[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
CONSORT Flowchart of Clinical Study
CONSORT Flowchart of Clinical Study

Includes update of analysis of clinical and molecular factors and patterns of postprotocol therapies in long-term (≥8 years) survivors. Conventional dose of imatinib mesylate indicates 400 mg/d; higher dose, 800 mg/d.

Figure 2.
Long-term Overall Survival (OS) and Progression-Free Survival (PFS)
Long-term Overall Survival (OS) and Progression-Free Survival (PFS)

Analyses include all 695 eligible patients.

Figure 3.
Overall Survival (OS) by Tumor Genotype
Overall Survival (OS) by Tumor Genotype

Genotype results include 395 eligible patients. Genotypes include KIT exon 11, KIT/PDGFRA wild-type, and KIT exon 9 mutations. Different classes of KIT exon 11 mutations include deletions, insertions/duplications, and point mutations.

Figure 4.
Survival of Patients With KIT Exon 11 vs Succinate Dehydrogenase (SDH) Mutations
Survival of Patients With KIT Exon 11 vs Succinate Dehydrogenase (SDH) Mutations

Analysis includes 294 eligible patients. OS indicates overall survival; PFS, progression-free survival.

Table.  
Characteristics of Patients With SDH Mutations vs Original Exon 11 Mutations
Characteristics of Patients With SDH Mutations vs Original Exon 11 Mutations
1.
Nilsson  B, Bümming  P, Meis-Kindblom  JM,  et al.  Gastrointestinal stromal tumors: the incidence, prevalence, clinical course, and prognostication in the preimatinib mesylate era: a population-based study in western Sweden.  Cancer. 2005;103(4):821-829.PubMedGoogle ScholarCrossref
2.
Mucciarini  C, Rossi  G, Bertolini  F,  et al.  Incidence and clinicopathologic features of gastrointestinal stromal tumors: a population-based study.  BMC Cancer. 2007;7:230.PubMedGoogle ScholarCrossref
3.
Steigen  SE, Eide  TJ.  Trends in incidence and survival of mesenchymal neoplasm of the digestive tract within a defined population of northern Norway.  APMIS. 2006;114(3):192-200.PubMedGoogle ScholarCrossref
4.
Tryggvason  G, Gíslason  HG, Magnússon  MK, Jónasson  JG.  Gastrointestinal stromal tumors in Iceland, 1990-2003: the Icelandic GIST study, a population-based incidence and pathologic risk stratification study.  Int J Cancer. 2005;117(2):289-293.PubMedGoogle ScholarCrossref
5.
Tran  T, Davila  JA, El-Serag  HB.  The epidemiology of malignant gastrointestinal stromal tumors: an analysis of 1458 cases from 1992 to 2000.  Am J Gastroenterol. 2005;100(1):162-168.PubMedGoogle ScholarCrossref
6.
Hirota  S, Isozaki  K, Moriyama  Y,  et al.  Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors.  Science. 1998;279(5350):577-580.PubMedGoogle ScholarCrossref
7.
Rubin  BP, Fletcher  JA, Fletcher  CD.  Molecular insights into the histogenesis and pathogenesis of gastrointestinal stromal tumors.  Int J Surg Pathol. 2000;8(1):5-10.PubMedGoogle ScholarCrossref
8.
Plaat  BE, Hollema  H, Molenaar  WM,  et al.  Soft tissue leiomyosarcomas and malignant gastrointestinal stromal tumors: differences in clinical outcome and expression of multidrug resistance proteins.  J Clin Oncol. 2000;18(18):3211-3220.PubMedGoogle ScholarCrossref
9.
DeMatteo  RP, Lewis  JJ, Leung  D, Mudan  SS, Woodruff  JM, Brennan  MF.  Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival.  Ann Surg. 2000;231(1):51-58.PubMedGoogle ScholarCrossref
10.
Dematteo  RP, Heinrich  MC, El-Rifai  WM, Demetri  G.  Clinical management of gastrointestinal stromal tumors: before and after STI-571.  Hum Pathol. 2002;33(5):466-477.PubMedGoogle ScholarCrossref
11.
Joensuu  H, Roberts  PJ, Sarlomo-Rikala  M,  et al.  Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor.  N Engl J Med. 2001;344(14):1052-1056.PubMedGoogle ScholarCrossref
12.
Demetri  GD, von Mehren  M, Blanke  CD,  et al.  Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors.  N Engl J Med. 2002;347(7):472-480.PubMedGoogle ScholarCrossref
13.
Verweij  J, Casali  PG, Zalcberg  J,  et al.  Progression-free survival in gastrointestinal stromal tumours with high-dose imatinib: randomised trial.  Lancet. 2004;364(9440):1127-1134.PubMedGoogle ScholarCrossref
14.
Blanke  CD, Rankin  C, Demetri  GD,  et al.  Phase III randomized, intergroup trial assessing imatinib mesylate at two dose levels in patients with unresectable or metastatic gastrointestinal stromal tumors expressing the kit receptor tyrosine kinase: S0033.  J Clin Oncol. 2008;26(4):626-632.PubMedGoogle ScholarCrossref
15.
Blanke  CD, Demetri  GD, von Mehren  M,  et al.  Long-term results from a randomized phase II trial of standard- versus higher-dose imatinib mesylate for patients with unresectable or metastatic gastrointestinal stromal tumors expressing KIT J Clin Oncol. 2008;26(4):620-625.PubMedGoogle ScholarCrossref
16.
Gastrointestinal Stromal Tumor Meta-Analysis Group (MetaGIST).  Comparison of two doses of imatinib for the treatment of unresectable or metastatic gastrointestinal stromal tumors: a meta-analysis of 1640 patients.  J Clin Oncol. 2010;28(7):1247-1253.PubMedGoogle ScholarCrossref
17.
von Mehren  M, Heinrich  MC, Jeonsuu  H,  et al.  Follow-up results after 9 years (yrs) of the ongoing, phase II B2222 trial of imatinib mesylate (IM) in patients (pts) with metastatic or unresectable KIT+ gastrointestinal stromal tumors (GIST) [abstract 10016].  J Clin Oncol. 2011;29(15)(suppl):10016. doi:10.1200/jco.2011.29.15_suppl.10016Google ScholarCrossref
18.
Heinrich  MC, Owzar  K, Corless  CL,  et al.  Correlation of kinase genotype and clinical outcome in the North American Intergroup Phase III Trial of imatinib mesylate for treatment of advanced gastrointestinal stromal tumor: CALGB 150105 Study by Cancer and Leukemia Group B and Southwest Oncology Group.  J Clin Oncol. 2008;26(33):5360-5367.PubMedGoogle ScholarCrossref
19.
Kaplan  EL, Meier  R.  Nonparametric estimation from incomplete observation.  J Am Stat Assoc. 1958;53(282):457-481.Google ScholarCrossref
20.
Grasso  C, Butler  T, Rhodes  K,  et al.  Assessing copy number alterations in targeted, amplicon-based next-generation sequencing data.  J Mol Diagn. 2015;17(1):53-63.PubMedGoogle ScholarCrossref
21.
Corless  CL, Fletcher  JA, Heinrich  MC.  Biology of gastrointestinal stromal tumors.  J Clin Oncol. 2004;22(18):3813-3825.PubMedGoogle ScholarCrossref
22.
Joensuu  H.  Risk stratification of patients diagnosed with gastrointestinal stromal tumor.  Hum Pathol. 2008;39(10):1411-1419.PubMedGoogle ScholarCrossref
23.
Miettinen  M, Lasota  J.  Gastrointestinal stromal tumors: review on morphology, molecular pathology, prognosis, and differential diagnosis.  Arch Pathol Lab Med. 2006;130(10):1466-1478.PubMedGoogle Scholar
24.
Rubin  BP, Heinrich  MC, Corless  CL.  Gastrointestinal stromal tumour.  Lancet. 2007;369(9574):1731-1741.PubMedGoogle ScholarCrossref
25.
Nannini  M, Astolfi  A, Urbini  M,  et al.  Integrated genomic study of quadruple-WT GIST (KIT/PDGFRA/SDH/RAS pathway wild-type GIST).  BMC Cancer. 2014;14:685.PubMedGoogle ScholarCrossref
26.
Urbini  M, Astolfi  A, Indio  V,  et al.  SDHC methylation in gastrointestinal stromal tumors (GIST): a case report.  BMC Med Genet. 2015;16:87.PubMedGoogle ScholarCrossref
27.
Bonvalot  S, Eldweny  H, Péchoux  CL,  et al.  Impact of surgery on advanced gastrointestinal stromal tumors (GIST) in the imatinib era.  Ann Surg Oncol. 2006;13(12):1596-1603.PubMedGoogle ScholarCrossref
28.
Bauer  S, Hartmann  JT, de Wit  M,  et al.  Resection of residual disease in patients with metastatic gastrointestinal stromal tumors responding to treatment with imatinib.  Int J Cancer. 2005;117(2):316-325.PubMedGoogle ScholarCrossref
29.
Le Cesne  A, Ray-Coquard  I, Bui  BN,  et al; French Sarcoma Group.  Discontinuation of imatinib in patients with advanced gastrointestinal stromal tumours after 3 years of treatment: an open-label multicentre randomised phase 3 trial.  Lancet Oncol. 2010;11(10):942-949.PubMedGoogle ScholarCrossref
30.
Park  SJ, Ryu  MH, Ryoo  BY,  et al.  The role of surgical resection following imatinib treatment in patients with recurrent or metastatic gastrointestinal stromal tumors: results of propensity score analyses.  Ann Surg Oncol. 2014;21(13):4211-4217.PubMedGoogle ScholarCrossref
31.
Bauer  S, Rutkowski  P, Hohenberger  P,  et al.  Long-term follow-up of patients with GIST undergoing metastasectomy in the era of imatinib—analysis of prognostic factors (EORTC-STBSG collaborative study).  Eur J Surg Oncol. 2014;40(4):412-419.PubMedGoogle ScholarCrossref
32.
Raut  CP, Posner  M, Desai  J,  et al.  Surgical management of advanced gastrointestinal stromal tumors after treatment with targeted systemic therapy using kinase inhibitors.  J Clin Oncol. 2006;24(15):2325-2331.PubMedGoogle ScholarCrossref
33.
Du  CY, Zhou  Y, Song  C,  et al.  Is there a role of surgery in patients with recurrent or metastatic gastrointestinal stromal tumours responding to imatinib? a prospective randomised trial in China.  Eur J Cancer. 2014;50(10):1772-1778.PubMedGoogle ScholarCrossref
34.
Boikos  SA, Pappo  AS, Killian  JK,  et al.  Molecular subtypes of KIT/PDGFRA wild-type gastrointestinal stromal tumors: a report from the National Institutes of Health Gastrointestinal Stromal Tumor Clinic.  JAMA Oncol. 2016;2(7):922-928.PubMedGoogle ScholarCrossref
Original Investigation
July 2017

