Shi Y, Zhang P, Li L, Yang HM, Li ZM, Zheng J, Yang L. Interpretable model based on multisequence magnetic resonance imaging radiomics for predicting the pathological grades of hepatocellular carcinomas. World J Radiol 2025; 17(12): 112911 [DOI: 10.4329/wjr.v17.i12.112911]
Corresponding Author of This Article
Lin Yang, MD, Department of Radiology, Interventional Medical Center, Science and Technology Innovation Center, Affiliated Hospital of North Sichuan Medical College, No. 63 Wenhua Road, Nanchong 637000, Sichuan Province, China. linyangmd@163.com
Research Domain of This Article
Imaging Science & Photographic Technology
Article-Type of This Article
Retrospective Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (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: http://creativecommons.org/licenses/by-nc/4.0/
Dec 28, 2025 (publication date) through Dec 29, 2025
Times Cited of This Article
Times Cited (0)
Journal Information of This Article
Publication Name
World Journal of Radiology
ISSN
1949-8470
Publisher of This Article
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Shi Y, Zhang P, Li L, Yang HM, Li ZM, Zheng J, Yang L. Interpretable model based on multisequence magnetic resonance imaging radiomics for predicting the pathological grades of hepatocellular carcinomas. World J Radiol 2025; 17(12): 112911 [DOI: 10.4329/wjr.v17.i12.112911]
Yue Shi, Peng Zhang, Jing Zheng, Lin Yang, Department of Radiology, Interventional Medical Center, Science and Technology Innovation Center, Affiliated Hospital of North Sichuan Medical College, Nanchong 637000, Sichuan Province, China
Li Li, Hui-Min Yang, Zu-Mao Li, Institute of Basic Medicine, Department of Pathology, Affiliated Hospital of North Sichuan Medical College, Nanchong 637000, Sichuan Province, China
Co-corresponding authors: Jing Zheng and Lin Yang.
Author contributions: Shi Y and Zhang P performed the data curation work and prepared the original draft of the manuscript; Zheng J and Yang L conceptualized the study; they contributed equally to this article and are the co-corresponding authors of this manuscript; Li L, Yang HM, and Li ZM performed the pathological grading analysis; and all authors contributed to manuscript revision and provided approval for publishing the final version of the manuscript.
Institutional review board statement: This study was approved by the Medical Ethics Committee of Affiliated Hospital of North Sichuan Medical College, No. 2025ER8-1.
Informed consent statement: Owing to the retrospective data analysis, informed consent from the patients was waived.
Conflict-of-interest statement: All the authors report no relevant conflicts of interest for this article.
Data sharing statement: The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.
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: Lin Yang, MD, Department of Radiology, Interventional Medical Center, Science and Technology Innovation Center, Affiliated Hospital of North Sichuan Medical College, No. 63 Wenhua Road, Nanchong 637000, Sichuan Province, China. linyangmd@163.com
Received: August 11, 2025 Revised: September 15, 2025 Accepted: December 3, 2025 Published online: December 28, 2025 Processing time: 138 Days and 22.4 Hours
Abstract
BACKGROUND
Despite the promising prospects of using artificial intelligence and machine learning (ML) for disease classification and prediction purposes, the complexity and lack of explainability of this method make it difficult to apply the constructed models in clinical practice. We developed and validated an interpretable ML model based on magnetic resonance imaging (MRI) radiomics and clinical features for the preoperative prediction of the pathological grades of hepatocellular carcinomas (HCCs). This model will help clinicians better understand the situation and develop personalized treatment plans.
AIM
To develop and validate an interpretable ML model for preoperative pathological grade prediction in HCC patients via a combination of multisequence MRI radiomics and clinical features.
