Adjuvant and neoadjuvant therapy for gastrointestinal stromal tumors

INTRODUCTION — Gastrointestinal stromal tumors (GISTs) are rare mesenchymal neoplasms of the gastrointestinal (GI) tract. GISTs that arise from the bowel wall typically present as subepithelial neoplasms in the stomach and small intestine; however, they can arise in any portion of the gastrointestinal tract and, occasionally, the omentum, mesentery, and peritoneum. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Clinical presentation'.)

Most GISTs harbor characteristic mutations in KIT or platelet-derived growth factor receptor-alpha (PDGFRA), which result in constitutive activation of the tyrosine kinase receptors KIT and PDGFRA, respectively. Other GISTs have no detectable mutations in either KIT or PDGFRA (ie, "wildtype" GISTs); a majority of these have molecular alterations in either succinate dehydrogenase (SDH) or other genes. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Molecular alterations'.)

Small molecule tyrosine kinase inhibitors (TKIs), such as imatinib, inhibit tumor proliferation by blocking KIT and PDGFRA signaling, and are used to treat advanced and metastatic disease. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors".)

TKIs can also be used to treat GISTs in the adjuvant setting (ie, after complete resection of a primary tumor at sufficient risk of recurrence) and the neoadjuvant setting (ie, to reduce tumor burden in locally advanced disease prior to resection). The approach to adjuvant and neoadjuvant therapy for GISTs will be discussed here. The clinical presentation, diagnosis, prognosis, local treatments, and the use of TKIs for advanced GISTs are discussed separately.

●(See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors".)

●(See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract".)

●(See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors".)

OVERVIEW OF APPROACH TO THE PATIENT — An approach to management of patients with apparently localized disease is summarized as follows:

Apparently localized disease

●Surgery versus surveillance – All apparently localized GISTs ≥2 cm in size should be resected. However, there is no consensus on the management of smaller tumors, and management must be individualized. Although small gastric GISTs may be followed endoscopically until they grow or become symptomatic, the optimal frequency of follow-up and specific risks of this strategy are uncertain. An algorithmic approach to management of GISTs based upon size and endoscopic ultrasound (EUS) appearance (algorithm 1) [1] has been adopted by the National Comprehensive Cancer Network (NCCN) [2] for gastric GISTs but not those at other sites in the GI tract. We agree with this approach. When endoscopic assessment is not possible, excision is the standard approach. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract", section on 'GIST and leiomyoma'.)

●Initial surgery versus neoadjuvant imatinib – The goal of surgical treatment is macroscopically complete resection, if possible. While surgical resection is the treatment of choice for potentially resectable tumors, initial neoadjuvant therapy with imatinib may be preferred for a locally advanced tumor without distant metastases where decreasing tumor burden with systemic therapy could facilitate resection and/or reduce surgical morbidity (eg, a GIST arising in the esophagus, esophagogastric junction, duodenum, or distal rectum) (algorithm 2). In such cases, a tumor biopsy should be performed to confirm the diagnosis and establish tumor genotype:

•We do not administer neoadjuvant imatinib to patients who have a platelet-derived growth factor receptor-alpha (PDGFRA) D842V mutation, or those with wildtype tumors (neither KIT nor PDGFRA mutations) such as a succinate dehydrogenase (SDH)-deficient or neurofibromatosis (NF)-related GIST, and instead proceed directly to surgery. (See 'Role of genotyping' below and "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'KIT/PDGFRA wild-type GISTs'.)

Although European Society for Medical Oncology (ESMO) guidelines suggest that neoadjuvant avapritinib could be considered for patients with PDGFRA D842V mutated tumors [3], there are no clinical trial data on the efficacy or safety of avapritinib in this setting. The use of avapritinib in advanced or metastatic GIST is discussed separately. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Avapritinib for PDGFRA D842V mutant tumors'.)

•For others, the usual dose of imatinib is 400 mg daily. For the rare patient who is being considered for neoadjuvant therapy and whose tumor harbors an exon 9 KIT mutation (the majority of these mutations are in small bowel primaries, which are uncommonly treated neoadjuvantly), which confers relative resistance to imatinib, an initial dose of 800 mg per day may be preferred, if tolerated.

The optimal duration of neoadjuvant imatinib is not established. In most cases, patients are treated to "maximal response," usually not exceeding 10 to 12 months. (See 'Duration of neoadjuvant therapy' below.)

●Adjuvant imatinib – For patients who undergo initial resection, rather than neoadjuvant imatinib, the decision to pursue adjuvant imatinib depends on an estimation of the risk of recurrence and molecular genotyping (algorithm 3):

•Estimating recurrence risk – Recurrence risk is typically based upon tumor size, mitotic index, location within the GI tract, the presence or absence of tumor rupture (either spontaneously or during surgery (table 1)), and completeness of resection. The AFIP prognostic model (table 2 and table 3) is most commonly used in the United States, although others are available. In this prognostic model, tumors are generally classified as no, low, intermediate, or high risk for recurrence. (See 'Estimation of recurrence risk' below.)

•Molecular genotyping – We also perform molecular analysis on all patients with resected GIST who are being evaluated for adjuvant therapy. We do not offer adjuvant therapy to patients with a PDGFRA D842V mutation, or those with a KIT and PDGFRA wildtype tumor (eg, an SDH-deficient or NF-related GIST), as these mutations are associated with primary resistance to imatinib. (See 'Molecular subtypes and primary resistance' below.)

•High-risk GIST – For patients with a completely resected primary high-risk GIST with a KIT or PDGFRA mutation other than D842V, we suggest adjuvant imatinib for a minimum of three years rather than a shorter duration of therapy or surveillance alone. (See 'SSG XVIII trial (High-risk GIST)' below.)

•Intermediate-risk GIST – For those with completely resected primary intermediate-risk GIST with a KIT or PDGFRA mutation other than D842V, we suggest adjuvant imatinib for a minimum of three years rather than shorter treatment duration or surveillance alone, after an informed discussion about the risks and benefits of therapy. Patients who decline or are ineligible for adjuvant imatinib may be offered surveillance. Patients with intermediate-risk disease were included in the PERSIST-5 trial, which demonstrated no recurrences in those with imatinib-sensitive tumors after five years of imatinib therapy. However, other trials demonstrating the benefits of adjuvant imatinib have excluded patients with intermediate-risk disease. (See 'Optimal duration of therapy' below.)

•No-, very low-, or low-risk GIST – For those with a completely resected primary GIST who are at no, very low, or low risk of recurrence, we offer surveillance.

•Dose – For most patients, the standard dose of adjuvant imatinib is 400 mg daily. However, for patients whose tumors harbor a KIT exon 9 mutation, we suggest higher-dose imatinib (800 rather than 400 mg daily), if tolerated. (See 'ACOSOG Z9001' below.)

•Duration of therapy – In patients receiving adjuvant therapy, adjuvant imatinib is administered for a minimum of three years after resection. The optimal duration of adjuvant treatment is not yet established. (See 'SSG XVIII trial (High-risk GIST)' below.)

For all patients undergoing neoadjuvant imatinib, imatinib should be continued postoperatively to complete a total of at least three years of imatinib therapy (combined preoperative and postoperative). Use of preoperative imatinib prohibits accurate assessment of recurrence risk, based upon analysis of the surgical resection specimen. (See 'Postoperative management' below.)

Metastatic disease

●Initial imatinib – Most patients with metastatic disease (even potentially resectable disease) are treated with imatinib rather than initial attempted resection. Patients who present with metastatic disease should undergo an initial biopsy to confirm the diagnosis and to establish the tumor genotype. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors".)

●Indications for surgery – There is no consensus as to the indications for surgical management in patients with metastatic GIST. Surgery may be considered in patients with potentially resectable metastases who do not develop generalized disease progression while receiving therapy with a tyrosine kinase inhibitor (TKI) and for the rare patient who becomes resectable after receiving neoadjuvant imatinib for initially unresectable metastatic disease at a limited number of sites. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract", section on 'Metastatic GIST with potentially resectable disease' and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Treatment of disease refractory to imatinib'.)

