Med Oncol (2014) 31:819 DOI 10.1007/s12032-013-0819-x
ORIGINAL PAPER
Prognostic value of mutational characteristics in gastrointestinal stromal tumors: a single-center experience in 275 cases Ming Wang • Jia Xu • Wenyi Zhao • Lin Tu • Weiqing Qiu • Chaojie Wang Yangyin Shen • Qiang Liu • Hui Cao
•
Received: 22 November 2013 / Accepted: 7 December 2013 / Published online: 14 December 2013 Ó Springer Science+Business Media New York 2013
Abstract The objective of this study was to investigate the impact of KIT/PDGFRA mutations on the prognosis of gastrointestinal stromal tumors (GISTs). In the present study, genotype analyses were performed based on GIST samples from 275 patients. The relationship between mutation and clinicopathological characteristics were explored. All factors were evaluated for their impacts on relapse-free survival (RFS). Briefly, the results of genotype analyses showed that mutations were identified in 258 (93.8 %) patients, and deletion was the most frequent type of mutation accounting for 47.3 % (122/258) of all mutation cases, followed by substitution (87/258, 33.7 %) and duplication (49/258, 19.0 %). Moreover, for KIT exon 11 mutation, the most frequently involved area was from codon 557 to 560. Deep analyses showed that the mutation types were correlated with tumor location (P = 0.005), tumor size (P = 0.022), mitosis rate (P \ 0.001), risk grade (P \ 0.001), and relapse (P = 0.004). Furthermore, delW557-K558 correlated with mitosis rate (P = 0.042) and relapse (P = 0.036), and delTyr568/570 correlated with tumor origin (P = 0.018). Most importantly, mitotic rate [HR = 2.901 (95 % CI 1.094–7.695), P = 0.032] and risk grade [HR = 9.629 (95 % CI 1.997–46.416), P = 0.005] would be the representative traditional prognostic factors, and deletion with [3 codons would be an
M. Wang J. Xu W. Zhao L. Tu W. Qiu C. Wang H. Cao (&) Department of General Surgery, Renji Hospital, Shanghai Jiao Tong University School of Medicine, P.O. Box 200127, Shanghai, China e-mail:
[email protected] Y. Shen Q. Liu Department of Pathology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
novel independent predictor of poor outcome for RFS in GIST patients with deletion mutation of KIT exon 11 [HR = 7.970 (95 % CI 1.774–35.803), P = 0.007]. All results indicated that mutation determined clinicopathological features and prognosis of GISTs, and more than three codons involvement may be a novel adverse indicator. Keywords Gastrointestinal stromal tumors KIT PDGFRA Mutation Survival
Introduction Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumor of the gastrointestinal tract, arising primarily in the wall of the gastric, small intestine, colon, and other sites within the abdominal cavity with an estimated incidence of 1.5/100,000/year [1]. Although the treatment of choice for localized GISTs is complete surgical excision, following surgical resection, GISTs often relapse locally, spread diffusely throughout the serosal surfaces of the abdomen, and/or metastasize to the liver [2]. Advanced disease is associated with metastases to distant sites, including to the lung and bone [3]. Furthermore, the prognosis for advanced GISTs was poor owing to their inherent resistance to both chemotherapy and radiation therapy [4]. As a result, the prediction of relapse in GISTs after surgery is the current foci area for clinicians [5, 6]. Based on the increased biological knowledge, the diagnosis of GIST has been facilitated considerably by the widespread application of KIT (CD117) immunohistochemistry [7], and the target therapy has also been explored [8, 9]. Unfortunately, prediction of biological behavior of primary GISTs remains challenging. Although the widely
123
819 Page 2 of 11
accepted National Institutes of Health (NIH) risk classification is helpful [10], there are still some cases whose biological behavior cannot be predicted with this classification. Presence or absence of KIT/PDGFRA mutation in GIST has an important value in predicting sensibility to imatinib therapy and choosing proper candidates and has highlighted the way to predict clinical outcomes accurately [1, 11]. Many studies tried to reveal the prognostic significance of such mutational features [12, 13]. KIT exon 11, exon 9, exon 13, and PDGFRA exon 12 and exon 18 mutations have been described to have essential role in GIST prognosis [14–16]. Apart from genetic status, mutation types have been shown to play important role in prognosis [17]. Moreover, certain mutation types with different codons involved could affect the prognosis of GIST. Deletion involving codons 557 and/or 558 of exon 11 of KIT (delW557-K558) is the most widely investigated mutation type, which was described as critical deletion mutation compared with other deletion mutation. Another deletion type, KIT exon 11 Tyr568 and/or Tyr570 deletion, has also been noted because these two tyrosine residues, phosphorylated during activation, are consensus binding sites of Src family kinases [18, 19], which are implicated in activation of different signaling pathways. Based on this understanding, the present study investigated the distribution of gene mutation in 275 patients with GIST. In addition, deep analyses were performed to explore the factors for prognosis prediction, and special attention was paid to the prognostic value of the mutation types and codons.