METHODS
MRI and clinical data derived from 125 patients with HCCs confirmed by postoperative pathological examinations were retrospectively analyzed. The patients were randomly split into training and validation groups (7:3 ratio). Univariate and multivariate logistic regression analyses were performed to identify independent clinical predictors. The tumor lesions observed on axial fat-suppressed T2-weighted imaging (FS-T2WI), arterial phase (AP), and portal venous phase (PVP) images were delineated in a slice-by-slice manner using 3D-slicer to generate volumetric regions of interest, and radiomic features were extracted. Interclass correlation coefficients were calculated, and least absolute selection and shrinkage operator regression were conducted for feature selection purposes. Six predictive models were subsequently developed for pathological grade prediction: FS-T2WI, AP, PVP, integrated radiomics, clinical, and combined radiomics-clinical (RC) models. The effectiveness of these models was assessed by calculating their area under the receiver operating characteristic curve (AUC) values. The clinical applicability of the models was evaluated via decision curve analysis. Finally, the contributions of the different features contained in the model with optimal performance were interpreted via a SHapley Additive exPlanations analysis.
RESULTS
Among the 125 patients, 87 were assigned to the training group, and 38 were assigned to the validation group. The maximum tumor diameter, hepatitis B virus status, and monocyte count were identified as independent predictors of pathological grade. Twelve optimal radiomic features were ultimately selected. The AUC values obtained for the FS-T2WI model, AP model, PVP model, radiomics model, clinical model, and combined RC model in the training group were 0.761 [95% confidence interval (CI): 0.562-0.857], 0.870 (95%CI: 0.714-0.918), 0.868 (95%CI: 0.714-0.959), 0.917(95%CI: 0.857-0.959), 0.869 (95%CI: 0.643-0.973), and 0.941 (95%CI: 0.857-0.945), respectively; in the validation group, the AUC values were 0.724 (95%CI: 0.625-0.833), 0.802 (95%CI: 0.686-1.000), 0.797 (95%CI: 0.688-1.000), 0.901(95%CI: 0.833-0.906), 0.865 (95%CI: 0.594-1.000), and 0.932 (95%CI: 0.812-1.000), respectively. The combined RC model demonstrated the best performance. Additionally, the decision curve analysis revealed that the combined RC model had satisfactory prediction efficiency, and the SHapley Additive exPlanations value analysis revealed that the “FS-T2WI-wavelet-HLL_gldm_Large Dependence High Gray Level Emphasis” feature contributed the most to the model, exhibiting a positive effect.
CONCLUSION
An interpretable ML model based on MRI radiomics provides a noninvasive tool for predicting the pathological grade of HCCs, which will help clinicians develop personalized treatment plans.
Core Tip: Despite the promising prospects of using artificial intelligence and machine learning for disease classification and prediction purposes, the complexity and lack of explainability of these methods make it difficult to apply the constructed models in clinical practice. This study aimed to develop and validate an interpretable machine learning model for conducting preoperative pathological grade prediction in hepatocellular carcinoma patients via a combination of multisequence magnetic resonance imaging radiomics and clinical features, which will help clinicians better understand the situation and develop personalized treatment plans.
Citation: Shi Y, Zhang P, Li L, Yang HM, Li ZM, Zheng J, Yang L. Interpretable model based on multisequence magnetic resonance imaging radiomics for predicting the pathological grades of hepatocellular carcinomas. World J Radiol 2025; 17(12): 112911
Hepatocellular carcinoma (HCC) is the 6th most common cancer and the 3rd leading cause of cancer-related death worldwide, and its incidence is increasing[1,2]. Tumor differentiation is an important factor affecting the prognosis of patients with HCC[3,4]. Accurate preoperative assessment of HCC pathological staging aids in decision-making for individualized treatment and improves patient outcomes[5]. Liver biopsy is an invasive examination that patients are hesitant to undergo, and the biopsy results may differ from those obtained from the tumor grading of surgical resection samples[6]. Radiomics, which has emerged in recent years, involves the detection of important features that cannot be identified by the human naked eye from existing imaging findings and achieves the classification and prediction of diseases through feature screening and model building[7-14]. Despite the promising prospects of artificial intelligence and machine learning (ML), the “black box” effect makes it difficult to explain the output of these models and makes them difficult to apply in clinical practice[15]. To improve the interpretability of the “black box” model, some researchers have recently proposed the integration of the SHapley Additive exPlanations (SHAP) method. SHAP is a game theory-based explanation method that reflects the contribution of each feature to the forecast result.