Correlation of Long-term Results of Imatinib in Advanced Gastrointestinal Stromal Tumors With Next-Generation Sequencing ResultsAnalysis of Phase 3 SWOG Intergroup Trial S0033

Author Affiliations
  • 1Division of Hematology and Medical Oncology, Portland Veterans Affairs Health Care System, Knight Cancer Institute, Oregon Health and Science University, Portland
  • 2SWOG Statistical Center, Seattle, Washington
  • 3Division of Hematology and Medical Oncology, SWOG Group Chair’s Office, Knight Cancer Institute, Oregon Health and Science University, Portland
  • 4Department of Medicine, Developmental Therapeutics and Gastrointestinal Malignancies Programs, Dana Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts
  • 5Cancer Division, Departments of Solid Tumor Oncology and Cancer Biology, Cleveland Clinic Foundation, Cleveland, Ohio
  • 6Department of Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania
  • 7Medical Oncology, Mount Sinai Hospital, Toronto, Ontario, Canada
  • 8Medical Oncology, Washington Cancer Institute, Medstar Washington Hospital Center, Washington, DC
  • 9Medical Oncology, Memorial Sloan-Kettering Cancer Center, Weill Cornell Medical College, New York, New York
  • 10Medical Oncology, Monter Cancer Center, Northwell Health, Lake Success, New York
  • 11Medical Oncology, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York
  • 12Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts
  • 13Department of Pediatrics, Brigham and Women’s Hospital, Boston, Massachusetts
  • 14Department of Biostatistics and Bioinformatics, Duke University, Durham, North Carolina
  • 15Division of Cancer Medicine, University of Texas, MD Anderson Cancer Center, Houston
  • 16Division of Hematology/Oncology, Department of Internal Medicine, University of Michigan, Ann Arbor
  • 17Department of Pharmacology, University of Michigan, Ann Arbor
JAMA Oncol. 2017;3(7):944-952. doi:10.1001/jamaoncol.2016.6728
Key Points

Question  What are the long-term clinical outcomes for patients with advanced gastrointestinal stromal tumors treated with front-line imatinib mesylate?

