Time interval-based indication for liver resection of metastasis from pancreatic cancer

Background Surgical indications for liver metastases from pancreatic ductal adenocarcinoma (PDAC) are lacking because outcomes are usually poor. However, liver resection and the recent progress in perioperative chemotherapy have been observed to improve survival. Methods We performed liver resection for liver metastases from PDAC only under the following criteria: (1) liver-only metastasis, (2) up to three tumors, and (3) no increase in the number of metastases during the 3-month observation period. No limitations were placed on the location or size of liver metastasis. In this study, we aimed to validate our surgical criteria and analyze factors affecting survival in patients with PDAC. Results Seventy-nine patients underwent curative resection for PDAC between 2005 and 2015. Seventy-one patients experienced recurrence, with liver-only recurrence in 17 patients. Among these, nine patients underwent liver resection and eight did not. The median survival time was significantly better for patients who underwent liver resection (55 months) than for those with other recurrences (17.5 months, p = 0.016). The median survival after liver recurrence was significantly better in the liver resection group (31 months) than in the non-liver resection group (7 months, p = 0.0008). The median disease-free interval (DFI) after pancreatectomy was significantly longer in the liver resection group (21 months; range, 3–44 months) than in the non-liver resection group (3 months; range, 2–7 months; p = 0.02). Conclusion Good indications for liver metastases from PDAC include solitary metachronous tumors and longer DFIs.

Recently, advances in perioperative management have allowed liver resection for non-colorectal cancers [9,10]. However, deducing the surgical indications for the resection of liver metastases from each type of cancer has been difficult owing to the limited number of participants. Moreover, liver resection after pancreaticoduodenectomy is complex because biliary reconstruction makes hepatic inflow occlusion during liver transection difficult [4,11,12]. Therefore, liver resection for PDAC has rarely been studied, and factors affecting survival have been poorly documented.
We have been performing liver resection for metastases from PDAC using our unique criteria over the past decade. Based on our observations, we have compiled this article to highlight the benefits of liver resection for liver-only metastasis from PDAC and validate the effectiveness of our criteria based on indications.

Patients
Between 2005 and 2015, data were collected from patients who underwent pathologically curative resection for PDAC. All patients were followed up every 4 months after pancreatic resection to measure tumor marker levels and underwent to monitor the tumor's growth using contrastenhanced computed tomography (CT). When a local or distant metastasis was suspected, magnetic resonance imaging or positron emission tomography-CT was additionally performed every month during follow-up. A recurrence was confirmed when two or more modalities or a recurrent lesion showed visible recurrence at two different time points.

Indications for liver resection
When metachronous liver metastases were found in up to three tumors, liver resection was considered after a 3-month observation period. Patients were classified into two groups: the liver resection and the non-liver resection groups. Indications for the liver resection group were as follows: (1) liver-only metastasis, (2) up to three tumors, and (3) no increase in the number of metastases during the 3-month observation period. There were no limitations on the location or size of the tumors. When the patient received adjuvant chemotherapy, liver resection was planned after an interval of ≥ 3 weeks from the last chemotherapy. All patients showed Child class A liver function and no contraindications for general anesthesia. The nonliver resection group comprised patients not meeting the above criteria; they received the standard systemic chemotherapy. Indications for a second liver resection were the same as those for the first liver resection. However, a second liver resection for liver-only metastasis was considered only for solitary recurrence after the first liver resection with a disease-free interval (DFI) ≥ 12 months.

Liver resection
Liver resection was performed using the clamp-crushing method, and perioperative care was carried out according to institutional protocols as described by Yamazaki et al. [13]. Non-anatomical partial resection is the first-line procedure for liver metastases. When hepatic inflow occlusion was impossible because of pancreaticoduodenectomy, liver resection was performed without Pringle's maneuver. The median total follow-up time after pancreatectomy was 55 months (range, 15-119 months).

Assessment
Patient characteristics were compared between the liver resection and non-liver resection groups for liver-only metastasis. To assess the benefits of liver resection, the overall survival was compared between the liver resection group and all other patients showing recurrence. Finally, to identify the impact of resection in liver-only recurrence, survival times from the time of recurrence were compared between the liver and non-liver resection groups. Complications were ranked using the Clavien-Dindo classification system as severe (grade IIIb or higher). The institutional review board attested all the protocols in this study (registration number RK-150414-4). Informed consent was obtained from all the patients.

