Survival curves according to the number of liver nodules for patients with synchronous metastasis and 0 to 3 lymph node metastases around the primary cancer (pN0 or pN1 by the Union International Contre le Cancer staging system).
Survival curves according to the number of liver nodules for patients with synchronous metastasis and 4 or more lymph node metastases around the primary cancer (pN2 by the Union International Contre le Cancer staging system).
Survival curves calculated from the time of initial colorectal resection for patients with synchronous metastasis according to the timing of hepatic resection.
Minagawa M, Yamamoto J, Miwa S, Sakamoto Y, Kokudo N, Kosuge T, Miyagawa S, Makuuchi M. Selection Criteria for Simultaneous Resection in Patients With Synchronous Liver Metastasis. Arch Surg. 2006;141(10):1006–1012. doi:10.1001/archsurg.141.10.1006
While simultaneous resection has been shown to be safe and effective in patients with synchronous metastasis, neoadjuvant chemotherapy followed by hepatectomy has gradually gained acceptance for both initially nonresectable metastasis and resectable metastasis. The boundary between these treatments is becoming unclear. We hypothesized that factors associated with colorectal cancer may play an important role in the prognosis of patients with synchronous metastasis and may be useful for identifying patients who can be expected to have adequate results following simultaneous resection.
Tertiary referral center.
From January 1980 to December 2002, 187 patients underwent curative resection for synchronous liver metastasis from colorectal cancer. One hundred forty-two patients received simultaneous resection, 18 underwent staged resection, and 27 underwent delayed hepatic resection. Twenty-one clinicopathological factors were analyzed, and long-term prognosis was assessed.
Main Outcome Measures
Prognostic factors and patient survival.
There was no in-hospital death. In a multivariate analysis, the factors that significantly affected the prognosis of synchronous metastasis were 4 or more lymph node metastases around the primary cancer (P<.001) and multiple liver metastases (P = .003). In patients with 3 or fewer lymph node metastases around the primary cancer, the 5-year survival rates of those with 1, 2 to 3, and 4 or more liver metastases were 63%, 33%, and 40%, respectively, but these rates were 15%, 22%, and 0%, respectively, in patients with 4 or more lymph node metastases around the primary cancer.
The results support the application of simultaneous resection in patients with 0 to 3 colorectal lymph node metastases. However, in patients with 4 or more colorectal lymph node metastases, biological selection by neoadjuvant chemotherapy may be more suitable.
Surgical resection of liver metastasis from colorectal cancer remains the only therapy with potential for cure.1 The 5-year survival rate after hepatic resection has been reported to be 14% to 40% in studies with 100 or more patients, and the incidence of synchronous metastasis has ranged from 23.0% to 46.8% (Table 1).1- 16
The optimal timing for surgical resection and the upper limits of surgical indications for synchronous metastasis have long been controversial. During the past decade, most studies5,14,17,18 have recommended a staged operation with initial resection of the colorectal primary cancer followed by hepatic resection 2 to 6 months later. However, the paradigm for the treatment of synchronous colorectal metastasis has recently begun to change. Similar results regarding mortality and morbidity in selected patients have been reported after simultaneous and staged hepatic resection; thus, simultaneous resection has gained acceptance in some institutions because of its safety and efficiency.16,19- 23 On the other hand, neoadjuvant chemotherapy has been found to be effective not only for initially nonresectable metastasis but also for resectable synchronous metastasis.24- 27 Thus, the boundary of simultaneous resection and neoadjuvant chemotherapy followed by hepatectomy is becoming unclear. Moreover, guidelines regarding the upper limits of operative indications for synchronous metastasis have not been defined, mainly because the distinctive prognostic factors of synchronous metastasis, not including metachronous metastasis, have not been fully evaluated in medium or large series.
The primary goal of this study was to elucidate the indications for simultaneous hepatic resection of synchronous colorectal metastasis from a prognostic point of view.
From January 1980 to December 2002, 388 patients with hepatic metastasis from colorectal cancer underwent liver resection at the Department of Surgery, National Cancer Center, Tokyo, Japan (1980-1990), the First Department of Surgery, Shinshu University, Matsumoto, Japan (1990-1994), and the Division of Hepato-Biliary-Pancreatic Surgery, Department of Surgery, University of Tokyo, Tokyo (1994-2002). One of us (M. Makuuchi) participated in all of these operations. Selection criteria for surgery were the possibility of an oncologically radical operation and the possibility of preserving at least 40% of the normal hepatic parenchyma. After January 1990, metastasis with hepatic lymph node metastasis was excluded. The total number of hepatic metastases, their unilateral or bilateral presentation, and the existence of extrahepatic metastasis were not considered exclusion criteria. Since January 1980, our policy for synchronous metastasis has been simultaneous resection regardless of the number and extent of liver metastases and the location of the primary cancer.
