PDF
Citation
Print
INTRODUCTION
Although cholangiocarcinoma (CCA) comprises only 3% of gastrointestinal malignancy, its incidence in Korea is relatively high; with more than 6 cases per 100,000 people, it is 20 times more frequent than in Western countries [1]. Along with incidence, mortality of CCA has increased up to 1–6 per 100,000 inhabitants worldwide [2] and is presenting as a global health problem. To increase the frequency of long-term survival, curative resection is necessary [3]. However, in the case of hilar CCA, in which the tumor is located at the hilum near pivotal structures such as the portal triad and the bile duct confluence with its proximity to the caudate lobe and liver parenchyma, attempts to perform curative resection may be challenging. To obtain a negative resection margin, perioperative interventions such as preoperative biliary drainage and portal vein embolization (PVE) are used widely. Surgical techniques of portal vein resection and anastomosis in cases of vascular invasion and concomitant liver resection including caudate lobectomy are performed to obtain a negative resection margin. However, even after aggressive surgical resection, the 5-year survival rates vary from 14% to 48% [4567]. More research is needed to determine the proper management of this devastating malignancy. The aim of this study is to evaluate the prognostic factors for long-term survival in hilar CCA in a high-volume center.
METHODS
Ethics statement
The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea (No. 2022-09-070). The patient's consent was waived due to the retrospective study design.
Study design
Patients who underwent curative resection for hilar CCA at Samsung Medical Center in Seoul, Korea, from January 2000 to December 2020 were included. Demographics, surgical and oncological outcomes, short-term complications, recurrence, pathologic results, and survival were collected retrospectively and analyzed. Demographics were compared between the R0 and R1 groups. Multivariate analysis was performed both for overall survival and recurrence. For 5-year survival, a subgroup analysis of patients from 2000 to 2017 was conducted.
Statistical analysis
All analyses were performed with IBM SPSS Statistics ver. 27.0 (IBM Corp.). The Student t-test was used for comparison between 2 groups for continuous variables and chi-square tests for categorical variables. Variables for which P < 0.05 were statistically significant.
Survival was analyzed using the Kaplan-Meier curve and compared between groups with the log-rank test. For prognostic factor analysis, only patients who underwent curative intent-to-treat surgery were included. Multivariate analysis for prognostic factors was performed using a Cox proportional hazards method. Two-sided P-values less than 0.05 were considered statistically significant, and P-values less than 0.1 were considered marginally significant. A multivariable regression was carried out with variables whose P-value was <0.1 in the univariable analyses, using a backward stepwise method.
RESULTS
Among the 467 patients, R2 resection cases were excluded. A total of 449 patients diagnosed with hilar CCA who underwent intent-to-treat resection at Samsung Medical Center from 2000 to 2020 were included in this study.
Demographics were analyzed in 2 groups, 406 patients in R0 and 43 patients in R1, as shown in Table 1. Across all patients, the R0 resection rate was 90.4%, among which 94.3% underwent liver resection. The average age was 64.78 years (median, 66 years), with a range from 43 to 89 years. Preoperative biliary drainage was performed in 73.3% of patients, while PVE was conducted in 17.1%. Bismuth types did not show a statistical difference between the 2 groups. The concomitant liver and biliary resection group comprised about 93.8% of all patients, 94.3% from the R0 group, and 88.4% in the R1 group, with no statistical difference. Portal vein reconstruction was performed in 12.7% with no statistical significance. Neither lymphovascular invasion nor perineural invasion showed a significant difference between the 2 groups. The rate of adjuvant therapy was 20.7% in the R0 group compared to 62.8% in the R1 group (P < 0.001). Recurrence across all groups was 59.9%, with a significantly higher 61.1% rate in the R0 group compared to 48.8% in the R1 group (P = 0.001).
The mean follow-up period was 36.78 months. Overall, major morbidity occurred in 180 patients (40.1%), and 90-day mortality occurred in 15 (3.3%). The overall stage breakdown was as follows: 0.2% (1 case) with stage 0, 7.8% with stage I, 52.1% with stage II, 4.9% with stage IIIA, 0.7% with stage IIIB, 22.5% with stage IIIC, and 8.0% with stage IVA, with no significant differences between any 2 groups. The Kaplan-Meier survival curve revealed that the median disease-free survival was 19 months and median survival was 40 months. One-, 3-, and 5-year survival rates were 82.8%, 53.7%, and 35.8%, respectively (Fig. 1).