Findings  In this follow-up study of a randomized clinical trial of 695 adults treated with imatinib mesylate, the estimated 10-year progression-free and overall survival rates were approximately 7% and 23%, respectively. The highest 10-year progression-free and overall survival results were obtained for patients with KIT exon 11–mutant tumors or tumors lacking KIT/PDGFRA mutations (primarily succinate dehydrogenase–mutant tumors).

Meaning  Imatinib front-line treatment of advanced and/or metastatic gastrointestinal stromal tumors leads to long-term survival (≥10 years) in a substantial minority of treated patients, especially those with KIT exon 11–mutant gastrointestinal stromal tumors.

Abstract

Importance  After identification of activating mutations of the KIT gene in gastrointestinal stromal tumor (GIST)—the most common sarcomaof the gastrointestinal tract—a phase 2 study demonstrated efficacy of imatinib mesylate in patients with metastatic GIST harboring a KIT exon 11 mutation. Initial results of long-term follow-up have found a survival benefit in this subgroup of patients.

Obective  To assess the long-term survival of patients with GIST who were treated in SWOG study S0033 and to present new molecular data regarding treatment outcomes.

Design, Setting, and Participants  In this follow-up of randomized clinical trial participants (from December 15, 2000, to September 1, 2001), patients were required to have advanced GIST that was not surgically curable. Postprotocol data collection occurred from August 29, 2011, to July 15, 2015. Using modern sequencing technologies, 20 cases originally classified as having wild-type tumors underwent reanalysis. This intergroup study was coordinated by SWOG, a cooperative group member within the National Clinical Trials Network, with participation by member/affiliate institutions. This follow-up was not planned as part of the initial study.

Interventions  Patients were randomized to 1 of 2 dose levels of imatinib mesylate, including 400 mg once daily (400 mg/d) vs 400 mg twice daily (800 mg/d), and were treated until disease progression or unacceptable toxic effects of the drug occurred.

Main Outcomes and Measures  The primary end point was overall survival. Updated survival data were correlated with clinical and molecular factors, and patterns of postprotocol therapies were enumerated and described in long-term survivors.

Results  Of 695 eligible patients (376 men [54.1%]; 319 women [45.9%]; mean [SD] age, 60.1 [14.0] years), 189 survived 8 years or longer, including 95 in the 400-mg/d dose arm and 94 in the 800-mg/d arm. The 10-year estimate of overall survival was 23% (95% CI, 20%-26%). Among 142 long-term survivors, imatinib was the sole therapy administered in 69 (48.6%), with additional systemic agents administered to 54 patients (38.0%). Resequencing studies of 20 cases originally classified as KIT/PDGFRA wild-type GIST revealed that 17 (85.0%) harbored a pathogenic mutation, most commonly a mutation of a subunit of the succinate dehydrogenase complex.

Conclusions and Relevance  A subset of patients with metastatic GIST experiences durable, long-term overall survival with imatinib treatment. Although this study provides guidance for management of GIST harboring the most common KIT and PDGFRA mutations, optimal management of other genotypic subtypes remains unclear.

Trial Registration  clinicaltrials.gov Identifier: NCT00009906

Introduction

Gastrointestinal stromal tumor (GIST) is the most common sarcoma of the gastrointestinal tract. GIST constitutes less than 1% of all gastrointestinal tract tumors and has an annual incidence of approximately 7 to 10 cases per million as determined by multiple population-based studies.1-5 A major breakthrough occurred with the discovery of activating mutations of the KIT gene (CCDS CCDS3496.1), resulting in oncogenic constitutive signaling in most GISTs and the subsequent use of KIT (CD117) immunostaining as the first diagnostic marker for GIST.6,7

Before 2000, GIST was documented to be highly resistant to cytotoxic chemotherapy, with no available effective treatment and a uniformly grim prognosis for patients with metastatic or unresectable disease.8-10 A brief report11 published in 2001 described the impressive effects of a pilot proof-of-concept protocol using the tyrosine kinase inhibitor imatinib mesylate in a patient with metastatic GIST harboring a KIT exon 11 mutation that had been previously refractory to chemotherapy. Since then, several phase 2 and 3 trials in metastatic disease12-16 were conducted, confirming the efficacy of imatinib in metastatic GIST.

A phase 2 study (B2222) initially reported in 200212 was the first multicenter trial designed specifically for advanced GIST to report that imatinib produced high response rates and lasting disease control. A follow-up report from this study15 in 2008 showed a median overall survival (OS) of 4.75 years for the 147 patients treated, with 41 patients (27.9%) continuing long-term drug therapy. The presence of a KIT exon 11 mutation was associated with better survival. Estimated median survival was 5.25 years for patients with KIT exon 11 mutations and 3.67 years for those with KIT exon 9 mutations.15 A subsequent analysis of this trial found that 26 (17.7%) of the total 147 patients entered in this study continued to receive imatinib therapy, with a median follow-up time of 9.4 years.17 The estimated 9-year OS rate for all patients was 35%. No data were provided about other therapeutic modalities, such as surgical resection of metastatic lesions or other postprotocol systemic therapies that might have contributed to these OS results.

A large SWOG-directed, randomized phase 3 intergroup study, S0033, was designed and conducted to compare the outcome of patients with metastatic and/or unresectable (ie, surgically incurable) GIST randomized to be treated with imatinib mesylate at an initial dose of 400 mg once daily (400 mg/d) or 400 mg twice daily (800 mg/d). Blanke et al14 previously reported a median OS of 4.58 years for 345 patients treated with conventional-dose imatinib and 4.25 years for 349 patients treated in the high-dose arm. Further long-term results of this study are described herein, along with additional analyses of GIST tumors originally classified as KIT and platelet-derived growth factor receptor α gene (PDGFRA [NM_006206]) wild-type (KIT/PDGFRA WT) genotype using next-generation sequencing techniques. We also have correlated GIST genotypes with clinical outcomes during treatment, including a cohort of patients with succinate dehydrogenase (SDH)–deficient GIST.

