Introduction

The most common sites of metastatic recurrence of gastric cancer are the peritoneum and lymph nodes, and the most common site of hematogenous metastasis is the liver [1–3]. This distinction is also true for recurrence after radical resection of gastric cancer; therefore, resection is rarely indicated for these lesions. Systemic chemotherapy is the standard treatment for patients with distant metastasis or recurrence; however, the reported median survival time (MST) is only 10–14 months [4, 5]. In contrast for patients with peritoneal or distant lymph node metastases, resection is indicated for some patients with liver metastases. Standard resection is performed in cases of resectable liver metastases of colorectal cancer [6], and 5-year survival rates exceeding 50 % have been reported [7, 8]. Because liver metastases of gastric cancer (LMGC) are often multifocal at the time of diagnosis [9], hepatectomy is not commonly recommended. Nevertheless, excellent outcomes of hepatectomy for LMGC have recently been reported from several high-volume centers, with a 5-year survival rate of 13–37 % and MST of 11–34 months [10–15]. However, the majority of these reports were based on single-center retrospective analyses of relatively small sample sizes collected over long periods and without clearly defining the indications for resection. Furthermore, diagnostic modalities of liver metastases were not precisely described in these reports. The efficacy of contrast-enhanced magnetic resonance imaging (MRI) using a liver-specific contrast agent for the diagnosis of liver metastasis has been recently reported [16].

To identify the optimal target subpopulation for resection in patients with LMGC, we retrospectively analyzed the treatment outcomes of hepatectomies performed at our hospital based on modern diagnostic technology.

Methods

Patients

The data for 28 patients who underwent hepatectomy for LMGC were retrospectively analyzed at the Shizuoka Cancer Center between December 2004 and October 2014. All patients were diagnosed by contrast-enhanced MRI using a liver-specific contrast agent. Liver resection was indicated for patients with LMGC if they met the following inclusion criteria: (1) no more than three metastases, (2) no noncurative factors other than liver metastases, (3) eligible for radical surgical resection, and (4) good general health condition. The liver metastasis was diagnosed by multiple imaging modalities such as contrast-enhanced computed tomography (CT), ultrasound sonography, and contrast-enhanced MRI using liver-specific contrast agent. With regard to the liver-specific contrast agent, we used superparamagnetic iron oxide until 2008 and then gadolinium ethoxybenzyl-diethylenetriamine pentaacetic acid after 2009.

Data collection

Clinicopathological data were collected from databases prospectively registered into electronic medical records. Histopathological descriptions were handled in accordance with the Japanese Classification of Gastric Carcinoma (Third English Edition) [17]. Liver metastases already present when gastric cancer had been diagnosed were defined as synchronous metastases, whereas metastases that occurred in the liver after the curative resection of primary gastric tumors were defined as metachronous metastases. Postoperative complications were defined as those of grade II or above according to the Clavien–Dindo classification [18]. Data collection and analysis were approved by the institutional review board of the Shizuoka Cancer Center.

Statistical analysis

Fisher’s exact test and the Mann–Whitney U test were used to compare the two groups. Survival curves were created using the Kaplan–Meier method. Univariate and multivariate analyses of factors affecting survival time were performed using the Cox proportional hazards model.

Statistical analysis was conducted using EZR version 1.24 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [19], and p values less than 5 % were considered to indicate a statistically significant difference.

Results

Patient characteristics

The clinicopathological characteristics of the 28 patients are presented in Table 1. The male-to-female ratio was 4.6:1. In the synchronous liver metastasis cases, the primary gastric tumor was resected at the same time as the liver, whereas in the metachronous liver metastasis cases, the primary gastric tumor was resected during the initial surgery. The median time from gastrectomy to hepatectomy in the 13 patients with metachronous liver metastases was 19 months (range, 8–38 months). Chemotherapy was selected for initially unresectable liver metastases, and hepatectomy was then performed for 3 patients (2 synchronous and 1 metachronous case) after a reduction in metastases was observed.

Table 1 Clinicopathological characteristics of 28 patients who underwent hepatectomy for liver metastases of gastric cancer
Full size table

Comparison of surgical outcomes between synchronous and metachronous liver metastases

The amount of intraoperative blood loss and the proportion of patients who received an intraoperative blood transfusion were greater in the synchronous liver metastasis group, although the difference was not statistically significant (Table 2). Postoperative complications in patients with complication grade II or higher (determined as per the Clavien–Dindo classification) occurred in ten synchronous patients and one metachronous patient, indicating the occurrence of significantly more complications in the former (p = 0.002). Anastomotic leakage occurred in two patients in the synchronous liver metastases group; however, hepatectomy-induced complications, such as bile leakage and cholangitis, were more common (four patients). All patients were macroscopically and microscopically negative for surgical margins. The number of patients with multiple metastases was greater in the synchronous liver metastasis group (p = 0.038). The length of postoperative hospital stay was significantly prolonged in the synchronous liver metastasis group (p = 0.003). No in-hospital deaths occurred; all patients were safely discharged.

