[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.146.179.146. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Survival and Conditional Survival Curves
Survival and Conditional Survival Curves

AJCC indicates American Joint Committee on Cancer.

Figure 2.
Conditional Survival Stratified by Different Variables
Conditional Survival Stratified by Different Variables
Table 1.  
Patients Who Reach a Certain Survival Point After Resection of Intrahepatic Cholangiocarcinoma Given That They Have Already Survived a Certain Amount of Timea
Patients Who Reach a Certain Survival Point After Resection of Intrahepatic Cholangiocarcinoma Given That They Have Already Survived a Certain Amount of Timea
Table 2.  
Actuarial Survival Rates of Patients in Relationship to Clinical and Tumoral Characteristics
Actuarial Survival Rates of Patients in Relationship to Clinical and Tumoral Characteristics
Table 3.  
Three-Year Conditional Survival Rates of Patients in Relationship to Clinical and Tumor Characteristics
Three-Year Conditional Survival Rates of Patients in Relationship to Clinical and Tumor Characteristics
1.
El-Serag  HB, Engels  EA, Landgren  O,  et al.  Risk of hepatobiliary and pancreatic cancers after hepatitis C virus infection: a population-based study of US veterans. Hepatology. 2009;49(1):116-123.
PubMedArticle
2.
Endo  I, Gonen  M, Yopp  AC,  et al.  Intrahepatic cholangiocarcinoma: rising frequency, improved survival, and determinants of outcome after resection. Ann Surg. 2008;248(1):84-96.
PubMedArticle
3.
Malhi  H, Gores  GJ.  Cholangiocarcinoma: modern advances in understanding a deadly old disease. J Hepatol. 2006;45(6):856-867.
PubMedArticle
4.
Shaib  YH, Davila  JA, McGlynn  K, El-Serag  HB.  Rising incidence of intrahepatic cholangiocarcinoma in the United States: a true increase? J Hepatol. 2004;40(3):472-477.
PubMedArticle
5.
Welzel  TM, McGlynn  KA, Hsing  AW, O’Brien  TR, Pfeiffer  RM.  Impact of classification of hilar cholangiocarcinomas (Klatskin tumors) on the incidence of intra- and extrahepatic cholangiocarcinoma in the United States. J Natl Cancer Inst. 2006;98(12):873-875.
PubMedArticle
6.
Khan  SA, Toledano  MB, Taylor-Robinson  SD.  Epidemiology, risk factors, and pathogenesis of cholangiocarcinoma. HPB (Oxford). 2008;10(2):77-82.
PubMedArticle
7.
Carpizo  DR, D’Angelica  M.  Management and extent of resection for intrahepatic cholangiocarcinoma. Surg Oncol Clin N Am. 2009;18(2):289-305, viii-ix.
PubMedArticle
8.
Patel  T.  Increasing incidence and mortality of primary intrahepatic cholangiocarcinoma in the United States. Hepatology. 2001;33(6):1353-1357.
PubMedArticle
9.
Shaib  Y, El-Serag  HB.  The epidemiology of cholangiocarcinoma. Semin Liver Dis. 2004;24(2):115-125.
PubMedArticle
10.
de Jong  MC, Nathan  H, Sotiropoulos  GC,  et al.  Intrahepatic cholangiocarcinoma: an international multi-institutional analysis of prognostic factors and lymph node assessment. J Clin Oncol. 2011;29(23):3140-3145.
PubMedArticle
11.
Hyder  O, Marques  H, Pulitano  C,  et al.  A nomogram to predict long-term survival after resection for intrahepatic cholangiocarcinoma: an Eastern and Western experience. JAMA Surg. 2014;149(5):432-438.
PubMedArticle
12.
Mavros  MN, Economopoulos  KP, Alexiou  VG, Pawlik  TM.  Treatment and prognosis for patients with intrahepatic cholangiocarcinoma: systematic review and meta-analysis [published online April 9, 2014]. JAMA Surg. doi:10.1001/jamasurg.2013.5137.
PubMed
13.
Edge  S, Byrd  DR, Compton  CC, Fritz  AG, Greene  FL, Trotti  A, eds. AJCC Cancer Staging Manual. 7th ed. New York, NY: Springer; 2010.
14.
Wang  Y, Li  J, Xia  Y,  et al.  Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol. 2013;31(9):1188-1195.
PubMedArticle
15.
Nathan  H, de Jong  MC, Pulitano  C,  et al.  Conditional survival after surgical resection of colorectal liver metastasis: an international multi-institutional analysis of 949 patients. J Am Coll Surg. 2010;210(5):755-764, 764-766.
