Based on their histology, GISTs were originally considered to be derived from smooth muscle. However, they rarely showed clear-cut features of complete muscle differentiation. Additionally, in many cases, their immunophenotypic profile differed from that of leiomyomas arising from other sites (e.g., the uterus or soft tissue ). The understanding of GIST biology changed significantly with the identification of the near-universal expression of the CD117 antigen, also known as proto-oncogene c-kit, in GISTs in the late 1980s [15]. At that time, it was shown that the interstitial cells of Cajal that are part of the autonomic nervous system of the intestine and that serve a pacemaker function in controlling motility express the KIT receptor [16]. Interstitial cells of Cajal have immunophenotypic and ultrastructural features of both smooth muscle and neuronal differentiation. Because GISTs, like interstitial cells of Cajal, express KIT, interstitial cells of Cajal are thought to be the cell of origin. Additionally, as two-thirds of GISTs express CD34, it is postulated that GISTs originate from CD34-positive stem cells within the gut wall differentiating toward the pacemaker cell phenotype with time [17, 18].

The differential diagnosis of a subepithelial tumor arising in the gastrointestinal tract is broad, and histologic findings observed on haematoxylin and eosin-stained sections are not specific for GIST . The cellular morphology of GISTs is mainly divided into three categories, namely the spindle cell type (70 %), epithelioid type (20 %) and mixed type (10 %) [14, 19]. Whereas gastric, small intestinal and colonic GISTs are mostly composed of spindle cell tumours, KIT-negative GISTs are more often of the epithelioid type [20]. The epithelioid variant may show discohesive, hypercellular, sarcomatous morphology with significant atypia and mitotic activity [21] (Fig. 31.1).

A significant breakthrough was the discovery that most GISTs show strong positivity for CD117 (KIT) in contrast to leiomyomas, true leiomyosarcomas and other spindle-cell tumors of the GI tract, which were typically CD117 negative [22]. CD117 is an antigen that is part of the KIT transmembrane receptor tyrosine kinase (RTK) family and is the product of the KIT proto-oncogene (also denoted c-kit). In more than 80 % of GISTs, a mutation in the KIT gene leads to a structural variant of the KIT protein, which is abnormally activated and plays an essential role in cell survival, proliferation and differentiation. When KIT binds to its ligand, it forms a dimer that activates its intrinsic tyrosine kinase activity that, in turn, phosphorylates and activates signal transduction molecules that propagate the signal in the cell (Fig. 31.2).

Immunohistochemically, most GISTs (>90 %) show strong positivity for CD117 and usually negativity for desmin and S-100, which are positive in smooth muscle and neural tumours [23]. Although KIT positivity is a major defining feature for GIST , its expression may not be sufficient for diagnosis. KIT-positive malignancies include metastatic melanoma, angiosarcoma, the Ewing’s sarcoma family of tumours, seminoma, and others [24]. Other commonly expressed markers of GIST include CD34 antigen (70 %), smooth muscle actin (SMA; 30–40 %), desmin (<5 %), and S100 protein (~5 %) [25]. In contrast to GIST, leiomyoma and leiomyosarcoma are positive for SMA and desmin and negative for KIT and CD34. Malignant melanoma exhibits diffuse immunoreactivity for S100 protein but can be focally positive for KIT. Schwannomas are strongly and diffusely immunoreactive for S100 protein and negative for KIT [26] (Fig. 31.3).

A small subset of GISTs lacks the characteristic KIT mutations [20, 27]. In a proportion of these tumours, activating mutations in the related RTK, PDGFRA, were detected [28]. Many of these PDGFRA-mutant GISTs have an epithelioid morphology. Immunostaining with PDGFRA was shown to be helpful in discriminating between KIT-negative GISTs and other gastrointestinal mesenchymal tumors [29, 30].

DOG1, a calcium-dependent, chloride channel protein, is another highly sensitive and specific marker that often reacts with CD117-negative GISTs [31]. DOG1 expression does not appear to be different between the KIT/PDGFRA mutant or wild-type GISTs. Hence, this marker can be used to diagnose KIT-negative tumour variants.

Inactivation of the succinate dehydrogenase (SDH) complex appears to be an event shared by sporadic and syndromic GISTs that lack mutations in KIT and PDGFRA [32]. Immunohistochemical loss of succinate dehydrogenase subunit B (SDHB) has been shown to be a practical marker to identify SDH-deficient GISTs [33].

