|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Prognostic Value of KIT Mutation Type, Mitotic Activity, and Histologic Subtype in Gastrointestinal Stromal TumorsByFrom the Departments of Pathology and Surgery, Brigham and Women’s Hospital; the Departments of Adult Oncology and Pediatric Oncology, Dana-Farber Cancer Institute, and the Departments of Pathology, Surgery, Pediatrics, and Medicine, Harvard Medical School, Boston, MA. Address reprint requests to Samuel Singer, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, M607, 1275 York Ave, New York, NY 10021; email: singers{at}mskcc.org
PURPOSE: Previous studies have reported clinical correlates for KIT mutations in GISTs, but in most of those studies the KIT mutations were found in less than 50% of the GISTs. The aim of this study was to evaluate the prognostic relevance for KIT mutations in a series of GISTs in which the mutations were evaluated intensively by genomic and cDNA sequencing. PATIENTS AND METHODS: A comprehensive clinical and pathologic analysis of 48 patients with GISTs who had snap-frozen tissue was performed. The median tumor size was 10 cm (range, 2 to 30 cm). Median follow-up for disease-free patients was 48 months. KIT genomic and cDNA was sequenced by using nucleic acid templates isolated from frozen tumors.
RESULTS: The overall 5-year recurrence-free survival was 41% ± 6%. Five-year recurrence-free survival for patients with tumors that had mitotic counts of three mitoses or fewer per 30 high-power fields (HPF), more than three to CONCLUSION: In this series of KIT-expressing GISTs, tumor mitotic activity and histologic subtype were the most important prognostic features. The majority of GISTs contain KIT-activating mutations with the type/location of mutation serving as an independent predictor for disease-free survival. These results suggest that KIT mutation and activation are important in GIST pathogenesis and also may provide important prognostic information.
GASTROINTESTINAL STROMAL tumors (GISTs), although relatively rare, comprising only 0.1% to 3% of all gastrointestinal neoplasms, are the most common mesenchymal tumors of the gastrointestinal tract.1 GISTs have been the subject of considerable controversy regarding their line of differentiation, classification, and prognosis. They are heterogeneous histologically, with most composed of spindle cells, but they sometimes have an epithelioid component. The term GIST was initially introduced as a histogenetically neutral term referring to the main group of mesenchymal tumors of the gastrointestinal, tract which could be verified neither as neurogenic nor of smooth muscle origin. Recent evidence suggests that most mesenchymal tumors of the gastrointestinal tract form a biologically distinct entity that seems to differentiate into cells resembling the interstitial cells of Cajal (ICC).1-3 The ICC form a network of cells in the muscle walls of the gastrointestinal tract which function as a pacemaker system controlling peristaltic activity. Expression of the KIT receptor tyrosine kinase plays a critical role in ICC differentiation and proliferation and is a specific feature of GISTs as compared with other neurogenic and smooth muscle tumors of the abdomen.4-6 Arguably the most biologically relevant, specific, and practical diagnostic criteria for GIST is KIT (CD117) expression by immunhistochemistry.7 Many of the published series to date have used clinical presentation and morphology as inclusion criteria for the GIST diagnosis and thus may contain patients that are not true GIST cases. Over the last 2 to 3 years, it has become evident that immunohistochemistry with the KIT antibody, along with a panel of antibodies to S-100 protein, smooth muscle actin, and desmin, is essential to help delineate true GISTs from smooth muscle tumors, neural tumors, desmoids, and other spindle-cell neoplasms.7 KIT is a type III receptor tyrosine kinase which is activated when bound by a ligand known as steel factor or stem-cell factor.8 Oncogenic mutations involving KIT exon 11 have been found in a subset of GISTs.9-13 KIT exon 11 encodes the KIT receptor juxtamembrane domain, located in the region between the transmembrane and tyrosine kinase domains. The KIT juxtamembrane domain is pivotal in KIT signal transduction, with mutations in this region resulting in constitutive dimerization without binding stem-cell factor. In prior studies, GISTs with KIT exon 11 mutations were typically higher grade, or associated with poorer outcome, than those that lacked such mutations.10,12,13 In this single-institution study, 48 GISTs, as determined by overexpression of KIT on immunohistochemistry, were evaluated for mutations throughout the KIT coding sequence. The mutational analysis was then combined with clinicopathologic factors to evaluate whether various categories of KIT mutations might be predictive of clinical behavior independent of conventional clinicopathologic factors, such as mitotic activity, size, and histologic subtype.