Resection may also benefit selected patients with more advanced disease who are responding to imatinib or sunitinib (ie, those who have a partial response, stable disease, or focal progression). The purpose of resection in this setting is to delay or prevent the development of resistant clones by reducing tumor burden. Surgery has little to offer those who experience generalized disease progression while receiving a TKI, and it should not be attempted.

All patients should resume therapy with a TKI for an indefinite period of time after resection of metastatic disease.

ADJUVANT THERAPY — The optimal selection of patients with resected disease who are at sufficiently high risk for recurrence to warrant adjuvant therapy is not established. Several risk stratification tools are available, based upon tumor size, mitotic rate, location, the presence or absence of tumor rupture, and completeness of resection. However, particularly for tools, such as nomograms, that quantify the risk of disease recurrence after complete resection as a continuous variable, it is not clear what cutoff for disease recurrence should be used to select patients for imatinib. Thus, each case must be approached individually, balancing the estimated likelihood of a disease recurrence (based upon anatomic site, size, mitotic rate, and presence or absence of tumor rupture) with the risks of therapy. (See 'Estimation of recurrence risk' below.)

Some centers, including that of the authors, routinely genotype all patients with resected GIST who are being considered for adjuvant imatinib. We do not offer adjuvant imatinib to patients with KIT and PDGFRA wildtype tumor (eg, succinate dehydrogenase (SDH)-deficient GIST, neurofibromatosis (NF)-related GIST), or a PDGFRA D842V GIST, as these mutations are associated with primary resistance to imatinib. For patients whose tumors harbor a KIT exon 9 mutation, higher-dose imatinib (800 rather than 400 mg daily) is a reasonable option, if tolerated, although there are no prospective data upon which to base a recommendation either for or against this practice.

The standard of care for patients with a primary resectable GIST is surgery, aiming for a macroscopically complete resection with negative microscopic margins. Complete resection is possible in the majority of localized GISTs, but only approximately one-half remain recurrence-free for five or more years with surgery alone. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract".)

The success of imatinib in the setting of advanced disease prompted interest in its use in the adjuvant setting after complete resection of a primary tumor or metastatic disease.

Estimation of recurrence risk — An estimation of recurrence risk following resection of a GIST is used to select patients for adjuvant imatinib. Recurrence risk is based on clinical characteristics including tumor size, mitotic rate, primary tumor site, the presence or absence of tumor rupture, and completeness of resection. The AFIP prognostic model (table 2 and table 3) is most commonly used in the United States, although others are available. In this prognostic model, tumors are generally classified as no-, low-, intermediate-, or high-risk for recurrence.

Several criteria have been proposed, originally to classify the malignant potential of a GIST. Although the terms "benign" and "malignant" are not applied to GIST since all but subcentimeter tumors are considered to have at least some potential to behave in a malignant fashion (either with unresectable invasive recurrence or with metastatic disease), tumor size, mitotic rate, and site of tumor origin have gained the greatest acceptance as being predictive of the risk of recurrence and/or metastases [4]. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Risk stratification and prognosis'.)

Risk stratification models, such as the original National Institutes of Health (NIH) consensus criteria, have been proposed to distinguish prognosis in resected GIST (table 4) [5]. In the series of 289 patients used to construct this model, the cumulative five-year disease-specific survival rates for GISTs classified as risk level I through IV were 100, 96, 67, and 25 percent, respectively. The prognostic importance of mitotic rate, tumor size, and location was confirmed in an analysis of the adjuvant imatinib trial ACOSOG Z9001 [6]. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Risk stratification and prognosis' and 'ACOSOG Z9001' below and "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Other risk factors'.)

Models such as these do not take into account the location of the primary GIST lesion. In general, tumors arising from the small bowel, colon, rectum, or mesentery are associated with less favorable outcomes than those arising from the stomach [7-9]. Other risk prediction models have taken site of GIST origin into account . As an example, the Armed Forces Institute of Pathology (AFIP) prognostic model (table 2 and table 3), which is most commonly used in the United States, represents the largest published experience with GISTs diagnosed and treated in the modern era for which long-term clinical follow-up is available.

Based on data from the AFIP model, a tumor, node, metastasis (TNM) staging system for GIST was developed by the American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control (UICC) and published in 2010 [10]. The 2017 version is depicted in the table (table 5) [11]. Although the T and N designations are the same for all disease sites, there are separate stage groupings for gastric/omental and small bowel/esophageal/colorectal/mesenteric and peritoneal primaries. Rates of disease progression for gastric, small bowel, and rectal GISTs, stratified by stage at diagnosis, are presented in the tables (table 6 and table 7 and table 8) [7]. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Tumor size and mitotic rate'.)

Although not included in the TNM staging system, tumor rupture [12,13] and incomplete resection are also independent risk factors that negatively impact disease-free survival. A modification of the NIH consensus criteria for risk stratification has been proposed that incorporates both site and tumor rupture as prognostic variables [14]. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Other risk factors'.)

As an alternative to the risk classification systems that stratify patients into discrete categories, others have quantified the risk of disease recurrence after complete resection as a continuous variable through the use of a GIST tumor nomogram [15]. Different nomograms have been developed by others [16,17].

The goal of all risk stratification schemes is to identify those patients who are at the highest (or lowest) risk for recurrence so that management (in particular, the use of adjuvant imatinib) can be individualized. However, there is no clear consensus from expert groups as to what cutoff might constitute the lowest "acceptable" level of risk for metastasis or recurrence that would justify the use of adjuvant imatinib. Furthermore, it is not clear that any one prognostication tool outperforms the others [12,18]. (See 'Patient selection' below.)

This issue was addressed in an analysis of data from 10 different population-based published series totaling 2560 patients with completely resected GIST, none of whom received adjuvant imatinib [12]. Tumors were classified according to the NIH consensus criteria (table 4) [5], a modification of the NIH consensus criteria (table 1) that includes site as well as tumor rupture [14], and the AFIP criteria (table 2 and table 9) [7]. Large tumor size, high mitotic counts, non-gastric location, presence of rupture, and male sex were all independent adverse prognostic factors. Most recurrences occurred within the first five years of follow-up, and most patients were cured by surgery alone (estimated 5-, 10-, and 15-year relapse-free survival [RFS] rates 71, 63, and 60 percent, respectively).

A comparison of the NIH, modified NIH, and AFIP criteria showed that all risk stratification schemes appropriately predicted RFS and identified high-risk patients. The proposed modified NIH criteria (table 1) [14] were best at identifying a single subgroup of patients at high-risk of recurrence [12]. With all of the different classification schemes, those patients who were identified as intermediate-risk had a clinical course that was similar to that of the low-risk group, suggesting that only the high-risk patients would likely benefit from adjuvant therapy. These investigators also developed a novel risk stratification scheme in which tumor size and mitotic counts were assessed as continuous non-linear variables; novel prognostic contour maps were generated based upon these data plus site and tumor rupture. These maps were better than conventional models at predicting 10-year risk for GIST recurrence and would be particularly useful for discussing individual risk with patients as they are graphic and easy to explain.

Whether the results of molecular testing should also be integrated into risk stratification schemes is unclear. There has been relatively little work to make predictions of risk based on genotypic subsets of primary resected GIST, although there are data to support the fact that certain genotypes (such as KIT wildtype GISTs that are SDH-deficient or NF-related, or platelet-derived growth factor receptor-alpha [PDGFRA] D842V-mutant tumors) are less aggressive with a lower risk of recurrence or metastasis following primary resection than are other molecular subtypes [19]. Furthermore, there are no data demonstrating the benefit of adjuvant imatinib in these subtypes nor basic scientific rationale for antitumor activity. (See 'Impact of molecular subtypes' below.)

