Prognostic nomogram for predicting cancer-specific survival in patients with resected hilar cholangiocarcinoma: a large cohort study

Introduction

Hilar cholangiocarcinoma (HCCA) occurs in epithelial cells within the biliary tree and comprises approximately 50% of all malignant bile duct tumors (1,2). Although curative resection potentially offers HCCA patients great chance for long-term survival, the 5-year overall survival (OS) rate remains ranging from 10% to 40%, far from being satisfactory (3-6). Moreover, approximately 50% to 70% HCCA patients who even underwent radical resection may have tumor recurrence within one year after surgery (7,8). Therefore, to stratify subgroups of the patients with resected HCCA at high risk is imperative, which may aid to optimize the therapeutic strategy. Despite that the American Joint Committee on Cancer (AJCC) TNM staging system is the widely acknowledged and used for HCCA, accurate discrimination of individual patients prognosis remains to be challenging (9,10).

Undoubtedly, the AJCC stage system merely incorporates a limited number of prognostic tumor-related factors and other crucial clinicopathological factors are not included, which may result in an unsatisfying predictive power. In contrast, prognostic nomograms have demonstrated significant advantages over the current staging systems in predicting survival of individual patients in variety of tumors due to the demand of precision medicine and individualized prediction, including lung cancer (11), nasopharyngeal carcinoma (12), colorectal cancer (13), intrahepatic cholangiocarcinoma (ICC) (14,15). For HCCA, in the existing available literature (16-19), most nomograms were derived from a relatively small sample size (16,17) and mainly concentrated on the overall survival after surgery (17-19). Furthermore, most of the existing nomograms were not constructed with controversial prognostic factors, such as tumor size (2,20,21), lymph node ratio (LNR) (22-24). The impact of tumor size and LNR on the prognosis for patients with resected HCCA has remained to be an open question. Until now, to the best of our knowledge, the integrated predictive power of nomogram consisting of age, tumor size, tumor grade, LNR and T stage has not been studied extensively and remains undefined in the resected HCCA.

Hence, the aim of this study was to develop and validate an effective prognostic model mainly based on demographic, tumor-related features and LNR of HCCA and to better predict CSS for HCCA patients after surgical treatment in a large cohort derived from Surveillance, Epidemiology, and End Results (SEER) database. We present the following article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-21-543/rc).

Methods

Patients and study design

The all diagnosed HCCA patients between 2004 and 2015 of the study were extracted from the SEER database (registry database (1975–2016), using the SEER*Stat software version 8.3.6. The inclusion criteria of the present study were as following: (I) the International Classification of Diseases for Oncology, third Edition (ICD-O-3), primary site code C24.0 (extrahepatic bile duct), histology codes (25): 8000–8152, 8154–8231, 8243–8245, 8250–8576, 8940–8950, and 8980–8981 with behavior code (three-malignant tumor); (II) patients having undergone surgical therapy (surgery of primary site codes 20–90) with pathologically confirmed and at least one lymph node examined as well as follow-up time no less than one month. The exclusion criteria were as follows: (I) patients diagnosed at the age younger than 18 years; (II) patients with second primary cancer or multiple primary cancer or distal cholangiocarcinoma or cystic duct carcinoma; (III) patients with unknown tumor size or tumor grade or positive lymph node count (PLNC) or total lymph node count (TLNC) examined; (IV) patients with unknown tumor stage. Two thirds of the patients were randomly grouped to a training cohort to construct nomogram, and the rest of patients were served as an internal validation cohort. To identify individual classifications of T, N, M and overall stage for each eligible patient, the extent of disease and collaborative staging codes provided by SEER as well as available information of PLNC were seriously referred under the guidance of Staging Manual, 6th-, 7th-, and 8th-edition. Considering the public availability of the original data, the current study did not require the institutional review board approval nor informed consent. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013)

Data collection

The following demographic and clinicopathological characteristics of enrolled patients in this study were as follows: age at diagnosis, race, gender, year of diagnosis, tumor size, tumor grade, histology type, surgery type, PLNC, TLNC, LNR, location of positive lymph nodes, new T stage, TNM staging. The LNR was calculated by a ratio of PLNC to TLNC in patients who had at least one lymph node involved. The new T stage (T1, T2a, T2b, T3 and T4) in this study was mainly based on AJCC 8th edition. The CSS was the primary endpoint, which was calculated from the date of diagnosis to the date of HCCA-related death or the latest follow-up. In addition, with regard to the ambiguous definition of T stage of tumors infiltrating organs nearby such as the gall bladder, pancreas or duodenum, we continued to assign the T3/T4 stage to this newly T stage in accordance to the AJCC 6th edition.