Patients Between 1990 and 2000, 48 patients with localized GIST underwent surgical resection at the Brigham and Women’s Hospital and had snap-frozen tissue available for mutational analysis. To meet inclusion criteria, all patients had to have a histologically confirmed diagnosis of GIST, determined either prospectively or retrospectively, and had snap-frozen tissue available from their tumors at the time of surgical resection. The criteria used for the pathologic definition of GIST was diffuse and strong KIT expression by immunohistochemical analysis. Forty-four patients had primary GISTs without evidence of peritoneal dissemination. Four patients presented with regional peritoneal dissemination, which was evident at the time of resection of the primary tumor. None of the patients had distant metastasis at the time of presentation. Standard evaluation of patients included a history and physical examination, complete blood count, and serum chemistry analyses. Computed tomography scans of the chest, abdomen, and pelvis were also obtained. Patient, tumor, treatment, and survival data were prospectively acquired and entered into our sarcoma database. Patient demographics included age at diagnosis and sex. Location of tumor was divided into stomach, small bowel, and intraperitoneal dissemination at presentation. Tumor size was based on the largest dimension of the primary tumor as reported at the time of initial surgical resection. Distant metastases were identified by imaging studies. The type of surgery performed (total or subtotal excision) was ascertained from the operating surgeon’s assessment as documented on the operative report. Specimen margins were analyzed for evidence of microscopic disease. None of the patients received radiation therapy. Thirteen patients received chemotherapy. Patients were observed in our soft tissue sarcoma program at 3- to 4-month intervals during the first 2 to 3 years and at 6-month intervals thereafter. Information obtained during follow-up included status (alive without disease, alive with recurrent disease, dead without evidence of recurrent disease, or dead as a result of sarcoma treatment or with sarcoma).
Pathology
KIT Sequence Analysis
Statistical Methods
Patient and Tumor Characteristics Patient characteristics are listed in Table 1. There were 27 men and 21 women, with a median age of 56 years (range, 24 to 84 years). The histologic grade was low risk for 10 patients (21%), intermediate risk for 10 patients (21%), and high risk for 28 patients (58%). The median tumor size was 10 cm (range, 2 to 30 cm). The most common tumor cell morphology was spindle cell (32 patients), followed by mixed spindle cell and epithelioid (13 patients) and pure epithelioid (three patients). The most common primary tumor sites were the stomach (28 patients), followed by the small bowel (16 patients) and peritoneal dissemination at presentation (four patients). Local treatment consisted of surgical resection for all patients, with a total excision accomplished in 42 patients and a subtotal resection in the other six patients. Margins were evaluated both grossly and microscopically in six dimensions (superior, inferior, medial, lateral, anterior, and posterior). Margins were categorized as clear, microscopically positive, or grossly positive. A clear margin indicated that there was no tumor  1 mm from the edge of the inked specimen; a microscopically positive margin indicated microscopically discernible extension of tumor to within less than 1 mm of the edge of the inked specimen. Thirty patients (62.5%) had clear margins, six (12.5%) had microscopically positive margins, and 12 (25%) had grossly positive margins.
Immunohistochemical analyses revealed strong and diffuse KIT (CD117) expression in each of the 48 GISTs, whereas CD34 was expressed in 38 cases. A minority of the tumors expressed smooth muscle actin, which was generally focal (n = 19), and one GIST expressed focal desmin, whereas no GIST expressed S-100 protein. A combination of genomic and cDNA sequencing revealed mutations in 44 of 48 GISTs. The mutations were located in exons 9, 11, 13, and 17, which encode aspects of the KIT extracellular region, the juxtamembrane domain, the first lobe of the split-kinase domain, and the phosphotransferase domain, respectively.15 A case by case listing of these mutations has been published recently.15 In brief, 34 GISTs (71%) had exon 11 mutations, six (13%) had exon 9 mutations, and two (4%) each had mutations in exons 13 or 17. The exon 11 mutations included missense mutations, in-frame deletions, and in-frame duplications. Four GISTs (8%) had no identifiable KIT sequence alteration. Thus, 44 (92%) of 48 GISTs in this series possessed demonstrable KIT mutations.