Methods
S0033 Study Population

The S0033 trial accrued patients from December 15, 2000, to September 1, 2001, from 4 cancer clinical trial cooperative groups (SWOG, Cancer and Leukemia Group B, Eastern Oncology Group, and National Cancer Institute of Canada Clinical Trials Group) and from the University of Texas MD Anderson Cancer Center. The trial protocol is available in Supplement 1. Data collection and analyses for the trial were performed by the SWOG Statistical Center (Figure 1). For the original analysis, approval by the institutional review board of each participating institution was obtained, with written informed consent obtained from each participant.

Patients were required to have biopsy-proven metastatic and/or unresectable GIST of visceral or abdominal origin and immunohistochemical demonstration of KIT expression documented by antibody staining (DAKO Corporation). Complete details and results from the clinical study were previously reported.14 Tumor samples were sent to the Oregon Health and Science University, Portland, where tumor genotyping was centrally assessed. Initial results of the KIT and PDGFRA genotyping and correlation with clinical outcomes were also previously published.18

Ten years after initiation of accrual on this study, investigators following up patients last known to be alive were contacted to update follow-up. Patients known to have lived 8 years or more were defined as long-term survivors. A 2-page data form was created to obtain additional information about these long-term survivors. Use of additional therapies after discontinuation of imatinib therapy in this study was tabulated for these long-term survivors. Follow-up data were collected from December 26, 2000, through July 15, 2015. A concerted effort to gather delinquent data as well as the postprotocol therapies started August 29, 2011. The primary aim of this report was to correlate updated survival with clinical and molecular factors and to enumerate and describe patterns of postprotocol therapies of the long-term survivors.

Statistical Analysis

The distribution of OS was estimated using the Kaplan-Meier method.19 We used proportional hazards regression models to investigate the prognostic effect of the following variables on OS: age; sex; performance status (0, 1, 2, or 3, with higher scores indicating more limited activity); time from the initial diagnosis (in years); primary disease site (small bowel vs other); maximum diameter of the largest tumor; prior surgery, chemotherapy, or radiotherapy; and baseline white blood cell count, hemoglobin level, absolute neutrophil count, platelet count, and levels of bilirubin, albumin, and creatinine. For baseline white blood cell count, hemoglobin level, absolute neutrophil count, platelet count, and bilirubin and creatinine levels, we used a log transformation in the regression models. Initially, each factor was assessed in a univariate fashion. In subsequent analyses, multivariable models were created using the factors found to be significant in the univariate models at P = .05 using a backward selection model. Associations between genotype and patient characteristics were tested using χ2 test or the Fisher exact test. P values are unadjusted for multiple comparisons.

Targeted Exome Sequencing Methods

Targeted exome sequencing analyses were performed as previously described.20 Additional details are included in eMethods in Supplement 2.

Results

A total of 695 eligible patients were randomized to the 400- and 800-mg/d imatinib mesylate dose arms (376 men [54.1%]; 319 women [45.9%]; mean [SD] age, 60.1 [14.0] years). The initial clinical report14 found no statistical difference in OS, progression-free survival (PFS), or response between the treatment arms. Similarly, neither the B2222 phase 2 study reported by Blanke et al14 nor the European Organization for Research and Treatment of Cancer phase 3 study using an identical study design reported by Verweij et al13,16 found a difference in OS outcomes between standard and higher doses of imatinib. However, Verweij et al13,16 reported a difference in PFS in the KIT exon 9–mutant subset in favor of the high-dose imatinib treatment arm. In our present study of 695 patients, 556 have died with a median OS of 52 (95% CI, 48-61) months (Figure 2A and eTable 1 in Supplement 2). The 10-year OS estimate was 23% (95% CI, 20%-26%). The 10-year PFS estimate was 7% (95% CI, 6%-10%) (Figure 2B and eTable 1 in Supplement 2).

Of the 346 patients initially assigned to the 400-mg/d dose arm, 95 (27.5%) were long-term survivors, including 72 who continued to receive low-dose imatinib and 23 who were crossed over to the 800-mg/d dose arm. A total of 130 patients were crossed over during the study conduct. Among the 349 patients initially assigned to the 800-mg/d dose arm, 94 (26.9%) were long-term survivors.

The following prognostic factors were identified by univariate analysis as statistically significant with respect to OS: age; sex; performance status; prior chemotherapy; maximum tumor diameter; and baseline platelet count, hemoglobin level, absolute neutrophil count, white blood cell count, and bilirubin and albumin levels. In the multivariable model, 551 of the 695 eligible patients had complete data for all baseline prognostic variables. Using backward selection, multivariable analyses showed that being younger and female and having a good performance status score, smaller tumor diameter, lower white blood cell count, and higher albumin value were associated with significantly improved OS. The P values and hazard ratios are given in eTable 2 in Supplement 2 for the univariate and multivariable analyses. The influence of these prognostic factors on the estimate of 10-year survival is listed in eTable 3 in Supplement 2. For example, the 10-year survival estimate for patients with a performance status of 0 to 1 at the time of treatment initiation was 26% (95% CI, 22%-30%) compared with 7% (95% CI, 2%-13%) for patients with a performance status of 2 to 3.

Genotype results from the original analysis were available for 395 eligible patients, of whom 282 (71.4%) had KIT exon 11 mutations, 67 (17.0%) had the KIT/PDGFRA WT genotype, 32 (8.1%) had KIT exon 9 mutations, and 14 (3.5%) had other KIT or PDGFRA mutations. A univariate analysis of OS found a median OS of 66 (95% CI, 57-78) months for patients with the KIT exon 11 mutation, 38 (95% CI, 28-47) months for the KIT exon 9 mutation, and 40 (95% CI, 32-63) months for KIT/PDGFRA WT genotypes. When adjusted for all prognostically significant factors found in the analysis of eligible patients, the OS for patients with a KIT exon 11 mutation was significantly longer than for those with KIT/PDGFRA WT genotypes (P = .004; Figure 3A and eTable 1 in Supplement 2). We also analyzed survival for different classes of KIT exon 11 mutations, including any type of KIT exon 11 deletion (median OS, 65 months; 95% CI, 51-78 months) vs insertion/duplication mutations (median OS, 73 months; 95% CI, 50-87 months) vs point mutations (median OS, 66 months; 95% CI, 52-105 months). No significant difference in OS was found for patients whose GIST harbored different subsets of KIT exon 11 mutations (Figure 3B and eTable 1 in Supplement 2).