PubMedArticle
16.
Henson  DE, Ries  LA, Carriaga  MT.  Conditional survival of 56,268 patients with breast cancer. Cancer. 1995;76(2):237-242.
PubMedArticle
17.
Merrill  RM, Henson  DE, Ries  LA.  Conditional survival estimates in 34,963 patients with invasive carcinoma of the colon. Dis Colon Rectum. 1998;41(9):1097-1106.
PubMedArticle
18.
Skuladottir  H, Olsen  JH.  Conditional survival of patients with the four major histologic subgroups of lung cancer in Denmark. J Clin Oncol. 2003;21(16):3035-3040.
PubMedArticle
19.
Sun  M, Trinh  QD, Karakiewicz  PI.  Conditional survival of patients with metastatic renal-cell carcinoma. Lancet Oncol. 2012;13(11):e462.
PubMedArticle
20.
Wang  SJ, Emery  R, Fuller  CD, Kim  JS, Sittig  DF, Thomas  CR.  Conditional survival in gastric cancer: a SEER database analysis. Gastric Cancer. 2007;10(3):153-158.
PubMedArticle
21.
Mayo  SC, Nathan  H, Cameron  JL,  et al.  Conditional survival in patients with pancreatic ductal adenocarcinoma resected with curative intent. Cancer. 2012;118(10):2674-2681.
PubMedArticle
22.
Cucchetti  A, Piscaglia  F, Cescon  M,  et al.  Conditional survival after hepatic resection for hepatocellular carcinoma in cirrhotic patients. Clin Cancer Res. 2012;18(16):4397-4405.
PubMedArticle
23.
Eloranta  S, Lambert  PC, Andersson  TM, Björkholm  M, Dickman  PW.  The application of cure models in the presence of competing risks: a tool for improved risk communication in population-based cancer patient survival. Epidemiology. 2014;25(5):742-748.
PubMedArticle
24.
Daudt  HM, van Mossel  C, Dennis  DL, Leitz  L, Watson  HC, Tanliao  JJ.  Survivorship care plans: a work in progress. Curr Oncol. 2014;21(3):e466-e479.
PubMedArticle
25.
Burnand  B, Kernan  WN, Feinstein  AR.  Indexes and boundaries for “quantitative significance” in statistical decisions. J Clin Epidemiol. 1990;43(12):1273-1284.
PubMedArticle
26.
Spolverato  G, Ejaz  A, Kim  Y,  et al.  Tumor size predicts vascular invasion and histologic grade among patients undergoing resection of intrahepatic cholangiocarcinoma. J Gastrointest Surg. 2014;18(7):1284-1291.
PubMedArticle
27.
Janssen-Heijnen  ML, Gondos  A, Bray  F,  et al.  Clinical relevance of conditional survival of cancer patients in Europe: age-specific analyses of 13 cancers. J Clin Oncol. 2010;28(15):2520-2528.
PubMedArticle
28.
Merrill  RM, Hunter  BD.  Conditional survival among cancer patients in the United States. Oncologist. 2010;15(8):873-882.
PubMedArticle
29.
Johnson  DR, Ma  DJ, Buckner  JC, Hammack  JE.  Conditional probability of long-term survival in glioblastoma: a population-based analysis. Cancer. 2012;118(22):5608-5613.
PubMedArticle
30.
Morimoto  Y, Tanaka  Y, Ito  T,  et al.  Long-term survival and prognostic factors in the surgical treatment for intrahepatic cholangiocarcinoma. J Hepatobiliary Pancreat Surg. 2003;10(6):432-440.
PubMedArticle
31.
Isa  T, Kusano  T, Shimoji  H, Takeshima  Y, Muto  Y, Furukawa  M.  Predictive factors for long-term survival in patients with intrahepatic cholangiocarcinoma. Am J Surg. 2001;181(6):507-511.
PubMedArticle
32.
Choi  SB, Kim  KS, Choi  JY,  et al.  The prognosis and survival outcome of intrahepatic cholangiocarcinoma following surgical resection: association of lymph node metastasis and lymph node dissection with survival. Ann Surg Oncol. 2009;16(11):3048-3056.
PubMedArticle
33.
Ohtsuka  M, Ito  H, Kimura  F,  et al.  Results of surgical treatment for intrahepatic cholangiocarcinoma and clinicopathological factors influencing survival. Br J Surg. 2002;89(12):1525-1531.
PubMedArticle
34.
Li  T, Qin  LX, Zhou  J,  et al.  Staging, prognostic factors and adjuvant therapy of intrahepatic cholangiocarcinoma after curative resection. Liver Int. 2014;34(6):953-960.
PubMedArticle
35.
Janssen-Heijnen  ML, Houterman  S, Lemmens  VE, Brenner  H, Steyerberg  EW, Coebergh  JW.  Prognosis for long-term survivors of cancer. Ann Oncol. 2007;18(8):1408-1413.
PubMedArticle
Original Investigation
June 2015

