Trends in Survival after Surgery for Cholangiocarcinoma: A 30-Year Population-Based SEER Database Analysis

Nathan H, Pawlik TM, Wolfgang CL, Choti MA, Cameron JL, Schulick RD (2007) Trends in survival after surgery for cholangiocarcinoma: a 30-year population-based SEER database analysis. J Gastrointest Surg, 11(11):1488-1496; discussion 1496-1487

Introduction
Data on long-term survival of patients after surgical resection are limited to single-institution case series. Reported 5-year survival rates in recent surgical series (irrespective of margin status) vary widely, from 17 to 40% for ICC6–14 and from 9 to 41% for ECC.


Methods
This study was a retrospective analysis of prospectively collected data from the SEER database maintained by the National Cancer Institute.28

Results
Comparisons of survival curves by stage at diagnosis (Fig. 3) were all highly significant (P<0.001). As expected, more advanced disease conferred a worse prognosis, with 5-year RS of 37.4% for localized disease, 14.7% for regional disease, and 5.3% for distant disease.

Univariate Cox proportional hazards analysis for ICC (Table 3) showed the following variables to be significantly associated with decreased survival (P value for likelihood ratio test, percent of data missing): age per year over 80 years (P=0.001, 0%), tumor stage (P<0.001, 7.6%), tumor grade (P=0.017, 34.4%), and year of diagnosis per year after 1992 (P<0.001, 0%). These variables, together with rural area of residence (0.2% missing), were entered into the multivariate model for ICC. This approach revealed a year-to-year improvement in ICC survival over the years 1992–2002 that remained significant in multivariate analysis. In the final model of ICC survival (Table 3), age per year over 80 years and stage at diagnosis were strong predictors of worse survival, but the effect of tumor grade was not statistically significant. This analysis demonstrated a significant year-to-year improvement in ICC survival after 1992, corresponding to a cumulative 34.4% increase in adjusted survival from 1992 through 2002. Table 3


Discussion
Historically, analyses of patient survival after surgery for cholangiocarcinoma have been restricted to single-institution series. Although such institutional data may offer great depth of clinical information, they may be limited by poor generalizability and potential selection bias. We found 5-year crude survival rates of 17.7% for ICC and 17.1% for ECC over the period 1973–2002. For ICC, recent single-institution surgical series have reported 5-year crude survival rates of 17–40%.6–14 Even the highest 5-year crude survival rate for ICC in our analysis, 19.7% in the decade 1993–2002, falls below all except one of these single-institution rates.6 For ECC (including perihilar tumors), recent single-institution surgical series have reported 5-year crude survival rates of 9–41%.10,14–27

The improvement in ICC survival may reflect improving patient selection over time, likely as the result of improvements in imaging technology, such as multidetector computed tomography, that allow better preoperative assessments of resectability. Also, improvements in the safety of hepatic resection39 have likely led to the increased utilization of aggressive hepatic resection for ICC, contributing to improved oncologic results and increased long-term survival. Unfortunately, details of the type of surgical resection in the SEER database are inconsistently available, and margin status information is absent, preventing us from further investigating these hypotheses.

This study is limited primarily by the depth of surgical data in the SEER database. In addition to the lack of margin status data, the level of detail and completeness of data on tumor size, lymph node involvement, and details of resection have varied since 1973, such that comparisons that account for these factors over all 30 years are not possible. Although we did have some data on radiation therapy, we did not have any information on the use of chemotherapy. Finally, the ICD-O-3 coding scheme used in the SEER database did not allow us to separate perihilar tumors from other ECC, limiting comparisons with other studies.

In conclusion, this population-based analysis demonstrates that survival after surgery for extrahepatic cholangiocarcinoma has dramatically improved since 1973. Patients with intrahepatic cholangiocarcinoma, however, have achieved an improvement in survival largely confined to more recent years. We suggest that improvements in imaging technology, patient selection, and surgical techniques are largely responsible for these improvements. The discrepancies between the survival rates we report and those reported in single-institution series deserve further investigation to determine whether they are the result of publication bias, patient selection, disease characteristics, or disparities in access to adequate care. Finally, these population-based survival statistics demonstrate that extrahepatic and intrahepatic cholangiocarcinoma continue to carry very poor prognoses. Despite incremental advances in the surgical therapy of these biliary tract malignancies over the last three decades, there remains much opportunity for improvement.