The experience with these novel immunomarkers (other than KIT) is currently limited, and problems exist concerning the quality and availability of the commercial antibodies used to stain for them.

Complete surgical removal (R0 excision) of localised GISTs is the mainstay of treatment for potentially resectable tumours with a size ≥2 cm [57]. Routine lymph node dissection should not be performed because lymph node metastasis is an extremely rare event [58]. Nevertheless, approximately 50 % of GISTs will recur [43]. Resection can be performed by traditional open surgery or laparoscopic surgery, although the latter approach should only be performed by surgeons with expertise in the laparoscopic management of cancer and mainly for gastric primaries [59]. The importance of achieving negative microscopic margins is a controversially discussed issue because a negative impact on OS in patients treated with adjuvant imatinib is lacking. However, R1 resection may be associated with a greater risk for recurrence [60]. A re-resection in a R1 situation is not mandatory but may be carried out if functional sequelae are not expected. Depending on the primary tumour site (oesophago-gastric junction, small intestine, rectum), neoadjuvant treatment with imatinib should be considered (see Chap. 7.2).

The natural history of small oesophago-gastric and duodenal lesions smaller than 2 cm in size regarding the growth rate and metastatic potential is difficult to anticipate. Many of these lesions will have a very low risk of tumour progression and a low metastatic potential. Endoscopic biopsy may be difficult, and tumour spillage remains a relevant risk. Hence, endoscopic ultrasound assessment and regular follow-up are reasonable in these cases. Should there be any feature of malignant behaviour on ultrasound a resection should also be performed. An algorithmic approach to the management of gastric GISTs based on size and endoscopic ultrasound (EUS) appearance has been proposed [49].

The aim of neoadjuvant systemic therapy is to reduce the size of a locally advanced GIST to increase the likelihood of complete resection, reduce surgical morbidity and eventually limit the risk of tumour rupture. Because there are no prospective randomised data, the recommendations on neoadjuvant imatinib therapy are largely based on a few prospective, non-randomised and mainly retrospective studies [61–64].

Eisenberg and colleagues published a prospective phase II RTOG0132/ACRIN6665 trial investigating the feasibility of neoadjuvant imatinib in KIT-positive, resectable ≥5-cm primary GIST , or resectable, recurrent GIST. Sixty-three patients received 600 mg/day of imatinib for 8–12 weeks prior to surgery and then continued imatinib for two additional years. Among the patients with localised primary disease, only 2 (7 %) had an objective response to preoperative imatinib, but stable disease was achieved in 25 (83 %) patients. In 77 % of these patients, complete resection could be performed. The present study confirmed the safety of administering imatinib neoadjuvantly, although the treatment period was quite short [61]. Another open-label, single-arm phase II study from Canada investigated neoadjuvant imatinib treatment with 400–600 mg daily in patients with locally advanced or metastatic GIST that was potentially resectable. Imatinib was administered for a maximum of 12 months to a maximal tumour response. Six of 14 patients showed a partial response, and 8 showed stable disease; no progressive disease was documented. The median treatment duration was 9 months. Therefore, the authors concluded that the optimal preoperative treatment duration should be between 6 and 12 months [64].

Taken all together, the data reveal that there is no consensus regarding the indications for neoadjuvant therapy because a particularly treatment benefit was not proven. However, preoperative therapy is a widely accepted concept, particularly in large, bulky tumours of any origin and notably in GIST arising in the oesophagus, oesophago-gastric junction, duodenum and distal rectum, to reduce significant surgical morbidity. Importantly, a biopsy to confirm the diagnosis and exclude imatinib-resistant mutations is mandatory. The treatment response to imatinib should be evaluated early during the treatment course to exclude tumour progression and prepone resection.

To date, questions regarding the imatinib dose in patients with exon 9 mutation and the duration of additive adjuvant treatment in this specific situation remain unanswered, but a total duration of 3 years appears reasonable.