Nomogram construction

In the training cohort, all clinicopathological variables were analyzed by univariate analysis of Cox proportional regression model. Then, variables with P value <0.05 in the univariate analyses were enrolled to multivariable Cox regression analysis with forward stepwise method (Forward: LR) to screen for independent prognostic factors. Based on the identified independent prognostic factors, a prognostic nomogram to predict CSS of resected HCCA at 1-, 3- and 5- year was formulated by R version.3.6.1. CSS curves for different independent prognostic factors were generated by Kaplan-Meier method and were compared by using log-rank test.

Nomogram validation

Both concordance indexes (C-index) and calibration plots were performed for the training cohort and validation cohorts, respectively. The C-index represented the discrimination efficiency with value of 0.5 indicating no discrimination power and 1 displaying perfect discrimination ability. Calibration curves were constructed to validate the accuracy of the nomogram for CSS at 1, 3 and 5 years by comparing the survival probability of nomogram predicted and actual observed determined by Kaplan-Meier method with 1000 bootstrap samples.

Clinical utility of nomogram

Decision curve analysis (DCA) was conducted to assess its clinical value between nomogram and AJCC 8th staging system through weighing the net benefits at different threshold probabilities in the training cohort and validation cohort (26,27). The net benefit was calculated by deducing the proportion of all false-positive events from the proportion of true-positive events (13). In addition, a clinical impact curve of the nomogram was employed to directly assess its significance in a given population where the number of high risk patients and the number of high-risk patients with CSS events under given threshold probabilities were analyzed separately (28).

Statistic analysis

Statistical significance was analyzed using analysis of student’s t test (two-tailed)/Mann-Whitney U tests and χ² test/Fisher’s exact test as appropriate. X-tile program based on the maximal Chi-square test was employed to calculate the optimal cutoff values for continuous variables (29). Survival curves were depicted using Kaplan-Meier method and compared by log-rank test. Variables with P value of 0.05 or less in the univariate analyses were subjected to multivariable Cox regression analysis. The nomogram was built with those independent prognostic factors identified in multivariate Cox regression analysis. The nomogram performances of discrimination and calibration power were measured by corresponding C-index and calibration curves. The DCA and clinical impact curves were performed to assess the clinical practice value of this nomogram and AJCC staging systems through weighing the net benefits at different threshold probabilities. Statistical analyses were carried out by SPSS.22 (IBM, NY, USA) and R version 3.6.1 (http://www.r-project.org/) along with rms, rmda and survival R packages. All statistical results with two-sided P value <0.05 were considered statistically significant.

Results

Clinicopathologic characteristics of patients

A total of 806 HCCA eligible patients were included in the present study. There were 536 patients in the training cohort and 270 patients in the internal validation cohort. The median age of the entire study was 67.5 years with 60.42% (n=487) patients being male. The median follow-up time was 84 months (range, 1–155 months), 85 months (range, 1–151 months) respectively in training cohort and validation cohort. During the whole period of follow-up, 74.07% (n=597) patients had died and only 25.93% (n=209) patients were alive. Flowchart of the selected study participants was presented in Figure 1. The details of demographic and clinicopathologic characteristics of HCCA patients in the two cohorts were listed in Table 1.

Figure 1 The flowchart of the study cohort.