Benefit of imatinib

Phase III trials — At least three phase III trials have evaluated the benefit of adjuvant imatinib; only two (ACOSOG Z9001 and European Organisation for Research and Treatment of Cancer [EORTC] 62024) had a no-treatment control arm.

ACOSOG Z9001 — Benefit for imatinib compared with surgery alone was shown in a phase III, double-blind, multicenter ACOSOG Z9001 trial [20]. In this trial, 713 adults with a completely resected primary gastrointestinal GIST at least 3 cm in maximal diameter and immunohistochemically positive for KIT protein were randomly assigned to one year of adjuvant imatinib (400 mg daily) or placebo. The primary endpoint was RFS.

The trial was stopped early when planned interim analysis disclosed that significantly fewer patients in the treated group recurred. At a median follow-up of 20 months, 30 patients in the imatinib group recurred or died versus 70 in the placebo group (8 versus 20 percent). The one-year RFS rate was 98 versus 83 percent, favoring imatinib, with a hazard ratio (HR) for RFS of 0.35, 95% CI 0.22-0.53 [20]. Once discontinued, adjuvant imatinib appeared to provide one additional year of protection, after which the rate of recurrence seemed to parallel that of the control arm.

Subgroup analysis revealed that RFS was significantly longer with imatinib in all risk categories (based upon size, mitotic rate, and location in the GI tract (table 2 and table 3)). As expected, in a later analysis, the absolute benefit was greatest in those with high-risk disease (relapse rate 47 versus 19 percent for placebo and imatinib, respectively); for moderate-risk disease it was 14 versus 5 percent, respectively [21].

Imatinib was well tolerated by most patients. The drug was discontinued because of adverse reactions in 16 versus 5 percent of the placebo group. The side effect profile overall was similar to that observed in other clinical trials of imatinib. There were no significant cardiac toxic effects. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors" and "Cardiotoxicity of cancer chemotherapy agents other than anthracyclines, HER2-targeted agents, and fluoropyrimidines", section on 'Agents targeting BCR::ABL1'.)

No statistically significant overall survival (OS) differences have emerged in favor of imatinib in the ACOSOG Z9001 trial. Among the possible reasons are the short duration of follow-up, the limited number of relapses, and the high degree of efficacy of imatinib in relapsed disease [22]. Furthermore, after the study was unblinded, all patients randomized to placebo were allowed to crossover to active treatment, thus obscuring any potential differences in OS between the groups.

Based upon these findings, imatinib was given accelerated approval in the United States in 2008 for adjuvant treatment of completely resected GISTs ≥3 cm in size, without definitive guidance as to the optimal duration of treatment or which patients are most likely to benefit.

Although imatinib was clearly effective at reducing disease recurrence in this trial, key questions remain. Two important issues are the value of longer duration imatinib following complete resection and the definition of subsets of patients who derive the most benefit from adjuvant imatinib. Patients with GIST of more than 3 cm are a highly heterogeneous population within which the risk of relapse and death varies considerably. Furthermore, as noted above, the risk of relapse is affected not only by size, but also by mitotic index, location of the primary site, and molecular factors. (See 'Estimation of recurrence risk' above.)

Refining the indications for adjuvant treatment remains a big task for future studies.

EORTC 62024 — In another phase III trial (Intergroup EORTC 62024), two years of adjuvant imatinib therapy improved recurrence free-survival in patients with intermediate- and high-risk resected GIST and demonstrated a trend towards improved imatinib failure-free survival (IFFS) in those with high-risk disease [23-25]. Although adjuvant imatinib did not confer an OS advantage in this study, these data still indirectly support the use of three years of adjuvant imatinib therapy, as suggested by the results of the SSG XVIII trial. (See 'SSG XVIII trial (High-risk GIST)' below.)

In an open-label multicenter phase III trial, 908 patients with intermediate- or high-risk resected GIST were randomly assigned to two years of imatinib or observation alone [23,25]. Disease risk was defined according to the 2002 NIH classification (table 9). The study also included patients with R0 and R1 resections (which included tumor rupture) [24]. The primary endpoint was originally OS, but was subsequently modified to IFFS (ie, the time to start of a new systemic treatment with or without imatinib, or death from any cause) as a surrogate for OS.

At median follow-up of 9.1 years, compared with observation, imatinib demonstrated the following results in the entire study population [25]:

●Similar IFFS (5-year IFFS 87 versus 83 percent, 10-year IFFS 75 versus 74 percent, HR 0.87, 95.7% CI 0.65-1.15) and OS (5-year OS 93 versus 92 percent, 10-year OS 80 versus 78 percent, HR 0.88, 95% CI 0.65-1.21).

●Improved RFS (5-year RFS 70 versus 63 percent; 10-year RFS 63 versus 61 percent, HR 0.71, 95% 0.57-0.89).

Among the 526 patients with high-risk disease, compared with observation, adjuvant imatinib demonstrated a trend towards higher IFFS (10-year IFFS 69 versus 61 percent) and RFS (10-year RFS 48 versus 43 percent).

SSG XVIII trial (High-risk GIST) — For patients with a completely resected primary high-risk GIST with a KIT or PDGFRA mutation other than D842V, we suggest adjuvant imatinib for a minimum of three years rather than a shorter treatment duration or surveillance alone, as this approach improved OS in a randomized trial.

The Scandinavian Sarcoma Group (SSG) XVIII trial was an open-label trial comparing 36 versus 12 months of adjuvant imatinib (400 mg daily) in 400 patients with high-risk resected GIST [26]. High-risk was defined according to the modified consensus criteria [27] as having at least one of the following: tumor size >10 cm, mitotic count >10 per 50 high-power fields (HPF), tumor size >5 cm and mitotic count >5, or tumor rupture. Approximately one-half of the enrolled patients had gastric primary tumors.

At a median follow-up of 54 months, prolonged treatment was associated with a significant improvement in RFS, the primary endpoint (five-year RFS 66 versus 48 percent, HR 0.46, 95% CI 0.32-0.65), as well as OS (five-year OS 92 versus 82 percent, HR 0.45, 95% CI 0.22-0.89) [26]. Benefits persisted in a later report with a longer median follow-up of 119 months; patients assigned to three years of imatinib continued to have significantly greater RFS at both five years (71 versus 53 percent) and 10 years (53 versus 42 percent) and OS at both five (92 versus 86 percent) and 10 years (79 versus 65 percent) [28]. Extending the duration of treatment to three years reduced the number of deaths during the first 10 years of follow-up after surgery by approximately 50 percent.

Treatment-related adverse effects were more common with the longer duration of treatment, including periorbital edema (74 versus 59 percent), diarrhea (54 versus 44 percent), and muscle cramps (49 versus 31 percent) [26]. However, most were grade 1 or 2; the number of grade 3 or 4 events was similar in both groups. Nevertheless, twice as many patients discontinued imatinib for reasons other than disease progression in the prolonged therapy group (26 versus 13 percent). Benefits persisted with longer follow-up.

These data established at least 36 months of adjuvant imatinib as a new standard for patients with high-risk GIST, but questions remain as to whether treatment should be continued for longer than three years. In both groups, within 6 to 12 months of discontinuing adjuvant imatinib, rates of disease recurrence were similarly increased [26]. This finding raises questions as to whether recurrences are truly being prevented or just delayed. (See 'Optimal duration of therapy' below.)

Furthermore, factors other than imatinib duration also influence prognosis. A multivariate analysis of data from the SSG XVIII trial identified four factors other than imatinib duration as being independently predictive of disease recurrence, including non-gastric location, high mitotic count, tumor rupture, and large size [13]. Patients with the highest risk scores had a very high risk of recurrence, despite adjuvant imatinib, and recurrences were frequent both during adjuvant imatinib and after its completion.

