Abstract

Background: In recent years, imatinib mesylate (STI 571), a tyrosine kinase inhibitor, has shown short-term clinical usefulness for gastrointestinal stromal tumor or gastrointestinal leiomyosarcoma (GIST). The value of surgical resection, including hepatectomy, for metastatic GIST remains unknown. Our aim was to evaluate the outcome of surgical resection, including hepatectomy, for metastatic GIST at a single institute.

Methods: Eighteen patients who underwent hepatectomy for metastatic GIST were identified and the clinicopathological data of these patients were analyzed retrospectively.

Results: The primary site of GIST included stomach in 10, duodenum in five, ileum in two and esophagus in one patient. A hemihepatectomy or greater resection was undertaken in eight patients. Six patients underwent simultaneous resection for primary and hepatic desease. There was no in-hospital mortality in this series. The post-hepatectomy 3- and 5-year survival rates were 63.7 and 34.0% respectively, with a median of 36 (17–227) months. Recurrence after the initial hepatectomy was documented in 17 patients (94%), and metastatic mass of the remnant liver developed in 15 of these 17 patients (88%). Three patients survived >5 years after the initial hepatectomy who underwent multiple surgical resections during this period. No clinicopathological characteristic was a significant predictive factor for survival.

Conclusions: Multiple surgical resections, including hepatectomy, may contribute to important palliation in selected patients with metastatic GIST. Surgical cure seems to be difficult due to the high frequency of repeat metastasis to various sites. Therefore, adjuvant therapy must be required in the treatment of metastatic GIST.

INTRODUCTION

The liver is a common metastatic site for gastrointestinal stromal tumors or gastrointestinal leiomyosarcomas (GIST). Many patients with liver metastasis from GIST are either unresectable due to diffuse intrahepatic disease or inoperable due to extrahepatic disease. Some reports describing surgical resection of liver metastases from various sites of primary sarcoma have been published (13). In a recent analysis of 331 patients with liver metastasis from sarcoma, of which 131 patients had a GIST, 34 underwent hepatectomy of all gross disease (2). The post-operative 5-year survival rate was 30%, with a median survival of 39 months. The time interval from treatment of the primary tumor to the development of liver metastasis was a significant predictive factor of survival (2).

Liver metastasis from GIST previously has been considered to be insensitive to chemotherapy or chemoembolization. Surgical resection is a possibly effective therapy and may provide a potential cure. However, imatinib mesylate (STI 571), currently being tested in clinical trials, has shown effectiveness for c-kit-positive GIST. STI 571 has demonstrated efficacy, minimal toxicity and a partial response rate of ∼69% (4). Although there has been no complete response, the rate of disease progression has been only 11%. STI 571 has thus influenced the treatment of GIST patients dramatically.

This study reports the outcome of our 18 year experience at the National Cancer Center Central Hospital (NCCH) in Japan, to determine the value of surgical treatment, including hepatectomy, for patients with liver metastasis from GIST.

METHODS

From January 1984 to October 2003, there were 18 patients with liver metastasis from GIST who underwent hepatic resection of liver lesions with curative intent at the NCCH. Patient demographics were recorded with the clinicopathological characteristics of the primary tumor and extent of intrahepatic and extrahepatic metastatic disease. Patients' date of treatment, recurrence and survival after hepatectomy were examined retrospectively in their medical records. The diagnosis of all liver metastases was confirmed by a pathologist at the NCCH. The long axis of the largest tumor was recorded as the tumor size. A synchronous metastasis was defined as the detection of a liver metastasis within 1 month of resection of the primary tumor.

Statistical analysis was performed with SPSS statistical softwear (Chicago, IL). Statistical significance was defined as P < 0.05. Cumulative survival was calculated using the Kaplan–Meier method. Univariate analysis was performed with Mann–Whitney U-test and log-rank test for survival.

RESULTS

Patient Characteristics

The clinicopathological variables of the patients with metastatic GIST to the liver are shown in Table 1. There were 10 (56%) males and eight (44%) females. The median age at diagnosis of the metastatic liver tumor was 58 years (range 33–65). There were 12 patients (67%) with metachronous liver lesions and six (33%) with synchronous lesions.

Table 1.