Optimal cutoff values of continuous variables for predicting CSS

The median of tumor size, PLNC and TLNC was 2.2 cm (range, 0.1–25 cm, mean: 2.83 cm), 0 (range, 0–19, mean: 1.35), 7 (range, 1–65; mean: 9.98). In order to facilitate the application of variables on tumor size, PLNC, TLNC and LNR in clinical practice, these continuous variables were converted into categorical variables according to its optimal cutoff value determined by X-tile program. In the training cohort, the optimal cutoff value of tumor size was 3 cm, 3 of PLNC, 7 of TLNC, 0.21 of LNR, respectively. The predictive performance of these transformed categorical variables was further validated in an independent internal cohort.

Independent prognostic factors of CSS

Univariate Cox regression analysis revealed that age, tumor size, tumor grade, PLNC, LNR, T stage were potential prognostic factors (P<0.05) associated with CSS in the Training cohort. Furthermore, these potential prognostic factors were enrolled into multivariate Cox regression analysis model with forward stepwise method (Forward: LR). Ultimately, elder age (≥65years), poor tumor grade (poor/undifferentiation), larger tumor size (≥3cm), high level of LNR (≥0.21) and advanced T stage were identified as independent prognostic factors for CSS (Table 2). In addition, the survival curves of CSS stratified by these single independent prognostic factors also demonstrated significant difference (Figure 2).

Figure 2 The cancer-specific survival of hilar cholangiocarcinoma (HCCA) patients who underwent surgery stratified by different independent prognostic factors: (A) age; (B) tumor size; (C) tumor grade; (D) lymph node ratio (LNR); (E) T stage.

Construction of nomogram for CSS

Based on those identified independent prognostic factors, nomogram was formulated to predict 1-, 3-, and 5-year CSS of patients with resected HCCA. All covariates in the nomogram were listed in Figure 3. The C-index for CSS prediction nomogram was 0.655 (95% CI: 0.626–0.0.684) in the training cohort, 0.626 (95% CI: 0.585–0.669) in the validation cohort.

Figure 3 Nomogram for predicting the 1-, 3- and 5-year probabilities of cancer-specific survival in resected hilar cholangiocarcinoma (HCCA) patients in the training cohort. The nomogram was constructed by age, tumor grade, tumor size, lymph node ratio (LNR) and T stage. CSS, cancer-specific survival.

Validation and performance of nomogram for CSS

The calibration plots for the probability of 1-, 3- and 5-year CSS presented optimal agreement between the nomogram prediction and actual observations in the training cohort and validation cohort (Figure 4). In the training cohort, the C-index in the nomogram was 0.655 slightly higher than that of 0.631 in AJCC 7th and 0.631 in AJCC 8th. In the validation cohort, the C-index of the nomogram was 0.626, while it was 0.623 in AJCC 7th and 0.626 in AJCC 8th. In addition, the value of C-index for single predictor of age, tumor grade, tumor size, LNR and for T stage were 0.53, 0.547, 0.582, 0.549, 0.595, respectively, which were significantly inferior to that in this nomogram model.

Figure 4 Calibration curves for predicting 1-, 3- and 5-year cancer-specific survival (CSS) in patients with hilar cholangiocarcinoma (HCCA) after surgery in the training cohort (A, B and C, respectively) and in the internal validation cohort (D, E and F, respectively). The X-axis represented the nomogram-predicted probability of CSS and the Y-axis represented actual observed survival. The diagonal grey line represents an ideal evaluation where the predicted probabilities were identical to that of actual observed.

Clinical application of nomogram for CSS

The decision curve demonstrated that by applying the established nomogram to predict CSS would obtain more net benefit compared with AJCC 8th staging system, when the threshold probability ranged from 0.4 to 0.7 in the training cohort (Figure 5A) or 0.45 to 0.65 in the internal validation cohort (Figure 5B). Additionally, the clinical impact curve of the nomogram demonstrated that lower cost might bring more benefit especially when the risk threshold was less than 0.9 in the training cohort and validation cohort (Figure 5C,5D).