Impact of molecular subtypes — A later exploratory analysis of data from the SSG XVIII trial also suggests that mutational status influences the benefit from extended duration of therapy, and that the duration of imatinib might modify the risk of GIST recurrence associated with some high-risk KIT mutations [29]. Of the 400 enrolled patients, 341 had mutation analysis for KIT and PDGFRA performed centrally using conventional sequencing, 274 had a KIT mutation, 43 had a PDGFRA mutation, and 24 had tumors that were wild type for these mutations. Patients with KIT exon 11 deletion or insertion-deletion mutations had better RFS when allocated to three years of therapy (five-year RFS 71 versus 41 percent), but no significant benefit from three years of treatment was seen in any other mutational subgroup, likely reflecting primary resistance to imatinib in these patients. (See 'Molecular subtypes and primary resistance' below.)

KIT exon 11 deletion mutations, deletions that involved codons 557 and/or 558 (which are associated with highly aggressive disease), and deletions that led to pTrp557_Lys558del were associated with poor RFS in the one-year group but not in the three-year group. In addition, in the subset with KIT exon 11 deletion mutations, tumors with mitotic counts that were higher than the median had unfavorable RFS with one year of therapy but not in the three-year group. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Other risk factors'.)

Optimal duration of therapy — In patients receiving adjuvant therapy, adjuvant imatinib is administered for a minimum of three years after resection. Whether treatment should be continued for longer than three years is not known. Rates of disease recurrence have been high within 6 to 12 months of discontinuing adjuvant imatinib for up to three years, and it is possible that imatinib is maintaining tumor dormancy rather than eradicating microdeposits. Patients with intermediate- or high-risk tumors who are tolerating adjuvant imatinib may choose to remain on therapy for a total of five years or longer, rather than be subject to an increased rate of recurrence following discontinuation of the drug.

Preclinical data suggest that imatinib is not curative, inducing cellular quiescence but not death [30]. As noted above in the SSG XVIII trial, in which patients were randomized to one versus three years of adjuvant imatinib, rates of disease recurrence were similarly increased in both groups within 6 to 12 months of discontinuing adjuvant imatinib [26]. (See 'SSG XVIII trial (High-risk GIST)' above.)

Furthermore, although OS was improved with three years, as compared with one year, of therapy, suggesting an actual increase in cure, the survival gap became smaller with long-term follow-up [31]. These findings raise questions as to whether recurrences are truly being prevented or just delayed, or stated a different way, whether adjuvant imatinib is truly eliminating residual cancer cells and curing a subset of patients or just stopping them from growing for a period of time [32].

One phase II trial studied five years of adjuvant imatinib in intermediate- and high-risk resected GIST (PERSIST 5); intermediate- or high-risk disease was defined as a primary GIST of any site ≥2 cm with a mitotic count ≥5 per 50 HPF, or a non-gastric primary GIST ≥5 cm. All patients received imatinib 400 mg daily for five years or until relapse, progression, or intolerance [33]. Among the 91 enrolled patients, the median treatment duration was 55.1 months, and only 50 percent of patients completed a full five years of treatment. The main reasons for stopping treatment early were patient choice and adverse drug effects. The five-year estimate of RFS was 90 percent, while the OS rate at five years was 95 percent. One patient recurred and died during adjuvant imatinib; this patient had an imatinib-insensitive PDGFRA D842V mutation. Six other patients recurred after imatinib discontinuation. These results suggest that five years of imatinib therapy is effective in preventing recurrence in patients with sensitive mutations. However, nearly one-half of patients discontinued study treatment before five years of therapy, and most recurrences occurred after treatment discontinuation.

With longer-term follow-up, if the same steep recurrence rates after treatment discontinuation are seen as were observed in the SSG XVIII trial, it will be likely that these patients had micrometastases that were not eradicated but remained under control for many years through drug therapy, and that even longer duration imatinib may be needed.

Given the lack of a control arm, the PERSIST-5 trial does not prove that therapy for longer than three years will be beneficial.

Support for longer than three-year duration of therapy for patients with high-risk tumors is also provided by a retrospective single-institution analysis of 234 patients who underwent complete (R0) resection and were treated with adjuvant imatinib [34]. For the entire group, long-term outcomes were significantly improved by longer duration of therapy. Five-year RFS rates for the one-year, one-to-three-year, three-to-five-year, and more than five-year groups were 52, 72, 73, and 93 percent, respectively, while the corresponding rates for five-year OS were 44, 82, 84, and 97 percent, respectively. Of the 190 patients who were classified as high risk according to the 2008 NIH risk classification scheme (table 1), adjuvant treatment was continued for one, one to three, three to five, and more than five years in 20, 33, 23, and 24 percent of cases, respectively. The RFS rate of patients receiving imatinib for longer than five years was significantly better than that of those receiving treatment for less than five years. In contrast, among the 44 with intermediate-risk disease (table 1), the differences in RFS for one year, one to three years, and more than three years of therapy showed improvement, but the differences were not statistically significant.

Imatinib dosing — For most patients, the standard dose of adjuvant imatinib is 400 mg daily. However, for patients whose tumors harbor a KIT exon 9 mutation, we suggest higher-dose imatinib, 800 rather than 400 mg daily, if tolerated.

The ACOSOG Z9001 trial tested only the 400 mg daily dose in the adjuvant setting. In randomized trials of patients with advanced metastatic and/or unresectable GIST, patients whose GIST harbored KIT exon 9 mutations exhibited improved outcomes with 800 mg daily doses of imatinib compared with the standard 400 mg daily dosing. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Influence of mutations on response to therapy'.)

Whether doses greater than 400 mg should be used in the adjuvant setting will require prospective study. Until further information becomes available, some centers, including that of the authors, routinely genotype all patients with GIST who are being considered for adjuvant imatinib. The vast majority of the exon 9 mutants come from small bowel GISTs, and we discuss with such patients the option to take 800 mg of imatinib daily rather than 400 mg daily, if tolerated.

Patient selection — The optimal selection of patients who are at sufficiently high risk for recurrence to warrant adjuvant imatinib is not established. Although risk stratification tools are available based upon tumor size, mitotic rate, location, and in some cases, the presence or absence of tumor rupture, it is not clear what cutoff for disease recurrence should be used to select patients for imatinib. Thus, each case must be approached individually, balancing the estimated likelihood of a disease recurrence (based upon anatomic site, size, mitotic rate, and mutation type, if available) with the risks of therapy. (See 'Estimation of recurrence risk' above.)

Several risk stratification schema are available (table 2 and table 3 and table 1).

There is no consensus as to what cutoff for disease recurrence should be used to select patients for imatinib, and practice is variable. The following information informs this debate:

●In 2008, the US Food and Drug Administration (FDA) granted accelerated approval for imatinib in the adjuvant setting for completely resected primary GIST ≥3 cm, without indicating the optimal length of therapy; labeling was updated in January 2012 to include the significantly prolonged survival seen with three years of therapy as compared with one year of adjuvant imatinib. However, whether all patients in this broad category have a high enough risk of recurrence to warrant adjuvant therapy is not established. The European Medicines Agency (EMA) has extended the licensed indications of imatinib to include adjuvant treatment of adult patients who are at "significant risk of relapse" after resection of a KIT-positive GIST but does not define these subsets further.

●The SSG XVIII trial defined high-risk as follows: a ruptured GIST, tumor size >10 cm, mitotic rate >10 per 50 HPF, or tumor size >5 cm and mitotic count >5 [26]. The risk of recurrence in these groups is approximately one-third or higher [35-37]. (See 'SSG XVIII trial (High-risk GIST)' above.)

●Consensus-based clinical practice guidelines from the National Comprehensive Cancer Network (NCCN) [2] suggest adjuvant imatinib for at least 36 months for patients with an intermediate- or high-risk (tumor >5 cm in size with high mitotic rate [>5 mitoses per 50 HPF] GIST or a risk of recurrence that is >50 percent after surgery).