Conditional Probability of Long-term Survival After Liver Resection for Intrahepatic CholangiocarcinomaA Multi-institutional Analysis of 535 Patients

Author Affiliations
  • 1Department of Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland
  • 2Institute for Digestive Diseases and Liver Transplantation Fundeni, Bucharest, Romania
  • 3Curry Cabral Hospital, Lisbon, Portugal
  • 4Ospedale San Raffaele, Milan, Italy
  • 5Medical College of Wisconsin, Milwaukee
  • 6University of Sydney, Sydney, Australia
  • 7University of Virginia, Charlottesville
  • 8Eastern Hepatobiliary Surgery Hospital, Shanghai, China
  • 9Stanford University, Palo Alto, California
  • 10Emory University, Atlanta, Georgia
JAMA Surg. 2015;150(6):538-545. doi:10.1001/jamasurg.2015.0219
Abstract

Importance  Whereas conventional actuarial overall survival (OS) estimates rely exclusively on static factors determined around the time of surgery, conditional survival (CS) estimates take into account the years that a patient has already survived.

Objective  To define the CS of patients following liver resection for intrahepatic cholangiocarcinoma (ICC).

Design, Setting, and Participants  Between January 1, 1990, and December 31, 2013, a total of 535 patients who underwent resection of ICC were identified from an international multi-institutional database. In this retrospective international study conducted from January to June 2014, clinicopathological characteristics, operative details, and long-term survival data were analyzed. Conditional survival estimates were calculated as the probability of survival for an additional 3 years.

Intervention  Resection of ICC.

Main Outcomes and Measures  Overall survival and CS.

Results  While actuarial OS decreased over time from 39% at 3 years to 16% at 8 years (P = .002), the 3-year CS (CS3) increased over time among those patients who survived. The CS3 at 5 years—the probability of surviving to postoperative year 8 after having already survived to postoperative year 5—was 65% compared with 8-year OS of 16% (P = .002). Factors that were associated with worse OS included larger tumor size (hazard ratio [HR], 1.02; 95% CI, 1.00-1.05; P = .05), multifocal disease (HR, 1.49; 95% CI, 1.19-1.86; P = .01), lymph node metastasis (HR, 2.21; 95% CI, 1.67-2.93; P < .01), and vascular invasion (HR, 1.39; 95% CI, 1.10-1.75; P = .006). The calculated CS3 exceeded the actuarial survival for all high-risk subgroups. For example, patients with lymph node metastasis had an actuarial OS of 11% at 6 years vs a CS3 of 49% at 3 years (Δ38%). Similarly, patients with vascular invasion had an actuarial OS of 15% at 6 years compared with a CS3 of 50% at 3 years (Δ35%).

Conclusions and Relevance  Conditional survival estimates may provide critical quantitative information about the changing probability of survival over time among patients undergoing liver resection for ICC. Therefore, such estimates can be of significant value to patients and health care professionals.

Introduction

Cholangiocarcinoma is the second most common primary liver malignancy, accounting for 5% to 30% of all primary liver tumors.16 Cholangiocarcinoma can be anatomically classified as intrahepatic, hilar, or distal based on the tumor location in the biliary tree.7 The incidence of intrahepatic cholangiocarcinoma (ICC) has been increasing worldwide, from 3.2 per 1 million in 1975 to 8.5 per 1 million in 2000, and it has continued to increase over the last decade.4,8,9 Complete surgical resection remains the only hope for long-term survival in patients with ICC and 5-year survival following surgery has been reported to be 30% to 35%.10

To our knowledge, data on the survival of patients with ICC have only been reported as actuarial survival time measured from either the date of diagnosis or the date of index operation. These actuarial survival estimates are typically stratified using a host of different clinicopathological risk factors including tumor number and size, lymph node metastasis, and vascular invasion.1012 In turn, these factors have been used to construct prognostic schemas including the American Joint Committee on Cancer (AJCC) staging manual and predictive nomograms.11,13,14 However, these prognostic factors have been called into question, especially among patients who have already survived for a time after surgery. Specifically, our group and others have noted that conventional overall survival (OS) estimates may rely too heavily on static factors determined around the time of surgery. Rather than being static, the risk for recurrence and survival estimates are more dynamic and change over time. For this reason, conditional survival (CS) estimates, which account for years that a patient has already survived after surgery, have been proposed as better predictors of long-term prognosis.1522 Conditional survival accounts for probably the most important prognostic factor—life-years already accrued—when predicting future survival time. Conditional survival estimates can be important among health care professionals to guide surveillance. In addition, as the emphasis on survivorship grows, patients increasingly desire more accurate information regarding their long-term prognosis based on the fact that they have already survived for a time.23,24

While CS has been investigated for a number of different malignancies, such as colon/rectum, breast, pancreas, and hepatocellular carcinoma, to our knowledge, there are no data on CS for patients undergoing resection of ICC.1522 Conditional survival is particularly relevant to patients with cancers characterized by poor long-term actuarial survival because CS can help identify different prognostic groups among survivors who otherwise would have been predicted to do poorly based on factors measured at the time of surgery. Conditional survival estimates for patients with ICC may be particularly important given the poor 5-year OS actuarial estimates associated with this malignancy.