Figure 5 Decision curve analysis (DCA) (A and B) for the nomogram and AJCC 8th staging system and clinical impact curves (C and D) of the nomogram for cancer-specific survival (CSS) in hilar cholangiocarcinoma (HCCA) patients after surgery in the training cohort (A and C) and validation cohort (B and D). In the DCA, X-axis represented threshold probability and Y-axis represented net benefit. The horizontal black line represented the assumption that all patients survived, and the grey line represented the assumption that all patients died.

Discussion

Although AJCC TNM staging systems is specifically developed for HCCA prognosis prediction which is now being used widely, the predictive ability of survival is relatively lower compared with previous nomograms constructed by the most relevant clinicopathological factors (16-19). Simultaneously, majority of elderly cholangiocarcinoma patients tend to suffer from competing events such as comorbidities, which in turn affects the OS of those patients and makes the OS difficult to determine (30). However, the CSS is mainly determined by tumor biology including tumor size, tumor grade and tumor histology, which is independent on the competing events. Therefore, it is a better choice to predict CSS of HCCA by nomogam for individualized therapy.

The new constructed nomogram, aimed to predict CSS of resected HCCA, with a C-index of 0.655 (95% CI: 0.626–0.0.684) was superior to AJCC TNM 7th staging system with a C-index of 0.631 (95% CI: 0.6016–0.6604) and AJCC TNM 8th staging system with a C-index of 0.631 (95% CI: 0.6016–0.6604) in the training cohort, though it was inferior to 0.73 in the Groot Koerkamp et al. (16) study and 0.66 of van der Gaag et al. (31) study. Given that the controversy over whether the higher C-index of nomogram represented better clinical usefulness (27,32), to comprehensively assess the benefits, DCA was recommended to evaluate the clinical practical value of a predictive model by quantifying its net benefit according to the threshold probability in previous studies (33,34). In the current study, the decision curve demonstrated that the established novel nomogram not only obtained more net benefit but also lowered cost/benefit ratio to some extent compared with AJCC 8th staging system. Furthermore, the performance of this developed nomogram was further validated by an independent internal cohort, where the C-index was 0.626 (95% CI: 0.585–0.669) and calibration plots represented an excellent match without overfitting, also the DCA and clinical impact curves confirmed the similar clinical application value provided by the nomogram. These results demonstrated that the constructed nomogram could outperform the AJCC 8th staging system to predict CSS of resected HCCA patients in clinical practice with better discrimination and accuracy.

In the present study, the established nomogram was constructed by 5 independent prognostic factors of CSS in a large cohort from SEER database including elder age (HR: 1.322, 95% CI: 1.068–1.636, P=0.01), poor tumor grade (HR: 1.305, 95% CI: 1.050–1.622, P=0.016), larger tumor size (HR: 1.288, 95% CI: 1.032–1.607, P=0.025), higher level of LNR (HR: 1.893, 95% CI: 1.489–2.406, P<0.001) and advanced T stage (P<0.001) identified in the multivariate Cox regression analysis. Compared with the established risk factors in previously reported nomograms, such as age (17,18), surgical margin (16,17), lymph node status (16-18,31), tumor grade (16,31) and tumor marker (17), this study was mainly focused on the controversial prognostic factors such as tumor size and LNR. Simultaneously, to the best of our knowledge, it was the first time that nomogram incorporating age, tumor size, tumor grade, LNR and T stage was applied to predict the CSS of HCCA after surgery, of which the objective was different from the study of Chen et al. (17) and Qi et al. (18), and by deriving from a large population-based cohort would reasonably achieve more reliable results than those from single center or smaller sample size to some extent.

