Retrospective Cohort Study Open Access
Copyright ©The Author(s) 2025. Published by Baishideng Publishing Group Inc. All rights reserved.
World J Gastroenterol. Mar 21, 2025; 31(11): 100911
Published online Mar 21, 2025. doi: 10.3748/wjg.v31.i11.100911
Preoperative prediction of textbook outcome in intrahepatic cholangiocarcinoma by interpretable machine learning: A multicenter cohort study
Ting-Feng Huang, Hong-Zhi Liu, Qi-Zhu Lin, Rui-Lin Fan, Shi-Chuan Tang, Yong-Yi Zeng, Department of Hepatobiliary Surgery, Mengchao Hepatobiliary Hospital of Fujian Medical University, Fuzhou 350025, Fujian Province, China
Cong Luo, Department of Hepatopancreatobiliary Surgery, The People’s Hospital of Zizhong County, Neijiang 540045, Sichuan Province, China
Luo-Bin Guo, Ke-Can Lin, Department of Hepatopancreatobiliary Surgery, Mengchao Hepatobiliary Hospital of Fujian Medical University, Fuzhou 350025, Fujian Province, China
Jiang-Tao Li, Department of General Surgery, The Second Affiliated Hospital Zhejiang University School of Medicine, Hangzhou 310009, Zhejiang Province, China
Wei-Ping Zhou, Department of the 3rd Liver Surgery, Eastern Hepatobiliary Surgery Hospital, The Second Military Medical University, Shanghai 200438, China
Jing-Dong Li, The Affiliated Hospital of North Sichuan Medical College, Nanchong 637000, Sichuan Province, China
ORCID number: Jing-Dong Li (0000-0002-6934-0148); Yong-Yi Zeng (0000-0001-7441-4271).
Co-first authors: Ting-Feng Huang and Cong Luo.
Co-corresponding authors: Shi-Chuan Tang and Yong-Yi Zeng.
Author contributions: Zeng YY, Huang TF and Tang SC conceptualized and designed the research; Li JT, Li JD Zhou WP and Zeng YY screened patients and acquired clinical data; Guo LB, Lin QZ, Fan RL, Lin KC and Liu HZ performed Data analysis; Huang TF wrote the paper; All the authors have read and approved the final manuscript. Huang TF proposed, designed and analysis the date, performed data analysis and prepared the first draft of the manuscript. Luo C was responsible for patient screening, enrollment, collection of clinical data. Both authors have made crucial and indispensable contributions towards the completion of the project and thus qualified as the co-first authors of the paper. Both Zeng YY and Tang SC have played important and indispensable roles in the experimental design, data interpretation and manuscript preparation as the co-corresponding authors. Zeng YY applied for and obtained the funds for this research project. Tang SC conceptualized, designed, and supervised the whole process of the project. He searched the literature, revised and submitted the early version of the manuscript with the focus on the interpretability of machine learning algorithms. Zeng YY was instrumental and responsible for data re-analysis and re-interpretation, figure plotting, comprehensive literature search, preparation and submission of the current version of the manuscript with a new focus on interpretability and external validation of machine learning. This collaboration between Zeng YY and Tang SC is crucial for the publication of this manuscript and other manuscripts still in preparation.
Supported by National Key Research and Development Program, No. 2022YFC2407304; Major Research Project for Middle-Aged and Young Scientists of Fujian Provincial Health Commission, No. 2021ZQNZD013; The National Natural Science Foundation of China, No. 62275050; Fujian Province Science and Technology Innovation Joint Fund Project, No. 2019Y9108; and Major Science and Technology Projects of Fujian Province, No. 2021YZ036017.
Institutional review board statement: The study was reviewed and approved by the Ethics Committee of the Mengchao Hepatobiliary Hospital of Fujian Medical University, and exempts the requirement of written informed consent. All procedures were performed in accordance with World Medical Association Declaration of Helsinki, (Approval No. 2024_041_01).
Informed consent statement: The need for patient consent was waived due to the retrospective nature of the study.
Conflict-of-interest statement: All the authors report no relevant conflicts of interest for this article.
STROBE statement: The authors have read the STROBE Statement—checklist of items, and the manuscript was prepared and revised according to the STROBE Statement—checklist of items.
Data sharing statement: The data where our results derived from were from Mengchao Hepatobiliary Hospital of Fujian Medical University. The original data were not publicly available and could only be shared with the permission of the ethics committee of Mengchao Hospital.
Open Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Yong-Yi Zeng, MD, PhD, Professor, Department of Hepatobiliary Surgery, Mengchao Hepatobiliary Hospital of Fujian Medical University, No. 312 Xihong Road, Fuzhou 350025, Fujian Province, China. lamp197311@126.com
Received: August 30, 2024
Revised: January 10, 2025
Accepted: February 13, 2025
Published online: March 21, 2025
Processing time: 195 Days and 8.1 Hours

Abstract
BACKGROUND

To investigate the preoperative factors influencing textbook outcomes (TO) in Intrahepatic cholangiocarcinoma (ICC) patients and evaluate the feasibility of an interpretable machine learning model for preoperative prediction of TO, we developed a machine learning model for preoperative prediction of TO and used the SHapley Additive exPlanations (SHAP) technique to illustrate the prediction process.

AIM

To analyze the factors influencing textbook outcomes before surgery and to establish interpretable machine learning models for preoperative prediction.

METHODS

A total of 376 patients diagnosed with ICC were retrospectively collected from four major medical institutions in China, covering the period from 2011 to 2017. Logistic regression analysis was conducted to identify preoperative variables associated with achieving TO. Based on these variables, an EXtreme Gradient Boosting (XGBoost) machine learning prediction model was constructed using the XGBoost package. The SHAP (package: Shapviz) algorithm was employed to visualize each variable's contribution to the model's predictions. Kaplan-Meier survival analysis was performed to compare the prognostic differences between the TO-achieving and non-TO-achieving groups.

RESULTS

Among 376 patients, 287 were included in the training group and 89 in the validation group. Logistic regression identified the following preoperative variables influencing TO: Child-Pugh classification, Eastern Cooperative Oncology Group (ECOG) score, hepatitis B, and tumor size. The XGBoost prediction model demonstrated high accuracy in internal validation (AUC = 0.8825) and external validation (AUC = 0.8346). Survival analysis revealed that the disease-free survival rates for patients achieving TO at 1, 2, and 3 years were 64.2%, 56.8%, and 43.4%, respectively.