In all of the cases, the preoperative diagnostic workup included ultrasonography and plain and contrast-enhanced computed tomography to stage the liver involvement as well as chest radiography, chest computed tomography, barium enema, colonoscopy, and bone scintigraphy to assess the presence or absence of extrahepatic disease. Intraoperative bimanual liver palpation and intraoperative ultrasonography were also carried out in all of the cases, and all of the resections were guided by intraoperative ultrasonography.
Of the 388 patients, resections were not radical in 19 because of gross residual disease within or outside the liver, and the remaining 369 patients received curative resection. In these 369 patients, 187 had synchronous liver metastasis and the remaining 182 had metachronous liver metastasis. In this article, synchronous metastasis is defined as nodules that had been diagnosed before the primary colorectal surgery or found at primary surgery. Of the 187 patients with synchronous metastasis, 148 were admitted to our hospital before primary colorectal surgery: 142 patients underwent simultaneous resection (simultaneous group; rate of simultaneous resection, 96%), 3 received a staged operation (staged group) because 1 had extensive peritoneal dissemination at primary surgery and 2 had general complications, and the remaining 3 had initially refused hepatic resection and then received local chemotherapy via the hepatic artery followed by delayed hepatic resection (delayed group). Thirty-nine patients were transferred to us after resection of the primary cancer. Of these patients, 15 had only undergone colorectal resection (staged group) and 24 had undergone resection of the primary cancer and chemotherapy or other treatment for hepatic metastasis (delayed group). Accordingly, of the 187 patients with synchronous metastasis, 142 received simultaneous resection (simultaneous group), 18 underwent primary colorectal resection followed by hepatic resection without treatment of the liver nodule prior to hepatectomy (staged group), and 27 underwent resection of the colorectal cancer and received treatment for the liver deposit followed by hepatic resection (delayed group). The treatment modalities prior to hepatectomy for those in the delayed group were local chemotherapy in 14, systemic chemotherapy in 8, combined systemic and local chemotherapy in 1, irradiation and systemic chemotherapy in 1, ethanol injection and local chemotherapy in 1, microwave coagulation and local chemotherapy in 1, and radiofrequency ablation and systemic chemotherapy in 1. The median interval between the resection of colorectal cancer and hepatectomy was 6.4 months (range, 2.3-26.8 months) in the delayed group and 2.0 months (range, 0.5-15.4 months) in the staged group. Based on the Couinaud anatomical classification of the liver, the types of resection in these patients and the location of the primary colorectal cancer according to the timing of hepatectomy are shown in Table 2.
Survival was measured from the time of hepatic resection, and death was the end point. Survival curves were constructed using the Kaplan-Meier product-limit method and compared using a log-rank test. Significant prognostic factors in a univariate analysis were entered into a Cox proportional hazards model using stepwise selection to identify independent predictors of death. Statistical significance was defined as P<.05. The SAS version 8 software (SAS Institute, Inc, Cary, NC) was used for statistical analyses.
The mean (95% confidence interval [CI]) follow-up period was 4.08 years (3.68-4.48 years). There was no in-hospital death. The 3-, 5-, and 10-year survival rates in the 187 patients with synchronous colorectal liver metastasis were 49%, 35%, and 25%, respectively, and the rates in the 182 patients with metachronous metastasis were 55%, 41%, and 28%, respectively. No significant differences were found (P = .19).
The effects of 13 clinicopathological factors at hepatic resection (Table 3) and 8 at primary colorectal resection (Table 4) on survival after curative hepatic resection were analyzed. Significant factors were the number of lymph node metastases around the primary cancer (P<.001), a carcinoembryonic antigen level of 50 ng/mL or higher (P = .002), a resection margin of the liver nodule smaller than 5 mm (P = .004), extrahepatic invasion (P = .005), the number of liver metastases (P = .03), the depth of wall invasion by the primary cancer (TNM classification of pT4 according to the Union Internationale Contre le Cancer) (P = .04), and lymphatic duct invasion by the primary cancer (P = .04). In a multivariate analysis using the Cox proportional hazards model, independent prognostic factors were 4 or more lymph node metastases around the colorectal cancer (relative risk, 1.58; 95% CI, 1.26-1.96; P<.001) and multiple liver metastases (relative risk, 1.49; 95% CI, 1.14-2.00; P = .003) (Table 5).