Table 2 Operative details and perioperative outcomes of hepatectomy
Full size table

Postoperative treatment

Twelve patients (43 %) underwent postoperative adjuvant chemotherapy after hepatectomy. The chemotherapy regimen was S-1 in 10 patients, S-1 + cisplatin in 1 patient, and capecitabine + cisplatin + trastuzumab in 1 patient.

Survival and prognostic factors

The median follow-up period after hepatectomy was 26 months, and the median overall survival time was 49 months, with a 5-year survival rate and relapse-free survival rate of 32 % and 29 %, respectively (Figs. 1, 2).

Fig. 1
figure 1

Kaplan–Meier analysis of overall survival in 28 patients who underwent hepatectomy for liver metastases of gastric cancer. The 5-year survival rate was 32 %

Full size image
Fig. 2
figure 2

Kaplan–Meier analysis of relapse-free survival in 28 patients who underwent hepatectomy for liver metastases of gastric cancer. The 5-year relapse-free survival rate was 29 %

Full size image

Univariate analysis of factors affecting survival time revealed that age 70 years or older (p = 0.030), synchronous liver metastases (p = 0.017), and postoperative complications (p = 0.042) resulted in a significantly poor survival outcome (Table 3). The 5-year survival rate after resection in the metachronous liver metastasis group was 59 % (Fig. 3).

Table 3 Univariate analysis of factors affecting survival
Full size table
Fig. 3
figure 3

The 5-year survival rates of patients with metachronous liver metastases (59 %) or synchronous liver metastases (13 %). There was a significant difference between the groups (log-rank test, p = 0.009)

Full size image

The variables that showed significant differences in the univariate analysis—age, timing of liver metastases, and postoperative complications—were subjected to multivariate analysis using the number of liver metastases and intraoperative blood transfusion amount, which showed differences of p < 0.1 in the univariate analysis, as covariates. None of the analyzed variables was identified as an independent prognostic factor.

Recurrence pattern

Recurrence occurred in 17 patients (61 %) during the follow-up period, 15 (83 %) of whom experienced recurrence in the liver remnant. In addition, brain and lymph node metastases occurred in 1 patient each. Eight patients underwent chemotherapy for recurrence, 6 received best supportive care, and 3 underwent repeat hepatectomy for recurrence in the liver remnant. Of the 3 patients who underwent repeat hepatectomy, 2 survived without recurrence for more than 4 years from the initial hepatectomy, and the remaining (1) patient experienced recurrence in the abdominal wall and ribs 10 months after repeat hepatectomy and died 19 months after surgery.

Discussion

Good survival outcomes of hepatectomy were obtained with MST of 49 months and a 5-year survival rate of 32 % in patients with LMGC who were diagnosed by contrast-enhanced magnetic resonance imaging (MRI) using liver-specific contrast agents. These results were relatively better than those of the previous reports, which included MST of 11–34 months and a 5-year survival rate of 13–37 %. In previous studies, solitary metastasis had been reported as an independent better prognostic factor [11–13], resulting in hepatectomy generally being considered only for patients with a solitary liver metastasis.

Oki et al. recently reported that solitary metastasis was an independent prognostic factor in a large-scale multi-institutional cohort study [20]. Patients with multiple liver metastases are generally poor candidates for hepatectomy. In the recent Japanese Clinical Oncology Group (JCOG) trial, which investigated the role of volume reduction surgery in patients with a single noncurable factor, it was reported that the presence of two to four liver metastases, distant lymph node metastasis, and peritoneal metastasis were factors associated with noncurable disease [21]. However, Kinoshita et al. had reported that survival was better in patients with fewer than three liver metastases in a large-scale multi-institutional retrospective cohort study with a large sample of 256 patients [15]. Furthermore, Takemura et al. also reported good results with a 5-year survival rate and MST of 37 % and 34 months, respectively, using the same criteria for hepatectomy as ours, namely, three or fewer liver metastases [14].