Tumor size is a powerful and reliable intuitive indicator (35), which has been widely accepted as a robust parameter of predicting prognosis of intrahepatic cholangiocarcinoma (ICC) and hepatocellular carcinoma (HCC) in AJCC TNM staging system. Even though HCCA is different from ICC and HCC in clinicopathological features and therapeutic strategies (20,36), whether tumor diameter is an independent risk factor for prognosis of HCCA remains controversial (2,20,37,38). In this study, tumor size was identified as an independent prognostic factor of CSS. The median CSS of 19 months for patients with tumor size larger than 3 cm was worse than 28 months for those with tumors smaller than 3 cm (P=0.001). In line with our research findings, tumor size (over 3 cm) was also regarded as a significant parameter and accepted as an independent risk factor influencing HCCA prognosis in the DeOliveira et al. (20) staging system and Roberts et al. (2) clinical staging system. An increasing number of evidence proved that tumor diameter was an independent risk factor for survival, though the consensus on the optimal cutpoint of tumor diameter remains controversial (2,20,39,40). By contrast, tumor grade has been established and considered as a measure index of biological aggressiveness of a tumor and a predictor of long-term survival (7,16,41). A study concerning the relationship of tumor-size and tumor differentiation for HCCA demonstrated larger tumors was prone to be poorly differentiated (39), which may account for the rationality of tumor size being a robust parameter. In addition, it was reported that the aggressiveness of tumor was positively correlated with tumor size in pancreatic ductal adenocarcinoma, which may underscore its’ role of measuring tumor burden in cancer patients (42).

Despite that LNR is a controversial prognostic factor, a growing number of studies have been recommended it as an alternative independent prognostic factor in HCCA (22-24). In present study, it was found that median CSS of 32 months for HCCA patients with LNR of 0.21 or less was significantly better than 15 months for patients with LNR greater than 0.21 (P<0.001), which was in line with the results of previous studies with LNR ranging from 0.20 in the study by Nagino et al. (24) to 0.25 in the study by Guglielmi et al. (23) to 0.37 in the study by Hakeem et al. (22). One of the possible reasons was that potential metastasized lymph nodes would only be detected with harvesting more lymph nodes, which may result in more precise tumor staging (43,44). Undoubtedly, the N-stage has been revised and re-classified by the amount of resected positive lymph nodes instead of by the location of LN in recently released AJCC 8th-edition, which clearly underlines the significance of PLNC. Whereas, it may ignore the fact that the PLNC was vulnerable to TLNC. Additionally, although the log odds of positive lymph nodes (LODDS) in the published studies indicated that it was considered to be the most accurate and effective LN staging system for resected HCCA (45,46), both LNR and LODDS exhibited similar discriminatory ability to predict prognosis among patients with over 4 lymph nodes retrieved (47). Of note, the HCCA was usually performed with regional lymphadenectomy even extensive lymphadenectomy, which resulted in more than 4 lymph nodes retrieved (47). Thus, the prognostic value of LODDS for HCCA remained to be further validated. Hence, we believed that merely PLNC may not completely reflect the disease status but the LNR might facilitate lymph nodes staging so as to have a more precise prediction of prognosis.

To further investigate the predictive value of the T stage for HCCA, we recoded the T stage in the AJCC 6th system by referring to the AJCC 7th/8th edition. In the new T stage, tumor infiltrating organs nearby including the gall bladder, pancreas or duodenum was continued to assign the T3/T4 stage according to the AJCC 6th edition. However, the newly released AJCC 8th staging system failed to emphasize the definite T stage classification of neighboring organs infiltration except the liver parenchyma (T2b), resulting in insufficient guidance for the clinical practices. Furthermore, the study conducted by Juntermanns et al. indicated that the more exact definition of tumor infiltration would facilitate the way of improving HCCA patients stratification particularly for intermediate tumor stages (48). Thus, it was very necessary to assign the definite T stage for the infiltrated neighboring organs, which would facilitate the individualized therapy strategy developing. Additionally, further studies are required to clarify the rationality of recoding T stage strategy.

There were still several limitations that must be considered in this study. For one thing, due to lacking details of surgical margin, Bismuth classification, surgical procedure, postoperative radiotherapy and chemotherapy, it was challenging to verify whether the other surgical approaches were equivalent to palliative procedure, which might partly explain that there was no statistically significant difference between radical surgery group and others in this study. For another, clinicopathological features of surgical margin status and vascular invasion as well as tumor marker such as CA19-9 were unavailable in SEER database, which resulted in the data derived from SEER database failing to work as an eligible external validated cohort to assess the predictive power of previously established nomograms and make a direct comparison with this novel model. Despite that the established nomogram demonstrated good accuracy for predicting CSS, the C-index of nomogram was not excellent. Simultaneously, due to lack of adequate samples and sufficient available data, randomly splitting a single data set into training and validation cohorts would lead to insufficient statistic power (49). In addition, external validation was not performed in this study. Taken together, those might affect the model’s repeatability and reliability to some extent.

