Clinicopathological analysis of recurrence patterns and prognostic factors for survival after hepatectomy for colorectal liver metastasis

Despite recent improvements in the diagnosis and management of colorectal cancer (CRC), which have enabled early detection followed by early treatment, many advanced cases with hepatic or peritoneal metastasis are still encountered. For further improvement of the prognosis of CRC, it is particularly important to prolong the survival of these advanced cases with distant metastasis.

Hepatectomy for liver metastasis from colorectal cancer (CRCLM) is recommended as the most effective therapy [1‐11]. However, to date, 50% to 75% of patients develop recurrence of the disease after curative resection of CRCLM. Moreover, the cure rate by initial hepatectomy is only 20% to 30% of cases [4‐6, 12]. Factors associated with recurrence and prognostic determinants after initial hepatectomy are still controversial. Recently, many studies reported that the duration of survival is markedly prolonged by effective chemotherapies, including the molecular target drugs, for unresectable or recurrent colorectal cancer [13, 14]. Therefore, it is necessary to elucidate predictive factors relevant to recurrence and survival in order to determine optimal treatment strategies for each patient with CRCLM.

The aim of this study was to retrospectively investigate surgical outcomes in a consecutive series of patients undergoing hepatectomy for CRCLM at a single institution, in order to analyze recurrence patterns and clinicopathological prognostic factors for survival.

While untreated CRCLM has a poor prognosis with median survival ranging from 6 to 12 months and is not expected to be over 5 years, long-term survival and potential for cure following surgical resection for CRCLM had been demonstrated in numerous uncontrolled studies [7, 8, 12]. These studies reported that the overall 5-year survival rates are in the range of 20% to 58%, and the median survival times are 24-46 months [4, 6‐8, 10‐12, 16, 17]. However, it has also been reported that 57% to 78% of those patients will develop a recurrence of the disease in the hepatectomy series, and intrahepatic recurrence occurs in approximately 50% of patients [4, 6, 9, 18]. It has been considered that occult metastasis surfacing from the primary CRC and residual lesion scattered from liver metastasis are the two major pathways through which hepatic recurrence of metastatic lesion occurs after initial hepatectomy [19, 20]. Therefore, treatment strategies including hepatic resection should be determined on the basis of these mechanisms for recurrence of metastasis. In Japan, surgical procedures for CRCLM are recommended if the following conditions are met: (1), the primary tumor was curatively resected; (2), metastasis is located only in the liver; (3), the patient is in a condition to be able to bear hepatectomy; and relative indications include: (4), extrahepatic metastasis can be controlled when present; and (5), reductive operation is a part of multimodal treatment [21]. However, the detailed consensus in terms of ideal timing, type, and extent of hepatectomy and optimal combination with perioperative chemotherapy are not currently provided with sufficient evidence. The aims of this study are to retrospectively evaluate the significant prognostic factors for survival and risk factors for recurrence in patients who underwent hepatectomy, and thereby to determine the optimal timing and method of hepatectomy with concurrent use of perioperative chemotherapy.

In the present study, prognostic factors on univariate analysis were the depth of the primary CRC, portal vein invasion and surgical margin at hepatectomy for CRCLM, and the presence of intra- and extrahepatic recurrence after the initial hepatectomy. Independent prognostic factors on multivariate analysis were the presence of intra- and extrahepatic recurrence. Of these factors, the portal vein invasion and surgical margin status at hepatectomy are important factors through which surgery can improve prognosis.

There have been several reports on the risk factors of intra- and extrahepatic recurrence [7, 22‐24]. In the present study, an independent risk factor for intrahepatic recurrence was the hepatic resection margin, while there were no risk factors regarding extrahepatic recurrence. Hematogenous dissemination of CRC was reported to be significantly associated with the size of liver metastatic nodules by the cascade theory [25]. In this theory, in accordance with the process of CRCLM metastasizing to the lung and subsequently to the other organs, the size of CRCLM is thought to increase. However, the present study did not include the size or the number of CRCLM as risk factors for intra- or extrahepatic recurrence, implicating the significance of surfacing micrometastasis from the primary CRC which could not be detected perioperatively.

