The consensus seems to be that there is not apparent value to surgery if residual disease remains, whether it is involvement of resection margins, other distant metastases, or peritoneal carcinosis.
A positive resection margin is also not an independent prognostic factor in colorectal liver metastases because of its strong relationship with the number of tumors resected. In approximately 70% of patients, recurrent disease developed after hepatic resection, most commonly in the liver. Recurrent tumors were more frequently distributed in both lobes than in the resected lobe, suggesting that liver recurrence is more probably derived from multiple metastatic foci from the primary disease than from intrahepatic remetastases of the liver lesion
The paper by Nomura [
46] underlines the role of intrahepatic micrometastases around the liver as a cause of recurrence of the disease. The survival rate of patients with micrometastases was significantly worse than that of patients without. In addition, recurrence after hepatic resection is more strongly associated with systemic spread through vessel infiltration than with local spread through lymphatic or serosal invasion of the primary tumor. A generous surgical margin may not be essential for curative hepatic resection of liver metastases, even if in the study by Ambiru [
9] a margin <10 mm is considered a poor prognostic factor for survival. Nevertheless a positive surgical margin should be avoided and the surgeon should strive to obtain an adequate margin, because this is the only prognostic factor on which the surgeon could have any influence over. According to the pattern of recurrence, relapse developed most commonly in the liver (70%, range 63.6% to 83.3%), indicating that the remaining liver should be a focus for relapse monitoring. Early or high recurrence rates translate into prolonged periods of systemic therapy and potentially reduced quality of life: these will mitigate any short-term advantage resulting from liver resection in LMGC. Data about recurrence made liver resection less appealing. However, the majority of patients analyzed did not receive adjuvant chemotherapy after liver resection. Given the biology and systemic nature of LMGC, this may partially account for the high recurrence rates. A sensible strategy for improving survival could be close observation for a second relapse in the liver and adjuvant chemotherapies after surgery. Trials using the FOLFOXIRI regimen (5-fluorouracil, leucovorin, oxaliplatin, and irinotecan) [
20] achieved a median overall survival of 15 months. We can assume that the recurrence rates could potentially decrease with these regimen in an adjuvant setting.
Timing of hepatic resection
Timing of hepatic resection has been reported to be a significant prognostic factor. In some paper synchronous hepatectomy was a significant poor prognostic factor. Ambiru [
9] reported significantly longer survival in patients with metachronous metastases than in those with synchrounous disease (29%
vs. 6% at 3 years). Bines [
38] commented that synchronous resection of hepatic metastases has little value, whereas metachronous resection of isolated lesions can produce meaningful long-term survival, when the procedure renders the patient disease-free. Some authors suggest that resective treatment may be indicated only for the patient with metachronous isolated metastases [
29,
42]. Other studies did not demonstrate any differences in terms of survival among the groups (Cheon) [
36]. Nevertheless, an analysis of the data reported in the recent literature showed that 22 of 48 5-year survivors underwent a synchronous hepatectomy. In fact in Sakamoto’s study [
43] three of five patients who survived more than 3 years had synchronous solitary metastases and Ochiai too reported three 5-year survivors with synchronous disease. Recently Tsujimoto [
41] confirm the absence of statistically difference between metachronous or synchronous nodules. Thus, synchronous hepatectomy should not be a contraindication for hepatic resection. With regard to perioperative morbidity, Bines [
38], observed that synchronous liver resection carries a higher risk. This may depend on the concern regarding the use of aggressive liver surgery in conjunction with the treatment of gastric cancer under synchronous conditions. The disease-free interval (DFI) between gastric and hepatic resection has been reported to be a prognostic factor. Fujii [
44] showed that a DFI > 1 year has a significant survival advantage, due to the slow growing nature of these tumors.
However, data concerning long-term survivors demonstrate that, if we exclude bilobar spread of metastases (H3), none of the previous cited predictive factors (alone or in combination) can deprive a patient of the possibility of long-term survival after hepatic resection, raising concern about the clinical value of prognostic factors emerging from small and selected populations submitted to liver resection.
