All GIST essentially look very similar histologically, and we cannot definitively predict whether a primary tumor will return just by looking at it under the microscope. Instead of having a binary benign vs malignant classification, we instead utilize a spectrum of risk so that some tumors have a less than 1% chance of recurrence after surgery while others will have a greater than 95% chance of recurrence after surgery. To get to those numbers, we look at 4 variables. 

The first variable is the abdominal organ in which the tumor developed. Approximately 70% of GIST arise in the stomach, and that is a less risky primary site than other sites. The next most common site is the small intestine. Any primary site other than the stomach is a bit riskier, and this information is incorporated into all of the current available risk stratification systems. Tumor size is the second variable, with larger tumors being riskier than smaller tumors. Some of the current classification systems use size cutoff values and others view size as a continuous variable. The third variable, which is likely the most important overall, is the mitotic index. The pathologist determines the mitotic index by examining a biopsy slide and counting the mitotic figures in a standardized area. In the case of GIST and many other tumors, it is 5 mm². The final variable is whether the tumor ruptured, either spontaneously or during manipulation in surgery. Historically, tumor rupture has been associated with an extremely high risk of recurrence (ie, >90%) in some cases. It is a variable that some of the risk classification systems do not even include, but we consider it very important. 

We emphasize that risk stratification should be done after surgery because a biopsy does not provide all of the required information. After surgery, the whole tumor is assessed for mitotic index, and then the region with the highest count is used. Molecular markers are starting to be considered somewhat, but, in most of the classification systems, the mutation status is not an independent predictor of outcome or the recurrence rate. As noted previously, GIST arising in the stomach are often less risky, and this is true despite the fact that they tend to harbor mutations that we think of as less actionable. 

Approximately 80% of SDH-deficient GIST are due to a germline inheritance, and this is helpful to know about because it becomes a family genetic counseling and testing situation. My view is that all gastric GIST should have routine SDH staining up front to identify those that are SDH deficient so that additional testing, including possible germline testing, can be performed. The findings will influence treatment choices and referrals for genetic counseling and testing.

In conclusion, it is likely that risk stratification in patients with GIST will continue to evolve in the coming years. Mutational status will continue to inform the selection of GIST therapy and to help guide the follow-up of patients who were treated initially with imatinib.