CONCLUSION

Child-Pugh classification, ECOG score, hepatitis B, and tumor size are preoperative predictors of TO. In both the training group and the validation group, the machine learning model had certain effectiveness in predicting TO before surgery. The SHAP algorithm provided intuitive visualization of the machine learning prediction process, enhancing its interpretability.

Key Words: Intrahepatic cholangiocarcinoma; Textbook outcome; Interpretable machine learning; Prediction; Prognosis

Core Tip: This study developed a machine learning model to preoperatively predict the Textbook outcome (TO), a measure of surgical quality and short-term prognosis, and utilized the SHapley Additive exPlanations technique to enhance model transparency. Based on the analysis of 376 intrahepatic cholangiocarcinoma patients from four Chinese medical institutions, logistic regression identified key preoperative factors, including Child-Pugh classification, Eastern Cooperative Oncology Group score, hepatitis B status, and tumor size. The EXtreme Gradient Boosting algorithm was used to construct the prediction model, while SHAP visualized its decision-making process. The model effectively stratified recurrence-free survival, demonstrating its utility in preoperative TO prediction.
INTRODUCTION

Intrahepatic cholangiocarcinoma (ICC), the second most prevalent primary liver tumor[1], is characterized by high malignancy and rapid progression[2,3], with a globally increasing incidence[4,5]. Although surgical resection remains the primary treatment for this condition[6], over 50% of patients experience relapse within 2 years post-surgery[7,8]. Recurrence serves as a significant risk factor for poor prognosis among ICC patients[9,10].

Good postoperative conditions are one of the prerequisites for long-term survival. Thus, alongside evaluating long-term survival, assessing short-term postoperative prognosis is equally critical[11,12]. Currently, single indicators such as postoperative complications and length of hospital stay are commonly used in clinical practice to evaluate short-term postoperative prognosis[13,14]. However, these isolated metrics fail to comprehensively reflect the quality of surgery and the overall diagnosis and treatment process[15]. As a composite index for short-term postoperative prognosis, the textbook outcome (TO) provides a more accurate and comprehensive evaluation of surgical quality and overall diagnosis and treatment[16]. TO encompasses several criteria, including negative surgical margins, absence of perioperative blood transfusions, absence of postoperative complications, a short hospital stay, no deaths within 30 days of surgery, and no readmissions within 30 days of discharge[17]. Previous studies have shown that achieving TO not only facilitates better psychological and physical recovery but also improves patients' quality of life, prolongs their long-term prognosis, and may enhance disease-free survival (DFS)[18,19]. Therefore, assessing whether a patient can achieve TO serves as an important preoperative evaluation. It not only alerts medical staff to improve preoperative preparation but also lays the foundation for analyzing patient prognosis[20].

In recent years, numerous preoperative predictive models have been developed to predict the prognosis of ICC patients. For example, Zhu et al[21] constructed a nomogram model based on the immune-inflammatory-nutritional score to predict the survival and recurrence, achieving a certain level of accuracy. Xin et al[22] utilized microbiome analysis to assess risks and predict prognosis in ICC patients. Other studies have employed radiomic differences between cancerous and peritumoral tissues to predict ICC prognosis[23]. While perioperative conditions significantly influence patients' quality of life and long-term outcomes, few studies have used preoperative indicators to predict perioperative outcomes and, subsequently, patient prognosis.

In summary, determining whether patients can achieve TO before surgery holds great clinical significance. Therefore, this study applies machine learning algorithms to predict TO, enabling accurate preoperative assessment of short-term postoperative prognosis. Additionally, the incorporation of the Shapley Additive Explanation (SHAP) algorithm enhances the interpretability of the model, offering better insights into the prediction process for clinical application.

MATERIALS AND METHODS
Patient and data collection

The study was approved by the Ethics Committee of the Mengchao Hepatobiliary Hospital of Fujian Medical University, and exempts the requirement of written informed consent. All procedures were performed in accordance with the World Medical Association Declaration of Helsinki. The database was retrospectively derived from patients with pathologically confirmed ICC who underwent hepatic resection at the following institutions: Mengchao Hepatobiliary Hospital of Fujian Medical University (39 cases), Eastern Hepatobiliary Surgery Hospital of Naval Medical University (287 cases), The Second Affiliated Hospital of Zhejiang University School of Medicine (26 cases), and the Affiliated Hospital of North Sichuan Medical College (24 cases).

Inclusion criteria: (1) Patients who underwent hepatectomy; (2) Pathological confirmation of ICC; (3) No concurrent malignant tumors; and (4) Complete clinicopathologic data and follow-up information.

Exclusion criteria: (1) Receipt of other anti-tumor treatments prior to surgery; (2) Patients who underwent palliative surgery or had incomplete macroscopic tumor resection; and (3) Missing clinicopathologic data or loss to follow-up after discharge.

Clinicopathologic variables

The evaluation criteria are as follows.

DFS: Defined as the time from surgery to death or recurrence from any cause.

Maximum tumor diameter: Defined as the preoperative tumor diameter.

Tumor burden: A solitary tumor is defined as a single lesion in the liver, whereas two or more lesions are classified as multiple tumors.

Included variables: The variables analyzed include gender, age, hepatitis B infection, Eastern Cooperative Oncology Group (ECOG) score (0, 1), Child-Pugh classification, carbohydrate antigen 19-9, alpha-fetoprotein, carcinoembryonic antigen, total bilirubin, direct bilirubin, perioperative blood transfusion, surgical margins (positive or negative), complications, perioperative death or readmission within 30 days after discharge, and prolonged hospitalization.

Definition and description of variables: (1) ECOG Performance Status Scale: 0: Fully active, able to carry out all pre-disease activities without any restrictions, 1: Capable of ambulatory and light physical activity, including light housework or office work, but unable to perform moderate to heavy physical activities; (2) Child-Pugh classification: Used to assess liver function for all patients; and (3) Perioperative blood transfusion: Refers to the administration of blood products (e.g., red blood cells, platelets, plasma) during the perioperative period.

Achieving TO: Defined as meeting all of the following criteria: Negative surgical margins, no perioperative blood transfusion, no complications, no prolonged hospitalization (hospitalization duration ≤ the median length of stay for the cohort), and no deaths within 30 days post-surgery.