The objective of the current study was to analyze CS in a large cohort of patients with ICC using an international multi-institutional database. Furthermore, we sought to investigate the impact of various clinicopathologic prognostic factors on OS and CS among patients who underwent curative intent liver resection for ICC.

Methods
Patient Population and Data Collection

Patients who underwent liver resection for ICC between January 1, 1990, and December 31, 2013, were identified from an international multi-institutional database (Johns Hopkins Hospital, Baltimore, Maryland; Fundeni Clinical Institute of Digestive Disease, Bucharest, Romania; Curry Cabral Hospital, Lisbon, Portugal; Hopitaux Universitaires De Geneve, Geneva, Switzerland; Ospedale San Raffaele, Milan, Italy; Medical College of Wisconsin, Milwaukee; Stanford University, Palo Alto, California; Eastern Hepatobiliary Surgery Hospital, Shanghai, China; University of Virginia, Charlottesville; Emory University, Atlanta, Georgia; and University of Sydney, Sydney, Australia). The institutional review board of each participating institution approved this study. Only patients with histologically confirmed ICC who received their curative treatment for primary ICC at a study center were included. Patients with metastatic disease (AJCC stage IVb) were excluded from the study cohort, whereas patients with AJCC T4N0M0 or any T and N1M0 (AJCC stage IVa) were included. Patients who died within 30 days from the date of surgery were excluded from the study. Patient consent was waived because inclusion of all consecutive eligible patients would not have been otherwise possible.

Standard demographic and clinicopathologic data were collected including sex, age, and primary tumor characteristics. Specifically, data were collected on primary tumor size, number, morphologic subtype, and presence of vascular invasion. Tumor size was defined as the largest diameter of the tumor in the resected specimen. Among patients with multiple tumors, the largest lesion was used as the index lesion. Histologic grade was categorized as well, moderately, or poorly differentiated; if tumor grade varied in a specific specimen, the worst grade was used as the index tumor grade. Data on treatment-related variables, such as type of liver resection, receipt of lymphadenectomy, and adjuvant therapy, were also obtained. Major hepatectomy was defined as the removal of 3 or more Couinaud segments. Resection margin and nodal status were ascertained based on final pathologic assessment. Date of last follow-up and vital status were collected for all patients. Mortality was calculated from the date of index operation.

Statistical Analysis

Summary statistics for the study population are presented as proportions for categorical variables or as median values with interquartile ranges for continuous variables. The Kaplan-Meier method was used to assess OS estimates for the entire study population calculated from the date of surgery to the date of last follow-up or death; differences in survival were assessed using the log-rank test. The association of relevant clinicopathologic variables with OS was assessed using Cox proportional hazards models; the prognostic power of covariates was expressed by calculating hazard ratios (HRs) with 95% CIs. The variables considered in our analyses included age, sex, number of tumors (solitary vs multiple), tumor size, vascular invasion, nodal status, AJCC stage, and presence of cirrhosis. The predicted risk for survival for the study cohort was also stratified based on the nomogram quartiles.11 Conditional survival estimates were calculated as the probability of survival for an additional 3 years (CS3), calculated as: CS3 = S(x +3)/S(x), where S indicates the number of years survived from the date of surgery. For example, CS3 among patients who had survived 2 years from the date of surgery was calculated by dividing the survival at 5 years by the survival at 2 years. Changes in CS3 over time were assessed using linear regression, and standardized differences (d) were used to assess the differences of CS between subgroups. Standardized differences can be used as the index to contrast 2 rates such as CS: d < 0.1 for very small differences; 0.1 ≤ d < 0.3 for small differences; 0.3 ≤ d < 0.5 for moderate differences, and d ≥ 0.5 for considerable differences.22,25 All analyses were carried out with Stata version 12.0 (StataCorp). All tests were 2-sided and P < .05 was considered statistically significant.

Results
Demographic and Clinicopathologic Characteristics

A total of 535 patients who underwent curative intent liver resection for ICC and met the inclusion criteria were identified. Baseline characteristics of the population are summarized in the eTable in the Supplement. The median age of the study population was 60 years (interquartile range, 51-69 years); roughly one-half of patients were female (n = 242, 45.7%) and non-Hispanic white (n = 377, 79.7%). A small subset of patients had a preoperative diagnosis of cirrhosis (n = 52, 10.0%). Most patients had solitary tumors (n = 366, 68.4%), whereas 169 patients (31.6%) had multiple tumors. The median tumor size was 6.5 cm (interquartile range, 4.7-9.0 cm). A subset of patients had bilateral disease (n = 161, 30.5%), whereas most patients had unilateral disease (n = 366, 69.5%). On pathology, most patients had mass-forming tumors (n = 401, 81.2%) that were moderately (n = 317, 60.8%) or poorly (n = 131, 25.1%) differentiated. Vascular invasion was present in 158 patients (29.5%), while perineural and biliary invasion were less common (n = 93, 18.2% and n = 78, 14.8%, respectively). About one-fifth of patients had lymph node metastasis (n = 106, 20.7%). Most tumors were T1 or T2 (n = 234, 46.1% and n = 178, 35.0%, respectively) and most patients had stage I or stage II disease (n = 142, 37.1% and n = 93, 24.3%, respectively).

