original article introduction article

Transcription

1 original article Annals of Oncology 21: , 21 doi:1.193/annonc/mdq47 Published online 15 March 21 Prognostic time dependence of deletions affecting codons 557 and/or 558 of KIT gene for relapse-free survival (RFS) in localized GIST: a Spanish Group for Sarcoma Research (GEIS) Study J. Martin-Broto 1 *, A. Gutierrez 1, X. Garcia-del-Muro 2, J. A. Lopez-Guerrero 3, J. Martinez- Trufero 4, L. M. de Sande 5, N. Lainez 6, J. Maurel 7, A. De Juan 8, F. Losa 9, R. Andres 1, A. Casado 11, P. G. Tejido 12, R. Blanco 13, J. Carles 14, J. Bellmunt 15, A. Gomez-España 16, R. Ramos 17, J. Martinez-Serra 18, A. Llombart-Bosch 19 & A. Poveda 2 1 Departments of Oncology and Hematology, Medical and Translational Hemato-Oncology Group, Hospital Universitario Son Dureta, Palma de Mallorca; 2 Department of Oncology, Institut Catala` d Oncologia, Barcelona; 3 Laboratory of Molecular Biology, Fundacio n Instituto Valenciano de Oncologıá, Valencia; 4 Department of Oncology, Hospital Miguel Servet, Zaragoza; 5 Department of Oncology, Complejo Hospitalario de Leo n, Leon; 6 Department of Oncology, Hospital Virgen del Camino, Pamplona; 7 Department of Oncology, Hospital Clıńic, Institut d Investigacions Biome`diques August Pi i Sunyer, Centro de Investigacio n Biome dica de Enfermedades Hepa ticas y Digestivas, Barcelona; 8 Oncology Department, Hospital de Valdecilla, Santander; 9 Department of Oncology, Hospital General, Hospitalet de Llobregat; 1 Department of Oncology, Hospital Clıńico Lozano Blesa, Zaragoza; 11 Department of Oncology, Hospital Universitario San Carlos, Madrid; 12 Department of Oncology, Hospital San Pedro, Logron o; 13 Department of Oncology, Hospital Mutua de Terrassa, Terrassa; 14 Department of Oncology, Hospital del Mar, Barcelona; 15 Department of Oncology, Hospital Vall d Hebron, Barcelona; 16 Department of Oncology, Hospital Reina Sofia, Co rdoba; 17 Department of Pathology, Hospital Universitario Son Dureta, Palma de Mallorca; 18 Laboratory of Molecular Biology, Hospital Universitario Son Dureta, Palma de Mallorca; 19 Department of Pathology, Universidad de Valencia, Valencia and 2 Department of Oncology, Fundacio n Instituto Valenciano de Oncologıá, Valencia, Spain original article Received 5 December 29; revised 1 February 21; accepted 2 February 21 Background: To assess whether deletions involving codons 557 and/or 558 (critical deletions) of exon 11 of KIT are relevant in the prognosis of relapse-free survival (RFS) in gastrointestinal stromal tumor (GIST) patients with a long follow-up. Patients and methods: A univariate and multivariate analysis for RFS were carried out on 162 localized GIST patients over the entire follow-up period and over the intervals 4 years and >4 years. Factors assessed among others were Fletcher/National Institutes of Health and Miettinen Lasota/Armed Forces Institute of Pathology (M-L/ AFIP) risk categories, critical deletions and non-deletion-type mutation (NDTM) within exon 11 of KIT. Results: Multivariate analyses revealed that M-L/AFIP [relative risk (RR) 11.45, confidence interval (CI) , for the high-risk subgroup and RR 5.97, CI , for the intermediate subgroup] and critical deletions (RR 3.5, CI ) were independent prognostic factors for RFS for the first 4 years and for the entire follow-up period. Beyond 4 years, the high-risk M-L/AFIP subgroup (RR 8.12, CI ) and NDTM (RR 6.42, CI ) were independent prognostic factors for RFS. The median follow-up was 84 months. Conclusion: Critical deletions represent a time-dependent prognostic factor limited to the first 4 years after surgery, which could help identify a subset with higher and earlier risk for relapse in GIST patients. Key words: KIT mutation, localized GIST, prognostic factors, relapse-free survival, time dependence introduction The recognition of the oncogenic role of certain mutations within the KIT or PDGFR-a genes in most gastrointestinal stromal tumor (GIST) patients and the advent of effective target-based therapies to treat this disease, such as imatinib and *Correspondence to: Dr J. Martin