The surgical margin of < 5 mm at hepatectomy was detected as a risk factor for subsequent intrahepatic metastasis in this study. The role of surgical margin status as a prognostic factor to predict posthepatectomy survival has been controversial [24, 26‐29], and in the largest single-center series, Are et al. reported that a margin width > 1 cm was an independent predictor for better survival and is optimal [28]. Nuzzo et al. also recently reported that a histological surgical margin of ≤ 5 mm was associated with lower overall and disease-free survival rates independent of other clinic-pathologic factors [29]. In contrast, Bodingbauer et al. demonstrated that the rate of recurrence at the surgical margin was low and a positive margin was not associated with an increased risk of recurrence either at the surgical margin or elsewhere, if hepatectomy was performed with ultrasonic dissector by experienced surgeons [24]. The present study is distinct from previous studies because macroscopic, instead of histological, surgical margin utilized as a parameter. A macroscopic surgical margin rather than microscopic margin was adopted, since 1) use of various dissection and coagulation devices, such as soft-coagulation system, ultrasonic dissector, and bipolar electrocautery during hepatic parenchymal transection [15], would potentially hinder an accurate assessment regarding pathological surgical margin status including R0 (no residual tumor) and R1 (microscopic residual tumor) status [24, 29]; and 2) the macroscopic surgical margin was considered a better parameter of clinical usefulness for patient and treatment selection. Namely, during preoperative evaluation, whether a surgical margin of ≥ 5 mm was macroscopically feasible or not, it was utilized as a practical marker to determine which was indicated first as an initial therapy for each patient: hepatectomy or neoadjuvant chemotherapy. If the surgical margin was estimated to be < 5 mm, preoperative chemotherapy is the best option [30]. It does need to be noted that resection should not be precluded whatever the width of the surgical margin, since no other single treatment modality can surpass hepatectomy even with a 0- to 1-cm surgical margin [28]. In the present study, surgical margin status was determined as a prognostic factor on univariate, but not on multivariate analysis. These findings are in accordance with the report by Kokudo et al. that found the surgical margin was correlated with tumor recurrence but not with survival rate [26]. However, when an insufficient surgical margin is suspected during hepatectomy, or postoperative pathology of the resected specimen reveals inadequate surgical margin, adjuvant chemotherapy should be started to prevent intrahepatic recurrence, though a prospective study on this approach is necessary to make definite conclusions.

Portal vein invasion in the resected liver was observed in a small number of patients (9.8%), but also showed significant negative impact on survival, suggesting it could reflect the grade of aggressiveness and invasiveness of each tumor or the potential for intrahepatic micrometastases, as has been implicated by some authors [29, 31] in the context of surgical margin status as a surrogate marker for malignant potential. Therefore, optimal surgery should include a strategy against portal vein invasion, such as anatomical hepatic resection including Glisson's sheath with a sufficient margin, and aggressive adjuvant chemotherapy for this subgroup of patients [32]. Further studies are needed to clarify its clinical significance.

Regarding the timing of hepatectomy for synchronous CRCLM, controversy remains on whether to perform simultaneous resection or to include an observation period for a couple of months. No differences between these two approaches were detected in the present study, suggesting that in synchronous situations, surgery should be planned at least before hepatic tumor(s) become unresectable and that patients need to be treated with adjuvant chemotherapy immediately, since this subgroup appeared to benefit from adjuvant chemotherapy.

A further point of concern is multiple bilobar CRCLM lesions which precipitate the most challenging situation for hepatic surgeons. Although some portion of this patient cohort would indeed benefit from aggressive chemotherapy and subsequent hepatectomy after successful down-sizing of those lesions, many of these patients are not candidates for surgical intervention. Appropriate application and adequate combination of modalities, such as radiofrequency ablation [33], percutaneous transhepatic portal embolization [34], and two-stage hepatectomy are expected to extend the surgical indication for these patients [35].

There is not yet a sufficient consensus as to whether the perioperative chemotherapy is significantly associated with disease-free survival or prognosis [36]. In the present report, a combination of surgical therapy and adjuvant chemotherapy after hepatectomy controlled intrahepatic recurrence and consequently improved the prognosis significantly, but only in the subgroup with synchronous CRCLM. Conversely, in the patient subgroup with synchronous CRCLM, adjuvant chemotherapy would be essential. As for metachronous disease, no benefits from chemotherapy were detected in this study. However, it should be noted that the chemotherapy in the present series mainly depended on UFT + LV, TS-1, FOLFOX and FOLFIRI, and the recently developed molecular targeting agents, such as bevacizumab and cetuximab, which have been reported to improve the prognosis of recurrent and unresectable CRC [14, 37‐39], were not included in this study. Since use of these newer chemotherapeutic agents before hepatectomy is expected to control recurrence by extermination of the micro cancer cell, a similar study is currently being conducted on those patients who received these newer chemotherapeutic agents.

In CRCLM, intra- and extrahepatic recurrence were independent prognostic factors, and independent risk factors for intrahepatic recurrence included macroscopic resection margin during hepatectomy. Combination of surgery and adjuvant chemotherapy for synchronous CRCLM could control intrahepatic recurrence and significantly improve prognosis. Considering the outcomes of treatment for CRCLM are not yet satisfying, extermination of the micro cancer cell should be achieved by introduction of more potent chemotherapeutic agents in combination with optimal surgery. Further studies are needed to clarify this matter.