●Risk stratification tools are available based upon tumor size, mitotic rate, location, and tumor rupture, and they all define a "high-risk" subset. The 2012 analysis of Joensuu, described above, found that all risk classification schema identified a group of patients with high-risk disease who had a significantly worse RFS than did those with intermediate- or low-risk disease [14]. The proposed modified NIH criteria (table 1) [14] were best at identifying a single subgroup of patients at high risk of recurrence. (See 'Estimation of recurrence risk' above.)

Patients who were identified as intermediate-risk had a clinical course that was similar to the low-risk group, suggesting that only the high-risk patients would likely benefit from adjuvant therapy. Although the concept that only high-risk patients derive benefit from adjuvant imatinib has not been prospectively validated, based upon these data, it would seem reasonable to offer adjuvant therapy to all patients who fall into a "high-risk" category, regardless of the risk stratification model used.

Molecular subtypes and primary resistance — We perform molecular analysis on all patients with resected GIST who are being evaluated for adjuvant therapy. There are certain mutations that are associated with primary resistance to imatinib, and we do not offer adjuvant therapy to patients with tumors that harbor a D842V mutation in the PDGFRA gene, or those with KIT and PDGFRA wildtype tumor (eg, SDH-deficient GIST, or an NF-related GIST). This approach is consistent with guidelines from the European Society for Medical Oncology (ESMO) [3]. For individuals whose tumors contain a KIT exon 9 mutations (which is associated with partial imatinib resistance, which may be overcome with higher-dose therapy), higher dose imatinib therapy (800 rather than 400 mg daily) may be preferred, if tolerated.

Some molecular features (ie, KIT exon 9 mutations, PDGFRA exon 18 D842V mutations, lack of detectable KIT or PDGFRA mutations [previously referred to as "wildtype" GISTs, now known most often to represent the biologically unique SDH-deficient subtype of GIST]) have been associated with inferior response to imatinib in the setting of advanced disease [6]. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'KIT/PDGFRA wild-type GISTs' and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Influence of mutations on response to therapy'.)

At least some data are available to suggest that this primary resistance to imatinib seen with certain mutations also applies to the adjuvant setting:

●In an analysis of molecular features associated with treatment benefit in the ACOSOG Z9001 study, described above, adjuvant imatinib was associated with higher RFS in patients with a KIT exon 11 deletion of any type but not a KIT exon 11 insertion or point mutation, KIT exon 9 mutation, PDGFRA mutation, or wildtype tumor [6]. (See 'ACOSOG Z9001' above.)

However, the patient groups were very small, especially those with "KIT/PDGFRA wildtype" tumors (nine in the placebo, six in the imatinib group) and exon 9 mutations (seven placebo, four imatinib). Because of these small numbers, the data cannot be considered definitive, but the therapeutic impact of adjuvant imatinib in these subsets certainly seems much smaller. As a result, whether patients with KIT exon 9 mutations, PDGFRA exon 18 D842V mutations, or "wildtype" GISTs should be identified prospectively and specifically counselled not to receive standard-dose adjuvant imatinib is controversial. At several institutions, including some of the authors', patients with D842V mutations, or SDH-deficient or NF-related GISTs are not considered for adjuvant imatinib, while a higher imatinib dose (ie, 800 mg rather than 400 mg daily) may be considered for those with an exon 9 KIT mutation, based upon data derived from patients with more advanced disease. (See 'Imatinib dose' below and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Influence of mutations on response to therapy'.)

●For patients with a PDGFRA D842V mutation or a KIT/PDGFRA wildtype GIST (usually SDH-deficient or NF-related GIST), the evidence from the SSG XVIII trial also suggests no benefit from adjuvant imatinib [26].

●In another report of 95 patients with a KIT/PDGFRA wildtype GIST (84 of which were SDH-deficient), only 1 of 49 patients treated with imatinib had a partial response [38].

NCCN [2] recommendations do not specifically exclude patients with these or any other molecular subset from adjuvant therapy with imatinib. However, we discourage the use of adjuvant imatinib in patients with a PDGFRA D842V mutation or a KIT/PDGFRA wildtype GIST (often SDH-deficient or NF1-related).

Among patients with a KIT mutation, the type of mutation correlates with tumor response to imatinib. Based upon an analysis of data from the ACOSOG Z9001 trial, for patients who harbor an exon 9 KIT mutation, which confers relative resistance to adjuvant imatinib, a dose of imatinib 800 mg per day may be preferred over the standard dose of 400 mg daily, if tolerated.

NEOADJUVANT THERAPY — Neoadjuvant imatinib (algorithm 2) is appropriate for patients who have locally advanced tumors without distant metastases where decreasing tumor burden with systemic therapy could facilitate resection and/or reduce surgical morbidity and for most patients with an anorectal GIST (unless the tumor is small and sphincter-preserving surgery is possible upfront), provided that the tumors harbor a KIT or platelet-derived growth factor receptor-alpha (PDGFRA) mutation other than D842V. If possible, such patients should be enrolled in a clinical trial.

For patients with a PDGFRA D842V mutation or those with wildtype tumors (neither KIT nor PDGFRA mutations), we do not use neoadjuvant imatinib and, instead, proceed directly to surgery. Although European Society for Medical Oncology (ESMO) guidelines suggest that neoadjuvant avapritinib could be considered for patients with PDGFRA D842V mutated tumors [3], there are no clinical trial data on the efficacy or safety of avapritinib in this setting. The use of avapritinib in advanced or metastatic GIST is discussed separately. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Avapritinib for PDGFRA D842V mutant tumors'.)

Indications for neoadjuvant therapy — There are several clinical scenarios for preoperative (neoadjuvant) therapy. In all cases, the goal of treatment is a reduction in tumor size that may facilitate complete surgical resection and/or increase the likelihood of organ preservation. Common indications for neoadjuvant therapy, which are generally consistent with clinical guidelines from the National Comprehensive Cancer Network (NCCN) and ESMO, include [2,3]:

●An unresectable or borderline resectable primary tumor.

●A potentially resectable tumor at risk for significant morbidity and/or extensive organ disruption.

●A local recurrence of locally advanced disease.

●An anorectal GIST where upfront sphincter-preserving surgery is not possible.

●A limited amount of potentially resectable metastatic disease. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract", section on 'Metastatic GIST with potentially resectable disease'.)

●Tumors where neoadjuvant therapy would permit a switch from an open to a minimally invasive (ie, laparoscopic or robotic) surgical approach, avoid complex multivisceral resection, or preserve the affected organ for tumors involving the esophagus, esophagogastric junction, duodenum, or rectum [39].

Role of genotyping — For most patients being evaluated for neoadjuvant imatinib, testing for specific molecular alterations can be used to guide selection and dosing of neoadjuvant therapy, as follows:

●We initiate imatinib at a dose of 400 mg daily.

●Patients with an exon 9 KIT mutation may be offered the option of a higher dose of imatinib (800 mg per day), although there are no prospective data upon which to base a recommendation either for or against this practice. Most of these patients will have a small bowel primary tumor and will be referred for upfront surgery rather than neoadjuvant imatinib.

●For patients with a PDGFRA D842V mutation or those with wildtype tumors (neither KIT nor PDGFRA mutations), we do not use neoadjuvant imatinib and, instead, proceed directly to surgery. Although ESMO guidelines suggest that neoadjuvant avapritinib could be considered for those with PDGFRA D842V mutated tumors [3], there are no clinical trial data on the efficacy or safety of avapritinib in this setting. The use of avapritinib in advanced or metastatic GIST is discussed separately. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Avapritinib for PDGFRA D842V mutant tumors'.)

●If genotyping for molecular alterations is not available or routinely pursued, it is wise to check for signs of primary resistance to neoadjuvant imatinib using computed tomography (CT), positron emission tomography (PET)-CT, or contrast-enhanced ultrasound, which is used in Europe. For example, some UpToDate contributors do not obtain further molecular testing for patients receiving neoadjuvant imatinib whose tumor histology and immunohistochemistry are consistent with either a KIT or PDGFRA mutation. For those with insufficient tissue samples for molecular testing, repeat biopsy is not required if patients are responding to imatinib. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Approach to testing'.)