The correct approach can be extrapolated from papers that addressed the topic analyzing unselected populations of gastric cancer patients presenting hepatic metastases as only site of metastatic disease (Table
4). From a cohort of 58 patients Cheon [
36] did not identify any primary-related or metastasis-related factor showing prognostic value. The same conclusion is done by group of Makino [
49] from Japan, who studied 63 patients. Ueda [
50] studied a cohort of 73 patients presenting synchronous metastases. Their data show that factors influencing survival were the extent of hepatic involvement (H1-2
vs. H3) and macroscopic peritoneal dissemination (P0
vs. P1) detected at surgical exploration. When focusing on the subgroup of H1-2 and P0 patients, they showed that number (1
vs. >1) and size of hepatic metastases and N status of gastric cancer (N0-1
vs. N2-3) were predictors of survival. An Italian survey performed under the auspices of the Italian Research Group on Gastric Cancer [
51] studied an unselected cohort of 73 patients presenting metachronous metastases after curative gastrectomy. This study demonstrated that the factors T, N, and G of the gastric primary, when rated T3b-T4, N +, and G3, independently display a clear negative prognostic value with an effect that is cumulative. All the above mentioned studies strongly suggest that the main factor influencing long-term survival (
P ranging from 0.01 to 0.001) is the therapeutic approach to the liver metastases, in particular when a surgical approach is performed. In the Italian study hepatectomy was associated to a five-fold increase in survival of less favorable patients (>1 negative prognostic factor) and achieved a 5-year survival rate of 20%. Comparing the treatment options (supportive care, chemotherapy, and liver resection) they found that 1-, 3-, and 5-year survival rates were as follows: 22%, 2%, and 0% for supportive care, 45%, 6%, and 0% after chemotherapy, and 81%, 20%, and 20% after liver resection, with a median survival as follow: non-treated patients 5 months, 12 if chemotherapy was employed and 23 months after surgery. Thus the therapeutic approach to liver metastases displayed independent significant association with survival. In the study of Ueda [
49] in which a group of patients with synchronous LMGC underwent surgery or other treatments (chemotherapy, supportive cures), 1-, 3-, and 5-year overall survival rates were 57%,43%, and 43%, 60%, 0%, and 0%, and 16.7%, 0%, and 0%, respectively. Furthermore in this study adding hepatic artery infusion (HAI) chemotherapy to liver surgery not seem to offer a survival benefit for the patients. The author suggests that only liver surgery, but not HAI, could significantly prolong the survival period in this cohort of patients. Moreover the degree of liver metastases (H1-H2
vs. H3) seems to be a very strong prognostic factor, with a median overall survival of 16.6, 10.2, and 4.4 months, respectively,despite the treatment.
Table 4
Prognostic factors from series considering unselected populations and related survivals
| 58 | Synchronous + metachronous | Overall: 16 | No hepatic resection: 29.4%; 0%; 0% | |
Hepatic resection RFA: 75.3%; 31.7%; 20.8% | Ro resection | | | | |
| 72 | Synchronous | NA | No hepatic resection: 36.4%; 0%; 0% | |
Hepatic resection (HAIC): 80%; 60%; 60% | H; P; R0 resection | | | | |
| 73 | Metachronous | Overall: 7 | | |
BST: 5 | | | | | |
Chemotherapy: 12 | | | | | |
Hepatic resection: 23 | BST: 22%; 2%; 0% | | | | |
Chemotherapy: 45%; 6%; 0% | | | | | |
Hepatic resection; 81%; 20%; 20% | T; N; G of primary | | | | |
R0 resection | | | | | |
| 73 | Synchronous | | | Stage of primary; H |
Extrahepatic disease; treatment of mets | | | | | |
| | Synchronous + metachronous | Overall: 16 | | |
Hepatic resection: 31.2 | No hepatic resection: 53.2%; 4.2%; 0% | | | | |
Hepatic resection: 82.3%; 46.4%; 37.1% | | | | | |
Furthermore, two studies [
36,
49] underlined that R0
vs. R1 operation affects long-term survival (overall 5-year survival rates of 60% and 20%, respectively). Multimodal treatments can further enhance survival rates, in particular if modern chemotherapy protocols are employed. An interesting 75% 5-year survival rate in a subgroup of eight patients submitted to radical surgery followed by hepatic artery infusion chemotherapy has been reported by Ueda. Radio-frequency ablation is another important strategy for the treatment of hepatic metastases from gastric cancer. This ablative technique can be employed either as alternative or in association to hepatectomy. It could be the approach of choice in case of poor general conditions contraindicating surgery. The number of reported procedures is low not allowing to draw any conclusion on its efficacy: moreover follow-up are short and data cannot always be effectively extrapolated from the series. However, Hwang [
52], considered 72 patients with metachronous metastases submitted to different treatments other than hepatectomy (Table
4). The paper showed that 15 patients without extrahepatic disease treated by RFA ± chemotherapy displayed a median survival of 22 months, with 3- and 5- year survival rates of 50% and 40%, respectively, similar to those reported in surgical series (Table
2). Yamakado [
53] and Cheon [
36] reported similar data: in their experience a subgroup of nine patients submitted to RFA compared favorably with 22 patients submitted to radical surgery, with a 4-year survival of 40% and 20%, respectively. Another paper of Kim
et al.[
54] report worse survival results similar to those of classic systemic chemotherapy alone.