One-third of the patients underwent less than a hemihepatectomy (n = 164, 31.2%). Margin status was microscopically positive (R1) in 85 patients (16.4%) and microscopically negative (R0) in 432 patients (83.6%). No patient had a grossly positive margin (R2). While 218 patients (41.2%) experienced some type of morbidity within 90 days after surgery, most complications were grades 1 or 2 (n = 136, 62.4%). In the postoperative setting, 255 patients (51.2%) received adjuvant chemotherapy.

Factors Associated With Overall Survival

The median OS was 27.4 months. Of note, most disease-specific deaths (n = 227, 65.6%) occurred within 24 months after surgery. Several factors were associated with worse OS including larger tumor size (HR, 1.02; 95% CI, 1.00-1.05; P = .05), multifocal disease (HR, 1.49; 95% CI, 1.19-1.86; P = .01), vascular invasion (HR, 1.39; 95% CI, 1.10-1.75; P = .006), lymph node metastasis (HR, 2.21; 95% CI, 1.67-2.93; P < .01), and advanced AJCC stage (referent stage I/II; stage III/IVa: HR, 1.59; 95% CI, 1.24-2.05; P < .001). Overall survival was also associated with patient stratification according to nomogram quartile. Specifically, when patients were divided into quartiles based on the nomogram, patients in the third and fourth quartiles had an increased hazard of death compared with patients in the first and second quartiles (HR, 1.86; 95% CI, 1.49-2.31; P < .001). On multivariable analysis, lymph node metastasis (HR, 2.79; 95% CI, 1.60-4.87; P < .001) and stratification into the third and fourth nomogram quartiles (HR, 2.21; 95% CI, 1.36-3.59; P = .001) remained associated with the risk for death.

Comparison of Overall and Conditional Survival

Overall actuarial survival after resection was 75% at 1 year, 39% at 3 years, and 25% at 5 years (Figure 1A). While actuarial OS decreased over time, from 39% at 3 years to 16% at 8 years (P = .002), the CS3 increased over time among those patients who survived (Figure 1B). Specifically, the CS3 at 5 years—the probability of surviving to postoperative year 8 after having already survived to postoperative year 5—was 65% compared with an actuarial 8-year OS of 16% (P = .002) (Figure 1B). Conditional survival estimates of the entire cohort based on time already survived after surgical resection are presented in Table 1. The survival probabilities of an additional 3 years, given that the patient had already survived for 1, 2, and 5 years, were 38%, 43%, and 44%, respectively.

Actuarial survival and CS3 rates among subgroups stratified by different clinicopathologic variables of known prognostic importance, such as age, number of lesions, tumor size, vascular invasion, AJCC stage (Figure 1C and D), lymph node metastasis, and cirrhosis, were examined. As expected, multiple lesions, tumor size 5 cm or greater, vascular invasion, lymph node metastasis, AJCC stage, and stratification into the third or fourth nomogram quartile were each associated with decreased actuarial survival (all P < .05; Table 2). For example, the 5-year actuarial OS of patients with stage I and II disease was 33% vs 16% for patients with stage III and IVa disease. The 5-year actuarial OS of patients in the first and second nomogram quartiles was 36% vs 30% for patients in the third and fourth nomogram quartiles. The calculated CS3 exceeded the actuarial survival for all corresponding subgroups. Furthermore, the difference between CS3 estimates and actuarial survival was more substantial among patients who were initially predicted to have the worse prognosis. For example, patients with lymph node metastasis had an actuarial OS of 11% at 6 years vs a CS3 of 49% at 3 years (Δ38%). Similarly, patients with vascular invasion had an actuarial OS of 15% at 6 years compared with a CS3 of 50% at 3 years (Δ35%). Patients categorized into the third and fourth nomogram quartiles had an actuarial OS of 14% at 6 years but a CS3 of 50% at 3 years (Δ36%). Patients characterized by fewer high-risk tumor characteristics had smaller differences between actuarial and CS estimates. For example, patients without lymph node metastasis had an actuarial OS of 29% at 6 years vs a CS3 of 49% at 3 years (Δ20%). Similarly, patients in the first and second nomogram quartiles had an actuarial OS at 6 years of 30% compared with a CS3 at 3 years of 59% (Δ29%). When CS was examined across a number of different risk strata, CS was noted to increase with time elapsed after surgery (Table 3). In addition, many of the differences of CS3 over time were most pronounced among patients with the worse initial prognostic features. For example, CS3 estimates increased over time in patients with tumors 5 cm or greater (36%-47%; Δ11%), lymph node metastasis (23%-50%; Δ27%), and AJCC stage III/IVa (28%-75%; Δ47%) (all P < .05).