Patients with primary or recurrent localized disease — There is no consensus as to the indications for neoadjuvant imatinib among patients with apparently localized GIST. We offer initial neoadjuvant treatment with imatinib (algorithm 2) for patients with locally advanced tumors without distant metastases where decreasing tumor burden with systemic therapy could facilitate resection and/or reduce surgical morbidity. We also suggest neoadjuvant imatinib, rather than initial surgery, for most patients with a localized rectal GIST, unless the tumor is small and sphincter-preserving surgery is possible upfront.

For patients who are bleeding, the approach taken depends on the severity of the bleed. If the bleeding is brisk and life threatening, the patient needs surgery. However, if the patient is too frail to survive surgery, treatment with imatinib can itself resolve the bleed. For patients with low-grade chronic bleeding, neoadjuvant imatinib is even more likely to be helpful.

An important point is that preoperative imatinib prohibits accurate assessment of recurrence risk based upon analysis of the surgical resection specimen. As a result, for patients undergoing neoadjuvant imatinib, we continue imatinib postoperatively to complete a total of three years of imatinib therapy (combined preoperative and postoperative).

Data supporting benefit of neoadjuvant imatinib are available from several case reports and small retrospective series, most of which include a mix of patients with borderline resectable and unresectable primary disease, as well as metastatic and locally recurrent disease that is potentially amenable to gross resection. In addition, a single phase II United States Intergroup trial of neoadjuvant imatinib has been completed, and preliminary results are available [40]; results are also available from a multi-institutional, Asian phase II trial of neoadjuvant therapy in large stomach tumors [41].

RTOG 0132/ACRIN 6665 trial — The multicenter Radiation Therapy Oncology Group (RTOG) 0132/American College of Radiology Imaging Network (ACRIN) 6665 trial was a prospective phase II trial in which 63 patients with KIT-positive GIST, and either a resectable primary ≥5 cm or resectable recurrent disease received preoperative imatinib 600 mg daily for 8 to 12 weeks [40]. Following surgery, all patients received at least two additional years of postoperative imatinib, while those with metastatic disease were treated until disease progression.

Thirty of the 52 analyzable patients had locally advanced primaries, and 22 had locally recurrent or potentially resectable metastatic disease. Among the patients with localized primary disease, only two (7 percent) had an objective response to preoperative imatinib (as assessed by CT scan), but stable disease was achieved in 25 (83 percent). The corresponding values for the group with metastatic disease were 5 and 91 percent, respectively.

In the latest update, at a median follow-up of 5.1 years, the estimated five-year progression-free and disease-specific survival rates for patients presenting with localized primary disease were 57 and 77 percent, respectively; the corresponding rates were 30 and 68 percent in those with recurrent or metastatic disease [42].

While this trial confirmed the safety of neoadjuvant imatinib, it tested a relatively brief period of preoperative treatment. The available data suggest that the earliest time to a partial response in patients treated with neoadjuvant imatinib is 16 weeks [43] and that maximal radiographic response to imatinib generally occurs after three to nine months of treatment. (See 'Response assessment' below and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors".)

Korean/Japanese trial in large gastric GIST — A multinational phase II trial enrolled 56 patients with large (≥10 cm) gastric GIST and administered six to nine months of neoadjuvant imatinib (400 mg daily) [41]. Overall, 53 were evaluable, and 46 completed six or more months of therapy. Of the patients who received at least six months of neoadjuvant imatinib, the earliest point at which maximal reduction of tumor size was observed was four weeks in one case (2 percent), 12 weeks in nine cases (20 percent), 24 weeks in 29 cases (63 percent), and 36 weeks in seven cases (15 percent). The objective response rate by Response Evaluation Criteria In Solid Tumors (RECIST) was 62 percent, and the complete (R0) resection rate was 91 percent (48 of 53). Of the 48 patients undergoing R0 resection, preservation of at least one-half of the stomach was achieved in 42. After R0 resection, all patients received imatinib at the same dose as was used preoperatively for at least one year. At a median follow-up of 32 months, the two-year overall and progression-free survival rates were 98 and 89 percent, respectively.

Retrospective series — Data from multiple retrospective series also support the benefit of initial imatinib therapy in patients with locally advanced GISTs [44-50]. The largest experience consisted of 161 patients who underwent surgery after imatinib; the majority of primary tumors were in the stomach (55 percent), followed by the rectum (20 percent), duodenum (10 percent), ileum/jejunum (10 percent), and esophagus (3 percent) [50]. The response to preoperative imatinib was partial response in 129 (80 percent) and stable disease in 30 (19 percent); only two patients had disease progression during neoadjuvant therapy. A complete (R0) resection was possible after a median 40 weeks of imatinib (range 6 to 190). At a median follow-up of 46 months, the five-year disease-specific survival and disease-free survival rates were 95 and 65 percent, respectively. Only 56 percent of patients continued imatinib after resection (for a median duration of 19 months [range 12 to 76]), and five-year disease-free survival was higher in this cohort (72 versus 57 percent). During follow-up, there were 37 recurrences, only three of which were local; the remainder were intraperitoneal dissemination or liver metastases.

Rectal GISTs — Neoadjuvant imatinib might be of particular benefit for rectal GISTs, which can be large, bulky tumors and require extensive surgery to achieve a surgical complete resection [51]. The benefits of perioperative imatinib in patients with a rectal GIST can be illustrated by several retrospective reports [52-56]; two of the largest are described in detail:

●One series included 47 patients with a localized primary rectal GIST treated at a single center from 1982 to 2016 and stratified by treatment before and after the year 2000, when imatinib became available [54]. In the imatinib era, 24 of the 30 resected patients had received perioperative imatinib. Among the 34 high-risk patients (ie, tumor size >5 cm or mitotic rate >5 per 50 high-power fields [HPF]), rates of organ preservation (92 versus 48 percent) and negative margins (69 versus 29 percent) were higher among the 13 patients who received neoadjuvant imatinib than among the 21 treated with initial surgery. Furthermore, high-risk patients who underwent neoadjuvant imatinib had greater five-year overall survival, disease-specific survival, local relapse-free survival, and distant relapse-free survival compared with those who had initial surgery (91, 100, 100, and 71 versus 47, 65, 74, and 41 percent, respectively).

●In another report of 32 patients with a rectal GIST, 22 received imatinib prior to surgery for a median of nine months; complete resection was possible in 17 (77 percent), and the median disease-free survival had not been reached with a median follow-up duration of 39 months [53]. The 10 patients who underwent initial surgery had smaller tumors (median 6 versus 9.3 cm before treatment) and a similar rate of R0 resection (7 of 10, 70 percent). Sphincter preservation was achieved in a higher number of those treated with neoadjuvant therapy, although the difference was not statistically significant (41 versus 30 percent, p = 0.57). Locoregional recurrence developed in 3 of 10 patients treated with initial surgery (30 percent, versus one patient in the neoadjuvant group, 5 percent), and distant metastases developed in seven (70 percent, versus two patients in the neoadjuvant group, 9 percent).

Given these data and the fact that most rectal GISTs have exon 11 KIT mutations (which are more sensitive to imatinib than are other KIT mutations, such as those affecting exon 9), we prefer neoadjuvant imatinib, rather than initial surgery, for most patients with a rectal GIST unless it is small and sphincter-preserving surgery is possible upfront. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract", section on 'Colon and rectum' and "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Other risk factors'.)

Imatinib dose — All of the retrospective series and the single prospective trial of neoadjuvant imatinib utilized a daily dose of imatinib of 400 mg per day, and this is the usual approach. However, as with adjuvant therapy, if a KIT exon 9 mutation is identified and neoadjuvant therapy is being considered, dose escalation to 800 mg per day is reasonable and is supported in the European Society for Medical Oncology (ESMO) guidelines [3]. They also note, however, that regulatory constraints may limit this practice, which is currently not supported in the adjuvant setting by controlled trials. (See 'Imatinib dosing' above.)