Clinicopathological characteristics

 Median survival
 
   

 
n
 
%
 
Months
 
P-value
 
Gender    0.52 
    Female 44 38  
    Male 10 56 38  
Age    0.62 
    ≤50 years 39 40  
    >50 years 11 61 38  
Primary site    0.75 
    Stomach 10 56 38  
    Duodenum 28 37  
    Ileum 11 23  
    Esophagus 38  
Size of liver metastasis    0.36 
    ≤5 cm 44 37  
    >5 cm 10 56 38  
No. of liver metastases    0.88 
    Solitary 33 40  
    Multiple 12 67 37  
Time to liver metastasis     
    Synchronous 33 30 0.52 
    Metachronous 12 67 38 0.31 
    <3 years 11 61 34  
    ≥3 years 39 40  
Extent of hepatectomy    0.32 
    Less than lobectomy 10 56 40  
    Lobectomy or more 44 34  
Radicality    0.48 
    Complete 15 83 38  
    Incomplete 17 39  
 Median survival
 
   

 
n
 
%
 
Months
 
P-value
 
Gender    0.52 
    Female 44 38  
    Male 10 56 38  
Age    0.62 
    ≤50 years 39 40  
    >50 years 11 61 38  
Primary site    0.75 
    Stomach 10 56 38  
    Duodenum 28 37  
    Ileum 11 23  
    Esophagus 38  
Size of liver metastasis    0.36 
    ≤5 cm 44 37  
    >5 cm 10 56 38  
No. of liver metastases    0.88 
    Solitary 33 40  
    Multiple 12 67 37  
Time to liver metastasis     
    Synchronous 33 30 0.52 
    Metachronous 12 67 38 0.31 
    <3 years 11 61 34  
    ≥3 years 39 40  
Extent of hepatectomy    0.32 
    Less than lobectomy 10 56 40  
    Lobectomy or more 44 34  
Radicality    0.48 
    Complete 15 83 38  
    Incomplete 17 39  

Tumor Characteristics

The distribution of primary tumor location included 10 cases in the stomach, five duodenal, two ileal and one in the esophagus. The size of the metastatic liver tumor was > 5 cm in 10 cases and ≤5 cm in eight cases (median 6.3 cm, range 1.6–24.0). Of 18 cases with hepatic metastases, 12 cases (67%) had multiple liver lesions. Five of these 12 cases had ≥5 liver lesions each. The time interval between the primary lesion and metastatic liver disease was ≥3 years in seven patients (median 53 months, 43–180), and <3 years in 11 patients including six patients with synchronous liver lesions (median 0 months, 0–35).

Treatment

Eight of 18 patients were treated with a lobectomy or greater resection, three with segmentectomies and seven with partial resection of the liver. Macroscopic complete resection of liver metastasis was achieved in 15 (83%) patients. No clinicopathological characteristic was a significant predictive factor for survival on univariate analysis (Table 1).

Survival and Recurrence

Of the 18 patients, 13 died from the primary disease, four were alive with disease, and only one was alive without recurrence, during a follow-up period of 35 months from the time of hepatic resection (Table 2). The median follow-up period after hepatectomy was 36 months (range 17–227). The post-hepatectomy 3- and 5-year survival rates were 63.7 and 34.0%, respectively (Fig. 1).

Figure 1.

Survival after initial hepatectomy for metastases from GIST.

Figure 1.

Survival after initial hepatectomy for metastases from GIST.

Table 2.

Survival and recurrence after hepatectomy


 
n
 
%
 
Survival status   
    No evidence of disease 
    Alive with disease 22 
    Died of disease 13 73 
Recurrence in the remnant liver 15* 88 
    Resection  
    Other treatment  
Recurrence of other site 11* 65 
    Resection  
    Other treatment  

 
n
 
%
 
Survival status   
    No evidence of disease 
    Alive with disease 22 
    Died of disease 13 73 
Recurrence in the remnant liver 15* 88 
    Resection  
    Other treatment  
Recurrence of other site 11* 65 
    Resection  
    Other treatment  
*

Nine patients had both recurrence of remnant liver and extrahepatic organs.

Recurrence following hepatectomy occurred in 17 patients (94%) including three patients who underwent macroscopic incomplete resection. The median time to first recurrence in those patients who underwent a macroscopic complete resection (n = 15) was 13.5 months (range 4–49). Fifteen patients (88%) developed recurrence within the remnant liver, of which six underwent a total of nine further hepatic resections (Table 2). Nine patients undertook other treatments for recurrent liver tumor including radiation in one, ethanol injection in two, radiofrequency ablation (RFA) in two, and chemotherapy in four including the use of STI 571 for three patients. Eleven patients (65%) developed recurrence within numerous other extrahepatic organs. Bone metastasis occurred in five patients, peritoneal disease in four and adrenal gland metastasis in two patients. Other metastatic sites included lung, chest wall, skin, soft tissue, brain and axillary lymph nodes. Six of these patients underwent a total of 18 further resections. Of these 17 patients with recurrent tumor after hepatectomy, nine cases (53%) had both intra- and extrahepatic recurrence.