Discussion

Although a relatively rare disease, the incidence of ICC is increasing in the United States and worldwide.4,8,9 Intrahepatic cholangiocarcinoma is a disease associated with a high case fatality and even patients who undergo curative intent surgical resection have a reported 5-year actuarial survival of only 30% to 35%.10 Furthermore, patients with certain clinicopathological characteristics (eg, multiple tumors and vascular invasion) are at particular risk for disease-specific death.10,12,26 While aggregate data on 5-year survival can be helpful, health care professionals and patients are often more interested in individual patient-specific data on prognosis. Therefore, some investigators have proposed nomogram-based means to stratify patients with regard to long-term survival.11,14 Population-based data and nomogram-based survival estimates are exclusively created, however, using data from the time of surgery and therefore can be inadequate. Specifically, standard survival curves at the time of diagnosis tend to be overly pessimistic because they are heavily weighted toward patients who die within the first few years.27 Rather, when patients have survived for a certain period after surgery, patients and health care professionals are more interested in future prognosis based on the specific time of the patient in the course of their disease. Conditional survival is a conditional probability estimate of long-term survival that is particularly relevant to patients with cancer who are alive for a period following treatment. While CS has been used to estimate the conditional probability of survival of patients with certain malignancies, such as glioblastoma and pancreas and colon cancer, no data currently exist concerning the prognosis of patients with ICC who have undergone curative intent surgery who have survived a year or more.17,28,29 The current study is important because we are, to our knowledge, the first group to define the conditional probability of survival for patients with ICC undergoing surgical resection. We have provided data on survival after surgical resection of ICC in a large, multi-institutional cohort and have demonstrated that CS may provide more accurate prognostic information for patients who have survived 1 year or longer. Specifically, we noted that CS estimates exceeded actuarial OS estimates among patients who had survived for a period following surgical resection (Figure 1B).

Tumor characteristics, including tumor size, multiple lesions, vascular invasion, lymph node involvement, and cirrhosis, have been associated with worse long-term survival.2,10,3034 We similarly identified several of these as adverse prognostic factors. However, of more interest were data comparing actuarial OS relative to CS among patients with each of these various tumor characteristics. While patients with ICC tumors characterized by any of these adverse prognostic factors were estimated to have worse actuarial survival, we noted the greatest improvements in CS among these patients. For example, over a 5-year period, patients with lymph node metastasis demonstrated a 27% increase in CS3 estimates vs only an 8% increase for patients without lymph node metastasis. Similarly, patients with multifocal disease had a 31% increase in CS3 compared with only a 13% increase for patients with a solitary lesion (Figure 2). The reason for this is in part owing to overly pessimistic actuarial OS estimates that are disproportionately influenced by patients who have these adverse clinical factors and die within the first few years.15,35 In turn, those patients who live longer and accrue years lived are in a sense out living the prognostic impact of those factors that were initially obtained at the time of surgery. The adverse prognostic factors at the time of surgery become increasingly less relevant as that specific patient accrues years lived. As such, our data suggest that CS may be a more valuable tool in the late postoperative period to estimate prognosis of patients who otherwise were predicted to have died based on initial actuarial estimates.

Relative OS estimates based on actuarial vs CS not only varied based on patient- and tumor-specific characteristics, but were also time dependent. In other words, relative CS improved the most over time, especially for those patients with the initial worse prognosis. As Figure 1B demonstrates, relative differences in actuarial and CS3 estimates were comparable at years 1 and 2 following surgery (Δ ≈ 10%-20%); however, by years 4 and 5, differences noted in CS relative to predicted actuarial survival were more considerable (Δ ≈ 40%-50%). These data suggest that traditional estimates of OS may be more accurate closer to the time of surgery but become increasingly less helpful as patients survive longer. Given that risk factors used in actuarial estimates are based on data derived from the perioperative period, these factors are most relevant near the time of surgery. As patients live longer, patients accrue survival time and self-select as individuals who have beaten the odds,15 and accrued survival time, which CS takes into consideration, becomes much more important relative to those factors that defined the tumor at the time of surgery in the past. As such, CS estimates provided herein can be used to assist clinicians in providing patients with ICC with more accurate information regarding prognosis over time.

The current study had several limitations that should be considered. As with all retrospective studies, there may have been a selection bias regarding the diagnosis and treatment of patients with ICC in the study. While the multicenter nature of the study is a considerable strength that conferred a larger sample size and more generalizability, it also undoubtedly led to some heterogeneity in the therapeutic approach to patients. Information on adjuvant therapy, and its potential impact on actuarial and CS estimates, was also not analyzed. However, the goal of the current study was to assess actuarial survival vs CS estimates and the lack of data on adjuvant therapy should not have impacted our ability to achieve this objective.

Conclusions

Survival estimates following surgical resection of ICC change accordingly to survival time accrued since surgical resection. Patients with advanced-stage disease at the time of surgery demonstrated a substantial increase in CS over time. Conditional survival estimates may provide critical quantitative information about the changing probability of survival over time among patients undergoing liver resection for ICC and therefore can be of significant value to patients and health care professionals.