Broto, Medical Oncology Department, Medical and Translational Hemato-Oncology Group, Son Dureta Hospital, Andrea Doria 55, 714 Palma de Mallorca, Spain. Tel: ; Fax: ; javier.martin@ssib.es other tyrosine kinase inhibitors, have led to consider GIST as an attractive model for the studies of several specialists including surgeons, radiologists, pathologists, molecular biologists and oncologists [1]. Mitotic count and size are the most recognized prognostic factors in localized GIST [2, 3]. Indeed, a risk stratification index of primary tumors was developed on the basis of these parameters and published in 22 by Fletcher et al. [Fletcher/ National Institutes of Health (F/NIH] [4]. More recently, new risk category groups were proposed by Miettinen and Lasota ª The Author 21. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please journals.permissions@oxfordjournals.org

2 Annals of Oncology [5], from Armed Forces Institute of Pathology (AFIP), on the basis of the long-term follow-up of >1 patients (M-L/AFIP) and incorporating tumor location as an important factor. Surgery is the mainstay treatment in localized disease. However, even when microscopically free surgical margins are obtained (R), recurrence ranges from 2% 55% [6 1]. Treatment with imatinib has dramatically changed expectations on response rates, progression-free survival and overall survival in metastatic patients [11, 12]. In recent years, there has been special interest in exploring prognostic factors at the molecular level for localized GIST and thus, some papers have assessed the potential relevance of mutations on prognosis [13 17]. Our group found, in a series of 162 localized GIST patients, that deletions involving codons (DEL ) within exon 11 of the KIT gene had an independent prognostic value for relapse-free survival (RFS) [18]. We planned three main objectives in the current study: to assess whether DEL still has prognostic relevance after a longer follow-up period, to evaluate whether DEL still serves as an independent prognostic factor for RFS in conjunction with the new M-L/AFIP risk categories, and finally, to explore whether there is a time-dependent prognostic variable for RFS in GIST patients. patients and methods patients and study design The selection of a final subset of 162 patients after a clinical and pathological review (paraffin blocks) of 357 cases has been explained previously [18]. The current update was carried out through a query-based task applied to each patient at risk of recurrence, and it involved data managers and clinical oncologists. The patients have been followed through clinical examination and abdominal computed tomography on a regular basis since 21. clinical and mutational prognostic variables The clinical and pathological variables considered in the RFS analysis included the F/NIH risk classification and M-L/AFIP risk classification (Table 1). To increase the sensitivity of the mutational analysis, those cases in which no mutation was detected by direct sequencing were then screened by denaturing high-performance liquid chromatography (DHPLC) in a Transgenomic WAVE DNA Fragment Analysis System (Transgenomic, Inc., Omaha, NE). Mutations were studied within the following groups: deletions involving codons 557 and/or 558 (henceforth referred to as critical mutations), presence of KIT gene mutations, deletions within exon 11 of the KIT gene and those cases with non-deletion-type mutation within exon 11 of the KIT gene (missense and duplication-type mutations). This latter aggregation was chosen for comparison and was justified as each type of mutation had a tendency to cluster. statistical methods The RFS was measured from the date of surgery and it was estimated according to the Kaplan Meier method [19]. A multivariate analysis was carried out using the Cox proportional hazards model [2]. Survival analysis was carried out over the entire follow-up period, over the interval covering the first 4 years and for that from 4 years after surgery. In this way, the time-dependence weighting of the variables could be evaluated. results clinical data