SEOM-GEIS clinical guideline for gastrointestinal stromal tumors (2022)

Gastrointestinal bleeding and abdominal pain are the most frequent symptoms at diagnosis. Bleeding can be chronic (anemia) or acute (hematemesis or melena) requiring an urgent intervention. Additionally, presentation can be as acute abdomen due to tumor rupture or small bowel perforation. In some cases, presentation is asymptomatic [6, 7].

GISTs are generally diagnosed with upper endoscopy (gastric/duodenum locations). Upper endoscopic ultrasound (EUS) is useful for the detection of small intramural lesions. Abdomen–pelvis contrast-enhanced computerized tomography (CT) scan with image acquisitions of the arterial and portal phases is indicated for evaluating tumor extension (III, A). GIST rarely metastasizes to thorax, so thorax CT-scan is not routinely recommended. MRI is the best imaging technique for rectum GIST or for characterizing uncertain liver lesions [I, A]. Positron Emission Tomography (PET)-CT is not routinely recommended, unless early prediction of response to first-line imatinib is anticipated during the initial work-up [8]. Serum tumoral markers are not required for GIST diagnosis [9]. A core- (preferable) or a fine-needle biopsy can be performed via EUS, and it is enough for the GIST diagnosis prior surgery. However, if radiological images are conclusive or highly suggestive of GIST, surgery can be performed without a prior histological assessment. This is the case of abdominal nodules or masses not amenable of endoscopic assessment and in which an external biopsy can lead to intraperitoneal tumor spillage or rupture [10]. Noteworthy, if neoadjuvant or first-line treatment is primarily indicated, a core needle biopsy is mandatory for diagnosis (III, A) [9]. In some cases, GISTs are smaller than 2 cm and surgical excision could be the only approach for histological diagnosis.

A core needle biopsy is the standard approach for histological and molecular assessment. The pathology evaluation should include the description of morphological features, mitotic count expressed in number of mitoses per 5 mm², and immunohistochemistry (IHQ) determinations. A recommended IHQ panel includes the following antibodies and rates of positivity: CD117 (95%), DOG1 (98%), CD34 (70–90%), actin (20–30%), S100 (8–10%), and desmin (2–4%) [11]. In rare cases with suspicion of GIST diagnosis but negative for CD117 and DOG1, molecular determination of KIT/PDGFRA mutations can be of aid [12].

In surgically resected specimens, the following features should be reported: site and size of the primary tumor, mitosis per 5 mm² in the most proliferative areas, margins, histologic sub-type (spindle cell, 77%; epithelioid, 8%; or mixed, 15%), presence of necrosis, and tumor rupture or perforation. In rare cases of resected lymph nodes under the surgical suspicion of tumoral involvement, their pathological description should be added. The same IHQ panel can be used here. Importantly, Ki-67 immunostain must not replace the mitotic count [11, 12].

Between 85 and 90% of GISTs are caused by activating mutations in the KIT or PDGFRA genes [12]. It is strongly recommended to perform the mutational analysis, in either localized or metastatic lesions, when medical treatment is indicated, as these analyses provide prognostic and predictive information for response to approved tyrosine kinase inhibitors (TKIs) (II, A). KIT/PDGFRA molecular studies are commonly performed in paraffine-embedded formalin-fixed tumors through Sanger or Next-Generation Sequencing (NGS). KIT primary mutations (75%) often emerge in exons 11 and 9, and less frequently in exons 13 and 17. Although all of them are sensitive to first-line imatinib, KIT exon 11-mutants are more sensitive than exon 9, and the latter commonly requires double dose of imatinib for metastatic disease [13]. If KIT exon 11 codons 557 and/or 558 are affected, there is a higher risk of relapse in intermediate-risk, surgically resected GISTs [14]. PDGFRA mutations are commonly found in exon 18 (5% of all GIST), being less frequent in exons 12 or 14. The PDGFRA exon 18 D842V substitution was formerly deemed resistant to all available TKIs until the recent activity shown by avapritinib [15].

Resistance to TKIs in GIST is commonly (> 90%) due to the polyclonal emergence of secondary mutations in KIT or PDGFRA [16]. Thus, routine determination of KIT/PDGFRA genotype is not recommended in imatinib-resistant GIST patients given the heterogeneity of resistance mutations (II, D). Likewise, there are no current validated data supporting the use of circulating tumor DNA to take clinical decisions, although it can be assessed with investigational purposes (III, C) [17].

Between 10 and 15% of all GISTs are wild type (WT) for KIT and PDGFRA mutations. In these cases, IHQ for SDHB can identify SDH-deficient GISTs, which represent the overwhelming majority of KIT/PDGFRA WT GIST. The hereditary implications of SDH-deficient GISTs have been discussed above. Of note, the Carney triad syndrome is a not heritable condition consisting of gastric GISTs, pulmonary chondromas, and paragangliomas. Driver alterations in KIT/PDGFRA WT GIST positive for SDHB are diverse and may affect RAS, BRAF, NF1, and NTRK [12].