The patients with synchronous metastasis were divided into those who had 0 to 3 (for pN0 and pN1) and 4 or more (for pN2) lymph node metastases around the primary cancer. Among patients with pN0 and pN1 cancer, the median (95% CI) survival times were 7.2 years (3.9 years to not calculated; survival curve remains above a survival rate of 50%) for 36 patients with 1 liver metastasis, 3.1 years (1.7-5.3 years) for 38 patients with 2 to 3 liver metastases, and 3.3 years (2.3-5.8 years) for 42 patients with 4 or more liver metastases (P = .02 for 1 vs 2-3 liver metastases; P = .06 for 1 vs ≥4 liver metastases; and P = .59 for 2-3 vs ≥4 liver metastases) (Figure 1). Among patients with pN2 cancer, the median (95% CI) survival times were 1.8 years (1.4-2.8 years) for 18 patients with 1 liver metastasis, 2.1 years (0.9- 2.3 years) for 20 patients with 2 to 3 liver metastases, and 1.6 years (0.7-2.3 years) for 17 patients with 4 or more liver metastases (P = .84 for 1 vs 2-3 liver metastases; P = .38 for 1 vs ≥4 liver metastases; and P = .36 for 2-3 vs ≥4 liver metastases) (Figure 2).
With regard to the timing of hepatectomy in patients with synchronous metastasis, there was no significant difference among the simultaneous, staged, and delayed groups. The median (95% CI) survival times, which were calculated from the time of colorectal resection, were 3.1 years (2.4-4.2 years) for the 142 patients in the simultaneous group, 2.6 years (2.3-7.9 years) for the 18 patients in the staged group, and 2.4 years (2.0-4.4 years) for the 27 patients in the delayed group (P = .95 for simultaneous group vs staged group; P = .36 for staged group vs delayed group; and P = .32 for simultaneous group vs delayed group) (Figure 3).
Synchronous colorectal liver metastases are found in 20% to 30% of patients at the time of the initial diagnosis of colorectal cancer.4 However, only 20% of these patients can receive curative hepatic resection.4 One of the reasons for this low rate must be the fact that surgical guidelines regarding the upper limit of surgical aggressiveness have not been defined. The surgical indications for this large group of patients and the optimal timing of hepatectomy are still controversial and widely debated.
Most studies on the surgical management of synchronous liver metastasis have recommended a staged operation. Nordlinger et al5 found that simultaneous resection should be limited when liver metastasis could be removed by minor hepatectomy and both of the lesions could be operated on through the same abdominal incision, as the mortality rate was 6.9% for simultaneous resection but 2.4% for a staged operation when major hepatectomy was necessary. Similar results were found by Bolton and Fuhrman14: the mortality rate was 23.5% for simultaneous resection of 4 or more unilobar metastases or multiple bilobar metastases and 0% for a staged operation. Based on these results, they recommended a staged operation for these hepatic lesions, with a delay of at least 3 months after colon resection. On the other hand, Jaeck et al22 proposed simultaneous resection especially for patients with a primary tumor of the ascending colon with metastases resectable by means of minor hepatectomy. With regard to colorectal cancer, it was generally considered that simultaneous resection was not suitable for left colon or rectal cancer because of a high rate of anastomotic complications. Elias et al28 showed that simultaneous resection could be made safe for patients with left colon or rectal cancer by making a temporary colostomy. Gradually, safety of simultaneous resection has been shown in selected patients and the indication for simultaneous resection in synchronous metastases has been extended.16,21,23 Weber et al19 described the safety of simultaneous resection in 35 patients compared with the safety of a staged operation in 62 patients. Their selection criterion was fewer than 4 unilobar metastases in the preoperative radiological assessment regardless of the location of the primary cancer. Martin et al20 described the safety and efficacy of simultaneous resection in patients with synchronous metastasis by comparing 134 patients who underwent simultaneous resection and 106 who underwent a staged operation. Simultaneous resection tended to be performed for right colon primary cancers and smaller and fewer metastases. Simultaneous resection for synchronous metastasis regardless of the location of the primary cancer and the extent of liver metastases has been our policy since January 1980. Based on this treatment policy, 142 of 148 patients with synchronous metastasis who were admitted to our hospital before resection of the primary cancer actually underwent simultaneous resection. No operative or in-hospital mortality was observed in our series.