When mutation status in long-term survivors was correlated with response (complete and partial), more responses were seen in the KIT exon 11 mutant genotype group (64 of 92 [69.6%]) than in the KIT/PDGFRA WT (10 of 21 [47.6%]) or the KIT exon 9 (2 of 4 [50.0%]) groups. Median PFS estimates by treatment arm and mutation status are summarized in eTable 1 in Supplement 2. The median PFS for patients with KIT exon 11 mutations was 25 (95% CI, 21-28) months compared with 17 (95% CI, 9-25) months for patients with KIT exon 9 mutations and 13 (95% CI, 8-18) months for those with the WT genotype (eFigureA in Supplement 2). As with our OS analysis, no difference in PFS was found among patients whose GIST had different classes of KIT exon 11 mutations, including any type of KIT exon 11 deletion (median PFS, 24 months; 95% CI, 20-29 months) vs insertion/duplication mutations (median PFS, 30 months; 95% CI, 17-34 months) vs point mutations (median PFS, 24 months; 95% CI, 17-28 months) (eFigureB and eTable 1 in Supplement 2).

Postprotocol Progression Therapies During Survival Follow-up

Descriptions of additional therapies received after discontinuation of imatinib treatment in this protocol were obtained for patients defined as long-term survivors (ie, those known to have lived for ≥8 years after enrollment in this trial). This additional follow-up was not planned as part of the initial study. Of the 189 long-term survivors, additional treatment information was obtained from 142 individuals. Of these 142 patients, imatinib was the sole therapy administered to 69 (48.6%), while 73 (51.4%) received other therapies. Fifty-four (38.0%) received additional systemic agents; sunitinib malate (41 of 142 [28.9%]) and sorafenib (17 of 142 [12.0%]) were the most commonly used additional therapies among the long-term survivors. Subsequent surgical procedures were also reported in 41 patients (28.9%); radiofrequency ablation, in 10 (7.0%); and radiotherapy, in 6 (4.2%) (eTable 4 in Supplement 2). Of the 395 eligible patients with a known GIST genotype, 117 were long-term survivors. Of these, 90 (76.9%) had information on additional nonprotocol therapy.

Resequencing Studies From a Subset of Cases Previously Identified as KIT/PDGFRA WT

At the time of study initiation, KIT mutations were the only known pathogenic abnormality associated with GIST. After the completion of enrollment of patients to this study (September 1, 2001), a number of additional pathogenic events in GIST were described, including PDGFRA, KRAS (CCDS8702.1), NRAS (CCDS877.1), HRAS (NM_005343.2), and BRAF (CCDS5863.1) mutations, loss of SDH complex iron sulfur subunit B protein expression owing to mutation or epimutation, and the association of GIST with neurofibromatosis.21-24 The KIT and PDGFRA mutations were the only analyses performed in the original report owing to limitations of testing that existed at that time. However, significant advances have improved our ability to sequence tumor DNA from formalin-fixed, paraffin-embedded tumors and to test for multiple genomic abnormalities in the same analysis. We identified 20 KIT/PDGFRA WT cases for which sufficient residual DNA remained from the original studies for reanalysis using a targeted exome panel for GIST-associated pathogenic abnormalities. We identified the putative causative mutation in 17 (85.0%) of the 20 cases (eTable 5 in Supplement 2). The most commonly mutated pathway in our analysis was the SDH complex, with a mutation in 12 (60.0%) of 20 cases (SDHA [CCDS3853.1] in 9, SDHB [CCDS176.1] in 2, and SDHC [CCDS1230.1] in 1). We also identified 2 cases with neurofibromin 1 gene (NF1 [CCDS42292.1]) loss as the presumed pathogenic mutation. Finally, we found 3 cases with KIT mutations that were missed in the original analysis (KIT exon 13 K642E, KIT exon 11 V559D, and KIT exon 9 K509I in 1 case each). Three cases (15%) of the original KIT/PDGFRA WT cohort did not contain an identifiable pathogenic mutation using our targeted panel. Owing to the lack of residual unstained slides, we could not perform SDHB immunohistochemistry in these 3 cases. Therefore, we do not know whether these cases represent “quadruple-WT GIST” or cases with SDH deficiency with alterations that could not be detected using our methods (eg, epimutation).25,26

Our collection of 12 patients with SDH-deficient GIST represents the largest series of such patients treated uniformly with imatinib as part of a prospective clinical study. The demographics and baseline characteristics of these patients are detailed in the Table and compared with patients with KIT exon 11–mutant GIST. As expected, the patients with SDH-mutant GIST were younger than the patients with KIT exon 11–mutant GIST, and all 11 (100%) of the primary tumors arose outside the small bowel (presumably all gastric) compared with 172 (65.6%) outside the small bowel for KIT exon 11–mutant tumors (P = .02). The response rate of the patients with SDH-mutant GIST was 8.3% (95% CI, 0%-48%. There were no complete responses (0%; 95% CI, 0%-26%) and 1 partial response (8.3%; 95% CI, 0%-48%). In comparison, the response rate included 185 (65.6%; 95% CI, 60%-71%) of 282 patients with KIT exon 11–mutant tumors, including a complete response in 18 (6.4%; 95% CI, 3%-10%) and a partial response in 167 (59.2%; 95% CI, 53%-65%) (P < .001). Median OS was similar for both groups, although the CI for the small SDH-deficient population was quite broad (116 [95% CI, 22-116] months vs 66 [95% CI, 57-78] months; P = .38) (Table and Figure 4A). Likewise, the median PFS was not significantly different for patients with SDH-deficient tumors compared with KIT exon 11–mutant GIST (9 [95% CI, 3-58] months vs 24 [95% CI, 21-28] months; P = .07, adjusted for all prognostically significant variables) (Table and Figure 4B).