However, to date, there have been no reports on the prediction of HCC pathological grade using an interpretable ML model based on magnetic resonance imaging (MRI) radiomic features combined with clinical factors. The main objective of this study was to develop and validate an interpretable ML model based on multisequence MRI radiomics and clinical features for the preoperative prediction of the pathological grades of HCCs, and on the basis of the constructed traditional nomogram, a dynamic nomogram was developed through a public platform. This will help clinicians better understand the model and develop personalized treatment plans.
MATERIALS AND METHODS
Research subjects
This study was approved by the Ethics Committee of the Affiliated Hospital of North Sichuan Medical College, approval No. 2025ER8-1. Owing to the retrospective data analysis, informed consent from the patients was waived. This study included HCC patients who underwent hepatectomy between November 2016 and September 2023 at the Affiliated Hospital of North Sichuan Medical College and were screened according to the following inclusion and exclusion criteria. The inclusion criteria were as follows: (1) Patients who had received hepatectomy and were pathologically confirmed to have HCC; (2) Patients who underwent enhanced MRI examination within 1 month before surgery; and (3) Patients who did not receive other antitumor treatments. The exclusion criteria were as follows: (1) Incomplete clinical data; (2) Complicated with other malignant tumors; (3) Poor image quality; and (4) No pathological classification. A total of 230 patients were included, and 125 eligible HCC patients with HCC were included in the study (Figure 1).
Figure 1 Flowchart of patient enrollment.
MRI: Magnetic resonance imaging; HCC: Hepatocellular carcinoma.
Clinical data collection
In accordance with previously published studies[16,17], the following common clinical data were collected from each patient as feature variables for constructing the model: Sex, age, maximum tumor diameter, portal vein tumor thrombus (yes/no), lymph node metastasis (yes/no), hepatitis B virus (HBV) (yes/no), liver cirrhosis (yes/no), alpha-fetoprotein (AFP) level, gamma-glutamyl transferase level, alanine transaminase levels, aspartate aminotransferase levels, platelet count, neutrophil count, lymphocyte count, monocyte (MONO) count, neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, lymphocyte-to-MONO ratio, albumin/globulin ratio, and albumin-bilirubin grade.
According to the World Health Organization pathological grading criteria for tumors, HCCs can be classified into three types: Poorly differentiated, moderately differentiated, and well-differentiated HCCs. Similarly, the patients in this study were divided into poorly differentiated and nonpoorly differentiated (well-differentiated and moderately differentiated) groups[18,19].
Magnetic resonance image acquisition
All patients underwent preoperative MRI scans using a Discovery 750 3.0T superconducting MRI scanner (GE, United States). The participants did not eat or drink within 4 hours before the examination. Patients underwent breathing training before the scan. The scan sequences included axial fat-suppressed (FS) T2-weighted imaging (T2WI) and axial dynamic enhanced 3D-LAVA sequences (Table 1). A high-pressure syringe was used to inject contrast medium through the dorsal vein of the hand at a speed of 2-2.5 mL/second. The injection dose ranged from 15–20 mL.
Table 1 Magnetic resonance imaging scan sequences and parameters.
Image segmentation and radiomic feature extraction: Original FS-T2WI, arterial phase (AP), and portal venous phase (PVP), magnetic resonance (MR) images of patients in digital imaging and communications in medical format were acquired from the picture archiving and communication system. Two radiologists with 5 years of work experience used a 3D slicer to manually sketch the axial FS-T2WI, AP, and PVP images in a layer-by-layer manner to obtain the volume of interest. For patients with multiple lesions, the largest lesion was selected for sketching. When sketching, the surrounding blood vessels were avoided (Figure 2). Finally, the radiomic features were extracted from the FS-T2WI, AP, and PVP images using 3D-Slicer.
Figure 2 Target volume sketch and 3D volumetric regions of interest view.
A: The volumetric regions of interest were obtained by manually sketching the region of interest layer by layer on the fat-suppressed T2-weighted imaging image; B: The 3D view was automatically generated from the volumetric regions of interest.
