In Chapter 33, Drs. Look Hong and Raut review gastrointestinal stromal tumor (GIST). Historically, metastatic GIST carried a median survival of 12 months, but this has improved to 5 years with the introduction of tyrosine kinase inhibitors (TKIs).^1,2 The landmark finding of the effectiveness of imatinib in GIST has transformed a disease for which little treatment beyond surgery existed into a chronic disease for which multimodality therapy can give patients prolonged survival with good quality of life and possibly even cure. Following the success of imatinib in GIST, the paradigm of targeted molecular therapy has rapidly expanded, with US Food and Drug Administration (FDA) approval of orally available targeted molecular therapy agents for multiple targets in multiple solid tumors. Initial success with targeting tyrosine kinase receptors in non–small-cell lung cancer (EGFR and later ALK) and renal cell carcinoma (VEGFR) has now been broadened to include approval of drugs for breast cancer, colorectal cancer, thyroid cancer, soft tissue sarcoma, and pancreatic neuroendocrine tumor. In some instances, just 1 or 2 kinases are targeted, whereas in others, multiple kinases are inhibited. Additional available agents are now also widely used for targets downstream of tyrosine kinases, including BRAF in melanoma; mTOR in breast cancer, angiomyolipoma, and pancreatic neuroendocrine tumors; CDK4/6 in breast cancer; PARP in ovarian cancer; and Hedgehog in basal cell carcinoma.³ While experience gained in the treatment of GIST has paved the way for the introduction of targeted therapy in many other cancers, it has also taught us valuable lessons as to complex issues that may arise. The simplicity of targeting KIT or PDGFRA in GIST has turned out to be far more complex than initially thought, with response to therapy dependent on the specific mutation and the nearly inevitable development of secondary mutations and other mechanisms of resistance in advanced and recurrent disease.

While molecular therapy is highly effective, it is important to recognize that surgery remains the only chance for cure in patients with GIST. We now know from the placebo arm of the American College of Surgeons Oncology Group (ACOSOG) Z9001 trial that 70% of patients with GISTs ≥3 cm appear to be cured by surgery with 74 months of follow-up.⁴ Given the diverse anatomic locations of primary and metastatic disease, a variety of technical issues must be mastered for appropriate and safe treatment of GIST. Although surgical resection is the central component of treating GIST with curative intent, the surgeon treating GIST must also become knowledgeable in 3 key elements of the molecular therapy for this disease, which we highlight here: (1) appropriate patient selection and duration of neoadjuvant therapy; (2) adjuvant therapy in an increasingly well-defined subset of intermediate- and high-risk patients; and (3) surgical resection in well-selected patients with recurrent or metastatic disease.

For the surgeon planning to resect a GIST, it is important to understand a number of unique surgical principles. GISTs are extremely friable, especially after neoadjuvant imatinib, and thus may rupture if not handled with great care. Tumor rupture must be avoided as it may lead to early peritoneal recurrence and is known to worsen outcome. As such, tumor handling during laparoscopy is best achieved by minimal direct contact with the tumor, and once resected, the tumor must be removed via a retrieval bag to prevent tumor seeding. GISTs are particularly vascular, with the potential for significant blood loss if the extensive venous and arterial collateral vessels are not carefully dissected and ligated. While GISTs do not typically invade into adjacent organs and instead push them aside, any organ adherent to a GIST should be removed at least partially en bloc.

Since GISTs arise in diverse locations, with highest frequency in the stomach followed by small intestine, rectum, and occasionally other sites, site-specific considerations are numerous. Laparoscopic (and possibly robotic) surgery is appropriate in each of these sites for small- to medium-sized tumors if the surgeon has adequate expertise and the appropriate equipment. When these tumors are small and have a low mitotic rate, cure by surgery alone is possible, particularly for gastric GISTs. During laparoscopic exploration, exophytic gastric tumors are readily identified, especially along the greater curvature and anterior stomach. Tumors that grow into the stomach lumen instead of exophytically can be identified by intraoperative endoscopy, or by filling the stomach with water and using laparoscopic ultrasound, although this is rarely necessary. Posterior gastric tumors are usually still amenable to minimally invasive resection but require extensive mobilization of the stomach, which is further facilitated by retracting the left lateral segment of the liver to the right. Formal anatomic gastrectomy is typically not required; rather, partially gastrectomy with a 1-cm margin is adequate.

Partial gastrectomy is frequently possible using surgical staplers, unless doing so will lead to unacceptable compromise of the gastric lumen. In this case, a gastrotomy is made and the tumor resected with a 1-cm margin using cautery, followed by hand-sewn closure of the defect. For GISTs arising from the gastroesophageal junction (GEJ), we prefer open surgery, especially for posterior tumors. We typically remove these tumors by gastrotomy and hand-sewn closure; however, if staplers are used, a bougie should be placed in the esophagus to avoid narrowing the GEJ. Total gastrectomy is rarely required for massive gastric tumors, and the postoperative complications should be considered and discussed with the patient preoperatively. Such large tumors may be intimately associated with other organs including the distal pancreas, spleen, and even the splenic flexure of the colon, possibly requiring en bloc resection. When the extent of the tumor is recognized preoperatively, neoadjuvant imatinib (as discussed below) is highly recommended and may result in organ preservation.

The small intestine is the next most frequent site of primary GIST. Jejunal and ileal GISTs are readily removed laparoscopically, unless very large. They are easily identified by systematically running the small bowel from the ligament of Treitz to the terminal ileum. Duodenal GISTs are perhaps the most complex to deal with. Tumors located in the second portion of the duodenum may require a pancreaticoduodenectomy, unless small and arising from the lateral wall. In that case, local resection may be possible either using primary closure of the defect or reconstruction with a Roux-en-Y jejunal limb. GISTs in the third or fourth portion of the duodenum can be resected after appropriate mobilization and reconstructed either with a direct anastomosis to jejunum or with closure of the distal duodenum and creation of a Roux-en-Y limb of jejunum to the lateral second portion of the duodenum. For a duodenal tumor that is larger than a few centimeters, neoadjuvant imatinib should be considered if it may reduce the extent of the operation required.

GIST rarely arises in the colon, but rectal GISTs make up approximately 5% of primary tumors and thus deserve discussion. Akin to adenocarcinoma of the rectum, an organ-preserving neoadjuvant approach should be employed. This may allow avoidance of abdominoperineal resection and permanent colostomy. Likewise, it may allow local transanal resection for very distal tumors.

Many patients with GISTs already have a tissue diagnosis by the time of referral. However, for tumors with a classic radiologic appearance, biopsy may not be necessary, although the patient should be counseled preoperatively that other tumors such as Schwannoma, leiomyoma, or even ectopic pancreas are possible. Biopsy should be performed if identification of a benign tumor may avoid an otherwise morbid resection, such as an esophagectomy for a GEJ tumor. Likewise, if neoadjuvant therapy is being considered for a large tumor, biopsy is generally recommended to confirm the tumor is a GIST.