Discussion

As previously noted, before the clinical development of imatinib, metastatic GIST was a uniformly fatal disease with a rapidly progressive course and a median OS of less than 2 years. The initial publication of the results from S0033 reported on median survival and molecular biomarkers associated with the response to imatinib. With a much longer follow-up, we now report that a significant subset of patients with GIST experience long-term survival with front-line imatinib treatment of advanced and metastatic disease. Estimated 8- and 10-year survival were 31% and 23%, respectively. Notably, patients with advanced GIST treated with imatinib alone have experienced survival without report of progression for periods that approach and even exceed 1 decade. In the subset of patients who have survived more than 8 years since study entry, 73 (51.4%) received additional nonprotocol therapy, with only 54 (38.0%) receiving further systemic therapies beyond imatinib.

In this study, tumor bulk at the time of initiation of imatinib therapy significantly influences PFS and OS. This finding supports previous studies of imatinib therapy of advanced GIST.27,28 The influence of tumor bulk (complete vs partial responses vs stable disease) on PFS was also noted in the randomized imatinib discontinuation studies reported by the French Sarcoma Group.29 Taken together, these observations raise the hypothesis that surgical tumor debulking may enhance the duration of imatinib response and disease control, at least in some patients. Several nonrandomized, retrospective surgical series30-32 have suggested a benefit to surgical removal of metastatic lesions in imatinib-treated patients. This hypothesis was tested in a randomized phase 3 study reported by Du et al.33 In their study, patients were randomized from 3 to 13 months after starting imatinib therapy to undergo surgery for residual disease with the goal of removing macroscopic disease as completely as possible or to continue imatinib therapy without surgical intervention. The primary end point was the 2-year PFS. Unfortunately, the study was closed early owing to poor accrual, with only 41 patients enrolled of a planned 210. No advantage to surgery was found in terms of PFS (2-year PFS, 88% for combined treatment vs 58% for medical therapy only; P = .09). However, the secondary end point of OS was significantly improved by combination treatment (median OS not reached for surgery vs 49 months for imatinib-only treatment; P = .02). These data suggest that surgical debulking may maximize the imatinib response for some patients, but how to optimally integrate surgery and imatinib treatment in the management of metastatic GIST remains unclear.

The resequencing studies reported herein confirm the hypothesis that most GISTs harbor identifiable pathogenic mutations. In the original report, the presumed causative mutation was identified in 85% of GIST (KIT and PDGFRA analysis only). Using next-generation sequencing technologies, such as those described above, we now estimate that 97.5% of GIST can be assigned a pathogenic genotype. The remaining 2.5% of GIST represent GIST with SDH deficiency due to mutations outside of the 4 SDH subunits that affect SDH complex function or GIST with undiscovered pathogenic mechanisms (quadruple-WT GIST). The objective response rate to imatinib treatment was significantly lower for SDH-mutant GIST when compared with KIT exon 11–mutant GIST (1 [8.3%] vs 185 [65.6%]; Fisher exact test, P < .001) based on post hoc analyses. The single responding patient with SDH-mutant GIST had a verified 62% decrease in tumor size that lasted for 4 years before progression was noted. These results are consistent with a recent retrospective case series that reported 1 partial response of 49 patients with SDH-deficient GIST treated with imatinib.34 Our results also identify GIST with KIT exon 11 mutations or lacking KIT/PDGFRA mutations (most of which are SDH deficient) as having a longer median OS when compared with KIT exon 9–mutant GIST. Given prior reports of cases of indolent behavior of untreated SDH-deficient metastatic GIST, whether imatinib alters the natural history of most cases of KIT/PDGFRA WT GIST is unclear.34 Further studies are required to better define the role of imatinib therapy in the treatment of this type of GIST.

Limitations

This study reports long-term treatment outcomes from a phase 3 study that enrolled patients from mid-December 2000 to early September 2001. Although we obtained long-term follow-up for most patients, we do not have data on the status and postprotocol therapies (if any) for all patients because we only attempted to collect this information for all patients surviving 8 years or longer. In addition, these reported treatment results reflect primary therapy with imatinib as part of our study and the potential influence of subsequent postprotocol therapies. The impact of postprotocol treatments on the expected OS of patients with metastatic GIST currently starting imatinib treatment cannot be determined from our data set. In addition, although this study provides guidance for management of GIST harboring the most common KIT and PDGFRA mutations, optimal management of other genotypic subtypes remains unclear.

Conclusions

Our results provide important data on the long-term outcomes of patients with metastatic GIST treated with imatinib. Further studies are needed to improve on these results using other strategies that might incorporate surgery, combination medical therapy (additional signaling pathway inhibitors or immune modulation agents), or some form of intermittent therapy. Alternatively, enhanced imaging and biomarker analysis to detect residual GIST cells in patients with long-term response may identify those patients who might benefit from less intense medical therapy.

Back to top
Article Information

Corresponding Author: Michael C. Heinrich, MD, Division of Hematology and Medical Oncology, Portland Veterans Affairs Health Care System, Knight Cancer Institute, Oregon Health and Science University, 3710 SW US Veterans Hospital Rd, Mail Code R&D-19, Portland, OR 97239 (heinrich@ohsu.edu).

Accepted for Publication: December 6, 2016.

Correction: This article was corrected on March 23, 2017, for a missing author initial in the byline.

Published Online: February 9, 2017. doi:10.1001/jamaoncol.2016.6728

Author Contributions: Dr Heinrich had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Heinrich, Blanke, Demetri, Borden, von Mehren, Crowley, Benjamin, Baker.

Acquisition, analysis, or interpretation of data: Heinrich, Rankin, Blanke, Demetri, Ryan, von Mehren, Blackstein, Priebat, Tap, Maki, Corless, Fletcher, Owzar, Crowley, Benjamin, Baker.

Drafting of the manuscript: Heinrich, Blanke, Demetri, Fletcher, Owzar, Baker.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Heinrich, Rankin, Owzar, Crowley.

Obtained funding: Heinrich, Demetri, Fletcher.

Administrative, technical, or material support: Heinrich, Blanke, Blackstein, Baker.

Study supervision: Blanke, Demetri, Borden, von Mehren, Blackstein, Priebat, Fletcher, Benjamin.