and the distribution of mutations in the patients Clinical and pathological characteristics are depicted in Table 1. Four novel mutations were found by the DHPLC screening resulting in a total of mutated cases. survival analyses for the entire time period The median follow-up for the patients that remained free of recurrence after the current update was 84 months, with an actuarial RFS rate at 7 years of 68%. There were 48 events of recurrence, and 16 patients died by causes unrelated to GIST and without evidence of relapse. Table 2 depicts in detail the outcome of the univariate analyses. Several aggregations of mutations were analyzed and those patients harboring a critical mutation displayed the worst 7-year RFS (34%; Table 2 and Figure 1). time-dependence survival analyses We chose the 4-year time cut-off as it marked clearer differences with a sufficiently large follow-up including at least one-third of the main relevant original mutational subsets. This defined two different time periods: the - to 4-year aftersurgery interval (41 recurrences in 162 patients at risk) and >4 years (7 recurrences in 15 patients at risk). Considering the 15 patients at risk beyond 4 years, the median follow-up of the nonrelapsing patients was 49 months (4 126). Nine of the initial 25 (36%) patients with critical mutations and 23 of the initial 35 (66%) patients with non-deletion-type mutations were at risk of recurrence after 4 years from surgery. Median follow-up beyond 4 years were 5 months (21 98) and 46 months (22 94) for critical and non-deletion-type mutations within exon 11, respectively. Considering the molecular variables, critical mutations appeared to be significant and strong predictors of RFS, either for the entire follow-up period or for the - to 4-year interval. Nevertheless, it was noteworthy that critical mutations were no longer of prognostic relevance beyond the 4-year time interval. Conversely, the patients carrying non-deletion-type mutations within exon 11 still developed recurrences beyond 3 or 4 years after surgery (Table 2 and Figure 1). Factors independently associated with recurrence for the entire follow-up period were identical to those seen in the first 4 years after surgery, i.e. the M-L/AFIP risk categories (intermediate and high) and critical deletions. Taking into account the multivariate analysis for the time interval beyond 4 years after surgery, only the high-risk M-L/AFIP category and duplication or missense mutation type within exon 11 of the KIT gene were independent prognostic factors for RFS (Table 3). discussion original article The M-L/AFIP risk criteria are powerful and they strongly discriminate three different subsets of patients according to their probability of recurrence, as demonstrated here and by others [21]. Volume 21 No. 7 July 21 doi:1.193/annonc/mdq

3 original article Annals of Oncology Table 1. Clinical and pathological characteristics of patients according to the presence of critical mutations or the presence of non-deletion-type mutations within exon 11 Characteristic All patients, n (%) Del557 and/or Del558, n (%) P Exon 11 nondeletion, n (%) P Yes No Yes No Number Sex NS NS Male 82 (51) 11 (44) 71 (52) 17 (49) 65 (51) Female (49) 14 (56) 66 (48) 18 (51 62 (49) Age NS NS 59 37% 44% 36% 16 (49%) 44 (35%) 62% 56% 64% 17 (51%) 83 (65%) Histological phenotype NS NS Spindle type 122 (75) 18 (75) (75) 27 (77) 95 (75) Epithelioid type 2 (12) 2 (8) 18 (13) 5 (14) 15 (12) Mixed type 2 (12) 4 (17) 16 (12) 3 (9) 17 (13) Cellularity. NS High 38 (23) 1 () 28 (21) 6 (17) 32 (26) Moderate paucicellular 124 (77) 15 () 16 (79) 29 (83) 92 (74) Tumor location NS NS Gastric 93 (57) 17 (68) 82 () 19 (54) (63) Nongastric 69 (43) 8 (32) 55 () 16 (46) 47 (37) KIT IHQ pattern NS NS Diffuse 137 (85) 22 (88) 115 (84) 31 (89) 16 (83) Golgi 9 (6) () 9 (7) 3 (9) 6 (5) Mixed 16 (9) 3 (14) 13 (9) 1 (3) 15 (12) PDGFR IHQ pattern NS NS Positive 142 (88) 21 (84) 116 (85) 27 (82) 11 (89) Negative 2 (12) 4 (16) 16 (12) 6 (18) 14 (11) Mitosis (per 5 HPF).25 NS <5 96 (59) 13 (52) 83 (61) 21 () 75 (59) (12) () 19 (14) 3 (9) 16 (13) >1 