Recurrence-Free Survival Analysis
Tumor size, mitotic rate, histologic type, margin of resection, sex, and KIT mutation type was entered into a multivariate model to identify independent predictors of recurrence-free survival (Table 2). The following factors were significantly associated with recurrence by Cox regression analysis: mitotic rate of more than 15 mitoses per 30 HPF (hazard ratio [HR] = 18; P = .0001), mixed cytomorphology (HR = 21; P = .0001), presence of deletion/insertion exon 11 mutation (HR = 4; P = .005), and male sex (HR = 3; P = .05). The point exon 11 mutation type was not significant in the multivariate model, likely because of its association with both mitotic activity and histologic type. Eleven percent of patients with tumors containing point exon 11 mutations had mitotic activity of more than 15/30 HPF, compared with 49% of patients with tumors that had other mutation types. Furthermore, 11% of patients with tumors containing point exon 11 mutations had mixed cytomorphology, compared with 31% of patients with tumors that had other mutation types.
GISTs are the most common mesenchymal tumors of the gastrointestinal tract, yet they continue to pose a challenge with regard to prediction of their behavior as well as to controlling their spread within the peritoneal cavity. One of the reasons for the difficulty in predicting the clinical outcome for GIST patients has been the lack of consistency in the definition of GIST and in setting up strict criteria for inclusion. In this study, GIST is defined by the expression of KIT by immunohistochemical analysis, because KIT expression is diffuse and strong in virtually all GISTs.3,20 This definition has resulted in a more consistent classification of these tumors, which often have substantial variation in histologic appearance, differentiating toward an ICC phenotype. In this series, on the basis of intensive screening by genomic and cDNA sequencing, we detected oncogenic KIT mutations in 44 (92%) of 48 GISTs. Thus, KIT activation seems to play a central role in GIST pathogenesis, seemingly serving as a requisite for neoplastic behavior in the majority of GISTs.15 The overall recurrence-free survival in this study was 49% ± 8%, similar to the results in other published series of GISTs.21,22 In these studies, for those patients who underwent complete surgical resection, the size of the primary tumor was predictive of disease-specific survival. In our study, both the size and the margin of resection were prognostic for disease-specific survival on univariate analysis; however, on multivariate analysis, after adjusting for mitotic activity and histologic type, they were no longer independent predictors of disease-specific survival. In fact, mitotic activity of more than 15 mitoses per 30 HPF and mixed spindle/epithelioid cytomorphology were found to be the two most important independent clinicopathologic prognostic factors for recurrence-free survival in GISTs. The presence of a high mitotic count is generally accepted as the most reliable indicator of malignancy22-29; however, considerable controversy remains in determining what criteria should be used to distinguish benign from malignant disease. Furthermore, mitotic counts are observer and sample dependent because they are generally taken from the block that subjectively appeared the most cellular and mitotic. Thus, there is a real need to develop new more objective molecular prognostic factors that can reliably predict clinical outcome. The type of KIT mutation and the mechanism for activation of the tyrosine kinase of KIT may have important clinical implications with regard to predicting GIST behavior. Patients whose tumors contained KIT exon 11 missense mutations had a 5-year recurrence-free survival of 89% ± 10%, compared with a 5-year recurrence-free survival of 40% ± 8% for GISTs with other mutation types (P = .03). The group of patients with tumors that had no detectable KIT mutations had a 5-year recurrence-free survival of 100%, with one of the four patients having a late recurrence at 94 months. However, when compared with the 5-year recurrence-free survival of 43% ± 8% for patients with tumors with detectable KIT mutations, this difference did not reach statistical significance (P = .15). With larger patient numbers, the tendency to improved recurrence-free survival in patients with GISTs lacking detectable KIT mutations may reach statistical significance. In a recent series of GISTs that were analyzed for mutations in exon 11 and 17 by using DNA extracted from paraffin-embedded material, the mutation-positive GISTs had more frequent recurrences and a higher mortality compared with mutation-negative GISTs. The authors concluded that the presence of a KIT mutation was an independent prognostic factor for overall and cause-specific survival of the patients with GISTs. However, this series contained a significant number of so-called GISTs (11%) that did not express the KIT protein, and only 57% of patients had detectable mutations in the KIT gene. Furthermore, in contrast to this study, only exon 11 KIT mutation types were identified. Thus, it still remains to be determined whether the presence or absence of a GIST mutation is of prognostic importance. A multivariate model was made comparing recurrence-free survival for patients whose GISTs had different