Materials and methods Patients and follow-up The records of patients who underwent surgical resection of GISTs at the Department of General Surgery, Renji Hospital of Shanghai Jiaotong University School of Medicine (Shanghai, China), from July 1998 to December 2012 were reviewed. Major demographic and clinicopathological characteristics were retrieved. The pathological type, mitotic rate, tumor size, tumor site, and surgical margins were reconfirmed histologically. Inclusion and exclusion criteria were defined as follows. Patients were included when histology confirmed GISTs, the survival data were available, and R0 resection has been obtained. Patients were excluded when histology identified a pathological type other than GISTs, and histopathological and survival data were incomplete. In addition, none of them underwent neoadjuvant imatinib therapy. And patients were excluded when the tumor was ruptured during surgery.
123
Med Oncol (2014) 31:819
Those patients underwent a routine follow-up with CT scan every 3–4 months for 3 years, then every 6 months until 5 years, and yearly afterward or until the patient died. Relapse-free survival (RFS), defined as the duration from operation to relapse or last follow-up, was used for prognosis evaluation. The median follow-up for the entire cohort was 23.5 months (range 5.0–139.0 months). The last follow-up was on September 31, 2012. The primary endpoint of this study was RFS, and patients alive without disease at the last follow-up were recorded as censored events. Relapse was determined based on CT scan or biopsy. Risk grade was defined according to National Institute of Health (NIH) consensus criteria [10, 20]. Ethics All patients provided written informed consent for their information to be stored in the hospital database, and we obtained separate consent for use of research. Study approval was obtained from independent ethics committees from Renji Hospital of Shanghai Jiaotong University School of Medicine. The study was undertaken in accordance with the ethical standards of the World Medical Association Declaration of Helsinki. Mutational analysis A total of 275 formalin-fixed paraffin-embedded (FFPE) samples from GIST patients were obtained for mutational analysis. Mutational analyses were performed on genomic DNA extracted from paraffin-embedded tumor tissue using a combination of PCR amplification. Briefly, DNA was first extracted from three to five sections of FFPE tissue using QIAamp DNA FFPE Tissue Kit (Qiagen, Deutschland) following the manufacturer’s instructions. Then mutational analysis of KIT exons 9, 11, 13, and 17 and PDGFRA exons 12 and 18 was performed. The primers used to amplify KIT and PDGFRA are listed in Table 1. Direct sequencing was conducted in forward and reverse direction. Finally, the results were compared with the standard sequences of KIT (NM_001093772) and PDGFRA (NM_006206) in National Center for Biotechnology Information (NCBI) database. GIST patients with very low risk, low risk, and intermediate risk were defined as ‘‘Not high risk.’’ For the codons involved, the cutoff point for the definition of subgroups (B3 and [3) was the mean value. Statistical analysis SPSS 17.0 software (IBM, Chicago, USA) was adopted for data analysis. Pearson v2 test was used for analysis of relationships between individual qualitative variables. Kaplan–Meier analysis was used for survival analysis, and
Med Oncol (2014) 31:819 Table 1 PCR primers for KIT and PDGFRA exons
Page 3 of 11
Gene