Approximately 70% of patients with GIST will be cured by surgery alone, and therefore, it constitutes the standard treatment for localized GIST ≥ 2 cm or GISTs with suspicious EUS features [12]. R0 resection (negative margins) with a 1-cm margin is the aim of surgery, although seemingly an R1 surgery (microscopic positive margins) does not involve higher risk of recurrence nor worse survival [20]. Peritoneal or liver spread is exceedingly more common than local recurrence. With the exception of SDH-deficient GISTs, routine local lymphadenectomy is not indicated, although nearby enlarged lymph nodes should be removed [21]. GISTs are prone to bleed, friable, and easy to rupture masses, which is an independent poor prognostic factor [19]. The most common location of GISTs is the stomach and a formal anatomic gastrectomy is rarely required. Conversely, segmental resection with a 1-cm radial margin is the most common approach [22]. Segmental resection is also the best choice in GIST from the small bowel. However, pancreaticoduodenectomy or Whipple resection can be required in duodenal periampullary tumors. Surgery for rectal GISTs with anastomosis to the remaining rectum can be done but in GIST close to the anal sphincter abdominoperineal resection with permanent colostomy is required. Novel transanal techniques to remove rectal GISTs have been described although mostly in case reports.

Laparoscopic surgery is indicated if the GIST is less than 5 cm and up to 8 cm in size can be removed laparoscopically with retrospective data favoring this approach regarding less hospital staying and wound complications [23]. Data from endoscopic resection for small GISTs (most of them < 2 cm) have mixed results, with potential for perforation. Enucleation techniques attempt to preserve organ’s function and can be suitable for polyp-like GISTs with a narrow connection to muscularis propria [12].

In SDH-deficient and NF1-associated GISTs presenting as multifocal disease, surgery is often indicated when one or a few tumors are growing faster than the others or have become symptomatic.

Three randomized phase III trials established the indication and duration of adjuvant treatment in GIST. The ACOSOG Z9001 trial showed superiority in terms of recurrence-free survival (RFS) for imatinib against placebo in resected GIST tumors larger than 3 cm [24]. Later on, the EORTC 62,024/GEIS-10 evaluated 2 vs 0 years in intermediate- and high-risk GIST patients showing a benefit in imatinib failure-free for treated high-risk [25]. This benefit was not statistically significant, but the observed trend supports the results of the Scandinavian/German trial SSGX-VIII/AIO in which 3 years of adjuvant imatinib proved superior regarding overall survival and RFS in high-risk (following NIH modifications) GIST patients [26]. After 10-year follow-up, overall survival at 10 years was 79% in the 3-year arm compared to 65.3% in the 1-year arm [27].

The optimal duration of the adjuvant treatment is still under evaluation. In the single-arm phase II trial PERSIST, no patient with an imatinib-sensitive GIST recurred during the 5-year treatment period [28]. Currently, a phase III randomized trial is evaluating 3- vs 5-year imatinib in high-risk GIST (NCT02413736).

Based on the previous studies, to date, standard adjuvant treatment for surgically resected, imatinib-sensitive high-risk GIST consists on imatinib 400 mg daily for a 3-year period (I,A).

Molecular testing for KIT and PDGFRA is mandatory if adjuvant treatment is indicated (II, A). Data from the metastatic setting, the MetaGIST metaanalysis, showed that KIT exon 9-mutant GISTs are more sensitive to the 800 mg dose [29]. However, no data support this practice in the adjuvant setting and the only available retrospective series suggests that the use of 400 mg daily already provides benefit in this molecular subset (IV, B) [30]. For intermediate-risk patients, molecular testing might help to tip the balance toward adjuvant treatment if KIT exon 11 557–558 codons are affected (III, B) [14]. Patients with imatinib-resistant mutations, such as PDGFRA D842V, should not receive adjuvant treatment (II, E) [15]. The same principle applies for all WT GIST (IV, D) (II, E) [12].

Neoadjuvant treatment can be considered in imatinib-sensitive GIST patients (II, B). Patients benefiting from preoperative treatment include those for which an R0 resection cannot be initially obtained, patients in need of a function-sparing surgery (i.e., rectal GIST), or those considered at risk of bleeding or tumor rupture in which prior imatinib will reduce these risks [31]. Molecular testing is mandatory before its initiation. Patients should be followed up closely at the beginning to confirm the benefit of imatinib. Surgery is commonly undertaken between 6 and 9 months after treatment initiation. The completion of 3 years of adjuvant imatinib after the surgical procedure is based on the risk criteria assessed by the CT-scan (size, location) and the tumor biopsy (mitotic count) prior imatinib initiation.

Surgery for metastatic GIST can be discussed by an expert team within a multidisciplinary tumor board. On the one hand, it can be indicated as a debulking procedure during imatinib response to reduce the number of resistant clones and delay tumor progression (IV, C). Several series have suggested its positive impact in overall survival, although all these studies were retrospective and thus likely biased [32]. On the other hand, surgery can be also considered in the case of unifocal progression (IV, C). This approach can benefit mainly patients on imatinib [33]. However, it can be considered in later lines depending on the time on drug. As all these patients have uncurable metastatic disease, the same TKI used before the surgical procedure must be continued after the resection, regardless of the type of procedure—cytoreductive surgery or resection of resistance unifocal progression.

Finally, the evidence for other approaches, such as embolization or radiofrequency, is lacking, and therefore, surgery should be prioritized. Nonetheless, these patients are metastatic, and therefore, it is important to avoid mutilating procedures, which in turn may leave some room for these alternative procedures discussed on an individual basis by multidisciplinary teams in sarcoma-expert centers (IV, D).