Assessment of consistency: The reproducibility of the features was evaluated using the interclass correlation coefficient (Figure 3). The features with an interclass correlation coefficient < 0.75 were excluded, and the radiomic features with high repeatability were retained for subsequent feature screening.
Feature selection: The Z score was normalized by using the zero-mean normalization method on the extracted features to remove the unit limitation of each feature dataset. Univariate analysis was performed to screen for characteristics for which P < 0.05. Afterward, the dimensionality reduction ability of least absolute selection and shrinkage operator (LASSO) regression was used to screen the radiomic features with nonzero coefficients, and the optimal features were screened out via 10-fold cross-validation (Figure 4) and subsequently used to calculate the radiomics score (Rad score).
Figure 4 Radiomics features screened via least absolute selection and shrinkage operator regression.
The vertical axis shows the model misclassification rate. The horizontal axis shows log(λ). The two vertical lines were drawn at the selected values using cross-validation, and the optimal values were obtained by applying the minimum criteria and 1 of the minimum criteria (1-standard error criteria). A: Coefficient change spectra of the radiomics features; B: The process of feature selection for determining the parameter (λ) in the least absolute selection and shrinkage operator model by 10-fold cross-validation.
Model construction and evaluation: The clinical factors associated with pathological classification were screened using univariate and multivariate analyses. Logistic regression was used to construct models of each single MRI sequence (T2WI, AP, and PVP), a combined radiomics model, a clinical model, and a combined radiomics-clinical model incorporating the optimal radiomics and clinical features. The efficacy of the model was evaluated by the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. The clinical validity of the model was evaluated via decision curve analysis. Finally, the model with the optimal performance was selected to construct a nomogram.
Model interpretation: The SHAP value of each feature in the optimal model was calculated and ranked, and the degree of contribution of each feature to the model prediction result was visually displayed.
Statistical analysis
R (4.4.2) and Python (3.13.1) software programs were used for data processing and statistical analysis. For continuous variables, the Shapiro-Wilk test was used to assess normality. Continuous variables with a normal distribution are described as the mean ± SD, and intergroup comparisons were performed using an independent samples t test; for continuous variables with a nonnormal distribution, the median (interquartile range) was used. The intergroup differences were analyzed via the Mann-Whitney test. Categorical variables are expressed as n (%), and intergroup comparisons were performed using the χ2 test or Fisher’s exact test. Data with P < 0.05 (two-sided) were considered statistically significant.
The “irr” package of the R software was used to calculate the intergroup correlation coefficient; the “glmnet” package was used for LASSO regression analysis. LASSO regression was performed to identify important variables; LASSO was used to select variables by shrinking the coefficients of the least-important variables in the logistic regression model to zero. The “pROC” and “e1071” packages were used to construct the prediction models and calculate the AUC for each model; the “Hmisc”, “lattice”, “survival”, “formula”, “rms” and “ggplot2” packages were used to construct the traditional nomogram. The “DynNom” and “rsconnect” packages in R were used to construct the dynamic nomogram. The “shap” package of Python software was used to analyze the SHAP values of the model features.
RESULTS
Clinical baseline data
A total of 125 HCC patients were enrolled in the study and randomly divided into a training group (87 patients) and a validation group (38 patients) at a 7:3 ratio. The clinical baseline data are shown in Table 2.
Table 2 Comparison of the clinical characteristics between the training and validation groups, n (%).
A total of 1130 radiomic features were extracted from the FS-T2WI, AP, and PVP images. Finally, the 3, 4, and 5 optimal features were screened out (Table 3).
Univariate analysis revealed that the maximum tumor diameter, HBV level, liver cirrhosis status, platelet count and MONO level were significant risk factors. Multivariate logistic regression analyses revealed that the maximum tumor diameter, HBV, and MONO were independent predictors of the preoperative pathological classification (Table 4).
Table 4 Results of univariate and multivariate logistic regression analyses.
