Benefit of imatinib adjuvant therapy for patients with intermediate risk gastrointestinal stromal tumor: a multi-center retrospective analysis in China

Background still under debate. Thus, the present study aims to investigate the impact of imatinib adjuvant therapy on the prognosis of intermediate risk gastrointestinal stromal tumor patients.Methods From January 2000 to April 2018, a total of 450 intermediate risk gastrointestinal stromal tumor patients received R0 resection from eight centers in China were enrolled. The clinicopathological features and prognosis of patients were analyzed. Discrete variables were analyzed using the Chi-square test or Fisher's exact test. Significant predictors for survival identified by univariate analysis and multivariate analysis. Disease-free survival was evaluated by Kaplan-Meier method.Results Among the 450 patients, 109 patients (24.2%) received imatinib treatment. The recurrence rate of patients with and without imatinib treatment was 7.3% and 5.3%, respectively. For the entire cohort, gene mutational status was the only independent risk factor for disease-free survival of patients though univariate and multivariate analysis (P=0.028). For patients without imatinib treatment, tumor location (P=0.025) and gene mutation status (P=0.040) were prognostic factors in univariate analysis. However, only gene mutational status was independent prognostic factor (P=0.037). The disease-free survival of gastrointestinal stromal tumor with exon 11 deletion mutation with and without imatinib treatment were comparable (P=0.531).Conclusions Imatinib adjuvant treatment could not improve the prognosis of intermediate risk gastrointestinal stromal tumors. Intermediate gastrointestinal stromal tumors with exon 11 deletion mutation or non-gastric location may be candidates for imatinib treatment. GISTs. However, these findings indicated that non-gastric intermediate risk GISTs may be candidates for imatinib treatment.


Results
The clinicopathological features of the entire cohort were summarized in Table 1. There were 215 male (47.8%) and 235 female (52.2%). The median age were 58 years (50 to 65 years). Most tumors were located in the stomach (440/450, 97.8%), the rest 10 tumors were located in esophagus (3), duodenal (3), small intestine (2) and rectum (2). Three hundred and seventy tumors displayed spindle cell morphology (370/418, 88.5%), 28 tumors displayed epithelioid morphology (28/418, 6.7%), and 20 tumors displayed mixed morphology (20/418, 4.8%). The positive rate of CD117, CD34 and DOG-1 were 95.4%, 97.2% and 93.5%. One hundred and sixty-six patients (36.9%) received c-kit/PDGFR gene mutation status examination. One hundred and nine patients (24.2%) received imatinib therapy. The median follow up time was 48 months (27 to 74 months). In total, 8 patients suffered from recurrence or metastasis in the treatment group (7.3%), and 18 patients suffered from recurrence or metastasis in the control group (5.3%). The most common location of metastasis was liver (15/26, 57.7%). The risk factors for the prognosis of entire cohort were analyzed using univariate and multivariate analysis and summarized in Table 2. The results showed that only gene mutational status was independent prognostic factor. Imatinib treatment was not associated with the prognosis of patients. The DFS of patients with and without imatinib treatment was comparable (Fig. 1). The DFS of patients with exon 11 deletion mutation was significantly lower than that of exon 11 nondeletion mutation and other gene mutational status (Fig. 2). Interestingly, the DFS of patients with non-gastric tumors was significantly lower than that with gastric tumors (Fig. 2).
However, in the univariate analysis, the prognostic value of tumor location was on the boundary of significance (P = 0.060). The risk factors for the prognosis of patients without imatinib treatment were analyzed using univariate and multivariate analysis and summarized in Table 3. The results showed that tumor location and gene mutational status were risk factors for the prognosis of patients. However, only gene mutational status was independent prognostic factor. Patients with exon 11 deletion mutation or non-gastric tumors were associated with poor DFS (Fig. 3).