Statistical analysis

All data analyses were performed by R software (version 4.1.0, http://www.r-project.org). Quantitative data conforming to a normal distribution were expressed as mean ± SD, while data not conforming to a normal distribution were presented as median [interquartile range]. Qualitative data were reported as numbers and percentages (n, %). Univariate logistic regression was conducted for each independent variable to evaluate its association with the dependent variable, producing odds ratios, confidence intervals, and P values. Variables with statistical significance (P ≤ 0.05) were selected for multivariate logistic regression, which assesses the effects of these variables while controlling for potential confounders. After multivariate regression analysis, variables with P values ≤ 0.05 were identified as independent influencing factors. The area under the curve (AUC) was used to evaluate the performance of the predictive model. The Xgboost package in R was employed to construct the XGBoost prediction model.

To enhance model interpretability, the SHAP algorithm was applied using the shapviz package. The SHAP algorithm enhances the interpretability of machine learning models by quantifying the contribution of each feature to individual predictions, using Shapley values derived from game theory. It calculates the marginal contributions of each feature to the prediction results. The SHAP value for each feature indicates its impact on a specific prediction: Positive values reflect an increase in the prediction, while negative values represent a decrease.

RESULTS
Baseline characteristics of patients

Among 376 patients, 287 were in the training group and 89 were in the validation group. Among all patients, 47 patients (12.5%) had perioperative blood transfusion, 18 patients (4.79%) had positive margins, 12 patients (3.19%) were readmitted within 30 days or perioperatively died, and 132 patients (35.11%) had prolonged hospitalization (Tables 1 and 2).

Table 1 Baseline characteristics of patients in training group, n (%).
Variables
Total (n = 287)
Textbook outcome (n = 150)
None-textbook outcome (n = 137)
P value
Age, M (Q1, Q3)56.00 (48.00, 63.00)55.00 (47.00, 62.00)59.00 (49.00, 63.00)0.118
Sex0.011
    Male189 (65.85)109 (72.67)80 (58.39)
    Female98 (34.15)41 (27.33)57 (41.61)
Hepatitis B0.002
    Negative179 (62.37)81 (54.00)98 (71.53)
    Positive108 (37.63)69 (46.00)39 (28.47)
ECOG score0.026
    085 (29.62)53 (35.33)32 (23.36)
    1202 (70.38)97 (64.67)105 (76.64)
Child Pugh0.005
    A175 (60.98)80 (53.33)95 (69.34)
    B112 (39.02)70 (46.67)42 (30.66)
Tumor size, M (Q1, Q3)6.40 (4.65, 8.90)6.10 (4.30, 8.40)7.00 (5.00, 9.20)0.098
Tumor number0.870
    Solitary192 (66.90)101 (67.33)91 (66.42)
    Multiple95 (33.10)49 (32.67)46 (33.58)
TBIL, M (Q1, Q3)20.00 (12.95, 31.90)22.90 (14.50, 32.90)17.10 (11.20, 28.10)0.013
DBIL, M (Q1, Q3)7.90 (4.60, 14.75)8.85 (5.30, 15.45)7.30 (3.50, 14.40)0.012
CA19-9, M (Q1, Q3)17.30 (8.95, 34.20)20.90 (13.90, 33.55)11.50 (3.90, 37.30)< 0.001
AFP, M (Q1, Q3)12.70 (3.30, 171.50)29.10 (3.62, 314.50)7.40 (3.20, 78.20)0.065
CEA, M (Q1, Q3)2.60 (1.60, 4.10)2.45 (1.50, 4.00)2.60 (1.60, 4.60)0.219
Blood transfusion< 0.001
    No247 (86.06)150 (100.00)97 (70.80)
    Yes40 (13.94)0 (0.00)40 (29.20)
Surgical margins< 0.001
    Negative276 (96.17)150 (100.00)126 (91.97)
    Positive11 (3.83)0 (0.00)11 (8.03)
Complications< 0.001
    No227 (79.09)150 (100.00)77 (56.20)
    Yes60 (20.91)0 (0.00)60 (43.80)
Perioperative death or readmission within 30 days after discharge0.008
    No279 (97.21)150 (100.00)129 (94.16)
    Yes8 (2.79)0 (0.00)8 (5.84)
Prolonged hospitalization< 0.001
    No187 (65.16)150 (100.00)37 (27.01)
    Yes100 (34.84)0 (0.00)100 (72.99)
M: Median; Q1: 1st quartile, Q3: 3st quartile; ECOG score: Eastern Cooperative Oncology Group score; TBIL: Total bilirubin; DBIL: Direct bilirubin; CA19-9: Carbohydrate antigen 19-9; AFP: Alpha-fetoprotein; CEA: Carcinoembryonic antigen.
Table 2 Baseline characteristics of patients in validation group, n (%).
Variables
Total (n = 89)
Textbook outcome (n = 49)
None-textbook outcome (n = 40)
P value
Age, M (Q1, Q3)58.00 (48.00, 65.00)59.00 (53.00, 64.00)56.50 (46.75, 65.00)0.501
Sex0.167
    Male53 (59.55)26 (53.06)27 (67.50)
    Female36 (40.45)23 (46.94)13 (32.50)
Hepatitis B0.190
    Negative64 (71.91)38 (77.55)26 (65.00)
    Positive25 (28.09)11 (22.45)14 (35.00)
ECOG score0.001
    032 (35.96)25 (51.02)7 (17.50)
    157 (64.04)24 (48.98)33 (82.50)
Child-Pugh0.017
    A57 (64.04)26 (53.06)31 (77.50)
    B32 (35.96)23 (46.94)9 (22.50)
Tumor size, M (Q1, Q3)6.50 (5.00, 8.40)6.70 (4.70, 8.40)6.50 (5.00, 8.27)0.853
Tumor number0.104
    Solitary66 (74.16)33 (67.35)33 (82.50)
    Multiple23 (25.84)16 (32.65)7 (17.50)
TBIL, M (Q1, Q3)17.20 (10.70, 33.10)19.00 (10.60, 41.50)16.00 (11.18, 28.75)0.668
DBIL, M (Q1, Q3)7.90 (4.60, 12.70)8.90 (5.10, 12.70)6.45 (4.33, 12.25)0.203
CA19-9, M (Q1, Q3)109.10 (13.90, 778.00)192.70 (17.60, 1000.00)83.95 (8.10, 485.52)0.073
AFP, M (Q1, Q3)6.20 (2.70, 97.60)5.10 (2.40, 97.60)6.95 (2.70, 74.53)0.738
CEA, M (Q1, Q3)2.70 (1.80, 4.79)2.50 (1.70, 3.70)3.10 (2.00, 10.67)0.030
Blood transfusion0.008
    No82 (92.13)49 (100.00)33 (82.50)
    Yes7 (7.87)0 (0.00)7 (17.50)
Surgical margins0.008
    Negative82 (92.13)49 (100.00)33 (82.50)
    Positive7 (7.87)0 (0.00)7 (17.50)
Complications< 0.001
    No70 (78.65)49 (100.00)21 (52.50)
    Yes19 (21.35)0 (0.00)19 (47.50)
Perioperative death or readmission within 30 days after discharge0.08
    No85 (95.51)48 (100.00)37 (90.00)
    Yes4 (4.49)0 (0.00)4 (10.00)
Prolonged hospitalization< 0.001
    No57 (64.04)49 (100.00)8 (20.00)
    Yes32 (35.96)0 (0.00)32 (80.00)
M: Median; Q1: 1st quartile; Q3: 3st quartile; ECOG score: Eastern Cooperative Oncology Group score; TBIL: Total bilirubin; DBIL: Direct bilirubin; CA19-9: Carbohydrate antigen 19-9; AFP: Alpha-fetoprotein; CEA: Carcinoembryonic antigen.
Influencing factors of TO