Other mutations are insensitive to imatinib altogether (eg, PDGFRA exon 18 D842V mutations, succinate dehydrogenase [SDH]-mutant, neurofibromatosis [NF]-related GIST), and we do not use neoadjuvant imatinib in these settings. If routine genotyping is not practiced, it is probably wise to check for an early response to neoadjuvant imatinib using CT, PET, or contrast-enhanced ultrasound [57]. (See 'Response assessment' below and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Influence of mutations on response to therapy'.)

Response assessment — Patients being considered for neoadjuvant treatment of a GIST should be managed in a center that has a strong multidisciplinary team approach. The usual schema is to administer imatinib for anywhere from 3 to 12 months, with frequent imaging studies and periodic reevaluation for surgical intervention.

The best method to assess response of a GIST to tyrosine kinase inhibitors (TKIs) is controversial. PET-CT scanning using fluorodeoxyglucose (FDG) is highly sensitive for detecting GIST, as for other types of tumors with a high glucose metabolism. Accumulating data support the view that very early evidence of tumor response (within days of instituting imatinib [58]) can be obtained by the use of metabolic imaging as compared with conventional CT scanning. A clinical scenario where obtaining a baseline and follow-up PET-CT scan after two to four weeks of TKI therapy might prove useful is a borderline resectable GIST (or a potentially resectable tumor that requires extensive organ disruption), in which there is a narrow window for moving to alternative therapy (eg, potentially function-impairing resection or alternative TKI) if imatinib were to be ineffective. This is particularly important if routine genotyping is not practiced. (See 'Imatinib dose' above and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Assessing response to therapy'.)

However, in most cases, periodic cross-sectional imaging using conventional CT scan of the abdomen and pelvis every eight to twelve weeks is sufficient to assess response. The following caveats must be kept in mind, however (see "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Cross-sectional imaging'):

●Radiographically, GISTs may actually increase in size during early treatment as a consequence of intratumoral hemorrhage or myxoid degeneration. A decrease in tumor density as seen on CT (the corollary of decreased FDG uptake on PET, see above) is an important early clinical marker of antitumor activity. Once tumors become hypodense (cystic), the size of the lesions may decrease slowly and eventually stabilize.

●Late responses are often seen in patients who initially have stable disease. Maximal response may take six months or even longer.

●As a result of these issues, response to imatinib is frequently defined as absence of progression at the time of the first formal disease reevaluation (typically two to three months after starting therapy) [59]. Clear-cut evidence of progression at this time point is considered initial (primary) resistance, and a switch to an alternative form of therapy is indicated.

Duration of neoadjuvant therapy — The optimal duration of neoadjuvant TKIs is not established. The decision as to how long to administer imatinib and when to operate (ie, at first resectability versus after achieving maximal response) must be individualized. For patients in whom tumor shrinkage would lessen the complexity of the surgical procedure, imatinib is often administered for up to 12 months, as long as a continued radiographic response is evident. We usually treat to "best tumor response" on cross-sectional imaging, trying to limit therapy to no more than 10 to 12 months.

Data on radiologic assessment of response to upfront imatinib come from a retrospective single-institution review of 20 patients who received neoadjuvant imatinib and then underwent surgery over a 12-year period [43]. The median duration of neoadjuvant imatinib was 32 weeks. The best response was a partial response in 16 and stable disease in four. The "best" response was seen at 28 weeks, irrespective of tumor size or location. Plateau response was seen at 34 weeks, beyond which further treatment was not beneficial.

At least some data suggest that extending neoadjuvant therapy beyond 10 to 12 months increases the likelihood of a disease recurrence after surgery, at least for primary GISTs [44,60]. Whether longer periods of neoadjuvant treatment are detrimental for patients being treated for a recurrent/metastatic GIST is unclear [60].

A provocative trial suggests that response rates to very brief periods of neoadjuvant imatinib (three to seven days), as assessed by FDG-PET and dynamic CT, are as high as 70 percent [61]. However, in this small prospective randomized phase II trial, there was no evidence of histologic cytoreduction (and, therefore, no potential benefit in terms of reduced tumor bulk) from ≤7 days of neoadjuvant imatinib and no suggestion that intraoperative blood loss was reduced, even though blood flow to the tumor was reduced, as measured by dynamic CT. Thus, the clinical benefit of very short periods of neoadjuvant imatinib (termed "nanoneoadjuvant therapy" [62]) is unproven.

Surgical resection — After completing neoadjuvant therapy, cytoreductive surgery may be offered to patients whose disease is stable or responding to tyrosine kinase inhibitor (TKI) therapy. Further details on the surgical management of GIST are discussed separately. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract".)

Postoperative management — For patients undergoing neoadjuvant imatinib, we continue imatinib postoperatively to complete a total of three years of imatinib therapy (combined preoperative and postoperative).

Mitotic count cannot be reliably assessed on a surgical specimen after neoadjuvant therapy. Specifically, a low mitotic count in a patient undergoing surgery after neoadjuvant imatinib may reflect the effect of treatment and not the intrinsic biological behavior of the tumor. Since risk of recurrence cannot be accurately stratified after neoadjuvant therapy, our practice is to continue imatinib postoperatively to complete a total of three years of imatinib therapy (preoperative and postoperative). In these situations, the size of the primary tumor (eg, >15 cm) often is justification enough for consideration of adjuvant therapy.

Potentially resectable metastatic disease — Patients with metastatic GIST and potentially resectable disease (including liver and peritoneal metastases) may be offered neoadjuvant systemic therapy with TKIs followed by surgery. This subject is discussed separately. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract", section on 'Metastatic GIST with potentially resectable disease'.)

POSTTREATMENT FOLLOW-UP — There are no evidence-based guidelines on what constitutes appropriate follow-up after treatment of a GIST, and there is no consensus on this issue. Relapses most often occur to the liver and/or peritoneum; other sites of metastases, including bone metastases, are rare. Guidelines from the National Comprehensive Cancer Network (NCCN) [2] suggest the following:

●For a completely resected GIST, history and physical examination every three to six months for five years, then annually. A CT scan is recommended every three to six months for three to five years, then annually.

●For patients with more locally advanced or metastatic disease who are receiving imatinib, history and physical examination, laboratory studies (complete blood count and monitoring of phosphate levels), as well as CT scan of the abdomen and pelvis are recommended every three to six months. (See "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Side effects and their management'.)

In contrast, posttreatment follow-up guidelines from the European Society for Medical Oncology (ESMO) emphasize the value of risk assessment in selecting the frequency and specific components of follow-up [3]. We stratify our recommendations for posttreatment follow-up according to risk but do not adhere closely to a specific timetable for patients without high-risk disease.

●For patients with a high enough risk to justify adjuvant imatinib, we suggest cross-sectional imaging with an abdominal CT scan or magnetic resonance imaging (MRI) every three to six months for three years during adjuvant therapy, every three months during the two years that follow treatment discontinuation, where the risk of disease recurrence is highest, then at 6- to 12-month intervals to complete 10 years of follow-up.

●For low-risk tumors, the utility of routine follow-up is not known. For patients who elect for surveillance, we obtain an abdominal CT or MRI every 6 to 12 months for five years.

●Very low-risk tumors do not require routine follow-up, although patients should be informed that recurrences are still possible. We generally offer one abdominal CT two years following treatment completion, as most recurrences are mesenteric and can be better visualized on CT; if there is no evidence of disease recurrence, we do not offer further imaging. Another option is imaging annually or every two years, although data are limited for this approach and patients should be counseled on the risks of radiation exposure.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Gastrointestinal stromal tumors".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Soft tissue sarcoma (The Basics)")

SUMMARY AND RECOMMENDATIONS

●General principles – Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms of the gastrointestinal tract.

