Discussion

On fine needle aspiration, gastrointestinal stromal tumors can resemble other spindle cell neoplasms, with clusters of spindle shaped cells with elongated nuclei, and occasional single cells in the background. These samples can vary from minimally to highly cellular, with loose to cohesive groups of tightly packed cells with a background of dispersed cells. The nuclear chromatin can vary from fine to coarsely granular, with inconspicuous to large nucleoli. Intranuclear inclusions may be seen. The stroma, if sampled, may be loosely fibrillar or hyalinized.

 

Cytoplasm is generally delicate with wispy extensions that may be bipolar. The background dispersed cells may be stripped of cytoplasm, but may occasionally have attached cytoplasm. The epithelioid GIST will have polygonal / epithelioid cells with rounded nuclei and more abundant cytoplasm. Epithelioid GIST may also contain multinucleated cells and have a plasmacytoid appearance. The cytomorphology of GIST make it difficult to distinguish them from other spindle cell neoplasms, including leiomyoma and leiomyosarcoma.

 

Other entities that are included in the differential are nerve sheath tumors. With a differential inclusive of lesions from benign to malignant, it is important to obtain enough material for a cell block, as immunohistochemical staining of the cell block sections is important to hone the differential. The immunophenotype of GIST varies based upon its anatomic location. In the stomach, most GI stromal tumors have strong positive staining for CD117. The staining pattern is most commonly membranous. CD34 is also usually positive in spindle cell GIST. DOG-1 is also a useful marker, especially for the small percentage of gastric GIST that are negative for CD117. Notably, the gastric GIST that are negative for CD117 are more likely to harbor platelet derived growth factor alpha (PDGFRA) mutations. The most common PDGFRA mutation occurs at exon 18. This subset of gastric GIST are also more likely to have epithelioid morphology and may also be resistant to imatinib.

 

Currently, there are no agreed upon cytologic criteria to determine which GIST will progress. Size and anatomic location have been shown to be important prognostic factors, with no recurrence or progression of disease in lesions less then 2 cm. Mitotic rate, while informative, does not adequately stratify the risk of disease in FNA specimens. Pleomorphism is occasionally seen in overtly malignant GIST, however the lack of pleomorphism is not indicative of risk of progression. While FNA is an excellent tool to make the diagnosis of GIST, risk stratification of the tumor is best assessed on the surgical resection specimen. The risk of progressive disease depends upon anatomic location, tumor size, and mitotic activity. For example, gastric GIST smaller than or equal to 2 cm and with 5 or fewer mitotic figures per 50 high power fields have no risk of progressive disease while those greater than 5 cm with more than 5 mitotic figures per 50 high power fields have between 55% to 85% risk of progressive disease. FNA biopsy of submucosal lesions seen on endoscopy is an important component of the preoperative workup. The differential diagnosis of submucosal lesions of the gastrointestinal tract is broad, and FNA is frequently diagnostic in many cases, and therefore able to guide surgical planning or medical management.

 

Differential Diagnosis:
Leiomyoma, Inflammatory fibroid polyp, Leiomyosarcoma, Glomus tumor, Schwannoma, Gangliocytic paraganglioma, Fibromatosis.