mutation types (point exon 11 mutation, deletion/insertion exon 11, exon 13, exon 17, no detectable KIT mutation) and for other known clinicopathologic predictors of GIST outcome. This multivariate analysis suggested that patients with GISTs containing deletion/insertion exon 11 mutations had a significantly reduced recurrence-free survival compared with those patients with GISTs that contained exon 11 missense mutations. These results suggest that the type of mutation is an independent predictor of GIST clinical behavior. The majority of GISTs contain one of various types of KIT mutations, and these mutations differ in their form (missense, deletion, or duplication) and in the involved protein domains. Recent work has demonstrated that activation of known KIT signaling pathways depends on the specific location and nature of the activating mutation (A. Duensing and J.A. Fletcher, personal communication, July 2002). These pathways have variable effects on cell proliferation, apoptosis, differentiation, and adhesion, and the mutation-associated, selective activation of such pathways undoubtedly influences GIST clinical behavior. It is also probable that the level of KIT activation, as judged by catalytic activity, varies depending on the mutation. We note that there is precedence for association between oncoprotein structure and clinical outcome. Examples include the different forms of SYT-SSX, EWS-FLI1, and BCR-ABL oncoproteins in synovial sarcoma, Ewing’s sarcoma, and leukemia, respectively.30-32 Our analyses, although requiring confirmation, suggest that KIT exon 11 missense mutations are found disproportionately in lower-grade, favorable-outcome GISTs. Patients with GISTs containing mutations involving deletion or duplication of multiple amino acids in exon 11 might have a worse clinical outcome compared with patients whose tumors contain KIT exon 11 missense mutations. Notably, cytogenetic correlations have revealed several potential tumor suppressor loci that are deleted during neoplastic progression in GISTs. Chromosomes 14 and 22 are deleted frequently, but at least four other chromosomal regions are deleted in a substantial minority of GISTs.33-35 Therefore, it is reasonable to expect that GIST clinical behavior is influenced by alterations of several different genes above and beyond the KIT mutations. The therapy of malignant GISTs with conventional chemotherapy and radiotherapy has been largely ineffective. The finding that KIT activation serves as a central oncogenic pathway in GISTs, together with the availability of the KIT/ABL/PDGFR kinase inhibitor, STI571, has enabled the development of targeted low-toxicity therapy based on the inhibition of KIT kinase function. In vitro studies indicate that STI571 inhibits oncogenic KIT activation and inhibits cell proliferation in human GIST cell lines.36 Preliminary data from a multicenter phase II trial of STI571 have demonstrated efficacy and minimal toxicity in patients with unresectable or metastatic GIST. The partial response rate in this study was 59%, with dramatic symptomatic improvement in those patients who responded to therapy.37 These promising early results validate the hypothesis that uncontrolled KIT tyrosine kinase activity is critical to GIST pathogenesis and that inhibition of KIT signal transduction is an important new low-toxicity therapy for the treatment of GIST. The survival correlations with KIT mutation types, as described in our study, were in patients who had not received treatment with ST1571. Notably, preliminary analyses of the STI571 GIST trials suggest that clinical response can be influenced by KIT mutation type. Therefore, the survival correlations reported herein might be different in patients who receive STI571. Various clinical quandaries have arisen because of the STI571 therapeutic successes, and these underscore the need for accurate predictors of GIST natural history in newly diagnosed patients. For example, our data suggest that KIT mutation type is of value, along with conventional clinicopathologic prognosticators, in predicting the risk of recurrence in patients after complete resection of GIST. We propose that the adjuvant use of KIT kinase inhibitors, eg, STI571, would ideally be considered in light of the known clinical associations for the KIT mutation in each patient’s GIST. The most appropriate candidates for adjuvant therapy would undoubtedly include patients whose GISTs have clinicopathologic and KIT mutation features indicating a high risk for recurrence and those in whom the KIT mutant oncoprotein is known to be STI571 sensitive. In summary, we have shown that activating mutations of KIT are found in the majority of GISTs and that KIT mutational status is an independent prognostic factor for recurrence-free survival in patients with GIST. It remains to be determined whether the various KIT mutations differ quantitatively or qualitatively with respect to autophosphorylation or phosphorylation/activation of downstream signaling pathways. However, that possibility is supported by the clinical correlations in this study and in others, which indicate a relationship between the specific type of KIT mutation and clinical outcome. Future studies with larger patient numbers will be essential to confirm the prognostic importance of KIT mutation type and site in patients with GIST.