Based on existing domestic and foreign research and previous clinical experience, 12 observation indicators were finally included for variable screening. Single-factor logistic regression showed that: Child-Pugh grade, ECOG score, hepatitis B, and tumor size were preoperative factors that affected the TO (P < 0.05). Multi-factor results showed that: Child-Pugh grade, ECOG score, hepatitis B, and tumor size were preoperative factors that affected the TO (P < 0.05), which were independent influencing factors that affect reaching the TO (Table 3).

Table 3 Influencing factors of textbook outcome.
Characteristics
OR
95%CI
P value
OR1
95%CI1
P value1
Female1.891.16-3.110.0111.590.95-2.690.08
Age1.010.99-1.030.41NANANA
Hepatitis B0.470.29-0.760.0020.580.35-0.980.041
Child-Pugh A0.510.31-0.820.0060.50.3-0.830.007
ECOG score = 11.791.07-3.010.0271.91.1-3.280.022
Tumor size1.081.01-1.170.0321.091.01-1.180.036
Multiple tumor1.040.64-1.70.87NANANA
AFP level11-10.486NANANA
CA19-9 level11-10.544NANANA
CEA level11-10.656NANANA
TBIL level10.99-10.228NANANA
DBIL level10.99-10.262NANANA
1The results of multifactor logistic regression.
OR: Odds ratio; ECOG score: Eastern Cooperative Oncology Group score; TBIL: Total bilirubin; DBIL: Direct bilirubin; CA19-9: Carbohydrate antigen 19-9; AFP: Alpha-fetoprotein; CEA: Carcinoembryonic antigen; NA: Not available.
XGboost model development

We used the XGboost package in R language to build the XGboost prediction model. To facilitate subsequent model training and verification, we randomly divided the data into two parts, with 70% of the data serving as the training set and 30% as the validation set using an algorithm. The XGboost prediction model was established using the variables screened by logistic regression: Child-Pugh grade, ECOG score, hepatitis B, and tumor size. The results showed that the Xgboost model has good prediction effect in train group (AUC = 0.882) (Figure 1A) and validation group (AUC = 0.834) (Figure 1B).

Figure 1
Open in New Tab   Full Size Figure   Download Figure
Figure 1 The area under the curve of EXtreme Gradient Boosting model. A: The results showed that the EXtreme Gradient Boosting (XGBoost) model has good prediction effect in the train group [area under the curve (AUC) = 0.882]; B: The XGboost model has good prediction effect in validation group (AUC = 0.834). AUC: Area under the curve.
XGBoost model for SHAP

To make the XGboost model interpretable, we combine the SHAP algorithm with the XGboost model. In the overall visualization, the four variables were ranked in importance. The SHAP histogram shows that the variables with weights from high to low in the model are: Tumor size, Child-Pugh grade, hepatitis B, and ECOG score (Figure 2). In the SHAP bee swarm plot, each point represents the SHAP value of a sample (Figure 3). The color of the point represents the original value of the feature. Red represents high values and blue represents low values. The results show that tumor size has the most important impact on model prediction.

Figure 2
Open in New Tab   Full Size Figure   Download Figure
Figure 2 Variable importance ranking. The SHapley Additive exPlanations histogram shows that the variables with weights from high to low in the model are: Tumor size, Child-Pugh grade, hepatitis B, and Eastern Cooperative Oncology Group score. ECOG score: Eastern Cooperative Oncology Group score; SHAP: The SHapley Additive exPlanations.
Figure 3
Open in New Tab   Full Size Figure   Download Figure
Figure 3 The SHapley Additive exPlanations bee swarm plot. The color of the point represents the original value of the feature. Red represents high values and blue represents low values. The results show that tumor size has the most important impact on model prediction. ECOG score: Eastern Cooperative Oncology Group score.

Figure 4 shows the prediction process of patients by the XGboost model. Taking No. 1 patient as an example, the patient has a tumor of 8 cm (-0.154), HBV negative (+0.197), and liver function Child A grade (+0.136). ECOG 0 points (-0.922). The overall score f (x) = -0.743 points, so the model believes that the patient cannot achieve the TO, which is consistent with the actual situation.

Figure 4
Open in New Tab   Full Size Figure   Download Figure
Figure 4 Machine learning prediction process diagram. Taking No. 1 patient as an example, the patient has a tumor of 8 cm (-0.154), hepatitis B virus negative (+0.197), and liver function Child A grade (+0.136). Eastern Cooperative Oncology Group 0 points (-0.922). The overall score f (x) = -0.743 points, so the model believes that the patient cannot achieve the textbook outcome, which is consistent with the actual situation. ECOG score: Eastern Cooperative Oncology Group score; SHAP: The SHapley Additive exPlanations.

SHAP values and scatter point colors reflect the influence relationship between variables. Different features interact with each other to predict the TO. For example: In hepatitis B-negative patients, the smaller the tumor, the higher the SHAP value, indicating that in hepatitis B-negative patients, the smaller the tumor, the easier it is to achieve the TO (Figure 5).