Although surgery remains the therapeutic modality of choice for localised GIST , the risk of recurrence following complete excision is still eminent. In a recently published analysis of a pool of 2,560 patients, including 10 different population-based published series, the estimated 5-, 10-, and 15-year relapse-free survival [RFS] rates were 71 %, 63 %, and 60 %, respectively [54]. This meaningful risk of recurrence is likely due to persistent microscopic disease following surgery. Therefore, the effect of adjuvant systemic treatment with imatinib has been explored subsequently to improve the likelihood of survival in patients with a high risk of recurrence. However, there is no clear consensus from expert groups regarding the level or cutoff of recurrence risk that would justify the use of adjuvant imatinib [44].

After a few phase II trials with very promising results, the benefit of adjuvant imatinib therapy has been evaluated in at least three randomised studies.

In the multicentre, randomised, double-blind and placebo-controlled US trial Z9001, 713 patients with a resected GIST and a tumour ≥3 cm in size were included and patients were randomly assigned to imatinib 400 mg/day or placebo for 1 year. The study was closed after the first interim analysis, which confirmed a significant reduction in recurrence-free survival that was subsequently the primary endpoint. After a median follow-up of 19.7 months, the 1-year RFS rate was 98 versus 83 % favouring imatinib, with a hazard ratio for RFS of 0.35 and a 95 % CI of 0.22–0.53. A benefit in terms of OS could not be confirmed most likely due to cross-over to active treatment and the short duration of follow-up. Imatinib was well tolerated and showed the known toxicity profile (see below) [65]. That pivotal study led to the accelerated approval of imatinib for the adjuvant treatment of completely resected GISTs ≥3 cm in size. Notably, patients were not stratified according to tumour site and mitotic rate.

The second practise-changing phase III trial was performed by the Scandinavian Sarcoma Group (SSG) XVIII comparing 12 versus 36 months of adjuvant imatinib treatment. Eligible patients were of high risk defined according to the modified consensus criteria as having at least one of the following: a tumour size >10 cm, a mitotic count >10/50 high-power fields (hpf), a tumour size >5 cm with a mitotic rate >5/hpf, or tumour rupture. After recruitment of 400 patients with a median follow-up of 54 months, patients in the 3-year arm showed a significant improvement in RFS, the primary endpoint (5-year RFS, 66 versus 48 %; HR, 0.46; 95 % CI, 0.32–0.65) as well as overall survival (OS, 92 versus 82 %; HR, 0.45; 95 % CI, 0.22–0.89). Subgroup analysis demonstrated that patients with exon 9 or PDGFRA mutation did not show a treatment benefit. In summary, these data established at least 36 months of adjuvant imatinib as a new standard for patients with high-risk GIST [66].

Recently, an abstract of the EORTC 62024 study randomising GIST patients between 2 years of adjuvant imatinib and no adjuvant treatment was presented and showed no significant benefit in the primary endpoint, which was imatinib-free survival, under the intermediate- and high-risk scenario [67]. These results per se implicate that progression of GIST may be delayed but survival might not be improved with the available TKIs.

A few outstanding questions need further investigation. First, whereas there is a consensus that PDGFRA D842V-mutated GISTs should not be treated with adjuvant therapy due to their lack of imatinib-sensitivity, the treatment dose in patients with exon 9 mutation is a matter of debate and 800 mg/day of imatinib may be used analogous to the evidence in the metastatic tumour stage. However, there are often regulatory problems limiting this practise. Additionally, we could not confirm whether “wild-type” GISTs also benefit from adjuvant therapy considering their lower sensitivity to imatinib and more indolent natural history [37, 38, 68].

Second, the question remains concerning the optimal treatment duration and whether treatment should be continued for longer than 3 years. In the Scandinavian trial from Joensuu et al., in both groups, within 6–12 months of discontinuation of adjuvant imatinib, the rates of disease recurrence were similarly increased [66]. Similarly, we know from the BFR-14 trial, in patients with advanced GIST , that some patients who had a complete response to imatinib relapsed even after 5 years of treatment when therapy was interrupted [62]. Hence, the latter findings raises questions as to whether recurrences are truly being prevented or just delayed and whether the duration of adjuvant therapy should be beyond 3 years. Currently, a phase II, non-randomised, open-label multicentre study is investigating 5 years of adjuvant imatinib therapy in patients at significant risk for recurrence following complete resection of primary GISTs (NCT00867113).

Additionally, the optimal treatment duration in the case of tumour rupture is unknown given the uncertainty concerning whether these patients should be viewed as virtually metastatic.

Finally, there is no consensus concerning the definition of high-risk GIST , which depends on different risk classifications.