In Fig. 2, a significant difference in univariate analyses is shown in the Kaplan-Meier survival curves of T stage (P < 0.001) and N stage (P < 0.001). On the other hand, resection margin status and adjuvant treatment failed to demonstrate any survival benefit in this study. Univariate and multivariate analyses of prognostic factors using Cox proportional hazards models are shown in Table 2. Biliary drainage, higher N stage, and poor tumor differentiation did show a poor prognosis.
Table 3 shows the possible risk factors for recurrence. Biliary drainage, T stage, N stage, CA19-9 level ≥37 U/mL, and tumor differentiation were significant in univariate analysis. Resection margin status and adjuvant treatment failed to demonstrate any significance for recurrence. In multivariate analysis, biliary drainage (P = 0.004), N stage (P < 0.001), and tumor differentiation (P = 0.005) had statistical significance for recurrence.
The 5-year survival analysis was performed with 380 patients from 2000 to 2017 as shown in Table 4. There were 109 patients who lived longer than 5 years. There were significantly more patients with well-differentiated tumor biology in the 5-year survivor group. Preoperative biliary drain did show a significant difference between the 2 groups, with fewer drainage patients in the longer than 5-year survival group. T stage tended to be lower in the longer than 5-year survival group with 15% in T stage 0 and 1; 74.8% in T2; and 10.3% in T3, compared to 6.7%, 79.7%, and 14.1%, respectively, in the less than 5-year survival group. The N stage also showed a significant difference between the 2 groups, with 79.2% of the N0 cases and 20.8% of the N1 cases being in the longer than 5-year survival group, compared to 63.2% of the N0 cases, 34.5% of the N1 cases, and 2.3% of the N2 cases in the shorter than 5-year survival group (P = 0.013). In multivariate analysis, only poor tumor differentiation and N2 status had a prognostic effect.
Overall survival depending on the node status in combination with adjuvant treatment was analyzed and shown in Kaplan-Meier survival curves in Fig. 3. Only the R0 group was included and further divided into 4 groups. The median survival of each group was as follows: node-negative without adjuvant therapy, 45 months; node-negative with adjuvant therapy, 60 months; node-positive without adjuvant therapy, 22 months; and node-positive with adjuvant therapy, 19 months. This survival graph failed to demonstrate the effect of adjuvant treatment within the node-negative and -positive groups. But between the node-negative and -positive groups, there was statistical significance between 2 groups.
DISCUSSION
CCA is relatively prevalent in East Asia, including Korea, where the incidence is up to 20 times greater than that of Western countries [1]. However, the treatment for CCA is intractable depending on the location of the tumor and the nature of longitudinal spread of disease along the bile duct. In particular, a clear resection margin is more difficult to achieve in hilar CCA, which is defined as cancer involving the confluence of the bile duct. With Bismuth type IV tumor, there is a tendency for more frequent positive resection margin after surgery. However, type IV does not preclude a patient from undergoing curative resection, with advances in surgical technique and perioperative management such as PVE facilitating R0 resection [891011121314].
When performing liver resection, severe complications such as liver failure may develop. To prevent such fatal complications, PVE is recommended for those whose future remnant liver is small. In this study, a total of 37.5% of the liver resection patients underwent PVE, with 9 patients in the liver resection group having a liver failure grade greater than 3 (2.1%). When curative resection is feasible, aggressive surgical resection is recommended with proper preoperative management.
Another factor that affects postoperative liver failure is preoperative hepatic dysfunction caused by obstruction. Thus, relieving the biliary obstruction with preoperative drainage minimizes the risk of fatal complications [81516] and improves survival [1217]. However, in the present study, preoperative biliary drainage acts as an adverse prognostic factor. There was a tendency for more frequent biliary drainage in poor tumor biology groups, as demonstrated in Supplementary Table 1 (P < 0.001). This suggests that more aggressive tumors may result in preoperative jaundice that requires biliary drainage. Though this study did not show a significant correlation between complication and overall survival (P = 0.052), further subgroup analysis may aid in finding modifiable factors to prevent complications and improve survival.
