Survival after Combined Hepatic Metastasectomy and Chemotherapy in Patients with Concurrent Hepatic and Extrahepatic Colorectal Metastases: a SEER Database Analysis

Purpose Colorectal cancer is the second leading cause of death among all cancers worldwide. Hepatic metastases exist in approximately 50% of colorectal cancer patients. The purpose of this study was to assess the effect of combined hepatic metastasectomy and chemotherapy on overall survival in patients with concurrent hepatic and extrahepatic disease. Methods A total of 2533 patients from the US Surveillance, Epidemiology, and End Results (SEER) database with concurrent colorectal liver metastasis (CRLM) and extrahepatic disease (EHD) between January 1, 2010, and December 31, 2014, were retrieved. Survival analysis with Kaplan-Meier and Cox regression analyses was performed to assess the effect of combined hepatic metastasectomy and chemotherapy on 5-year survival.

However, with the advances that medicine has witnessed over the past few decades, the median survival of CRLM patients has successfully jumped to forty months with the implementation of liver resection [7].
Meanwhile, the presence of a concurrent extrahepatic disease (EHD) in patients with CLRM has long been viewed as an absolute contraindication for liver resection. Nevertheless, more patients with EHD are becoming eligible for liver resection with the development of better surgical techniques [8,9]. Certain clinicopathological features have been identi ed that suggest which patients could bene t from liver resection. For instance, a study [10] proposed ve factors that indicate a poor prognosis, including rightsided colon cancer, at least six liver metastases, EHD location other than lung metastases, EHD concomitant with CLM recurrence, and a carcinoembryonic antigen (CEA) level ≥ 10 ng/mL. They found that patients had a higher 5-year survival rate in the absence of those factors. Therefore, this study aimed to retrospectively assess the consequences of liver resection added to chemotherapy on the overall survival of patients with CRLM and concurrent EHD and to identify the characteristics of the patients who would bene t from such a combination.
Patients And Methods

Data Source
This retrospective cohort study made use of the Surveillance, Epidemiology, and End Results (SEER) database. After obtaining the SEER permission to use patient data, we surveyed the SEER*Stat database (version 8.3.6; National Cancer Institute) and collected the data of 2533 CRLM patients with concurrent EHD.

Patient Population
We selected the patients based on the following selection criteria: Cases diagnosed between January 1, 2010, and December 31, 2014. Data on the speci c site of distant metastasis have only become available on the SEER database since 2010, and therefore, patients diagnosed before 2010 were excluded. We also excluded patients diagnosed after 2015 to allow for a follow-up interval of at least 60 months.
Patients with liver metastasis (CS mets at DX-liver), along with an extrahepatic disease, including lung (CS mets at DX-lung), brain (CS mets at DX-brain), bone (CS mets at DX-bone), and peritoneal carcinomatosis (CS Mets at DX: code 36).
All selected patients received systemic chemotherapy, either alone or combined with hepatic resection/ablation (nonprimary surgical procedure to distant site). Those who received radiation were excluded.
Patients alive with no survival time were excluded; these included 80 patients. None of our patients were diagnosed by autopsy or reported on a death certi cate.
The length of survival was measured from the date of diagnosis until the time of death or the last followup. According to SEER data, all these patients had been adequately followed-up at the time of study cutoff (December 31, 2014). However, the actual follow-up protocol was not stated.

Statistical Analysis:
All statistical analyses were performed with IBM SPSS statistical software (version 25). Frequency distributions (No. and %) were used to describe categorical variables (e.g., age at diagnosis, sex, race, primary site, grade, and sites of extra-hepatic metastases, hepatic metastasectomy). Mean and the standard deviation was used to describe age at diagnosis as a continuous variable. Association between categorical variables and hepatic metastasectomy were tested for statistical analysis by Chi-square test or Fisher's exact test (if > 20% of expected values were less than 5). Difference in the mean age (a continuous variable) by the hepatic metastasectomy status (Yes/No) was tested for statistical signi cance by the independent-samples T-test. Overall survival time (in months) was described as the median survival. Survival function was presented in the Kaplan-Meier curve, while the difference in the survival distributions for "hepatic metastasectomy" and "no surgery" subgroups was tested for the statistical signi cance by the Log-rank test. Survival data were further modeled using Cox proportionalhazards regression. The Cox models were used to assess the association of the study covariates (age at diagnosis, sex, race, primary site, grade, and sites of extra-hepatic metastases) and surgical intervention (hepatic metastasectomy) with the 5-year mortality. The hepatic metastasectomy model (model 1) was then adjusted for sex, age at diagnosis (< 60 years, or ≥ 60 years), and race (model 2); and for the primary site, grade, and sites of extra-hepatic metastases (model 3). Cox proportional hazard ratios and its 95% con dence intervals were presented for each covariate and each hepatic metastasectomy model. A pvalue < 0.05 was regarded as statistically signi cant.

Results
Out of 2533 patients who had a diagnosis of CRLM and concurrent EHD, 57.2% were male and 42.8% were female ( Table 1). The mean age of the studied sample was 60.13 years (± 12.51) and ranged from 16 to 95 years. Fifty-three percent were aged 60 years old or older. The majority were white (73.5%). Approximately half of the patients presented with grade II colorectal cancer (53.8%), while 15.3% were at grade III, and 25% were of unknown grade. Among the different sites of primary colorectal tumors, the sigmoid colon was the most common (31.1%), closely followed by the rectum NOS (21.4), while splenic exure was the least common (2.9%). Approximately two-thirds of the patients had lung metastases at  1). In Table 2, CRLM patients who underwent hepatic metastasectomy were signi cantly associated with younger age at diagnosis (< 60 years), female sex, primary site, tumor grade, and site of extrahepatic metastases.  Table 3 shows that older age at diagnosis (≥ 60 years) and American Indian/Alaska Native race were signi cantly associated with a higher 5-year mortality hazard compared to earlier age at diagnosis (< 60 years) and white race, respectively. No statistically signi cant sex difference was detected in the 5- year mortality hazard ratio. Furthermore, primary sites at the rectosigmoid colon, sigmoid colon, and descending colon, grade III, and the presence of bone metastases were all signi cantly associated with a higher 5-year mortality hazard compared to rectum (or NOS), grade I, and lung metastases. However, surgical intervention by hepatic metastasectomy was associated with a 33.8% lower 5-year mortality hazard compared to no surgical interventions (i.e., chemotherapy only).