Statistical analysis
Data are expressed as medians and ranges or as absolute values and percentages. Student's t test, the χ 2 test, and Fisher's exact test were used for univariate analysis. Survival curves were obtained using the Kaplan-Meier method, and distributions were compared using the log-rank test. P values < 0.05 were considered indicative of statistical significance. All statistical analyses were performed using the software, JMP Statistical Discovery version 13.2 (SAS Institute, Cary, NC).

Pancreatectomy and liver metastases
Between 2005 and 2015, 79 patients underwent pathologically curative resection for PDAC. The baseline characteristics of all patients are shown in Table 1. Twenty patients (25.3%) received adjuvant chemotherapy after pancreatectomy with S-1 or gemcitabine. This study excluded pathologically diagnosed intraductal papillary mucinous carcinoma because its biological recurrence rate differed from that of PDAC. Of the 79 patients in our study, 71 experienced recurrences during follow-up ( Fig. 1), including 17 patients with liver-only metastasis. Of the 17, liver resection was performed only for nine patients, the liver resection group, classified based on indications. The remaining eight, the non-liver resection group, did not undergo liver resection; four had ≥ 4 tumors, and four exhibited increasing tumor numbers during the observation period.

Characteristics of patients with liver-only recurrence
Comparing the 17 patients with liver-only metastasis, we observed that the median DFI after pancreatectomy was significantly longer in the liver resection group than in the non-liver resection group (21 months, range [3-44 months]; 3 months, range [2-7 months], respectively; p = 0.017) ( Table 2). The number of tumors was significantly higher in the non-liver resection group owing to our criteria for resection (based on indications). No other significant differences in patient characteristics were observed. Although liver resection was performed without Pringle's maneuver in 6 patients with pancreaticoduodenectomy, intraoperative blood loss was minimal (median, 246.5 ml; range, 53-508 ml), and the operation time (median, 313 min; range, 88-413 min) was not extended. No patient had severe complications according to the Clavien-Dindo classification.

Survival
The median overall survival was significantly better in the liver resection group than in all the other patients who had a recurrence (55 months vs. 17.5 months, p = 0.016) (Fig. 2). Regarding liver-only metastasis, the median survival time after liver recurrence was significantly better in the liver resection group than in the non-liver resection group (31 months vs. 7 months, p = 0.0008) (Fig. 3). In the liver resection group, the median survival after liver resection was 25 months (range, 10-70 months) ( Table 3). Two patients underwent two liver resections based on the same criteria as those defined for the first resection.

Discussion
This study revealed that patients with up to three metachronous liver-only metastases from PDAC may be good candidates for liver resection. The observation time and DFI were key to optimizing the selection of patients for liver resection.
Regarding the timing of the operation, a constant number of tumors during the 3-month observation period and a long DFI may be indispensable indications. Four of the eight patients in the non-liver resection group presented an increase in the number of liver  However, multiple liver metastases were found during the subsequent 3 months. One patient was found with synchronous liver-only metastasis (discovered incidentally during the operation). Besides, studies in the past have revealed that liver resection for synchronous liver metastasis has not been beneficial to patients [14][15][16]. Moreover, a significantly longer DFI was observed in the liver resection group than in the non-liver resection group. This may reflect tumor aggressiveness and malignant potential in the primary PDAC [17,18]. Therefore, a longer DFI appears to be an important factor for improving our selection criteria. Regarding the participants who underwent the operation, no reports have mentioned long-term survivors after two liver resections from PDAC. In this study, two  patients underwent two liver resections and survived for more than 7 years after pancreatectomy. They also had a longer DFI after pancreatectomy and the first liver resection. The indications for the second liver resection were the same as those for the first one. Therefore, our criteria seem mandatory for the optimal selection of patients for a second liver resection.
In general, operations after pancreaticoduodenectomy are complex because the occlusion of blood inflow is difficult. This may result in a higher blood loss during liver transection. In our study, none of the cases showed an overly long operation time or excessive intraoperative blood loss. Recent technical advances in liver resection have allowed safer liver transection in patients with prior pancreaticoduodenectomy [11,12]. To avoid huge blood loss, we performed total liver mobilization and lifting of the liver transection site to the ventral side when the tumor was located in the right lobe. This method was effective in controlling intraoperative blood loss. Therefore, this study showed that liver resection for PDAC metastasis was reasonably safe.
A limitation of this study, however, was the small sample size attributed to the rarity of the pathological entity. Therefore, this study may include selection bias in terms of treatment. Therefore, further data collection is needed to analyze the effectiveness of liver resection for PDAC metastases.

Conclusions
Liver resection improves survival in patients with recurrent liver-only metastasis from PDAC. We believe that our criteria for selecting patients for liver resection based on indications work reasonably well.