Results
The incidence of GIST ranges from 11 to 15 per million per year. Growing evidence indicates the incidence is considerably underestimated. The number of GIST patients admitted to our center is on the rise. In the past two year, it has approached 100 cases a year (Figure 1).
(Enlarge Image)
Figure 1.
Annual number of newly admitted GIST patient since 1997.
Clinical and Pathological Characteristics
Total 497 GIST patients were involved in present study, with a median age of 60 years (range 23–90) and 55.9% was male. Stomach and small bowel were the most common sites of primary disease (59.0% and 22.5%, respectively). The most common clinical presentation was abdominal discomfort, followed by GI bleeding. Distribution of risk groups: 8.0%, very low; 36.4%, low; 15.7%, intermediate; and 39.8%, high risk. IM adjuvant therapy was given to 96 of the patients to prevent disease relapse. Recurrence or metastases were observed in 89 patients during the follow-up period. Among which, IM was used to control disease in 46 patients.
Of all the cases, 87.3% was CD117 (+); 80.3%, CD34 (+); 23.6%, SMA (+); and 21.5%, S-100 (+). DOG1 was a newly developed IHCA marker, which was positive in 139 out of 149 cases (93.3%). The diagnosis of GIST in patients presented as both CD117 and DOG1 negative was confirmed by detection of mutation in c-kit/PDGFRA gene. Their clinical and pathological characteristics are listed in Table 2.
Lymph node metastasis was detected in 5 out of 497 cases (1.01%); clinical and pathological characteristics of these 5 cases were described in Table 3.
Survival Analysis on Patients Without IM Adjuvant Therapy
Given the fact that imatinib is an effective drug on GIST, the first survival analysis was based on the population of patients who were not given IM adjuvant therapy. Therefore, 401 patients with operable GIST were enrolled in the cohort, with a median duration of 50 months (range, 7–187 months). Recurrence or metastasis occurred in 79 patients (19.7%). The abdominopelvic cavity was the most common site of metastases (51 cases), followed by liver (22 cases), lung (3 cases), vertebral column (1 case), umbilicus (1 case), and fossa axillaris (1 case). Forty-five patients died of GIST progression, and 4 died of other diseases. The 1-, 3-, 5-year OS of 401 GISTs was 97.7%, 92.6% and 84.8%, respectively; The 1-, 3-, 5-year RFS was 93.2%, 82.1% and 77.4%, respectively.
The 1-, 3-, 5-year OS according to risk grade was: 100%, 100%, 100% (very low risk); 100%, 100%, 100% (low risk); 100%, 97.8%, 89.6% (intermediated risk); 93.5%, 80.8%, 65.9% (high risk), respectively (Figure 2).
(Enlarge Image)
Figure 2.
Overall survival in 401 GIST patients according to risk class.
The 1-, 3-, 5-year RFS according to risk group was: 100%, 100%, 100% (very low risk); 100%, 100%, 98.1% (low risk); 100%, 93.8%, 90.9% (intermediated risk); 80.6%, 53.1%, 44.5% (high risk), respectively (Figure 3).
(Enlarge Image)
Figure 3.
Relapse-free survival in 401 GIST patients according to risk class.
Univariate analysis revealed that male gender, non-gastric origin, larger tumor size, higher mitotic rate, higher risk grade, CD34 negative expression, and adjacent organ involvement contributed to poorer outcome (lower OS and RFS), whereas age and expression of CD117, SMA, and S-100 were not associated with prognosis (see Table 4 and Additional files 1 and 2 http://www.biomedcentral.com/1471-2482/14/93/additional).
Multivariate analysis by Cox proportional hazards regression (Forward LR) model indicated that tumor size, mitotic rate, and risk grade were independent risk factors to OS for GISTs, and that mitotic rate, risk grade, and adjacent involvement were independent risk factors to RFS (Table 5 and Table 6).
Survival Analysis of Patients Received IM Therapy
From 2007 to 2012, 87 patients with intermediate-high risk GIST received IM adjuvant therapy after radical resection (Adjuvant group). Compared with those patients who were with same risk GIST (intermediate-high risk) while were not given IM adjuvant therapy (Non-adjuvant group, n = 188), adjuvant group had better 5-year RFS (82.3% vs. 56.3%, P < 0.001) and 5-year OS (94.9% vs. 72.1%, P = 0.001) (Figure 4). In addition, there was no statistical difference in other clinicopathological features (sex, age, tumor site, tumor size, mitotic rate, risk grade, etc.) between the two groups (see Additional file 3 http://www.biomedcentral.com/1471-2482/14/93/additional), indicating that these features had no impact on the effect of IM.
(Enlarge Image)
Figure 4.
Effect of treatment with or without postoperative IM adjuvant therapy on OS and RFS in 275 intermediate-high risk GIST patients (A: overall survival; B: relapse-free survival).
In the cohort, 87 patients developed recurrence of metastasis after surgery for the primary disease. Among them, c-kit/PDGFRA mutation status was screened in 39 patients. Their mutational characteristics were demonstrated in our previous report. Mutations in c-kit exon 11, c-kit exon 9, and PDGFRA exon 18 were identified in 29, 4, and 1 patients, respectively. And the rest 5 GISTs showed c-kit and PDGFRA wild type. Among all the 87 advanced GIST patients, 45 (including 33 c-kit mutant GISTs, 5 wild-type GISTs, and 7 GISTs with unknown mutation type) were treated with IM, and the other 42 didn't undergo any TKI therapy (10 due to personal reasons and the rest were cases prior to 2005). There was significant difference in outcome between the two groups: patients underwent postoperative IM treatment had better 1-, 3-year OS than those untreated with IM (97.6% and 75.6% vs. 58.7% and 6.8%, respectively, P < 0.001). IM therapy also improved 1-year progression-free survival (PFS) of these patients (87.6% vs. 12.4%, P < 0.001) (Figure 5).
(Enlarge Image)
Figure 5.
Effect of treatment with or without post-operation IM therapy on OS and PFS in 87 advanced GIST cases (A: overall survival; B: progression-free survival).