Figure 5
Open in New Tab   Full Size Figure   Download Figure
Figure 5 Diagram of variable interaction. A: Hepatitis B; B: Eastern Cooperative Oncology Group score; C: Child-Pugh classification; D: Tumor size. Different features interact with each other to predict the textbook outcome (TO). For example: In hepatitis B-negative patients, the smaller the tumor, the higher the SHapley Additive exPlanations value, indicating that in hepatitis B-negative patients, the smaller the tumor, the easier it is to achieve the TO. ECOG score: Eastern Cooperative Oncology Group score; SHAP: The SHapley Additive exPlanations.
The correlation analysis between TO and DFS

Survival analysis shows that the DFS of patients reaching the TO in 1, 2, and 3 years were 64.2%, 56.8%, and 43.4%. The DFS of patients that did not meet the TO in 1, 2, and 3 years were 44.7%, 32.5%, and 25.2%. There was a difference in DFS between the two groups (P < 0.05; Figure 6A). In order to evaluate the impact of the model prediction results on prognosis, patients were divided into the predicted TO group and the predicted not to meet the TO group based on the model prediction results. The results showed that the predicted outcomes were successfully stratify DFS (Figure 6B).

Figure 6
Open in New Tab   Full Size Figure   Download Figure
Figure 6 Disease-free survival. A: Disease-free survival (DFS) of patients who achieved textbook outcome (TO) and those who did not. Survival analysis shows that the DFS of patients reaching the TO in 1, 2, and 3 years were 642%, 56.8%, and 43.4%; B: DFS chart of patients predicted to achieve TO and those predicted not to achieve. The DFS of patients that did not meet the TO in 1, 2, and 3 years were 447%, 32.5%, and 25.2%. TO: Textbook outcome.
DISCUSSION

Achieving TO may improve patient survival and DFS and serves as an effective indicator of medical quality[24]. In this study, we used logistic regression to screen influencing factors based on 12 preoperative clinical indicators and developed a predictive model using machine learning algorithms. An XGBoost preoperative prediction model with high accuracy was trained, and the SHAP algorithm was employed to visualize the model. This approach intuitively demonstrated the prediction process from individual cases to overall trends, thereby elucidating the previously opaque operations of machine learning.

Postoperative complications and prolonged hospital stays are the main challenges faced by patients after surgery. In this study, 21.01% of patients experienced complications postoperatively. These complications can prolong hospital stays and reduce quality of life[25]. Some studies have shown that older age, cirrhosis, and low liver reserve function are potential risk factors for complications[14,26]. Therefore, identifying high-risk patients and implementing early intervention strategies may reduce complications and improve the rate of achieving TO. The median hospitalization time in this study was 10 days, and prolonged hospital stays were associated with complications and poor postoperative recovery[27]. Active preoperative preparation and the adoption of Enhanced Recovery After Surgery protocols may help reduce hospitalization durations[28-30], thereby increasing the proportion of patients achieving TO.

Our study identified several preoperative factors influencing the achievement of TO, including Child-Pugh classification, ECOG score, hepatitis B, and tumor size. Child-Pugh classification is an important indicator of liver reserve function[31]. Patients with poor liver reserve and low compensatory ability are more likely to experience postoperative complications and delayed recovery. Some studies have demonstrated the prognostic significance of the preoperative Child-Pugh classification[32] and highlighted the higher risk of complications after lymph node dissection (LND) in patients with liver cirrhosis[25]. The ECOG score is another key indicator of a patient's general health and daily functioning[33]. It has been used as a prognostic factor in several studies[34]. Accurate preoperative assessment of the ECOG score can provide valuable guidance for predicting postoperative outcomes. Hepatitis B is a major contributor to the development of ICC. Surgery and postoperative adjuvant therapy can potentially reactivate the hepatitis B virus, negatively affecting both short-term and long-term prognoses[35]. Research suggests that standardized preoperative and postoperative antiviral treatments can improve surgical outcomes and patient survival[36,37]. Tumor size is an independent risk factor affecting patient prognosis. Larger tumors are associated with more complex surgeries, longer operation times, and a higher likelihood of complications and surgical accidents[7]. Some studies have pointed out that larger tumors tend to exhibit lymph node involvement, which can increase surgical time and complications[38]. Our findings underscore the importance of early detection, diagnosis, and treatment to improve outcomes for patients with smaller tumors.

Preoperative judgment of the patient’s post-operative status is particularly important for perioperative preparation and accident prevention[39]. Predicting whether a patient will achieve TO prior to surgery can boost surgeons' confidence and guide preoperative planning. Conversely, identifying patients unlikely to achieve TO preoperatively can prompt enhanced preparations and more cautious surgical approaches. While many predictive models have been developed to assess the prognosis of ICC patients[40], most focus on postoperative pathological factors. This study addresses a gap by establishing a machine learning model to predict postoperative status and short-term outcomes using the composite TO index. By incorporating the SHAP algorithm, we addressed the interpretability challenge of machine learning models, enabling clinicians to more intuitively evaluate the likelihood of achieving TO. Furthermore, DFS analysis revealed that patients achieving TO had better DFS outcomes than those who did not, consistent with prior research[41]. Achieving TO benefits patients not only in the short term but also by delaying tumor recurrence.

This study has certain limitations. As a retrospective study, selection bias is inevitable. Additionally, the number of variables included was relatively small, and factors such as preoperative body mass index, hypertension, and diabetes were not considered. Including these variables could improve model accuracy and will be an avenue for future research. Furthermore, the high prevalence of hepatitis B in China may limit the model's generalizability to other populations. Expanding the sample size to include Western populations will be essential for future studies. Lastly, our study could not clarify the relationship between tumor size and hepatitis B status. Previous studies suggest that a tumor size cutoff of 5 cm stratifies prognosis in hepatitis B-related liver cancer, but this correlation has not been observed in non-hepatitis B patients[42]. HBV-associated ICC may originate from hepatocytes, leading to lower invasiveness and a better prognosis[43]. However, the impact of tumor size in this context remains uncertain and warrants further investigation.

It is important to emphasize that failure to achieve TO does not imply inadequate treatment. Similarly, preoperative predictions of failure to achieve TO do not mean patients have lost their chances of survival. Physicians should provide extra attention and care to patients predicted to fall short of achieving TO, striving to increase the proportion of patients reaching this benchmark.