Forward primer
819
Reverse primer
KIT exons 9
GCCACATCCCAAGTGTTTTATG
GAGCCTAAACATCCCCTTAAATTG
11
CCAGAGTGCTCTAATGACTG
TTATGTGTACCCAAAAAGGTGACA
13
GCTTGACATCAGTTTGCCAG
GACAGACAATAAAAGGCAGCTTG
17
TACAAGTTAAAATGAATTTAAATGGT
AAGTTGAAACTAAAAATCCTTTGC
PDGFRA exons
Table 2 Patient’s characteristics
demographic
12
CTCTGGTGCACTGGGACTTT
GCAAGGGAAAAGGGAGTCTT
18
CTTGCAGGGGTGATGCTATT
AGAAGCAACACCTGACTTTAGAGATTA
and
clinicopathological
Items
Value (%)
Age (M ± SD, years)
60.4 ± 11.9
Results Clinicopathological characteristics
B10
212 (77.1)
[10
63 (22.9)
Among the 275 cases included in this study were 154 (56.0 %) males and 121 (44.0 %) females, ranging in age from 25 to 90 yr (Mean ± SD 60.40 ± 11.96 yr). The most common site of tumor was stomach (n = 151, 54.9 %), followed by small bowel (n = 89, 32.4 %), rectum (n = 6, 2.2 %), and colon (n = 2, 0.7 %). In the rest of the 27 cases (9.8 %), tumors were located in extragastrointestinal sites, including retroperitoneum, mesenterium, and omentum. Patient’s demographic and clinicopathological characteristics are provided in Table 2.
B5/50 HPF
196 (71.3)
Genotype analysis
[5/50 HPF
79 (28.7)
Gender Male (%)
154 (56.0)
Female (%)
121 (44.0)
Tumor location Gastric
151 (54.9)
Non-gastrica
124 (45.1)
Tumor diameter (cm)
Mitotic rate
Risk Very low
15 (5.5)
Low Intermediate
90 (32.7) 53 (19.3)
High
117 (42.5)
CD117 Positive
250 (90.9)
Negative
25 (9.1)
Relapse Yes
39 (14.2)
No
236 (85.8)
Imatinib therapy Yes
85 (30.9)
No
190 (69.1)
a
Non-gastric included small bowel (n = 89, 32.4 %), rectum (n = 6, 2.2 %), colon (n = 2, 0.7 %), and others (n = 27, 9.8 %). In the others, tumors were located in extra-gastrointestinal sites, including retroperitoneum, mesenterium, and omentum
significance among patients’ subgroups was calculated by log-rank test. The Cox regression model was used to perform multivariate analysis. Two-sided P \ 0.05 was considered as statistically significant.
Overall, mutations were identified in 258 (93.8 %) patients, and the rest of the 17 (6.2 %) cases were wild type without detectable mutations. Of 245 GIST patients with KIT mutation, there were 30 (12.2 %) cases, 194 (86.5 %) cases, and 3 (1.2 %) cases detected with mutation of KIT exons 9, 11, and 13, respectively. Mutation in KIT exon 17 was not found in this study. Of 13 GIST patients with PDGFA mutations, there were ten (76.9 %) cases and three (23.1 %) cases detected with mutation of PDGFA exons 18 and 12, respectively. In addition, mutation types were explored in 258 patients with KIT or PDGFRA mutations. Deletion was the most frequent type of mutation accounting for 47.3 % (122/258) of all mutation cases, followed by substitution (87/258, 33.7 %) and duplication (49/258, 19.0 %). Data were shown in Table 3. Further analyses were focused on detailed codons involved in GIST gene mutation (Fig. 1). For KIT exon 11 mutations, the most frequently involved area was from codons 557 to 560 both in deletion and in substitution mutation. To be noted, the delW557-K558 and delTyr568/ 570 deletions were found in major KIT exon 11 mutations (n = 77). For KIT exon 9 mutation, the mutation type was unitary duplication of codon 502–503. K642E substitution
123
819 Page 4 of 11
Med Oncol (2014) 31:819