The initial hepatectomy resulted in macroscopic incomplete resection in three patients. One patient died of disease 19 months after hepatectomy. The other two patients were still alive at 39 and 45 months, receiving STI 571 administration following the hepatectomy.

Of all patients who underwent hepatectomy for liver metastasis from GIST, three patients survived >5 years after the initial hepatectomy (Table 3). Two of them underwent multiple resections for both recurrent extrahepatic disease and further hepatectomy for recurrent liver tumors. One patient underwent a total of 17 sessions of tumor resection: the liver twice, the chest wall twice, the lung four times, laminectomy twice and abdominal disease, left kidney and left iliac bone disease once each. The other patient underwent a total of five resections: the liver twice, the skin twice and soft tissue once. The time to recurrence after the initial hepatectomy in these three patients was 23, 35 and 49 months, respectively. This was a significantly longer interval compared with the remaining patients, excluding the three patients with incomplete initial hepatectomy (median 15 months, 4–24) (P = 0.01 using Mann–Whitney U-test).

Table 3.

Five-year survivors after complete resection in three cases

Case Age/sex Primary site Time of recurrence from hepatectomy No of surgical inteventions
 
 Status Follow-up (months) 

 

 

 

 
Liver
 
Other site
 

 

 
33/M Ileum 49 DOD 227 
63/F Stomach 35 DOD 61 
43/ Stomach 23 DOD 84 
Case Age/sex Primary site Time of recurrence from hepatectomy No of surgical inteventions
 
 Status Follow-up (months) 

 

 

 

 
Liver
 
Other site
 

 

 
33/M Ileum 49 DOD 227 
63/F Stomach 35 DOD 61 
43/ Stomach 23 DOD 84 

DOD, died of disease.

DISCUSSION

Hepatectomy for metastasis from GIST has been reported previously (13), although the largest series included not only GIST but also soft tissue sarcoma or leiomyosarcoma. The post-operative 5-year survival rate in this patient group who underwent complete resection was 30%, with a median of 39 months (2). The present series focuses only on metastatic liver tumors from GIST, with an overall 5-year survival rate (including three patients with incomplete resection) of 34.0% and a median survival of 36 months, from the time of hepatic resection. This is similar to previously reported series. To know the natural history of the patient with metastatic GIST, it is crucial to evaluate the treatment strategy. However, as we had aggressively performed hepatectomy for metastatic GIST before the use of STI 571 was introduced, the clinical course of unresectable cases had not been followed in our institute. Therefore, we could not make a comparison between the treatment results of with/without hepatectomy. The reported response rates for chemotherapy of the tumor are poor, with a duration of response of only a few weeks or months (5,6).

There were no other prognostic factors arising from the clinicopathological characteristics, including complete resection of all gross metastases, in the present series. Moreover, following hepatectomy for liver metastasis, there was a high proportion of recurrence in the remnant liver in addition to other organs. This result may show that surgical intervention for metastatic GIST is only palliative. However, three cases in our series had a long survival time of >5 years following repeated hepatectomy and tumor resection. In these three cases, the time to recurrence after the initial hepatectomy was statistically longer than that of the other patients. The longest survivor in our series was 227 months after the initial hepatectomy. Therefore, the only significant prognostic factor following hepatectomy on univariate analysis was a >5 year period to the development or recurrence (2). Other series have shown that metachronous metastases of ≥2 years from a complete resection of all gross disease was associated with a better prognosis on multivariate analysis. In all reports, the time to development of liver metastasis was invariantly an important prognostic factor in the surgical treatment. This time factor may be important in the selection of patients for treatment of recurrent tumor after hepatectomy.