CONCLUSION

The machine learning model predicts TO with good accuracy, and the DFS of patients who achieve the TO is better than that of patients who do not. The SHAP algorithm enables the visualization of the machine learning prediction model.

Footnotes

Provenance and peer review: Unsolicited article; Externally peer reviewed.

Peer-review model: Single blind

Corresponding Author's Membership in Professional Societies: Member of the Liver Cancer Quality Control Expert Committee of the National Cancer Center.

Specialty type: Gastroenterology and hepatology

Country of origin: China

Peer-review report’s classification

Scientific Quality: Grade B, Grade B, Grade C, Grade C

Novelty: Grade B, Grade B, Grade B, Grade B

Creativity or Innovation: Grade B, Grade B, Grade B, Grade B

Scientific Significance: Grade B, Grade B, Grade B, Grade B

P-Reviewer: Guo SB; Ling Y; Lv X S-Editor: Li L L-Editor: A P-Editor: Li X

References
1.  Bridgewater J, Galle PR, Khan SA, Llovet JM, Park JW, Patel T, Pawlik TM, Gores GJ. Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma. J Hepatol. 2014;60:1268-1289.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 862]  [Cited by in RCA: 1035]  [Article Influence: 94.1]  [Reference Citation Analysis (0)]
2.  Deng L, Chen B, Zhan C, Yu H, Zheng J, Bao W, Deng T, Zheng C, Wu L, Yang Y, Yu Z, Wang Y, Chen G. A Novel Clinical-Radiomics Model Based on Sarcopenia and Radiomics for Predicting the Prognosis of Intrahepatic Cholangiocarcinoma After Radical Hepatectomy. Front Oncol. 2021;11:744311.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 4]  [Cited by in RCA: 4]  [Article Influence: 1.0]  [Reference Citation Analysis (0)]
3.  Razumilava N, Gores GJ. Cholangiocarcinoma. Lancet. 2014;383:2168-2179.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1072]  [Cited by in RCA: 1323]  [Article Influence: 120.3]  [Reference Citation Analysis (1)]
4.  A S, Wu H, Wang X, Wang X, Yang J, Xia L, Xia Y. Value of glycogen synthase 2 in intrahepatic cholangiocarcinoma prognosis assessment and its influence on the activity of cancer cells. Bioengineered. 2021;12:12167-12178.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
5.  Nathan H, Pawlik TM, Wolfgang CL, Choti MA, Cameron JL, Schulick RD. Trends in survival after surgery for cholangiocarcinoma: a 30-year population-based SEER database analysis. J Gastrointest Surg. 2007;11:1488-1496; discussion 1496.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 172]  [Cited by in RCA: 184]  [Article Influence: 10.2]  [Reference Citation Analysis (0)]
6.  Kelley RK, Bridgewater J, Gores GJ, Zhu AX. Systemic therapies for intrahepatic cholangiocarcinoma. J Hepatol. 2020;72:353-363.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 144]  [Cited by in RCA: 238]  [Article Influence: 47.6]  [Reference Citation Analysis (0)]
7.  Kawashima J, Sahara K, Shen F, Guglielmi A, Aldrighetti L, Weiss M, Bauer TW, Alexandrescu S, Poultsides GA, Maithel SK, Marques HP, Martel G, Pulitano C, Cauchy F, Koerkamp BG, Matsuyama R, Endo I, Pawlik TM. Predicting risk of recurrence after resection of stage I intrahepatic cholangiocarcinoma. J Gastrointest Surg. 2024;28:18-25.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
8.  Chun YS, Javle M. Systemic and Adjuvant Therapies for Intrahepatic Cholangiocarcinoma. Cancer Control. 2017;24:1073274817729241.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 59]  [Cited by in RCA: 73]  [Article Influence: 10.4]  [Reference Citation Analysis (0)]
9.  Mavros MN, Economopoulos KP, Alexiou VG, Pawlik TM. Treatment and Prognosis for Patients With Intrahepatic Cholangiocarcinoma: Systematic Review and Meta-analysis. JAMA Surg. 2014;149:565-574.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 380]  [Cited by in RCA: 562]  [Article Influence: 56.2]  [Reference Citation Analysis (0)]
10.  Holzner ML, Mazzaferro V, Busset MDD, Aldrighetti L, Ratti F, Hasegawa K, Arita J, Sapisochin G, Abreu P, Schoning W, Schmelzle M, Nevermann N, Pratschke J, Florman S, Halazun K, Schwartz ME, Tabrizian P. Is Repeat Resection for Recurrent Intrahepatic Cholangiocarcinoma Warranted? Outcomes of an International Analysis. Ann Surg Oncol. 2024;31:4397-4404.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Reference Citation Analysis (0)]
11.  Shen Z, Tao L, Cai J, Zheng J, Sheng Y, Yang Z, Gong L, Song C, Gao J, Ying H, Xu J, Liang X. Safety and feasibility of laparoscopic liver resection for intrahepatic cholangiocarcinoma: a propensity score-matched study. World J Surg Oncol. 2023;21:126.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in RCA: 3]  [Reference Citation Analysis (0)]
12.  Xu L, Shao Z, Huang H, Li D, Wang T, Atyah M, Zhou W, Yang Z. Impact of Frailty on Short-Term Outcomes of Hepatic Lobectomy in Patients with Intrahepatic Cholangiocarcinoma: Evidence from the US Nationwide Inpatient Sample 2005-2018. Dig Surg. 2024;41:42-52.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Cited by in RCA: 1]  [Article Influence: 1.0]  [Reference Citation Analysis (0)]
13.  Kim KD, Lee JE, Kim J, Ro J, Rhu J, Choi GS, Heo JS, Joh JW. Laparoscopic liver resection as a treatment option for intrahepatic cholangiocarcinoma. Updates Surg. 2024;76:869-878.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
14.  Endo Y, Moazzam Z, Woldesenbet S, Araujo Lima H, Alaimo L, Munir MM, Shaikh CF, Guglielmi A, Aldrighetti L, Weiss M, Bauer TW, Alexandrescu S, Poultsides GA, Kitago M, Maithel SK, Marques HP, Martel G, Pulitano C, Shen F, Cauchy F, Koerkamp BG, Endo I, Pawlik TM. Predictors and Prognostic Significance of Postoperative Complications for Patients with Intrahepatic Cholangiocarcinoma. World J Surg. 2023;47:1792-1800.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in RCA: 4]  [Reference Citation Analysis (0)]