Discussion
Microscopic complete resection without rupture is the optimal standard treatment for localized GISTs. Although received R0 resection, over 40% of patients will develop recurrence or metastasis after surgery [5]. With demonstrated effect of imatinib in advanced GISTs, it has been recommended as adjuvant treatment of GIST patients [6]. Although Z9001 trial demonstrated that 1 year of adjuvant imatinib therapy could improve RFS of GIST patients [7], and SSGXVIII/AIO trial found that 3 years of imatinib therapy could improve RFS and overall survival of high risk Besides tumor size and mitotic index, location is also associated with the prognosis of GISTs, and non-gastric GISTs were reported to be more aggressive than gastric GISTs in clinical course [12,13]. In the study reported by Fu et al. [10], the prognosis of non-gastric GISTs was significantly worse than gastric GISTs. However, this has not been confirmed by multivariate analysis. Wu et al. demonstrated that nonstomach location was independently associated with poor prognosis of GISTs through multivariate analysis [9]. In our present study, although the prognosis of non-gastric GISTs was worse than gastric ones, the correlation disappeared in the multivariate analysis. This could be partially explained by the different constituent ratio of gastric and non-gastric GISTs among the three studies. In the published two studies, the proportion of non-gastric GISTs were 37.5% [9] and 27.1% [10], respectively. However, the proportion of non-gastric GISTs were only 2.2% (10/450) in our present study. Although DFS of gastric GISTs was significantly better than non-gastric ones, the 95% confidence interval could not be calculated out in the multivariate analysis due to the extremely small sample size and extremely low constituent ratio of non-gastric GISTs.
It was reported that the recurrence rate of was 3.6% for gastric GISTs with tumor size from 5.1 to 10 cm and mitotic index less than or equal to 5/50 HPF, 16.0% for gastric GISTs with tumor size from 2.1 to 5 cm and mitotic index exceed 5/50 HPF, and 50-54% for non-gastric GISTs with tumor size less than or equal to 2 cm and mitotic index exceed 5/50 HPF [14,15]. However, the above mentioned population included a proportion of high risk GISTs (mitotic index exceed 10/50 HPF) according to the modified NIH consensus classification system [11]. Thus, the recurrence rate of intermediate risk GIST in the natural clinical course was still unclear. In the study reported by Wu et al. [9], the recurrence rate was 3.6% for gastric GISTs and 18.4% for non-gastric GISTs. In our present study, the recurrence rate for gastric and nongastric GISTs was 4.8% and 22.2%, which was consistent with previous study. As the prognosis of non-gastric GISTs was significantly worse than gastric GISTs, Wu et al. [9] further evaluated the impact of imatinib adjuvant therapy on non-gastric GISTs, and found that imatinib treatment could significantly improve RFS of non-gastric intermediate risk GISTs. In our present study, the association between tumor location and prognosis was not identified by multivariate analysis and the benefit of imatinib treatment for non-gastric GISTs could not be evaluated due to the extremely low sample size of non-gastric GISTs. However, these findings indicated that non-gastric intermediate risk GISTs may be candidates for imatinib treatment.
The genetic subtypes of GISTs were also associated with different prognostic relevance [16,17]. It was reported that GISTs with KIT exon 11 deletions are associated with poor prognosis [18], especially in GIST with deletions affecting codons 557-558 [19][20][21]. However, in the Z9001 trial, the association between KIT exon 11 deletion and poor prognosis was not confirmed through multivariable analysis [22]. In the study reported by Wu et al. [9], exon 11 deletion mutation was also not associated with the prognosis of imatinib naïve GIST. However, in our present study, KIT exon 11 deletion mutation was independently associated with poor prognosis of GIST patients. Further, we wonder whether imatinib treatment could improve the prognosis of GISTs with exon 11 deletion mutation or not. Thus, the DFS of GIST with exon 11 deletion mutation with or without imatinib treatment were compared. The results showed that imatinib therapy could not improve the prognosis of GIST with exon 11 deletion mutation. This should be interpreted cautiously due to the relatively small sample size. However, imatinib therapy could significantly improve the survival of GISTs with exon 11 deletion mutation in the study reported by Wu et al. [9]. These findings indicated that intermediate risk GISTs