In conclusion, a novel nomogram for predicting CSS of resected HCCA patients based on histological variables was generated and validated by a internal validation cohort, which may facilitate the process of identifying resected HCCA patients with high risk and optimizing therapeutic benefit.

Acknowledgments

Funding: This work was supported by Grant [2013] 163 from Key Laboratory of Malignant Tumor Mechanism and Translational Medicine of Guangzhou Bureau of Science and Information Technology; Grant KLB 09001 from the Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institutes; National Natural Science Foundation of China (Nos. 81572407, 81672405); Key project of Natural Science Foundation of Guangdong Province, China (No. 4210016041); Guangdong Basic and Applied Basic Research Foundation (No. 2015A030313096, 2019A1515011418); Natural Science Foundation of Guangzhou, China (No. 4250016043); Grant from Guangdong Science and Technology Department (Nos. 2017B030314026, 2020B1212060018).

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-21-543/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-21-543/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Considering the public availability of the original data, the current study did not require the institutional review board approval or informed consent. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Khan SA, Davidson BR, Goldin RD, et al. Guidelines for the diagnosis and treatment of cholangiocarcinoma: an update. Gut 2012;61:1657-69. [Crossref] [PubMed]
2. Chaiteerakij R, Harmsen WS, Marrero CR, et al. A new clinically based staging system for perihilar cholangiocarcinoma. Am J Gastroenterol 2014;109:1881-90. [Crossref] [PubMed]
3. Razumilava N, Gores GJ. Cholangiocarcinoma. Lancet 2014;383:2168-79. [Crossref] [PubMed]
4. Nagino M. Perihilar cholangiocarcinoma: a surgeon's viewpoint on current topics. J Gastroenterol 2012;47:1165-76. [Crossref] [PubMed]
5. Soares KC, Kamel I, Cosgrove DP, et al. Hilar cholangiocarcinoma: diagnosis, treatment options, and management. Hepatobiliary Surg Nutr 2014;3:18-34. [PubMed]
6. Hemming AW, Reed AI, Fujita S, et al. Surgical management of hilar cholangiocarcinoma. Ann Surg 2005;241:693-9; discussion 699-702. [Crossref] [PubMed]
7. Saxena A, Chua TC, Chu FC, et al. Improved outcomes after aggressive surgical resection of hilar cholangiocarcinoma: a critical analysis of recurrence and survival. Am J Surg 2011;202:310-20. [Crossref] [PubMed]
8. Regimbeau JM, Fuks D, Pessaux P, et al. Tumour size over 3 cm predicts poor short-term outcomes after major liver resection for hilar cholangiocarcinoma. By the HC-AFC-2009 group. HPB (Oxford) 2015;17:79-86. [Crossref] [PubMed]
9. Kwon W, Jang JY, Chang YR, et al. Suggestions for improving perihilar cholangiocarcinoma staging based on an evaluation of the seventh edition AJCC system. J Gastrointest Surg 2015;19:666-74.
10. Ebata T, Kosuge T, Hirano S, et al. Proposal to modify the International Union Against Cancer staging system for perihilar cholangiocarcinomas. Br J Surg 2014;101:79-88. [Crossref] [PubMed]
11. Liang W, Zhang L, Jiang G, et al. Development and validation of a nomogram for predicting survival in patients with resected non-small-cell lung cancer. J Clin Oncol 2015;33:861-9. [Crossref] [PubMed]
12. Tang LQ, Li CF, Li J, et al. Establishment and Validation of Prognostic Nomograms for Endemic Nasopharyngeal Carcinoma. J Natl Cancer Inst 2016;108:djv291. [Crossref] [PubMed]