Since simultaneous resection has been shown to be safe, surgeons have had many options for synchronous metastasis: simultaneous resection, staged operation, and neoadjuvant chemotherapy followed by hepatectomy. However, the timing of hepatectomy for synchronous metastasis that is most beneficial for the patient's survival has not been fully evaluated. In our series, no difference in survival was observed between 142 patients who underwent simultaneous resection and 18 who underwent a staged operation. Three retrospective studies19,29,30 showed similar survival rates in a simultaneous resection group and a staged resection group. In patients who underwent resection of colorectal cancer in another hospital or who had severe general complications, a delay of 2 to 3 months until hepatectomy may not have significantly affected survival.
In patients with unresectable liver metastases, effectiveness of preoperative chemotherapy has been shown; among 1104 patients with unresectable metastases, 138 showed good response to preoperative chemotherapy and received liver resection, and the 5-year survival rate of those patients was 33%.25 In patients with resectable liver metastases, the role of neoadjuvant chemotherapy has still been controversial. In an article by Allen et al27 concerning the role of neoadjuvant chemotherapy in 106 patients with resectable synchronous metastasis, no difference in survival was observed between 52 patients who did and 54 patients who did not receive neoadjuvant chemotherapy. On the other hand, Tanaka et al24 showed that neoadjuvant chemotherapy improved the survival of patients with multiple bilateral metastases. In our study, neoadjuvant chemotherapy followed by hepatectomy was not a significant prognostic factor. Of course, to justify the role of neoadjuvant chemotherapy in these patients, randomized controlled studies are necessary.
In liver metastasis as a whole, including synchronous and metachronous nodules, significant prognostic factors have varied between studies. However, all of the studies have agreed that lymph node involvement of the hepatic hilum is a significant factor in a poor prognosis and that sex does not influence survival.13 These results may be owing to the different proportions of synchronous and metachronous metastases in these studies. The rate of synchronous metastasis in studies1- 16 with 100 or more patients has ranged from 23.0% to 46.8%, and this value was 50.7% in our series (Table 1). If the prognostic factors for synchronous and metachronous metastases are different, it is essential to define the significant factors in patients with synchronous metastasis to establish operative indications for synchronous metastasis. Sugawara et al31 described the prognostic factors for synchronous and metachronous metastases separately. The resection margin was the only significant factor in patients with synchronous metastasis, but the factors for primary cancer were not assessed. Fujita et al32 analyzed the data of 97 patients with synchronous metastasis and found that the most significant factor was lymph node metastasis of the primary cancer, followed by a positive pathological margin of liver metastasis. Interestingly, the number, size, and distribution of liver nodules were not significant prognostic factors in these articles. These results are partly compatible with our own. In our series, patients with single nodules had a favorable prognosis, with a median survival time of 4.0 years, although 63 patients with 2 to 3 nodules and 66 with 4 or more lesions showed almost the same survival rates and the diameter and distribution of liver nodules did not influence the prognosis. The patient's prognosis in our series was much more influenced by factors regarding the primary cancer: 55 patients with 4 or more lymph node metastases around a colorectal cancer showed a significantly worse prognosis than 47 patients without this condition and 69 patients with 1 to 3 metastases. The depth of wall invasion and lymphatic duct invasion of the primary cancer were also significant factors. These results suggest that the factors regarding liver metastasis are insufficient for considering the operative indications for patients with synchronous colorectal metastasis. Factors regarding the primary cancer play a more important role.
It is generally considered that major hepatectomy is not indicated for patients with peritoneal dissemination. Sugarbaker and Jablonski33 showed that peritoneal carcinomatosis was a treatable condition in selected patients; patients who received complete resection and intraperitoneal chemotherapy had a 3-year survival rate of 65% to 99%. Importance of complete cytoreductive surgery was shown in patients with peritoneal carcinomatosis and distant metastasis.34,35 In our series, 25 patients with peritoneal carcinomatosis had a median survival of 2.2 years with curative resection (Table 4). This condition was not a significant poor prognostic factor. If the number of peritoneal disseminated nodules is relatively small and they are curatively resectable, we usually perform hepatic resection with removal of disseminated nodules; of course, major hepatectomy can be indicated if it is necessary.