Conflict of Interest Disclosures: Dr Heinrich reports stock or other ownership with MolecularMD; honoraria from Novartis and Pfizer; a consulting or advisory role with Novartis, Ariad, and Blueprint; research funding from Blueprint, Inhibikase, Ariad, and Deciphera; patent or intellectual property with Novartis (via his institution); and expert testimony for Novartis. Ms Rankin reports employment with Pathology Associates Medical Laboratories. Dr Borden reports stock or other ownership with Alios BioPharma and patent or intellectual property through Medical College of Wisconsin, University of Wisconsin, and Cleveland Clinic. Dr Tap reports a consulting or advisory role with Novartis. Dr Maki reports honoraria from Novartis; a consulting or advisory role with Novartis; and research funding through his institution from Novartis. Dr Corless reports a consulting or advisory role with Roche and Asuragen; research funding through his institution from Roche; and travel, accommodations, or expenses from Roche. Dr Owzar reports patent or intellectual property through Duke University. Dr Benjamin reports a consulting or advisory role with Novartis. Dr Baker reports a consulting or advisory role with Tevan and Morphotek. No other disclosures were reported.

Funding/Support: This study was supported in part by grants CA180888, CA180819, CA180801, CA180846, CA180835, CA180818, CA180834, CA180828, CA180830, CA180820, CA180821, and CA180863 from the Public Health Service, Department of Health and Human Services, the National Cancer Institute (NCI), National Clinical Trials Network; by grants CA189953, CA189952, CA189954, CA189860, CA189804, CA189822, CA189971, CA189830, CA189858, CA189853, CA189957, CA189854, and CA189856 from the NCI Community Oncology Research Program; by grants CA13612, CA46282, CA63850, CA35119, CA74811, CA45450, CA58686, CA35262, CA46368, CA04919, CA68183, CA58348, CA16385, CA76447, CA46113, CA58416, CA12644, CA37981, and CA22433 from the National Institutes of Health (NIH), NCI legacy; by grants 5P50CA127003-08 and U54CA168512-04 from the NCI Specialized Programs of Research Excellence program; by grant 021039 from the Canadian Cancer Society; by Novartis; by Merit Review grants 1I01BX000338-01 and 2I01BX000338-05 from the Department of Veterans Affairs (Dr Heinrich); by funding from the GIST Cancer Research Fund (Dr Heinrich); and by the Life Raft Group (Dr Heinrich).

Role of the Funder/Sponsor: This study was sponsored by NCI/Cancer Therapy Evaluation Program and conducted by 4 cancer clinical trial cooperative groups (SWOG, Cancer and Leukemia Group B, Eastern Oncology Group, and National Cancer Institute of Canada Clinical Trials Group) and by the University of Texas MD Anderson Cancer Center. Data collection and analyses for the trial were performed by the SWOG Statistical Center. The authors are solely responsible for preparation and approval of the manuscript and decision to submit the manuscript for publication. Funding sources for individual investigators and/or their institutions are listed elsewhere. These funding sources had no role in the design or conduct of the study; collection, management, or analysis of the data; preparation or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or any of the funding organizations.