13. Huang YQ, Liang CH, He L, et al. Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol 2016;34:2157-64. [Crossref] [PubMed]
14. Wang Y, Li J, Xia Y, et al. Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol 2013;31:1188-95. [Crossref] [PubMed]
15. Jing CY, Fu YP, Huang JL, et al. Prognostic Nomogram Based on Histological Characteristics of Fibrotic Tumor Stroma in Patients Who Underwent Curative Resection for Intrahepatic Cholangiocarcinoma. Oncologist 2018;23:1482-93. [Crossref] [PubMed]
16. Groot Koerkamp B, Wiggers JK, Gonen M, et al. Survival after resection of perihilar cholangiocarcinoma-development and external validation of a prognostic nomogram. Ann Oncol 2015;26:1930-5. [Crossref] [PubMed]
17. Chen P, Li B, Zhu Y, et al. Establishment and validation of a prognostic nomogram for patients with resectable perihilar cholangiocarcinoma. Oncotarget 2016;7:37319-30. [Crossref] [PubMed]
18. Qi F, Zhou B, Xia J. Nomograms predict survival outcome of Klatskin tumors patients. PeerJ 2020;8:e8570. [Crossref] [PubMed]
19. Buettner S, van Vugt JL, Gani F, et al. A Comparison of Prognostic Schemes for Perihilar Cholangiocarcinoma. J Gastrointest Surg 2016;20:1716-24. [Crossref] [PubMed]
20. Deoliveira ML, Schulick RD, Nimura Y, et al. New staging system and a registry for perihilar cholangiocarcinoma. Hepatology 2011;53:1363-71. [Crossref] [PubMed]
21. Zhang Y, Wu Z, Wang X, et al. Development and external validation of a nomogram for predicting the effect of tumor size on survival of patients with perihilar cholangiocarcinoma. BMC Cancer 2020;20:1044. [Crossref] [PubMed]
22. Hakeem AR, Marangoni G, Chapman SJ, et al. Does the extent of lymphadenectomy, number of lymph nodes, positive lymph node ratio and neutrophil-lymphocyte ratio impact surgical outcome of perihilar cholangiocarcinoma? Eur J Gastroenterol Hepatol 2014;26:1047-54. [Crossref] [PubMed]
23. Guglielmi A, Ruzzenente A, Campagnaro T, et al. Prognostic significance of lymph node ratio after resection of peri-hilar cholangiocarcinoma. HPB (Oxford) 2011;13:240-5. [Crossref] [PubMed]
24. Aoba T, Ebata T, Yokoyama Y, et al. Assessment of nodal status for perihilar cholangiocarcinoma: location, number, or ratio of involved nodes. Ann Surg 2013;257:718-25. [Crossref] [PubMed]
25. Mao K, Liu J, Sun J, et al. Patterns and prognostic value of lymph node dissection for resected perihilar cholangiocarcinoma. J Gastroenterol Hepatol 2016;31:417-26. [Crossref] [PubMed]
26. Vickers AJ, Cronin AM, Elkin EB, et al. Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers. BMC Med Inform Decis Mak 2008;8:53. [Crossref] [PubMed]
27. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making 2006;26:565-74. [Crossref] [PubMed]
28. Kerr KF, Brown MD, Zhu K, et al. Assessing the Clinical Impact of Risk Prediction Models With Decision Curves: Guidance for Correct Interpretation and Appropriate Use. J Clin Oncol 2016;34:2534-40. [Crossref] [PubMed]
29. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004;10:7252-9. [Crossref] [PubMed]
30. Antwi SO, Mousa OY, Patel T. Racial, Ethnic, and Age Disparities in Incidence and Survival of Intrahepatic Cholangiocarcinoma in the United States; 1995-2014. Ann Hepatol 2018;17:604-14. [Crossref] [PubMed]
31. van der Gaag NA, Kloek JJ, de Bakker JK, et al. Survival analysis and prognostic nomogram for patients undergoing resection of extrahepatic cholangiocarcinoma. Ann Oncol 2012;23:2642-9. [Crossref] [PubMed]