Back to top
Article Information

Corresponding Author: Timothy M. Pawlik, MD, MPH, PhD, Department of Surgery, Johns Hopkins Hospital, 600 N Wolfe St, Blalock 688, Baltimore, MD 21287 (tpawlik1@jhmi.edu).

Accepted for Publication: September 29, 2014.

Published Online: April 1, 2015. doi:10.1001/jamasurg.2015.0219.

Author Contributions: Dr Pawlik had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Spolverato and Kim contributed equally to this project.

Study concept and design: Spolverato, Kim, Pawlik.

Acquisition, analysis, or interpretation of data: Spolverato, Kim, Ejaz, Alexandrescu, Marques, Aldrighetti, Gamblin, Pulitano, Bauer, Shen, Sandroussi, Poultsides, Maithel.

Drafting of the manuscript: Spolverato, Kim.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kim, Ejaz, Shen.

Administrative, technical, or material support: Spolverato, Gamblin, Pulitano, Bauer, Sandroussi, Poultsides, Pawlik.

Study supervision: Spolverato, Marques, Aldrighetti, Sandroussi, Pawlik.

Conflict of Interest Disclosures: None reported.

Additional Contributions: We acknowledge Gilles Mentha, MD (Hôpitaux Universitaires de Genève, Geneva, Switzerland), for important and significant contributions to the study concept and design and data acquisition. He did not receive compensation for his contributions.