References
1.
Nilsson  B, Bümming  P, Meis-Kindblom  JM,  et al.  Gastrointestinal stromal tumors: the incidence, prevalence, clinical course, and prognostication in the preimatinib mesylate era: a population-based study in western Sweden.  Cancer. 2005;103(4):821-829.PubMedGoogle ScholarCrossref
2.
Mucciarini  C, Rossi  G, Bertolini  F,  et al.  Incidence and clinicopathologic features of gastrointestinal stromal tumors: a population-based study.  BMC Cancer. 2007;7:230.PubMedGoogle ScholarCrossref
3.
Steigen  SE, Eide  TJ.  Trends in incidence and survival of mesenchymal neoplasm of the digestive tract within a defined population of northern Norway.  APMIS. 2006;114(3):192-200.PubMedGoogle ScholarCrossref
4.
Tryggvason  G, Gíslason  HG, Magnússon  MK, Jónasson  JG.  Gastrointestinal stromal tumors in Iceland, 1990-2003: the Icelandic GIST study, a population-based incidence and pathologic risk stratification study.  Int J Cancer. 2005;117(2):289-293.PubMedGoogle ScholarCrossref
5.
Tran  T, Davila  JA, El-Serag  HB.  The epidemiology of malignant gastrointestinal stromal tumors: an analysis of 1458 cases from 1992 to 2000.  Am J Gastroenterol. 2005;100(1):162-168.PubMedGoogle ScholarCrossref
6.
Hirota  S, Isozaki  K, Moriyama  Y,  et al.  Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors.  Science. 1998;279(5350):577-580.PubMedGoogle ScholarCrossref
7.
Rubin  BP, Fletcher  JA, Fletcher  CD.  Molecular insights into the histogenesis and pathogenesis of gastrointestinal stromal tumors.  Int J Surg Pathol. 2000;8(1):5-10.PubMedGoogle ScholarCrossref
8.
Plaat  BE, Hollema  H, Molenaar  WM,  et al.  Soft tissue leiomyosarcomas and malignant gastrointestinal stromal tumors: differences in clinical outcome and expression of multidrug resistance proteins.  J Clin Oncol. 2000;18(18):3211-3220.PubMedGoogle ScholarCrossref
9.
DeMatteo  RP, Lewis  JJ, Leung  D, Mudan  SS, Woodruff  JM, Brennan  MF.  Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival.  Ann Surg. 2000;231(1):51-58.PubMedGoogle ScholarCrossref
10.
Dematteo  RP, Heinrich  MC, El-Rifai  WM, Demetri  G.  Clinical management of gastrointestinal stromal tumors: before and after STI-571.  Hum Pathol. 2002;33(5):466-477.PubMedGoogle ScholarCrossref
11.
Joensuu  H, Roberts  PJ, Sarlomo-Rikala  M,  et al.  Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor.  N Engl J Med. 2001;344(14):1052-1056.PubMedGoogle ScholarCrossref
12.
Demetri  GD, von Mehren  M, Blanke  CD,  et al.  Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors.  N Engl J Med. 2002;347(7):472-480.PubMedGoogle ScholarCrossref
13.
Verweij  J, Casali  PG, Zalcberg  J,  et al.  Progression-free survival in gastrointestinal stromal tumours with high-dose imatinib: randomised trial.  Lancet. 2004;364(9440):1127-1134.PubMedGoogle ScholarCrossref
14.
Blanke  CD, Rankin  C, Demetri  GD,  et al.  Phase III randomized, intergroup trial assessing imatinib mesylate at two dose levels in patients with unresectable or metastatic gastrointestinal stromal tumors expressing the kit receptor tyrosine kinase: S0033.  J Clin Oncol. 2008;26(4):626-632.PubMedGoogle ScholarCrossref
15.
Blanke  CD, Demetri  GD, von Mehren  M,  et al.  Long-term results from a randomized phase II trial of standard- versus higher-dose imatinib mesylate for patients with unresectable or metastatic gastrointestinal stromal tumors expressing KIT J Clin Oncol. 2008;26(4):620-625.PubMedGoogle ScholarCrossref
16.
Gastrointestinal Stromal Tumor Meta-Analysis Group (MetaGIST).  Comparison of two doses of imatinib for the treatment of unresectable or metastatic gastrointestinal stromal tumors: a meta-analysis of 1640 patients.  J Clin Oncol. 2010;28(7):1247-1253.PubMedGoogle ScholarCrossref
17.
von Mehren  M, Heinrich  MC, Jeonsuu  H,  et al.  Follow-up results after 9 years (yrs) of the ongoing, phase II B2222 trial of imatinib mesylate (IM) in patients (pts) with metastatic or unresectable KIT+ gastrointestinal stromal tumors (GIST) [abstract 10016].  J Clin Oncol. 2011;29(15)(suppl):10016. doi:10.1200/jco.2011.29.15_suppl.10016Google ScholarCrossref
18.
Heinrich  MC, Owzar  K, Corless  CL,  et al.  Correlation of kinase genotype and clinical outcome in the North American Intergroup Phase III Trial of imatinib mesylate for treatment of advanced gastrointestinal stromal tumor: CALGB 150105 Study by Cancer and Leukemia Group B and Southwest Oncology Group.  J Clin Oncol. 2008;26(33):5360-5367.PubMedGoogle ScholarCrossref
19.
Kaplan  EL, Meier  R.  Nonparametric estimation from incomplete observation.  J Am Stat Assoc. 1958;53(282):457-481.Google ScholarCrossref
20.
Grasso  C, Butler  T, Rhodes  K,  et al.  Assessing copy number alterations in targeted, amplicon-based next-generation sequencing data.  J Mol Diagn. 2015;17(1):53-63.PubMedGoogle ScholarCrossref
21.
Corless  CL, Fletcher  JA, Heinrich  MC.  Biology of gastrointestinal stromal tumors.  J Clin Oncol. 2004;22(18):3813-3825.PubMedGoogle ScholarCrossref
22.
Joensuu  H.  Risk stratification of patients diagnosed with gastrointestinal stromal tumor.  Hum Pathol. 2008;39(10):1411-1419.PubMedGoogle ScholarCrossref
23.
Miettinen  M, Lasota  J.  Gastrointestinal stromal tumors: review on morphology, molecular pathology, prognosis, and differential diagnosis.  Arch Pathol Lab Med. 2006;130(10):1466-1478.PubMedGoogle Scholar
24.
Rubin  BP, Heinrich  MC, Corless  CL.  Gastrointestinal stromal tumour.  Lancet. 2007;369(9574):1731-1741.PubMedGoogle ScholarCrossref
25.
Nannini  M, Astolfi  A, Urbini  M,  et al.  Integrated genomic study of quadruple-WT GIST (KIT/PDGFRA/SDH/RAS pathway wild-type GIST).  BMC Cancer. 2014;14:685.PubMedGoogle ScholarCrossref
26.
Urbini  M, Astolfi  A, Indio  V,  et al.  SDHC methylation in gastrointestinal stromal tumors (GIST): a case report.  BMC Med Genet. 2015;16:87.PubMedGoogle ScholarCrossref
27.
Bonvalot  S, Eldweny  H, Péchoux  CL,  et al.  Impact of surgery on advanced gastrointestinal stromal tumors (GIST) in the imatinib era.  Ann Surg Oncol. 2006;13(12):1596-1603.PubMedGoogle ScholarCrossref
28.
Bauer  S, Hartmann  JT, de Wit  M,  et al.  Resection of residual disease in patients with metastatic gastrointestinal stromal tumors responding to treatment with imatinib.  Int J Cancer. 2005;117(2):316-325.PubMedGoogle ScholarCrossref
29.
Le Cesne  A, Ray-Coquard  I, Bui  BN,  et al; French Sarcoma Group.  Discontinuation of imatinib in patients with advanced gastrointestinal stromal tumours after 3 years of treatment: an open-label multicentre randomised phase 3 trial.  Lancet Oncol. 2010;11(10):942-949.PubMedGoogle ScholarCrossref
30.
Park  SJ, Ryu  MH, Ryoo  BY,  et al.  The role of surgical resection following imatinib treatment in patients with recurrent or metastatic gastrointestinal stromal tumors: results of propensity score analyses.  Ann Surg Oncol. 2014;21(13):4211-4217.PubMedGoogle ScholarCrossref
31.
Bauer  S, Rutkowski  P, Hohenberger  P,  et al.  Long-term follow-up of patients with GIST undergoing metastasectomy in the era of imatinib—analysis of prognostic factors (EORTC-STBSG collaborative study).  Eur J Surg Oncol. 2014;40(4):412-419.PubMedGoogle ScholarCrossref
32.
Raut  CP, Posner  M, Desai  J,  et al.  Surgical management of advanced gastrointestinal stromal tumors after treatment with targeted systemic therapy using kinase inhibitors.  J Clin Oncol. 2006;24(15):2325-2331.PubMedGoogle ScholarCrossref
33.
Du  CY, Zhou  Y, Song  C,  et al.  Is there a role of surgery in patients with recurrent or metastatic gastrointestinal stromal tumours responding to imatinib? a prospective randomised trial in China.  Eur J Cancer. 2014;50(10):1772-1778.PubMedGoogle ScholarCrossref
34.
Boikos  SA, Pappo  AS, Killian  JK,  et al.  Molecular subtypes of KIT/PDGFRA wild-type gastrointestinal stromal tumors: a report from the National Institutes of Health Gastrointestinal Stromal Tumor Clinic.  JAMA Oncol. 2016;2(7):922-928.PubMedGoogle ScholarCrossref
×