Discussion
The presence of extrahepatic disease (EHD) in patients with CLRM was previously regarded as an absolute contraindication for liver resection [11,12]. However, with the development of better surgical techniques that allow for complete resection and the emergence of more effective regimens of chemotherapy that shrink both intra-and extrahepatic disease, more patients with EHD are becoming eligible for liver resection [11,12]. It depends on the possibility of resecting all diseases, including the primary tumor, liver metastases, and EHD [13,14]. This study revealed that, in CRLM patients with concurrent EHD, resection of liver metastases following systemic chemotherapy is superior to chemotherapy alone, with median OS times of 24 months and 21 months, respectively. Chua et al.
clari ed that surgery removes the tumor mass, while chemotherapy targets micrometastatic disease [15]. Therefore, combining both modalities would lead to better outcomes. We found that the survival of CRLM patients with EHD who received combined regimens was not in uenced by their sex; however, patients with a younger age at diagnosis and white race had a lower 5-year mortality hazard. Meanwhile, a study by Aoki et al. indicated that patient age and sex were insigni cant prognostic factors [16]. Surprisingly, Adam et al. reported impressive survival rates of elderly CRLM patients undergoing liver resection and concluded that patients' age should not be regarded as a barrier to surgery [17].
Our study showed that patients had outcomes that were more favorable when their primary tumors were grade I, located at the rectum, or when they had an EHD located at the lungs. Similarly, a study observed more reduced survival rates following liver resection in patients with liver metastases from right-sided colon cancer [18]. In contrast, Aoki et al. found no signi cant differences in overall survival between different sites of the primary tumor [16]. Meanwhile, it has been consistently reported in the literature that the location of EHD is an essential prognostic factor [10,19,20]. A study found that following complete resection of the EHD, patients with lung metastasis had a better prognosis compared to those with peritoneal or portal and para-aortic nodal metastases [21]. Interestingly, a study by Adam et al. proposed ve factors for poor prognosis: right colon cancer, at least 6 liver metastases, EHD location other than lung metastases, EHD concomitant with CLM recurrence, and a carcinoembryonic antigen (CEA) level ≥ 10 ng/mL [10]. The authors observed a higher 5-year survival rate in patients with none of these factors compared to that of patients with three or more factors. Other studies have also addressed further negative prognostic factors, including R1 margin status, largest CRLM greater than 3 cm, portal or retroperitoneal nodal involvement, multiple EHD sites, and high tumor burden in the liver [19,[22][23][24].
Furthermore, Chua et al identi ed the Peritoneal Cancer Index (PCI) as a prognostic indicator of survival in CRLM patients with peritoneal carcinomatosis [25]. Lower PCI in these patients made it possible to perform complete cytoreductive surgery for peritoneal disease and thus improved their survival.
To date, the possibility of liver resection to offer a potential cure for patients with EHD is still a debate. Some studies found that disease recurrence is the rule in these patients [19,26]. For instance, a retrospective review reported a very high rate of disease recurrence following resection, with 90.2% of patients experiencing recurrence at a median of 8 months and 85% of them experiencing systemic recurrence [19]. The patients were rarely cured, and therefore, they considered liver resection in this setting as a non-curative intervention [19,27]. In contrast, some have postulated that combining targeted molecular therapies with chemotherapy or following a more aggressive approach may offer a potential cure in CRLM patients with concomitant EHD [11,28]. Such an approach consists of perioperative chemotherapy, resection of all metastatic sites, and re-surgery in case of recurrence [11,28]. However, even in the absence of curative intent, complete resection is needed to achieve better outcomes. Patients who did not complete two-stage hepatectomy had similar survival rates to patients treated with chemotherapy only [29]. Moreover, the survival of patients undergoing combined resection of CRLM and EHD was better than that of patients undergoing liver resection only or those receiving chemotherapy only [23,30]. For instance, Hwang et al. indicated that the 5-year overall survival of patients signi cantly jumped from 0-28% after resecting both hepatic and concurrent EHD [23].
The current study had several limitations. The data on SEER*Stat only report whether the patient has received chemotherapy or not. Further details of the chemotherapy regimen are not available in SEER, including the given agents and their combinations, the number of cycles, and the sequence of chemotherapy with surgery, preoperative, postoperative, or both. Another limitation is the lack of details about hepatic metastases, including their size and number. Such a shortage in detail has limited our ability to analyze the effect of various chemotherapy protocols and liver disease burden on patient outcomes following liver resection. Moreover, SEER*Stat describes liver resection with resection/ablation without further details on the resection technique.

Conclusion
This study revealed that CRLM patients with EHD could bene t from combined liver metastases and chemotherapy in terms of better survival. Future studies should address the most appropriate chemotherapy regimen in this setting and assess the effect of combined CRLM and EHD resection at different sites on patient survival. Cancer patient data were collected from the SEER database after approval. A data use agreement form was signed, and the authors had to use a username and password to log in and collect the data.

Consent for publication
Not applicable.

Availability of data and material
The ndings of this study were based upon data obtained from the SEER database; restrictions apply to the availability of these data, which were used under license for the current study and so are not available