1. Sakurai S, Fukasawa T, Chong JM, et al: C-kit gene abnormalities in gastrointestinal stromal tumors (tumors of interstitial cells of Cajal). Jpn J Cancer Res 90: 1321-1328, 1999[CrossRef][Medline]
2. Robinson TL, Sircar K, Hewlett BR, et al: Gastrointestinal stromal tumors may originate from a subset of CD34-positive interstitial cells of Cajal. Am J Pathol 156: 1157-1163, 2000 3. Kindblom LG, Remotti HE, Aldenborg F, et al: Gastrointestinal pacemaker cell tumor (GIPACT): Gastrointestinal stromal tumors show phenotypic characteristics of the interstitial cells of Cajal. Am J Pathol 152: 1259-1269, 1998[Abstract] 4. Huizinga JD, Thuneberg L, Kluppel M, et al: W/kit gene required for interstitial cells of Cajal and for intestinal pacemaker activity. Nature 373: 347-349, 1995[CrossRef][Medline] 5. Kluppel M, Huizinga JD, Malysz J, et al: Developmental origin and Kit-dependent development of the interstitial cells of cajal in the mammalian small intestine. Dev Dyn 211: 60-71, 1998[CrossRef][Medline] 6. Torihashi S, Nishi K, Tokutomi Y, et al: Blockade of kit signaling induces transdifferentiation of interstitial cells of cajal to a smooth muscle phenotype. Gastroenterology 117: 140-148, 1999[CrossRef][Medline] 7. Fletcher CD, Berman J, Corless C, et al: Diagnosis of gastrointestinal stromal tumors: A consensus approach. Hum Pathol 33: 459-465, 2002[CrossRef][Medline] 8. Fleischman RA: From white spots to stem cells: The role of the Kit receptor in mammalian development. Trends Genet 9: 285-290, 1993[CrossRef][Medline]
9. Hirota S, Isozaki K, Moriyama Y, et al: Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 279: 577-580, 1998 10. Ernst SI, Hubbs AE, Przygodzki RM, et al: KIT mutation portends poor prognosis in gastrointestinal stromal/smooth muscle tumors. Lab Invest 78: 1633-1636, 1998[Medline] 11. Moskaluk CA, Tian Q, Marshall CR, et al: Mutations of c-kit JM domain are found in a minority of human gastrointestinal stromal tumors. Oncogene 18: 1897-1902, 1999[CrossRef][Medline]
12. Lasota J, Jasinski M, Sarlomo-Rikala M, et al: Mutations in exon 11 of c-Kit occur preferentially in malignant versus benign gastrointestinal stromal tumors and do not occur in leiomyomas or leiomyosarcomas. Am J Pathol 154: 53-60, 1999
13. Taniguchi M, Nishida T, Hirota S, et al: Effect of c-kit mutation on prognosis of gastrointestinal stromal tumors. Cancer Res 59: 4297-4300, 1999 14. Newman PL, Wadden C, Fletcher CD: Gastrointestinal stromal tumours: Correlation of immunophenotype with clinicopathological features. J Pathol 164: 107-117, 1991[CrossRef][Medline]
15. Rubin BP, Singer S, Tsao C, et al: KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res 61: 8118-8121, 2001 16. Kaplan E, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53: 457-481, 1958[CrossRef] 17. Armitage P, Berry G: Statistical Methods in Medical Research, ed 2 . Oxford, UK, Blackwell, 1987 18. Peto R, Peto J: Asymptomatically efficient rank invariant test procedures. J Roy Stat 135: 185-206, 1972 19. Cox D: Regression models and life tables. J R Stat Soc 34: 187-220, 1972 20. Miettinen M, Sobin LH, Sarlomo-Rikala M: Immunohistochemical spectrum of GISTs at different sites and their differential diagnosis with a reference to CD117 (KIT). Mod Pathol 13: 1134-1142, 2000[CrossRef][Medline] 21. DeMatteo RP, Lewis JJ, Leung D, et al: Two hundred gastrointestinal stromal tumors: Recurrence patterns and prognostic factors for survival. Ann Surg 231: 51-58, 2000[CrossRef][Medline] 22. Emory TS, Sobin LH, Lukes L, et al: Prognosis of gastrointestinal smooth-muscle (stromal) tumors: Dependence on anatomic site. Am J Surg Pathol 23: 82-87, 1999[CrossRef][Medline] 23. Carrillo R, Candia A, Rodriguez-Peralto JL, et al: Prognostic significance of DNA ploidy and proliferative index (MIB-1 index) in gastrointestinal stromal tumors. Hum Pathol 28: 160-165, 