was the mutation type for KIT exon 13. D842V and V561D substitution was the mutation type for PDGFRA exon 18 and exon 12, respectively. Clinicopathological correlation analysis The relationship between gene status and clinicopathological characteristics was analyzed. The mitotic rate was higher in wild-type GISTs than in KIT or PDGFRA-mutant GISTs, and the differences were statistically significant (P = 0.048). The CD117 positive rate was higher in KIT mutation than in wildtype and PDGFRA-mutant GISTs; the differences were statistically significant (P \ 0.001). In addition, the KIT exon 9
Table 3 The genomic features of 275 GIST patients Features
Number (%)
Mutation type Deletion N (%)
Substitution N (%)
Duplication N (%)
Wild type
17 (6.2)
0
0
0
Mutation
258 (93.8)
122 (44.4)
87 (31.6)
49 (17.8)
Kit Exon
245 (89.1)
121 (44.0)
75 (27.3)
49 (17.8)
9
30 (10.9)
0
0
30 (10.9)
11a
212 (77.1)
121 (44.0)
72 (26.2)
19 (6.9)
13
3 (1.1)
0
3 (1.1)
0
17
0
0
0
0
13 (4.7)
1 (0.4)
12 (4.3)
0
12
3 (1.1)
1 (0.4)
2 (0.7)
0
18
10 (3.6)
0
10 (3.6)
0
PDGFRA
a
There was one case harboring synchronous deletion and non-adjacent site substitution
Fig. 1 Frequency of KIT exon 11 codons involved in mutation
123
mutations were more frequently in non-gastric than in gastric areas; the difference was statistically significant (P \ 0.001). Deep analyses showed that the mutation types were correlated with tumor location (P = 0.005), tumor size (P = 0.022), mitosis rate (P \ 0.001), risk grade (P \ 0.001), and relapse (P = 0.004) (Table 4). Considering the delW557-K558 and delTyr568/570 deletions accounts for more than half of deletion mutants (77 cases). Clinical and pathological features of GISTs with KIT deletion mutation were analyzed according to the presence of delW557-K558 or delTyr568/570 deletion. The delW557K558 was not associated with tumor site, tumor size, or risk grade. But, such mutation tended to have higher mitotic rate (P = 0.042) and more opportunities to develop relapse (P = 0.036), an association with higher risk grade (P = 0.045). GISTs with delTyr568/570 were more likely to origin from an extra-gastric site (P = 0.018) (Table 5). Survival analysis Given the well-known function of imatinib in GIST patients, only 189 cases without postoperative imatinib therapy were included in the current survival analyses. For 189 cases, disease relapse was observed in 23 (12.2 %) cases; the median RFS was 24.0 (range 5.0–139.0) months, and the 1-, 3-, and 5-year survival rates were 96.3, 83.5, and 79.6 %, respectively (Fig. 2a). As expected, these traditional factors were associated with GIST patients’ RFS, such as gender, tumor size, mitotic rate, and risk grade (P \ 0.05 for all). In addition, wild-type GISTs correlated with lower 5-year RFS (P = 0.008) (Fig. 2b, c). Multivariate survival analysis revealed that the risk for relapse increased in male patients
Med Oncol (2014) 31:819 Table 4 The relationship between gene mutations and clinicopathological characteristics among 258 GIST patients with mutations
Page 5 of 11
Clinicopathological characteristics
Mutation types Deletion
Substitution
Duplication
B60
64 (52.5)
40 (46.0)
23 (46.9)
[60
58 (47.5)
47 (54.0)
26 (53.1)
Age (years)
0.937
Male (%)
70 (57.4)
48 (55.2)
27 (55.1)
Female (%)
52 (42.6)
39 (44.8)
22 (44.9)
Gastric
68 (55.7)
57 (65.5)
18 (36.7)
Non-gastric
54 (44.3)
30 (34.5)
31 (63.3)
Tumor location
0.005
Tumor diameter (cm)
0.022 90 (73.8) 32 (26.2)
76 (87.4) 11 (12.6)
34 (69.4) 15 (30.6)
B5/50 HPF
74 (60.7)
75 (86.2)
39 (79.6)
[5/50 HPF
48 (39.3)
12 (13.8)
10 (20.4)
High
57 (46.7)
66 (75.9)
28 (57.1)
Not higha
65 (53.3)
21 (24.1)
21 (42.9)
Positive
117 (95.9)
79 (90.8)
46 (93.9)
Negative
5 (4.1)
8 (9.2)
3 (6.1)
Yes
23 (18.9)
3 (3.4)
7 (14.3)
No
99 (81.1)
84 (96.6)
42 (85.7)
Mitotic rate