Achieving a negative resection margin has shown survival benefits [918192021]. However, the present study failed to demonstrate any survival benefit in the R0 group. In multivariate analysis of prognostic factors, only N stage, tumor biology, and preoperative biliary drainage demonstrated a prognostic effect.
Even after achieving a negative resection margin, the prognosis is often dismal, with recurrence rates as high as 70% [2223]. In this study, overall recurrence was 59.9% among those who underwent intent-to-treat surgical resection. From the comparison of R0 vs. R1 group demographics, significantly more patients from the R1 group received adjuvant treatment (P < 0.001), resulting in a lower recurrence rate than that of the R0 group (R0 61.1% vs. R1 48.8%, P = 0.001). Risk factor analysis for recurrence revealed preoperative biliary drainage resulting in a higher recurrence rate, along with higher N stage and poorer tumor differentiation, but no difference was noted for resection margin status or adjuvant treatment. This implies that selection bias may play a role in nullifying the unfavorable prognostic effect of a positive resection margin with the survival benefit of performing adjuvant treatment.
The effect of adjuvant treatment is controversial. While some randomized controlled studies demonstrated no survival benefit with adjuvant chemotherapy [2425], other clinical studies showed that adjuvant chemotherapy can improve the overall survival of biliary tract cancer [262728]. The overall trend is that adjuvant treatment would provide survival benefits. Subgroup analysis of the R0 group divided into adjuvant treatment and no adjuvant treatment groups was performed to identify any significant difference between the 2 groups (Supplementary Table 2). As a retrospective study, this comparison may demonstrate some criteria considered for adjuvant therapy candidates and also possible selection bias. Along with the pathologic results requiring adjuvant treatment such as positive resection margin, presence of lymphovascular invasion, perineural invasion, higher T stage, node-positive group, and poor differentiation, the age and general condition of patients were also considered. Younger patients with better general conditions tend to receive adjuvant treatment. Unrelated to the nature of disease itself, these factors may affect overall survival and may have confounded the prognostic effect of negative resection margin status and adjuvant treatment.
Subgroup analysis in the R0 group was performed in search of any association between node status and adjuvant treatment. Only 30 of 264 patients (11.4%) from the N0 group underwent adjuvant treatment, while 51 of 125 patients (40.8%) from the node-positive group underwent adjuvant treatment. Due to the small sample size, the N0 with adjuvant treatment group did not demonstrate a significant difference in survival compared to the N0 without adjuvant treatment group, but the median survival was 60 months compared to 45 months (5-year survival rate: 52.7% in the node-negative and adjuvant group vs. 42.3% in the node-negative with no adjuvant group, P = 0.238) (Fig. 3). Not surprisingly, the node-positive group showed unfavorable survival compared to the N0 group regardless of adjuvant treatment. Yet this may suggest that even node-negative patients may benefit from adjuvant treatment. Further research is needed with a larger study population.
Due to the nature of this retrospective study, there may be missing information and selection bias from data collection. Though this study is meaningful by showing actual survival in a long-term follow-up period from a single center, over the past decades, standard management of hilar CCA has evolved, and the treatment strategies were consistent among the patients.
In this intent-to-treat group, there were only a few margin-positive patients. Owing to the small sample size, a comparison between the margin-negative and margin-positive groups may not have allowed a valid conclusion. Further analysis with a larger margin-positive group is needed.
More recent data on routine application of adjuvant treatment and analysis between comparable groups may reveal the effect of adjuvant treatment on survival. Further analysis of adjuvant treatment in node-negative patients may affirm its role in the management of hilar CCA. Also, a more detailed analysis of biliary drainage patient characteristics in association with infection and complications may disclose modifiable factors for prognosis.
Despite the risk of extensive surgery, which can lead to various complications including hepatic failure, hilar CCA is one of the most poorly understood diseases, and factors of the clinical prognosis in resected hilar CCA remain unclear. The present study investigated factors that could affect the prognosis of patients with hilar CCA who underwent surgical resection at a large-volume institute over a 20-year period and examined the characteristics of patients who survived for more than 5 years. This 5-year survival rate was 35.8%, which is somewhat higher than previous results. In addition, the prognostic factors were slightly different from the literature results, However, in patients who survived for more than 5 years, N2 stage and tumor biology played an important role.
XML Download