32. Balachandran VP, Gonen M, Smith JJ, et al. Nomograms in oncology: more than meets the eye. Lancet Oncol 2015;16:e173-80. [Crossref] [PubMed]
33. Wang ZX, Qiu MZ, Jiang YM, et al. Comparison of prognostic nomograms based on different nodal staging systems in patients with resected gastric cancer. J Cancer 2017;8:950-8. [Crossref] [PubMed]
34. Hijazi Z, Oldgren J, Lindbäck J, et al. The novel biomarker-based ABC (age, biomarkers, clinical history)-bleeding risk score for patients with atrial fibrillation: a derivation and validation study. Lancet 2016;387:2302-11. [Crossref] [PubMed]
35. Chun YS, Pawlik TM, Vauthey JN. 8th Edition of the AJCC Cancer Staging Manual: Pancreas and Hepatobiliary Cancers. Ann Surg Oncol 2018;25:845-7.
36. Mansour JC, Aloia TA, Crane CH, et al. Hilar cholangiocarcinoma: expert consensus statement. HPB (Oxford) 2015;17:691-9. [Crossref] [PubMed]
37. Jarnagin WR, Fong Y, DeMatteo RP, et al. Staging, resectability, and outcome in 225 patients with hilar cholangiocarcinoma. Ann Surg 2001;234:507-17; discussion 517-9. [Crossref] [PubMed]
38. Ito F, Agni R, Rettammel RJ, et al. Resection of hilar cholangiocarcinoma: concomitant liver resection decreases hepatic recurrence. Ann Surg 2008;248:273-9. [Crossref] [PubMed]
39. Hu HJ, Zhou RX, Shrestha A, et al. Relationship of tumor size with pathological and prognostic factors for hilar cholangiocarcinoma. Oncotarget 2017;8:105011-9. [Crossref] [PubMed]
40. DeOliveira ML, Cunningham SC, Cameron JL, et al. Cholangiocarcinoma: thirty-one-Year Experience With 564 Patients at a Single Institution. Ann Surg 2007;245:755-62. [Crossref] [PubMed]
41. Jang JY, Kim SW, Park DJ, et al. Actual long-term outcome of extrahepatic bile duct cancer after surgical resection. Ann Surg 2005;241:77-84. [Crossref] [PubMed]
42. He C, Zhong L, Zhang Y, et al. Development and validation of a nomogram to predict liver metastasis in patients with pancreatic ductal adenocarcinoma: a large cohort study. Cancer Manag Res 2019;11:3981-91. [Crossref] [PubMed]
43. Ito K, Ito H, Allen PJ, et al. Adequate lymph node assessment for extrahepatic bile duct adenocarcinoma. Ann Surg 2010;251:675-81. [Crossref] [PubMed]
44. Li X, Lin H, Sun Y, et al. Prognostic Significance of the Lymph Node Ratio in Surgical Patients With Distal Cholangiocarcinoma. J Surg Res 2019;236:2-11. [Crossref] [PubMed]
45. Li P, Song L. A novel prognostic nomogram for patients with surgically resected perihilar cholangiocarcinoma: a SEER-based study. Ann Transl Med 2021;9:54. [Crossref] [PubMed]
46. Conci S, Ruzzenente A, Sandri M, et al. What is the most accurate lymph node staging method for perihilar cholangiocarcinoma? Comparison of UICC/AJCC pN stage, number of metastatic lymph nodes, lymph node ratio, and log odds of metastatic lymph nodes. Eur J Surg Oncol 2017;43:743-50. [Crossref] [PubMed]
47. Bagante F, Tran T, Spolverato G, et al. Perihilar Cholangiocarcinoma: Number of Nodes Examined and Optimal Lymph Node Prognostic Scheme. J Am Coll Surg 2016;222:750-759.e2. [Crossref] [PubMed]
48. Juntermanns B, Sotiropoulos GC, Radunz S, et al. Comparison of the sixth and the seventh editions of the UICC classification for perihilar cholangiocarcinoma. Ann Surg Oncol 2013;20:277-84.
49. Moons KG, Altman DG, Reitsma JB, et al. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med 2015;162:W1-73. [Crossref] [PubMed]