1997[CrossRef][Medline] 24. Cooper PN, Quirke P, Hardy GJ, et al: A flow cytometric, clinical, and histological study of stromal neoplasms of the gastrointestinal tract. Am J Surg Pathol 16: 163-170, 1992[Medline] 25. el-Naggar AK, Ro JY, McLemore D, et al: Gastrointestinal stromal tumors: DNA flow-cytometric study of 58 patients with at least five years of follow-up. Mod Pathol 2: 511-515, 1989[Medline] 26. Franquemont DW, Frierson HF Jr: Proliferating cell nuclear antigen immunoreactivity and prognosis of gastrointestinal stromal tumors. Mod Pathol 8: 473-477, 1995[Medline] 27. Kim CJ, Day S, Yeh KA: Gastrointestinal stromal tumors: Analysis of clinical and pathologic factors. Am Surg 67: 135-137, 2001[Medline]
28. Panizo-Santos A, Sola I, Vega F, et al: Predicting metastatic risk of gastrointestinal stromal tumors: Role of cell proliferation and cell cycle regulatory proteins. Int J Surg Pathol 8: 133-144, 2000 29. Rudolph P, Gloeckner K, Parwaresch R, et al: Immunophenotype, proliferation, DNA ploidy, and biological behavior of gastrointestinal stromal tumors: A multivariate clinicopathologic study. Hum Pathol 29: 791-800, 1998[CrossRef][Medline]
30. Kawai A, Woodruff J, Healey JH, et al: SYT-SSX gene fusion as a determinant of morphology and prognosis in synovial sarcoma. N Engl J Med 338: 153-160, 1998
31. de Alava E, Kawai A, Healey JH, et al: EWS-FLI1 fusion transcript structure is an independent determinant of prognosis in Ewing’s sarcoma. J Clin Oncol 16: 1248-1255, 1998 32. Hermans A, Heisterkamp N, von Linden M, et al: Unique fusion of bcr and c-abl genes in Philadelphia chromosome positive acute lymphoblastic leukemia. Cell 51: 33-40, 1987[CrossRef][Medline]
33. el Rifai W, Sarlomo-Rikala M, Miettinen M, et al: DNA copy number losses in chromosome 14: An early change in gastrointestinal stromal tumors. Cancer Res 56: 3230-3233, 1996 34. el Rifai W, Sarlomo-Rikala M, Andersson LC, et al: DNA copy number changes in gastrointestinal stromal tumors: A distinct genetic entity. Ann Chir Gynaecol 87: 287-290, 1998[Medline]
35. el Rifai W, Sarlomo-Rikala M, Andersson LC, et al: DNA sequence copy number changes in gastrointestinal stromal tumors: Tumor progression and prognostic significance. Cancer Res 60: 3899-3903, 2000 36. Tuveson DA, Willis NA, Jacks T, et al: STI571 inacti-vation of the gastrointestinal stromal tumor c-KIT oncopro-tein: Biological and clinical implications. Oncogene 20: 5054-5058, 2001[CrossRef][Medline] 37. Blanke CD, von Mehren M, Joensuu H, et al: Evaluation of the safety and efficacy of an oral molecularly targeted therapy, STI571, in patients with unresectable or metastatic gastrointestinal stromal tumors expressing c-kit. Proc Am Soc Clin Oncol 20: 1a, 2001 (abstr 1) Submitted March 19, 2002; accepted May 30, 2002. This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
15 mitoses per 30 HPF, and more than 15 mitoses per 30 HPF were 89% ± 7%, 49% ± 12%, and 16% ± 6%, respectively (P = .0001). The 32 patients with spindle-cell histology had a 49% ± 7% 5-year recurrence-free survival; in contrast, the 16 patients with epithelioid or mixed histology had a 23% ± 11% 5-year recurrence-free survival (P = .01). Five-year recurrence-free survival for patients with tumors less than 5 cm, 5 to 10 cm, and more than 10 cm were 82% ± 12%, 45% ± 9%, and 27% ± 8%, respectively (P = .03). Prognostic associations were found with particular KIT mutation types, and patients with missense exon 11 mutations had a 5-year recurrence-free survival of 89% ± 11% compared with 40% ± 8% for GISTs with other mutation types (P = .03). The independent predictors for disease-free survival were the presence of deletion/insertion exon 11 mutations (hazard ratio [HR] = 4; P = .006), more than 15 mitoses per 30 HPF (HR = 18; P = .0001), mixed histology (HR = 21; P = .0001), and male sex (HR = 3; P = .05). 
2 = 53.2)