15.  Pavicevic S, Reichelt S, Uluk D, Lurje I, Engelmann C, Modest DP, Pelzer U, Krenzien F, Raschzok N, Benzing C, Sauer IM, Stintzing S, Tacke F, Schöning W, Schmelzle M, Pratschke J, Lurje G. Prognostic and Predictive Molecular Markers in Cholangiocarcinoma. Cancers (Basel). 2022;14:1026.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 9]  [Cited by in RCA: 21]  [Article Influence: 7.0]  [Reference Citation Analysis (0)]
16.  Hobeika C, Cauchy F, Fuks D, Barbier L, Fabre JM, Boleslawski E, Regimbeau JM, Farges O, Pruvot FR, Pessaux P, Salamé E, Soubrane O, Vibert E, Scatton O; AFC-LLR-2018 study group. Laparoscopic versus open resection of intrahepatic cholangiocarcinoma: nationwide analysis. Br J Surg. 2021;108:419-426.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 11]  [Cited by in RCA: 32]  [Article Influence: 8.0]  [Reference Citation Analysis (0)]
17.  Merath K, Chen Q, Bagante F, Alexandrescu S, Marques HP, Aldrighetti L, Maithel SK, Pulitano C, Weiss MJ, Bauer TW, Shen F, Poultsides GA, Soubrane O, Martel G, Koerkamp BG, Guglielmi A, Itaru E, Cloyd JM, Pawlik TM. A Multi-institutional International Analysis of Textbook Outcomes Among Patients Undergoing Curative-Intent Resection of Intrahepatic Cholangiocarcinoma. JAMA Surg. 2019;154:e190571.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 87]  [Cited by in RCA: 161]  [Article Influence: 26.8]  [Reference Citation Analysis (0)]
18.  Xu Z, Lv Y, Zou H, Jia Y, Du W, Lu J, Liu Y, Shao Z, Zhang H, Sun C, Zhu C. Textbook outcome of laparoscopic hepatectomy in the context of precision surgery: A single center experience. Dig Liver Dis. 2024;56:1368-1374.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Cited by in RCA: 1]  [Article Influence: 1.0]  [Reference Citation Analysis (0)]
19.  Atyah MM, Xu L, Yang Z. Novel definition of textbook outcome in biliary system cancers and its influence on patients' survival and quality of life. Cancer Med. 2024;13:e7186.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
20.  Bobrzynski L, Sędłak K, Rawicz-Pruszyński K, Kolodziejczyk P, Szczepanik A, Polkowski W, Richter P, Sierzega M. Evaluation of optimum classification measures used to define textbook outcome among patients undergoing curative-intent resection of gastric cancer. BMC Cancer. 2023;23:1199.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
21.  Zhu J, Wang D, Liu C, Huang R, Gao F, Feng X, Lan T, Li H, Wu H. Development and validation of a new prognostic immune-inflammatory-nutritional score for predicting outcomes after curative resection for intrahepatic cholangiocarcinoma: A multicenter study. Front Immunol. 2023;14:1165510.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in RCA: 13]  [Reference Citation Analysis (0)]
22.  Xin HY, Zou JX, Sun RQ, Hu ZQ, Chen Z, Luo CB, Zhou ZJ, Wang PC, Li J, Yu SY, Liu KX, Fan J, Zhou J, Zhou SL. Characterization of tumor microbiome and associations with prognosis in intrahepatic cholangiocarcinoma. J Gastroenterol. 2024;59:411-423.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
23.  Fiz F, Rossi N, Langella S, Conci S, Serenari M, Ardito F, Cucchetti A, Gallo T, Zamboni GA, Mosconi C, Boldrini L, Mirarchi M, Cirillo S, Ruzzenente A, Pecorella I, Russolillo N, Borzi M, Vara G, Mele C, Ercolani G, Giuliante F, Cescon M, Guglielmi A, Ferrero A, Sollini M, Chiti A, Torzilli G, Ieva F, Viganò L. Radiomics of Intrahepatic Cholangiocarcinoma and Peritumoral Tissue Predicts Postoperative Survival: Development of a CT-Based Clinical-Radiomic Model. Ann Surg Oncol. 2024;31:5604-5614.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Reference Citation Analysis (0)]
24.  Tsilimigras DI, Sahara K, Moris D, Mehta R, Paredes AZ, Ratti F, Marques HP, Soubrane O, Lam V, Poultsides GA, Popescu I, Alexandrescu S, Martel G, Workneh A, Guglielmi A, Hugh T, Aldrighetti L, Weiss M, Bauer TW, Maithel SK, Pulitano C, Shen F, Koerkamp BG, Endo I, Pawlik TM. Assessing Textbook Outcomes Following Liver Surgery for Primary Liver Cancer Over a 12-Year Time Period at Major Hepatobiliary Centers. Ann Surg Oncol. 2020;27:3318-3327.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 60]  [Cited by in RCA: 65]  [Article Influence: 13.0]  [Reference Citation Analysis (0)]
25.  Zhou YM, Sui CJ, Zhang XF, Li B, Yang JM. Influence of cirrhosis on long-term prognosis after surgery in patients with combined hepatocellular-cholangiocarcinoma. BMC Gastroenterol. 2017;17:25.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 18]  [Cited by in RCA: 18]  [Article Influence: 2.3]  [Reference Citation Analysis (0)]
26.  Kauppila JH, Johar A, Lagergren P. Postoperative Complications and Health-related Quality of Life 10 Years After Esophageal Cancer Surgery. Ann Surg. 2020;271:311-316.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 28]  [Cited by in RCA: 44]  [Article Influence: 8.8]  [Reference Citation Analysis (0)]
27.  Aliseda D, Sapisochin G, Martí-Cruchaga P, Zozaya G, Blanco N, Goh BKP, Rotellar F. Association of Laparoscopic Surgery with Improved Perioperative and Survival Outcomes in Patients with Resectable Intrahepatic Cholangiocarcinoma: A Systematic Review and Meta-Analysis from Propensity-Score Matched Studies. Ann Surg Oncol. 2023;30:4888-4901.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 7]  [Cited by in RCA: 5]  [Article Influence: 2.5]  [Reference Citation Analysis (0)]
28.  Peng LH, Wang WJ, Chen J, Jin JY, Min S, Qin PP. Implementation of the pre-operative rehabilitation recovery protocol and its effect on the quality of recovery after colorectal surgeries. Chin Med J (Engl). 2021;134:2865-2873.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Cited by in RCA: 25]  [Article Influence: 6.3]  [Reference Citation Analysis (0)]