47 (29) 12 (48) 35 (25) 11 (31) 36 (28) Size NS NS 2 5 cm 89 (55) 12 (48) 77 (57) 17 (48) 72 (57) 5 1 cm 35 (22) 6 (24) 29 (21) 8 (23) 27 (21) >1 cm 17 (23) 7 (28) 3 (22) 1 (29) 27 (21) F/NIH NS NS Low 64 (39) 7 (28) 57 (42) 11 (31) 52 (41) Intermediate 27 (17) 3 (12) 24 (17) 7 (2) 2 (16) High 71 (44) 15 () 56 (41) 17 (49) 54 (43) M L/AFIP NS NS Very low and low 64 (39) 8 (32) 56 (41) 1 (29) 54 (42) Intermediate 37 (23) 6 (24) 31 (23) 12 (34) 25 (2) High 61 (38) 11 (44) 5 (36) 13 (37) 48 (38) NS, non significant; F/NIH, Fletcher/National Institutes of Health; IHQ, immunohistochemical; HPF, high-power field; M-L/AFIP, Miettinen Lasota/Armed Forces Institute of Pathology. Our results indicate the independent statistical value of critical mutations with a median follow-up of 7 years for localized GIST patients, the longest follow-up published to date that analyzed this molecular variable. Furthermore, we have demonstrated the independent significance of such mutations after confronting this factor with the current best risk classification criteria for locally resected GIST patients: the M-L/AFIP risk categories. Wardelmann et al. [22] found a significant correlation between metastases and the presence of critical mutations. Similarly, Iesalnieks et al. [23] detected an independent prognostic value of such critical deletions, in a multivariate analysis on a series of 35 patients in whom GIST was completely removed. More recently, Tzen, in a series of 11 GIST patients, concluded that there were no significant prognostic factors that involved the presence of any type of mutation. However, the analysis was of progression-free survival rather than RFS due to the inclusion of 43 patients (42% of cases) that either were unresectable or had metastatic tumors [24]. DeMatteo, in another recent article of 127 GIST patients submitted to complete gross surgical resection and followed up for a median of 5.2 years, found that a significant worse prognosis was associated with those patients carrying deletions involving codons 557 or 558. This result did not reach an independent value in the multivariate model despite 1554 Martin-Broto et al. Volume 21 No. 7 July 21

4 Annals of Oncology original article Table 2. Univariate analysis of prognostic factors for RFS in the entire follow-up period, the first 4-year interval and beyond the 4 years after surgery Variable All intervals, % (range) Initial period ( 4 years), % (range) >4 years, % (range) 7-year RFS P RFS P RFS P Sex Male 67 (55 78)) 7 (59 ) 86 (72 ) Female 72 (61 82) 75 (65 84) 95 (89 ) Age (5 75) 68 (56 ) 91 (82 ) 73 (64 82) 75 (66 84) 91 (82 ) Size < cm 83 (75 91) 84 (77 92) 94 (86 ) 5 1 cm 67 (51 83) 7 (54 85) 95 (87 ) >1 cm 41 (24 58) 49 (32 65) 71 (42 ) Mitosis <.1 <.1 < ( 94) 87 ( 94) (53 96) 74 (53 96) >1 33 (18 47) 44 (29 58) 51 (2 82) Fletcher/NIH <.1 <.1.6 Low 95 (89 ) 95 (89 ) Intermediate (64 96) (64 96) High 44 (32 57) 52 ( 64) 71 (5 92) M-L/AFIP <.1 <.1.15 Very low and low 95 (9 ) 95 (9 ) Intermediate 73 (58 88) 77 (63 9) 86 (68 ) High (27 53) 46 (33 59) 78 (58 98) Localization Gastric 76 (67 85) (72 88) 91 (82 ) Nongastric 58 (46 71) 61 (48 73) 91 (79 ) Mutational status No mutation 79 (69 9) 83 (74 93) 98 (85 ) Any mutation 63 (52 73) 66 (56 75) 87 (76 98) KIT Mutated (49 71) 63 (52 74) 86 (72 ) Wild type 79 (69 88) 82 (74 91) 95 (89 ) Exon 11 deletion Yes 57 (41 72) 57 (42 72) No 74 (65 82) 78 (71 86) 89 ( 98) Del 557 and/or 558 <.1 <.1.37 Yes 34 (15 53) 34 (15 53) No 76 (68 83) (73 87) 82 (65 99) Exon 11 nondeletion Yes 66 (5 83) 73 (58 88) 72 (47 97) No 7 (62 78) 72 (64 ) 97 (92 ) RFS, relapse-free survival; M-L/AFIP, Miettinen Lasota/Armed Forces Institute of Pathology; F/NIH, Fletcher/National Institutes of Health. showing a markedly poorer RFS curve, probably due to a high number of rectal cases included [25]. To our knowledge, few studies have addressed the time dependence of prognostic factors in sarcomas and none in the context of GIST [26 28]. The present analysis demonstrates that the prognostic significance of critical mutations is time dependent. Thus, their predictive value for