•Molecular alterations – Molecular alterations that are commonly identified in GISTs include mutations in KIT, platelet-derived growth factor receptor-alpha (PDGFRA), and the family of succinate dehydrogenase (SDH) genes. (See "Clinical presentation, diagnosis, and prognosis of gastrointestinal stromal tumors", section on 'Molecular alterations'.)

•Tyrosine kinase inhibitors – Small molecule tyrosine kinase inhibitors (TKIs) that target KIT and PDGFRA, such as imatinib, are used in the adjuvant (postoperative) and neoadjuvant (preoperative) setting. (See 'Overview of approach to the patient' above.)

•Surgical resection – The standard treatment for patients with a primary resectable GIST is surgery, aiming for a macroscopically complete resection with negative microscopic margins (algorithm 1). Complete resection is possible in the majority of localized GISTs, but only approximately one-half remain recurrence-free for five or more years with surgery alone. (See "Local treatment for gastrointestinal stromal tumors, leiomyomas, and leiomyosarcomas of the gastrointestinal tract".)

●Approach to adjuvant therapy – In patients with localized resected GIST, the approach to adjuvant therapy is as follows (algorithm 3) (see 'Apparently localized disease' above):

•Estimating recurrence risk – An estimation of recurrence risk following resection of a GIST is used to select patients for adjuvant imatinib. Recurrence risk is based on clinical characteristics including tumor size, mitotic rate, primary tumor site, the presence or absence of tumor rupture, and completeness of resection. The AFIP prognostic model (table 2 and table 3) is most commonly used in the United States, although others are available. In this prognostic model, tumors are generally classified as no, low, intermediate, or high risk for recurrence. (See 'Estimation of recurrence risk' above.)

•Molecular genotyping – We perform molecular analysis on all patients with resected GIST who are being evaluated for adjuvant therapy. We do not offer adjuvant therapy to patients with a PDGFRA D842V mutated tumor or those with a KIT and PDGFRA wildtype tumor (eg, SDH-deficient GIST, neurofibromatosis [NF]-related GIST), as these mutations are associated with primary resistance to imatinib. (See 'Molecular subtypes and primary resistance' above.)

•High-risk GIST – For patients with a completely resected primary high-risk GIST with a KIT or PDGFRA mutation other than D842V, we suggest adjuvant imatinib for a minimum of three years rather than a shorter treatment duration or surveillance alone (Grade 2B). (See 'SSG XVIII trial (High-risk GIST)' above.)

•Intermediate-risk GIST – For those with completely resected primary intermediate-risk GIST with a KIT or PDGFRA mutation other than D842V, we suggest adjuvant imatinib for a minimum of three years rather than shorter treatment duration or surveillance alone (Grade 2C). Patients who decline or are ineligible for adjuvant imatinib may be offered surveillance. (See 'Optimal duration of therapy' above.)

•No-, very low-, or low-risk GIST – For those with a completely resected primary GIST who are at no, very low, or low risk of recurrence, we offer surveillance, as indicated. (See 'Apparently localized disease' above.)

•Dose of adjuvant imatinib – For most patients, the standard dose of adjuvant imatinib is 400 mg daily. However, for patients whose tumors harbor a KIT exon 9 mutation, we suggest higher-dose imatinib, 800 rather than 400 mg daily (Grade 2C), if tolerated. (See 'Imatinib dosing' above and 'ACOSOG Z9001' above and "Tyrosine kinase inhibitor therapy for advanced gastrointestinal stromal tumors", section on 'Influence of mutations on response to therapy'.)

•Duration of therapy – In patients receiving adjuvant therapy, adjuvant imatinib is administered for a minimum of three years after resection. Whether treatment should be continued for longer than three years is not known. Rates of disease recurrence have been high within 6 to 12 months of discontinuing adjuvant imatinib for up to three years, and it is possible that imatinib is maintaining tumor dormancy rather than eradicating microdeposits. Patients with intermediate- or high-risk tumors who are tolerating adjuvant imatinib may choose to remain on therapy for a total of five years or longer, rather than be subject to an increased rate of recurrence following discontinuation of the drug. (See 'Optimal duration of therapy' above.)

●Approach to neoadjuvant therapy (KIT or non-D842V PDGFRA mutated tumors) – For most patients with GIST, standard initial treatment is with surgery rather than neoadjuvant systemic therapy. Exceptions are made for the following subsets of patients with tumors with either a KIT or a non-D842V PDGFRA mutation:

•Rectal tumors – For such patients with a rectal tumor, neoadjuvant imatinib (algorithm 2) rather than upfront resection is appropriate. An exception would be for patients with a small tumor in whom a rectal sparing surgery could be used upfront. (See 'Rectal GISTs' above.)

•Locally advanced tumors without distant metastases – For patients with a nonmetastatic locally advanced or borderline resectable GIST, we suggest initial therapy with imatinib rather than attempted resection (Grade 2C). For such patients, neoadjuvant imatinib prior to surgery may be helpful to facilitate resection and/or reduce surgical morbidity. If possible, such patients should be enrolled in a clinical trial. (See 'Neoadjuvant therapy' above.)

•Rationale – The rationale for this approach is that upfront surgery in the above cases may lead to unnecessary morbidity; imatinib is an effective systemic therapy that can improve subsequent surgical outcomes. Neoadjuvant therapy may permit patients to switch from an open to a minimally invasive surgical approach, avoid complex multivisceral resection, or preserve the affected organ in those with disease involving the esophagus, esophagogastric junction, duodenum, or rectum. (See 'Indications for neoadjuvant therapy' above.)

•Molecular genotyping – For most patients being evaluated for neoadjuvant imatinib, testing for specific molecular alterations can be used to guide the selection and dosing of neoadjuvant therapy (algorithm 2). (See 'Role of genotyping' above.)

•Dosing – For most patients, we initiate neoadjuvant imatinib at a dose of 400 mg daily. If an exon 9 mutation is identified in a rare patient being considered for neoadjuvant imatinib, the option of a higher daily dose (800 mg per day) should be discussed with patients, although there are no prospective data upon which to base a recommendation either for or against this practice. Most of these patients will have a small bowel primary tumor and be referred for upfront surgery rather than neoadjuvant imatinib. (See 'Imatinib dose' above.)

•Response assessment – If genotyping for molecular alterations is not available or routinely pursued, it is wise to check for an early response to neoadjuvant imatinib using CT, positron emission tomography (PET), or contrast-enhanced ultrasound. (See 'Response assessment' above.)

•Treatment duration – The optimal duration of neoadjuvant TKIs is not established. The decision as to how long to administer imatinib and when to operate (ie, at first resectability versus after achieving maximal response) must be individualized and based upon drug tolerance, tumor location and extent, the results of periodic radiographic assessment, and the urgency of surgical treatment. In general, we treat to maximal response and try to limit neoadjuvant treatment to no more than 10 to 12 months. (See 'Duration of neoadjuvant therapy' above.)

For patients undergoing neoadjuvant imatinib, we continue imatinib postoperatively to complete a total of at least three years of imatinib therapy (combined preoperative and postoperative). (See 'Postoperative management' above.)

●Approach to neoadjuvant therapy (PDGFRA D842V or KIT/PDGFRA wildtype tumors) – For patients whose tumors have a PDGFRA D842V mutation or those with wildtype tumors (neither KIT nor PDGFRA mutations), we do not use neoadjuvant imatinib and, instead, proceed directly to surgery. Although European Society for Medical Oncology (ESMO) guidelines suggest that neoadjuvant avapritinib could be considered for patients with PDGFRA D842V mutated tumors, there are no clinical trial data on the efficacy or safety of avapritinib in this setting. (See 'Neoadjuvant therapy' above.)

ACKNOWLEDGMENT — The editorial staff at UpToDate acknowledges George Demetri, MD, and Jeffrey Morgan, MD, who contributed to earlier versions of this topic review.