References
1.
El-Serag  HB, Engels  EA, Landgren  O,  et al.  Risk of hepatobiliary and pancreatic cancers after hepatitis C virus infection: a population-based study of US veterans. Hepatology. 2009;49(1):116-123.
PubMedArticle
2.
Endo  I, Gonen  M, Yopp  AC,  et al.  Intrahepatic cholangiocarcinoma: rising frequency, improved survival, and determinants of outcome after resection. Ann Surg. 2008;248(1):84-96.
PubMedArticle
3.
Malhi  H, Gores  GJ.  Cholangiocarcinoma: modern advances in understanding a deadly old disease. J Hepatol. 2006;45(6):856-867.
PubMedArticle
4.
Shaib  YH, Davila  JA, McGlynn  K, El-Serag  HB.  Rising incidence of intrahepatic cholangiocarcinoma in the United States: a true increase? J Hepatol. 2004;40(3):472-477.
PubMedArticle
5.
Welzel  TM, McGlynn  KA, Hsing  AW, O’Brien  TR, Pfeiffer  RM.  Impact of classification of hilar cholangiocarcinomas (Klatskin tumors) on the incidence of intra- and extrahepatic cholangiocarcinoma in the United States. J Natl Cancer Inst. 2006;98(12):873-875.
PubMedArticle
6.
Khan  SA, Toledano  MB, Taylor-Robinson  SD.  Epidemiology, risk factors, and pathogenesis of cholangiocarcinoma. HPB (Oxford). 2008;10(2):77-82.
PubMedArticle
7.
Carpizo  DR, D’Angelica  M.  Management and extent of resection for intrahepatic cholangiocarcinoma. Surg Oncol Clin N Am. 2009;18(2):289-305, viii-ix.
PubMedArticle
8.
Patel  T.  Increasing incidence and mortality of primary intrahepatic cholangiocarcinoma in the United States. Hepatology. 2001;33(6):1353-1357.
PubMedArticle
9.
Shaib  Y, El-Serag  HB.  The epidemiology of cholangiocarcinoma. Semin Liver Dis. 2004;24(2):115-125.
PubMedArticle
10.
de Jong  MC, Nathan  H, Sotiropoulos  GC,  et al.  Intrahepatic cholangiocarcinoma: an international multi-institutional analysis of prognostic factors and lymph node assessment. J Clin Oncol. 2011;29(23):3140-3145.
PubMedArticle
11.
Hyder  O, Marques  H, Pulitano  C,  et al.  A nomogram to predict long-term survival after resection for intrahepatic cholangiocarcinoma: an Eastern and Western experience. JAMA Surg. 2014;149(5):432-438.
PubMedArticle
12.
Mavros  MN, Economopoulos  KP, Alexiou  VG, Pawlik  TM.  Treatment and prognosis for patients with intrahepatic cholangiocarcinoma: systematic review and meta-analysis [published online April 9, 2014]. JAMA Surg. doi:10.1001/jamasurg.2013.5137.
PubMed
13.
Edge  S, Byrd  DR, Compton  CC, Fritz  AG, Greene  FL, Trotti  A, eds. AJCC Cancer Staging Manual. 7th ed. New York, NY: Springer; 2010.
14.
Wang  Y, Li  J, Xia  Y,  et al.  Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol. 2013;31(9):1188-1195.
PubMedArticle
15.
Nathan  H, de Jong  MC, Pulitano  C,  et al.  Conditional survival after surgical resection of colorectal liver metastasis: an international multi-institutional analysis of 949 patients. J Am Coll Surg. 2010;210(5):755-764, 764-766.
PubMedArticle
16.
Henson  DE, Ries  LA, Carriaga  MT.  Conditional survival of 56,268 patients with breast cancer. Cancer. 1995;76(2):237-242.
PubMedArticle
17.
Merrill  RM, Henson  DE, Ries  LA.  Conditional survival estimates in 34,963 patients with invasive carcinoma of the colon. Dis Colon Rectum. 1998;41(9):1097-1106.
PubMedArticle
18.
Skuladottir  H, Olsen  JH.  Conditional survival of patients with the four major histologic subgroups of lung cancer in Denmark. J Clin Oncol. 2003;21(16):3035-3040.
PubMedArticle
19.
Sun  M, Trinh  QD, Karakiewicz  PI.  Conditional survival of patients with metastatic renal-cell carcinoma. Lancet Oncol. 2012;13(11):e462.
PubMedArticle
20.
Wang  SJ, Emery  R, Fuller  CD, Kim  JS, Sittig  DF, Thomas  CR.  Conditional survival in gastric cancer: a SEER database analysis. Gastric Cancer. 2007;10(3):153-158.
PubMedArticle
21.
Mayo  SC, Nathan  H, Cameron  JL,  et al.  Conditional survival in patients with pancreatic ductal adenocarcinoma resected with curative intent. Cancer. 2012;118(10):2674-2681.
PubMedArticle
22.
Cucchetti  A, Piscaglia  F, Cescon  M,  et al.  Conditional survival after hepatic resection for hepatocellular carcinoma in cirrhotic patients. Clin Cancer Res. 2012;18(16):4397-4405.
PubMedArticle
23.
Eloranta  S, Lambert  PC, Andersson  TM, Björkholm  M, Dickman  PW.  The application of cure models in the presence of competing risks: a tool for improved risk communication in population-based cancer patient survival. Epidemiology. 2014;25(5):742-748.
PubMedArticle
24.
Daudt  HM, van Mossel  C, Dennis  DL, Leitz  L, Watson  HC, Tanliao  JJ.  Survivorship care plans: a work in progress. Curr Oncol. 2014;21(3):e466-e479.
PubMedArticle
25.
Burnand  B, Kernan  WN, Feinstein  AR.  Indexes and boundaries for “quantitative significance” in statistical decisions. J Clin Epidemiol. 1990;43(12):1273-1284.
PubMedArticle
26.
Spolverato  G, Ejaz  A, Kim  Y,  et al.  Tumor size predicts vascular invasion and histologic grade among patients undergoing resection of intrahepatic cholangiocarcinoma. J Gastrointest Surg. 2014;18(7):1284-1291.
PubMedArticle
27.
Janssen-Heijnen  ML, Gondos  A, Bray  F,  et al.  Clinical relevance of conditional survival of cancer patients in Europe: age-specific analyses of 13 cancers. J Clin Oncol. 2010;28(15):2520-2528.
PubMedArticle
28.
Merrill  RM, Hunter  BD.  Conditional survival among cancer patients in the United States. Oncologist. 2010;15(8):873-882.
PubMedArticle
29.
Johnson  DR, Ma  DJ, Buckner  JC, Hammack  JE.  Conditional probability of long-term survival in glioblastoma: a population-based analysis. Cancer. 2012;118(22):5608-5613.
PubMedArticle
30.
Morimoto  Y, Tanaka  Y, Ito  T,  et al.  Long-term survival and prognostic factors in the surgical treatment for intrahepatic cholangiocarcinoma. J Hepatobiliary Pancreat Surg. 2003;10(6):432-440.
PubMedArticle
31.
Isa  T, Kusano  T, Shimoji  H, Takeshima  Y, Muto  Y, Furukawa  M.  Predictive factors for long-term survival in patients with intrahepatic cholangiocarcinoma. Am J Surg. 2001;181(6):507-511.
PubMedArticle
32.
Choi  SB, Kim  KS, Choi  JY,  et al.  The prognosis and survival outcome of intrahepatic cholangiocarcinoma following surgical resection: association of lymph node metastasis and lymph node dissection with survival. Ann Surg Oncol. 2009;16(11):3048-3056.
PubMedArticle
33.
Ohtsuka  M, Ito  H, Kimura  F,  et al.  Results of surgical treatment for intrahepatic cholangiocarcinoma and clinicopathological factors influencing survival. Br J Surg. 2002;89(12):1525-1531.
PubMedArticle
34.
Li  T, Qin  LX, Zhou  J,  et al.  Staging, prognostic factors and adjuvant therapy of intrahepatic cholangiocarcinoma after curative resection. Liver Int. 2014;34(6):953-960.
PubMedArticle
35.
Janssen-Heijnen  ML, Houterman  S, Lemmens  VE, Brenner  H, Steyerberg  EW, Coebergh  JW.  Prognosis for long-term survivors of cancer. Ann Oncol. 2007;18(8):1408-1413.
PubMedArticle
×