Hepatectomy for potentially resectable metastatic colorectal tumor is now considered to be the first line therapy and its safety has been increasing, with an operative mortality and morbidity rate of 1.8 and 5%, respectively, in a recent large series (7). This series also reported a 5-year survival rate of up to 39%; other larger series have also reported much the same survival benefit (8). Overall, these data may imply that metastatic liver tumor from colorectal carcinoma via the portal venous system is not a reflection of systemic disease in the selected patients. In contrast, liver metastases from sarcomatous tumors reach the liver through the systemic circulation and therefore liver disease may merely reflect systemic disease. GIST are thought to be drained by portal vein blood flow as is the case in colorectal carcinoma. It may be important to determine the results of surgical treatment by focusing only on the liver metastasis from GIST. Unfortunately, the present study clearly reaffirms the difference between liver metastasis of GIST and colorectal carcinoma. The high rate of recurrence after hepatectomy for metastatic GIST implies that the disease had already become a systemic disease.

STI 571 is an inhibitor of the tyrosine kinase activity of c-kit and has shown good activity against GIST in clinical reports (4,9). It has been reported in a phase II trial that the partial response rate to STI 571 was 59% and that only 13% of 86 patients with GIST had progressive disease (10). However, in this trial, no complete response was obtained and therefore it is too early to confirm the duration of response in patients with STI 571-sensitive GIST. To test the benefit of adjuvant STI 571 in patients after complete resection of high-risk primary GIST, a phase II trial is being conducted by the American College of Surgeons Oncology Group (11). A prospective evaluation of neoadjuvant STI 571 therapy may be warranted. Two patients out of three in the present series, resulting from incomplete hepatectomy for metastatic GIST, received STI 571 for residual disease. They demonstrated a prolongation of survival, in keeping with a partial response.

Hepatectomy for metastatic GIST can be performed safely, and multiple surgical resections including hepatectomy may contribute to important palliation in selected patients with slow-growing metastatic GIST. However, surgical cure seems unlikely due to the high frequency of subclinical metastases, and hepatectomy may form only part of the therapy for metastatic GIST. Therefore, adjuvant therapy must be evaluated prospectively in the treatment of metastatic GIST.

This research was supported in part by a Grant-in-Aid for the Second Term Comprehensive 10-year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare, Japan.

References

1.
Lang H, Nußbaum KT, Kaudel P, Fruhauf N, Flemming P, Raab R. Hepatic metastases from leiomyosarcoma: a single-center experience with 34 liver resections during a 15-year period.
Ann Surg
 
2000
;
231
:
500
–5.
2.
DeMatteo RP, Shah A, Fong Y. Results of hepatic resection for sarcoma metastatic to liver.
Ann Surg
 
2001
;
234
:
540
–8.
3.
Elias D, Cavalcanti de Albuquerque A, Eggenspieler P, Plaud B, Ducreux M, Spielmann M, et al. Resection of liver metastases from a noncolorectal primary: indications and results based on 147 monocentric patients.
J Am Coll Surg
 
1998
;
187
:
487
–93.
4.
van Oosterom AT, Judson I, Verweij J, Stroobants S, Donato di Paola E, Dimitrijevic S, et al. Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumors: a phase I study.
Lancet
 
2001
;
358
:
1421
–3.
5.
Edmonson JH, Ryan LM, Blum RH, Brooks JS, Shiraki M, Frytak S, et al. Randomized comparison of doxorubicin alone versus ifosphamide or mitomycin, doxorubicin, and cisplatin against advanced soft tissue sarcomas.
J Clin Oncol
 
1993
;
11
:
1269
–75.
6.
Casper ES, Christman KL, Schwartz GK, Johnson B, Brennan MF, Bertino JR. Edatrexate in patients with soft tissue sarcoma: activity in malignant fibrous histiocytoma.
Cancer
 
1993
;
72
:
766
–70.
7.
Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases.
World J Surg
 
1995
;
19
:
59
–71.
8.
Fong Y, Fortner J, Sun RL, Brennan MF, Blumgart LH. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases.
Ann Surg
 
1999
;
230
:
309
–21.
9.
Joensuu H, Roberts PJ, Sarlomo-Rikala M, Andersson LC, Tervahartiala P, Tuveson D, et al. Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor.
N Engl J Med
 
2001
;
344
:
1052
–6.
10.
Blanke CD, von Mehren M, Joensuu H, et al. Evaluation of the molecularly targeted therapy STI571 in patients with unresectable or metastatic gastrointestinal stromal tumors expressing KIT.
Proc Am Soc Clin Oncol
 
2001
;
20
:
2
.
11.
American College of Surgeons Oncology Group. www.acosog.org
2001
.