In a multivariate analysis in our series, the most important factor in the prognosis of patients with synchronous metastasis was 4 or more lymph node metastases around the primary cancer (pN2) (P<.001), and this was followed by having multiple liver metastases (P = .003). Patients with synchronous metastasis and 3 or fewer lymph node metastases around the primary cancer (pN0 or pN1) showed a favorable prognosis regardless of the number of liver metastases: 5-year survival rates were 63% for 1 metastasis, 33% for 2 to 3 metastases, and 40% for 4 or more metastases (Figure 1). The survival rate did not tend to decrease with an increase in the number of liver nodules. Such patients should undergo hepatic resection, but the magnitude of the operative procedure must be carefully considered based on the physiological condition of the patient as well as on the skills and experience of the operative center. Resection with reconstruction of the inferior vena cava and preoperative portal embolization can be applied.36- 38 Among patients with synchronous metastasis and 4 or more lymph node metastases around the primary cancer (pN2), 18 patients with 1 liver metastasis had a 5-year survival rate of 15% and a median survival time of 22 months, 20 patients with 2 to 3 liver metastases had a 5-year survival rate of 22% and a median survival time of 25 months, and 17 patients with 4 or more liver metastases had a 5-year survival rate of 0% and a median survival time of 19 months (Figure 2). Recent articles39 have shown a median survival time of 14.8 months by intrahepatic arterial chemotherapy and 14.7 months by systemic chemotherapy. Saltz et al40 also reported a median survival time of 14.8 months by irinotecan hydrochloride plus fluorouracil and leucovorin calcium. The median survival time was found to be 7.5 months for 484 untreated patients but 12.7 months and 11.1 months for patients who underwent regional and systemic chemotherapy, respectively.41 With 3 or fewer synchronous metastases, patients with pN2 primary cancer showed a better prognosis with hepatic resection than with chemotherapy, which supports surgical resection. However, in patients with 4 or more synchronous metastases and pN2 primary cancer, the prognosis by resection was better than the natural history but was almost the same as that with chemotherapy. Thus, the boundary of surgical indications for colorectal liver metastasis must be located around here.
To execute this selection criterion, evaluation of the nodal status around colorectal cancer is essential. By preoperative computed tomography, the nodal stage was accurately diagnosed in 80% to 85% of patients.42- 44 In addition, intraoperative manual palpation will be helpful; consequently, most patients will be correctly diagnosed at the nodal stage at primary operation. Of course, the patients who are diagnosed as having 3 or fewer lymph node metastases should receive simultaneous resection, but the patients who are suspected to have 4 or more lymph node metastases should undergo staged resection because the staged operation does not deteriorate the prognosis. After the pathological diagnosis is disclosed, the patients who have 3 or fewer lymph node metastases (pN0 or pN1) should receive hepatic resection approximately 2 weeks after the primary resection. However, in patients with 4 or more colorectal lymph node metastases (pN2), biological selection by neoadjuvant chemotherapy may be more suitable.
For patients with synchronous metastases, a single operation will be preferred if it is safe and superior in a prognostic aspect. Simultaneous resection could be done regardless of the location of the primary cancer and the extent of liver metastasis without in-hospital mortality. Our results support the application of simultaneous resection in patients with 0 to 3 colorectal lymph node metastases (pN0 or pN1).
Correspondence: Masatoshi Makuuchi, MD, PhD, Divisions of Hepato-Biliary-Pancreatic Surgery and Artificial Organ and Transplantation, Department of Surgery, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan (firstname.lastname@example.org).
Accepted for Publication: August 11, 2005.
Author Contributions:Study concept and design: Minagawa, Kokudo, and Miyagawa. Acquisition of data: Minagawa, Yamamoto, Miwa, Sakamoto, and Kosuge. Analysis and interpretation of data: Minagawa and Makuuchi. Drafting of the manuscript: Minagawa, Miwa, Sakamoto, Kokudo, and Kosuge. Critical revision of the manuscript for important intellectual content: Minagawa, Yamamoto, Miyagawa, and Makuuchi. Statistical analysis: Minagawa and Miwa. Obtained funding: Minagawa and Sakamoto. Administrative, technical, and material support: Minagawa, Yamamoto, Sakamoto, and Kosuge. Study supervision: Kokudo, Miyagawa, and Makuuchi.