29.  Wang LH, Zhu RF, Gao C, Wang SL, Shen LZ. Application of enhanced recovery after gastric cancer surgery: An updated meta-analysis. World J Gastroenterol. 2018;24:1562-1578.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in CrossRef: 54]  [Cited by in RCA: 58]  [Article Influence: 8.3]  [Reference Citation Analysis (0)]
30.  Huang Y, Xie Q, Wei X, Shi Q, Zhou Q, Leng X, Miao Y, Han Y, Wang K, Fang Q. Enhanced Recovery Protocol Versus Conventional Care in Patients Undergoing Esophagectomy for Cancer: Advantages in Clinical and Patient-Reported Outcomes. Ann Surg Oncol. 2024;31:5706-5716.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Reference Citation Analysis (0)]
31.  Wen X, Yao M, Lu Y, Chen J, Zhou J, Chen X, Zhang Y, Lu W, Qian X, Zhao J, Zhang L, Ding S, Lu F. Integration of Prealbumin into Child-Pugh Classification Improves Prognosis Predicting Accuracy in HCC Patients Considering Curative Surgery. J Clin Transl Hepatol. 2018;6:377-384.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 7]  [Cited by in RCA: 8]  [Article Influence: 1.1]  [Reference Citation Analysis (0)]
32.  Li H, Li J, Wang J, Liu H, Cai B, Wang G, Wu H. Assessment of Liver Function for Evaluation of Long-Term Outcomes of Intrahepatic Cholangiocarcinoma: A Multi-Institutional Analysis of 620 Patients. Front Oncol. 2020;10:525.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 8]  [Cited by in RCA: 10]  [Article Influence: 2.0]  [Reference Citation Analysis (0)]
33.  Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, Mcfadden ET, Carbone PP. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5:649-656.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 7038]  [Cited by in RCA: 7803]  [Article Influence: 185.8]  [Reference Citation Analysis (0)]
34.  Garajová I, Gelsomino F, Salati M, Mingozzi A, Peroni M, De Lorenzo S, Granito A, Tovoli F, Leonardi F. Second-Line Chemotherapy for Intrahepatic Cholangiocarcinomas: What Is the Real Gain? Life (Basel). 2023;13:2170.  [PubMed]  [DOI]  [Cited in This Article: ]  [Reference Citation Analysis (0)]
35.  Huang G, Lai EC, Lau WY, Zhou WP, Shen F, Pan ZY, Fu SY, Wu MC. Posthepatectomy HBV reactivation in hepatitis B-related hepatocellular carcinoma influences postoperative survival in patients with preoperative low HBV-DNA levels. Ann Surg. 2013;257:490-505.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 76]  [Cited by in RCA: 99]  [Article Influence: 8.3]  [Reference Citation Analysis (0)]
36.  Liang Y, Zhong D, Zhang Z, Su Y, Yan S, Lai C, Yao Y, Shi Y, Huang X, Shang J. Impact of preoperative antiviral therapy on the prognosis of hepatitis B virus-related hepatocellular carcinoma. BMC Cancer. 2024;24:291.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 1]  [Cited by in RCA: 4]  [Article Influence: 4.0]  [Reference Citation Analysis (0)]
37.  Wang XH, Hu ZL, Fu YZ, Hou JY, Li WX, Zhang YJ, Xu L, Zhou QF, Chen MS, Zhou ZG. Tenofovir vs. entecavir on prognosis of hepatitis B virus-related hepatocellular carcinoma after curative resection. J Gastroenterol. 2022;57:185-198.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 6]  [Cited by in RCA: 19]  [Article Influence: 6.3]  [Reference Citation Analysis (0)]
38.  Chen YX, Zeng ZC, Tang ZY, Fan J, Zhou J, Jiang W, Zeng MS, Tan YS. Prediction of the lymph node status in patients with intrahepatic cholangiocarcinoma: analysis of 320 surgical cases. Front Oncol. 2011;1:42.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 10]  [Cited by in RCA: 13]  [Article Influence: 0.9]  [Reference Citation Analysis (0)]
39.  Guo SB, Hu LS, Huang WJ, Zhou ZZ, Luo HY, Tian XP. Comparative investigation of neoadjuvant immunotherapy versus adjuvant immunotherapy in perioperative patients with cancer: a global-scale, cross-sectional, and large-sample informatics study. Int J Surg. 2024;110:4660-4671.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 2]  [Reference Citation Analysis (0)]
40.  Mashiko T, Carreras J, Ogasawara T, Masuoka Y, Ei S, Takahashi S, Nomura T, Mori M, Koyanagi K, Yamamoto S, Nakamura N, Nakagohri T. Intrahepatic cholangiocarcinoma with arterial phase hyperenhancement and specialized tumor microenvironment associated with good prognosis after radical resection: A single-center retrospective study. Surgery. 2024;176:259-266.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 4]  [Reference Citation Analysis (0)]
41.  Clocchiatti L, Marino R, Ratti F, Pedica F, Casadei Gardini A, Lorenzin D, Aldrighetti L. Defining and predicting textbook outcomes for perihilar cholangiocarcinoma: analysis of factors improving achievement of desired postoperative outcomes. Int J Surg. 2024;110:209-218.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in RCA: 6]  [Reference Citation Analysis (0)]
42.  Hwang S, Lee YJ, Kim KH, Ahn CS, Moon DB, Ha TY, Song GW, Jung DH, Lee SG. The Impact of Tumor Size on Long-Term Survival Outcomes After Resection of Solitary Hepatocellular Carcinoma: Single-Institution Experience with 2558 Patients. J Gastrointest Surg. 2015;19:1281-1290.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in Crossref: 93]  [Cited by in RCA: 119]  [Article Influence: 11.9]  [Reference Citation Analysis (0)]
43.  Song Z, Lin S, Wu X, Ren X, Wu Y, Wen H, Qian B, Lin H, Huang Y, Zhao C, Wang N, Huang Y, Peng B, Li X, Peng H, Shen S. Hepatitis B virus-related intrahepatic cholangiocarcinoma originates from hepatocytes. Hepatol Int. 2023;17:1300-1317.  [PubMed]  [DOI]  [Cited in This Article: ]  [Cited by in RCA: 3]  [Reference Citation Analysis (0)]