can be summarized as follows¹:

• KIT positive-nearly 100% of the time (the term c-Kit or CD117 may be used instead of KIT).
• CD34 positve-60% to 70% of the time.
• SMA postive-30% to 40% of the time.
• Desmin positive-very rare
• S-100 positive-5%+

Two other stains (tests) that are expressed in almost all GIST but that are not yet commercially available include:

• PKC theta (PKCθ)
• DOG-1

Prognostic Factors and Risk Assessment

Note: Pediatric GIST (occurring in patients under 18 years of age) may have a very different potential for metastases and these tables should not be used for estimating risk for pediatric GIST.

Many attempts have been made to classify GISTs as to their potential for malignant behavior. Pathologists who are GIST experts, currently think it most prudent to classify GISTs based on risk assessment, rather than try to classify them as benign or malignant. At least some GIST experts think it is unwise to use the term "benign" with GIST and that almost all GISTs should be considered as having some malignant potential.

Although many factors have been suggested to contribute to malignant potential, the two most commonly quoted factors are size and mitotic activity. Recently tumor location has been added as a third important prognostic factor.

In July, 2007, the Journal of the National Comprehensive Cancer Network issued a "NCCN Task force Report: Optimal Management of Patients with Gastrointestinal Stromal Tumor (GIST)-Update of the NCCN Clinical Practice Guidelines²." These guidelines contain the most current method of determining risk (defined as metastasis or tumor-related death). See the table below. The complete guidelines are available on the NCCN website.

Abbreviations: GIST, gastrointestinal stromal tumor; hpf, high power field; Insuff, insufficient.
Adapted from Miettinen and Lasota, 2006. Data are based on long-term follow-up of 1055 gastric, 629 small intestinal,
144 duodenal, and 111 rectal GISTs. (Miettinen et al. 2001, 2005, and 2006).
*Defined as metastasis or tumor-related death.
†Denotes small numbers of cases.

The following table is older (2002) and was developed from the Gastrointestinal Stromal Tumor (GIST) Workshop. It may still be particularly useful in cases of where the primary site is different than those listed in table 1 or the primary site is unknown.

Proposed Approach for Defining Risk of Aggressive Behavior in GISTS1
	Size*	Mitotic Count+
Very low risk	<2cm	<5/50 HPF
Low risk	2-5cm	<5/50 HPF
Intermediate risk	>5cm	6-10/50 HPF
	5-10cm	<5/50 HPF
High risk	>5cm	>5/50 HPF
	>10cm	Any mitotic rate
	Any size	>10/50 HPF
Abbreviation: HPF < high-power field.

*Size represents the single largest dimension. Admittedly this may vary somewhat between prefixation and postfixation and between observers. There is a general but poorly defined sense that perhaps the size threshold for aggressive behavior should be 1 to 2 cm less in the small bowel than elsewhere.

+Ideally, mitotic count should be standardized according to surface area examined (based on size of high-powered fields), but there are no agreed-on definitions in this regard. Despite inevitable subjectivity in recognition of mitoses and variability in the area of high power fields, such mitotic counts still prove useful.

Note: The authors of this paper (Diagnosis of Gastrointestinal Stromal Tumors) also conclude that the risk categories as they define them in their paper "...should prove clinically useful, and in light of the uncertainties expressed herein and the well-recognized tendency of these troublesome tumors to pursue an indolent clinical course with a significant risk of late relapse, we also strongly advocate that all patients with GIST be carefully and regularly followed up for an indefinite period."

Singer et el., risk-assessment

Another paper published in the Journal of Clinical Oncology ⁴ is also interesting. The tumor samples from the 48 patients studied in this group were well characterized as GIST tumors (some older studies contain some samples that are probably not true GISTs). In this series of KIT-expressing GISTs, tumor mitotic activity and histologic subtype were the most important prognostic features. The type/location of KIT mutation also served as an independent predictor for disease-free survival. Tumors with a spindle cell histology (n=32) had longer recurrence-free survival than patients with epithelioid cell tumors (n=3), and patients with a mixed spindle/epithelioid (n=13) cell type had the lowest recurrence-free survival times. In this study, 3 factors were used to divide the tumors into different grades/risk categories. The factors were: mitotic count per 30 HPF, histologic subtype (spindle, epithelioid, or mixed cell type), and the presence or absence of atypia.

Singer et el., Risk-Assessment
	Mitotic count	Cell type	Presence of Atypia
Low risk	<2/30 HPF	Spindle	None
Low risk	0/30 HPF	Epithelioid
High risk	>5/30 HPF	Spindle	None
High risk	3 to 5/30 HPF	Spindle	With frank atypia
High Risk	>2/30 HPF	Epithelioid

Tumors with parameters falling between low risk and high risk were classified as intermediate-risk tumors (according to grade). This paper has some nice charts showing recurrence-free survival as a function of: grade, tumor size, histologic subtype, mitotic activity, presence or absence of missense exon 11 mutation in KIT, and presence or absence of deletion/insertion exon 11 mutation in KIT.

Other important factors for a high risk of recurrence include:

• Rupture of the tumor either prior to or during surgery.
• Failure to obtain clear margins during surgery.

References:

1. Diagnosis of Gastrointestinal Stromal Tumors: A Consensus Approach-Christopher D. M. Fletcher, MD, FRCPATH, Jules J. Berman, MD, PhD, Christopher Corless, MD, PhD, Fred Gorstein, MD, Jerzy Lasota, MD, PhD, B. Jack Longley, MD, Markku Miettinen, MD, Timothy J. O'Leary, MD, PhD, Helen Remotti, MD, Brian P. Rubin, MD, Phd, Barry Shmookler, MD, Leslie H. Sobin, MD, and Sharon W. Weiss, MD
2. NCCN Task Force Report: Optimal Management of Patients with Gastrointestinal Stromal Tumor (GIST)—Update of the NCCN Clinical Practice Guidelines
George D. Demetri, MD; Robert S. Benjamin, MD; Charles D. Blanke, MD; Jean-Yves Blay, MD, PhD; Paolo Casali, MD; Haesun Choi, MD; Christopher L. Corless, MD, PhD; Maria Debiec-Rychter, MD, PhD; Ronald P. DeMatteo, MD; David S. Ettinger, MD; George A. Fisher, MD, PhD; Christopher D.M. Fletcher, MD, FRCPath; Alessandro Gronchi, MD;
Peter Hohenberger, MD, PhD; Miranda Hughes, PhD; Heikki Joensuu, MD; Ian Judson, MD, FRCP; Axel Le Cesne, MD;
Robert G. Maki, MD, PhD; Michael Morse, MD; Alberto S. Pappo, MD; Peter W.T. Pisters, MD; Chandrajit P. Raut, MD, MSc;
Peter Reichardt, MD, PhD; Douglas S. Tyler, MD; Annick D. Van den Abbeele, MD; Margaret von Mehren, MD; Jeffrey D. Wayne, MD; and John Zalcberg, MBBS, PhD
3. Gastrointestinal Stromal Tumors of the Stomach. A Clinicopathologic, Immunohistochemical, and Molecular Genetic Study of 1765 Cases With Long-term Follow-up. American Journal of Surgical Pathology. 29(1):52-68, January 2005.
Miettinen, Markku MD *; Sobin, Leslie H MD +; Lasota, Jerzy MD

4. Prognostic Value of KIT Mutation Type, Mitotic Activity, and Histologic Subtype in Gastrointestinal Stromal Tumors. By Samuel Singer, Brian P. Rubin, Marcia L. Lux, Chang-Jie Chen, George D. Demetri, Christopher D. M. Fletcher, and Jonathon A. Fletcher