The controversies of these reports may be partly because of the long accumulation period and various diagnostic modalities for liver metastasis. In the present study, we determined the criteria for surgery using modern diagnostic modalities. It has been reported that the diagnostic ability of contrast-enhanced MRI is more sensitive than that of conventional enhanced CT scan, particularly in the detection of small lesions. In this study, we occasionally observed that several metastases were detected using contrast-enhanced MRI that were not detected using a conventional CT scan, which only reported solitary metastases. Thus, small multiple metastases might have been present in previous studies that reported a survival advantage for the resection of solitary metastasis. Therefore, it is reasonable that relatively better survival was demonstrated in multiple metastases up to three lesions when diagnosis was made using sensitive contrast-enhanced MRI. We speculate that hepatectomy can be appropriately performed for as many as three metastatic lesions if diagnosis is made using modern technologies.

Although the multivariate analysis revealed no independent prognostic factors, the univariate analysis of survival-related factors showed that outcomes were poorer in the synchronous liver metastasis group. Ambiru et al. [10], Okano et al. [11], and Schildberg et al. [13] have also reported poorer survival outcomes in patients with synchronous liver metastases than in those with metachronous liver metastases. For liver metastases of colorectal cancer, Tsai et al. [22] reported that patients with synchronous liver metastases who underwent resection are likely to have a greater number of metastases. Generally, they have metastases in both lobes and a shorter relapse-free survival time compared with patients with metachronous liver metastases. Although the biological differences between synchronous and metachronous metastases of gastric cancer remain unknown, the present study also found that many cases of resectable synchronous metastasis had multiple metastases. This finding suggests that synchronous liver metastases may be more malignant than metachronous LMGC.

Differences were seen in survival outcomes depending on the presence or absence of postoperative complications. In various diseases, postoperative complications have been reported to exacerbate survival outcomes. In a previous paper [23], we reported the association between intraabdominal infectious complications after surgery for gastric cancer and survival outcomes. Hayashi et al. [24] and Kubota et al. [25] also reported that complications following surgery for gastric cancer shorten relapse-free and overall survival rates. In the present study, survival outcomes may have been affected by the increased complications resulting from the concurrent performance of gastrectomy and hepatectomy.

Elder patients also demonstrated poor survival in the present study, in contrast to previous reports. The possible reason is that the average age was relatively higher in the present study than in previous studies. In addition, in younger patients (less than 70 years old), all of 6 patients who recurred after hepatectomy received chemotherapy or repeat hepatectomy, whereas 6 of 11 elder patients (70 years old or more) who recurred did not receive active treatment. This difference in treatment modality after recurrence may cause the difference of survival between elder patients and non-elder patients.

Liver recurrence in 15 (54 %) of the 28 patients highlighted the importance of controlling the high liver recurrence rate after hepatectomy; however, no established treatment strategy is available for recurrence in the liver remnant. Takemura et al. [26] reported a 5-year survival rate and MST of 47 % and 31 months, respectively, in 14 patients who underwent repeat hepatectomy for recurrence in the liver remnant. Two of the three patients who underwent repeat hepatectomy in our study also achieved long-term survival, suggesting that resection may be considered if the aforementioned indications are met.

The usefulness of radiofrequency ablation (RFA) had been also demonstrated as a local therapy for liver metastasis. Guner et al. compared the results for LMGC between surgery and RFA and reported that similar results were obtained in the surgery group and RFA group [27]. RFA may be considered as a treatment option in patients for whom surgery is difficult for certain reasons even if the number of metastases is small.

This study had some limitations. First, it was based on retrospective analysis of a small sample size from a single institution, and multivariate analysis could not identify any independent prognostic factors. Most previous studies of hepatectomy for patients with LMGC were performed at a single institute over a relatively long period. Conversely, increased diagnostic ability such as contrast-enhanced MRI using liver-specific contrast agent has allowed detection of small metastases that were not previously detected. The results of this study are derived from successive cases treated over a short period under a fixed diagnostic modality and treatment policy. Therefore, this study, based on current diagnostic methods, is significant in that we were able to demonstrate good treatment outcomes in patients with three or fewer liver metastases. Second, the present study included some patients with a short follow-up period owing to the small number of patients who met the eligibility criteria during the 10 years of this study in a single institution. Further patient accumulation and follow-up are needed to demonstrate the true benefits of surgical therapy for LMGC.

In conclusion, hepatectomy has significance if indications are appropriately limited, particularly in cases of metachronous liver metastases, where active resection should be considered. However, treatment selection by modern diagnostic methods such as contrast-enhanced MRI using liver-specific contrast agent and improved surgical techniques to prevent postoperative complications are required.