recurrence is limited to the first 4 years after surgery and beyond 4 years, no patient harboring such mutations suffered a recurrence. Conversely, patients harboring non-deletion-type mutations within exon 11 had an adverse prognostic effect beyond the 4 years from surgery by maintaining their rate of relapses. Moreover, the high-risk category according to the M-L/AFIB criteria kept its negative impact value with the same relative risk (8.12) as that compared with the first 4-year time period (8.94). Two further observations are worthy of consideration: on the one hand, tumors with DEL557 and/or 558 mutations showed no primary resistance in a series of 22 cases treated with imatinib [29], while additionally, the frequency of DEL557 and/ or 558 mutations in localized GIST series is not inconsiderable, ranging from 15% % of the cases [18, 22, 25]. In conclusion, our findings could contribute to identify a subset of GIST patients with higher and earlier risk of relapse and irrespective of clinical prognostic factors. This would be the most suitable target population in which adjuvant tyrosine kinase inhibitors should be assessed to look for a significant impact on RFS. Volume 21 No. 7 July 21 doi:1.193/annonc/mdq

5 original article Annals of Oncology Low risk: 95% (89-) Very low & low risk: 95% (9-) Intermediate risk: % (64-96) High risk: 44% (32-57) Intermediate risk: 73% (58-88) High risk: % (27-53) Fletcher/NIH P <.1 M-L/AFIB P < Months form surgery Months from surgery No Critical 557 &/or 558 deletions: 76% Any other: 7% Critical 557 &/or 558 deletions: 34% Exon11 non-deletion: 66% P <.1 P < Months from surgery Months from surgery No Critical 557 &/or 558 deletions: % Exon11 non-deletion: 73% Critical 557 &/or 558 deletions: 34% Any other: 72% P < Months from surgery (-4 years time-period) P < Months from surgery (-4 years time-period) Critical 557 &/or 558 deletions: % Any other: 97% No Critical 557 &/or 558 deletions: 82% P <.37 Exon11 non-deletion: 72% P = Months beyond 4 years of follow-up Months beyond 4 years of follow-up Figure 1. Kaplan Meier curves for main prognostic scores and mutations considering the entire follow-up time, the first 4 years after surgery and beyond the 4 years after surgery Martin-Broto et al. Volume 21 No. 7 July 21

6 Annals of Oncology Table 3. Time-dependent relapses rates (A) and multivariate analysis (B) A Relapse, n (%) Mutation (exon 11) Global period 4 years >4 years Critical mutations (n = 25) 15 (%) 15/25 (%) /9 (%) Non-deletion type (n = 35) 14 (%) 9/35 (26%) 5/23 (22%) B Relative risk 95% CI P Entire time period M-L/AFIP risk categories Intermediate High <.1 DEL 557 and/or to 4-year time period M-L/AFIP risk categories Intermediate High <.1 DEL 557 and/or <.1 >4-year time period M-L/AFIP risk categories High Exon 11 non-deletion type CI, confidence interval; M-L/AFIP, Miettinen Lasota/Armed Forces Institute of Pathology. acknowledgements Andrés Meana (Hospital General de Alicante), Isabel Sevilla (Hospital Clínico de Málaga), Inmaculada Maestu (Hospital Virgen de los Lirios), Vicente Valentí (Hospital Sant Joan de Reus), Carmen Balañà (Hospital Germans Trias I Pujol), José Luis Carrasco (Hospital Xeral-Cíes), Ricardo Cubedo (Hospital Puerta de Hierro), Javier Cassinello (Hospital de Guadalajara), Fernando Molano (Hospital Insular Gran Canaria), Jordi Rubió (Institut Català d Oncologia Girona), Isabel Bover (Hospital Son Llàtzer), Fina Cruz (Hospital Universitario de Canarias), and Antonio López-Pousa (Hospital Sant Pau) also contributed to this work. We thank the pathologists associated with GEIS for providing us with the tissues. Finally, thanks to Marta Fernández for data management. references 1. Demetri GD. OncoloGIST, BioloGIST, RadioloGIST: the big impact on the field of oncology of a molecularly-targeted therapy designed to treat a rare disease. Eur J Cancer 23; 39(14): Emory TS, Sobin LH, Lukes L et al. Prognosis of gastrointestinal smooth-muscle (stromal) tumors: dependence on anatomic site. Am J Surg Pathol 1999; 23(1): Aparicio T, Boige V, Sabourin JC et al. Prognostic factors after surgery of primary resectable gastrointestinal stromal tumours. Eur J Surg Oncol 24; 3(1): Fletcher CD, Berman JJ, Corless C et al. Diagnosis of gastrointestinal stromal tumors: a consensus approach. Hum Pathol 22; 33(5): Miettinen M, Lasota J. Gastrointestinal stromal tumors: pathology and prognosis at different sites. Semin Diagn Pathol 26; 23(2): DeMatteo RP, Lewis JJ, Leung D et al. Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival. Ann Surg 2; 231(1): original article 7. Lin SC, Huang MJ, Zeng CY et al. Clinical manifestations and prognostic factors in patients with gastrointestinal stromal tumors. World J Gastroenterol 23; 9(12): Nilsson B, Bumming 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 25; 13(4): Ahmed I, Welch NT, Parsons SL. Gastrointestinal stromal tumours (GIST) 17 years experience from Mid Trent Region (United Kingdom). Eur J Surg Oncol 28; 34(4): Hassan I, You YN, Shyyan R et al. Surgically managed gastrointestinal stromal tumors: a comparative and prognostic analysis. Ann Surg Oncol 28; 15(1): Demetri GD, von Mehren M, Blanke CD et al. Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors. N Engl J Med 22; 347(7): Verweij J, Casali PG, Zalcberg J et al. Progression-free survival in gastrointestinal stromal tumours with high-dose imatinib: randomised trial. Lancet 24; 364(94): Ernst SI, Hubbs AE, Przygodzki RM et al. KIT mutation portends poor prognosis in gastrointestinal stromal/smooth muscle tumors. Lab Invest 1998; 78(12): Lasota J, Jasinski M, Sarlomo-Rikala M, Miettinen M. 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 1999; 154(1): Taniguchi M, Nishida T, Hirota S et al. Effect of c-kit mutation on prognosis of gastrointestinal stromal tumors. Cancer Res 1999; 59(17): Singer S, Rubin BP, Lux ML et al. Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol 22; 2(18): Kim TW, Lee H, Kang YK et al. Prognostic significance of c-kit mutation in localized gastrointestinal stromal tumors. Clin Cancer Res 24; 1(9): Martin J, Poveda A, Llombart-Bosch A et al. Deletions affecting codons of the c-kit gene indicate a poor prognosis in patients with completely resected gastrointestinal stromal tumors: a study by the Spanish Group for Sarcoma Research (GEIS). J Clin Oncol 25; 23(25): Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: Cox DR. Regression models and life-tables. J R Stat Soc 1972; 34: Goh BK, Chow PK, Yap WM et al. Which is the optimal risk stratification system for surgically treated localized primary GIST? Comparison of three contemporary prognostic criteria in 171 tumors and a proposal for a modified Armed Forces Institute of Pathology risk criteria. Ann Surg Oncol 28; 15(8): Wardelmann E, Losen I, Hans V et al. Deletion of Trp-557 and Lys-558 in the juxtamembrane domain of the c-kit protooncogene is associated with metastatic behavior of gastrointestinal stromal tumors. Int J Cancer 23; 16(6): Iesalnieks I, Rummele P, Dietmaier W et al. Factors associated with disease progression in patients with gastrointestinal stromal tumors in the pre-imatinib era. Am J Clin Pathol 25; 124(5): Tzen CY, Wang MN, Mau BL. Spectrum and prognostication of KIT and PDGFRA mutation in gastrointestinal stromal tumors. Eur J Surg Oncol 28; 34(5): Dematteo RP, Gold JS, Saran L et al. Tumor mitotic rate, size, and location independently predict recurrence after resection of primary gastrointestinal stromal tumor (GIST). Cancer 28; 112(3): Lewis JJ, Leung D, Casper ES et al. Multifactorial analysis of long-term follow-up (more than 5 years) of primary extremity sarcoma. Arch Surg 1999; 134(2): Stojadinovic A, Leung DH, Allen P et al. Primary adult soft tissue sarcoma: time-dependent influence of prognostic variables. J Clin Oncol 22; 2(21): Engellau J, Anderson H, Rydholm A et al. Time dependence of prognostic factors for patients with soft tissue sarcoma: a Scandinavian Sarcoma Group Study of 338 malignant fibrous histiocytomas. Cancer 24; (1): Bachet J-B, Hostein I, Le Cesne A et al. Prognosis and predictive value of KIT exon 11 deletion in GISTs. Br J Cancer 29; 11(1): Volume 21 No. 7 July 21 doi:1.193/annonc/mdq