The AUC values of the FS-T2WI model, AP model, PVP model, combined radiomics model, clinical model, and combined radiomics-clinical model in the training group were 0.761, 0.870, 0.868, 0.917, 0.869, and 0.941, respectively; in the validation group, they were 0.724, 0.802, 0.797, 0.901, 0.865, and 0.932, respectively. Among the six models, the combined radiomic-clinical model had the optimal performance in both the training group and the validation group (Table 5, Figure 5). Decision curve analysis revealed that the combined radiomic–clinical model had greater net clinical benefit than the clinical model and the combined radiomics model did (Figure 6).
On the basis of the comprehensive performance evaluation of the models, the combined radiomics-clinical model with the optimal model performance in the training group was selected to construct the traditional nomogram (Figure 7). Traditional nomograms can visually display the predicted value of the probability of poorly differentiated HCC, but manual calculation is needed, which increases the prediction error of the results and reduces the practicability and operability. Therefore, on the basis of the traditional nomogram, we developed a dynamic nomogram (Figure 8), and the user only needs to select the relevant predictive factor values on the web interface and click the “Predict” button to calculate the accurate predicted value. The dynamic nomogram can be accessed through the uniform resource locator https://mylogistics.shinyapps.io/dynnomapp/.
Figure 8 Dynamic nomogram based on the combined radiomics-clinical model via the web interface.
A dynamic nomogram was developed for the training group by combining the following four parameters: Rad score, tumor size, hepatitis B virus status, and monocyte count; and an example of predicting the probability of poor differentiation in hepatocellular carcinoma patients is shown. HBV: Hepatitis B virus; MONO: Monocyte count.
SHAP interpretability analysis
The combined radiomic-clinical model with the optimal performance was selected and used to perform feature importance analysis via SHAP analysis (Figure 9). According to the SHAP analysis, the radiomics feature “FS-T2WI_wavelet-HLL_gldm_Large Dependence High Gray Level Emphasis” contributed the most to the prediction of the HCC pathological classification results, indicating a positive contribution.
Figure 9 SHapley Additive exPlanations analysis diagram of each feature in the optimal model.
A: SHapley Additive exPlanations (SHAP) value bar graph, showing the mean absolute SHAP value of all the features, arranged in descending order of importance; B: SHAP value scatter graph, showing the average degree of influence of each feature in the model. The color of the dot represents the original value of the feature, with blue indicating a low value and red indicating a high value, and the position of the dot represents the SHAP value. A positive SHAP value indicates an increased risk of predicting poorly differentiated hepatocellular carcinoma, and vice versa. A higher value corresponds to a higher risk. Each point corresponds to a prediction for each patient; C: Waterfall diagram of a random sample. The calculation process of the degree of contribution of each feature in a single random sample to the prediction result and the final predicted value, which is represented by bars of different colors, with blue representing a negative value and red representing a positive value. SHAP: SHapley Additive exPlanations; FS-T2WI: Fat-suppressed T2-weighted imaging; PVP: Portal venous phase; AP: Arterial phase; LLL: Low-low-low frequency band; HBV: Hepatitis B virus; MONO: Monocyte count.
DISCUSSION
In recent years, several studies have discussed the value of MRI radiomics in predicting the pathological grade of HCC. Wu et al[17] reported that a radiomics model based on noncontrast-enhanced MR images could distinguish poorly differentiated HCC from nonpoorly differentiated HCC; the AUC values of the T1-weighted imaging + T2WI combined model in the training group and the validation group were 0.849 and 0.742, respectively; and after the independent risk factor (AFP) was included in the model, the AUC values of the training group and the validation group increased to 0.909 and 0.800, respectively. Moreover, Hu et al[19] and Yan et al[20] reported the potential of pathological classification prediction models for HCC based on the radiomic features of enhanced MR images. Among all the models, the combined model integrating radiomic features and clinical risk factors exhibited optimal performance; however, these studies did not include T2W sequence data.
Recently, the SHAP analysis method introduced by some researchers has increased the interpretability of the output of radiomics ML models[21-23]. Ye et al[21] developed 5 computed tomography radiomics-based ML models to predict the pathological classification of the neuroendocrine pancreas, and the best model was subjected to SHAP analysis. The results showed that the radiomics-based interpretable ML model was highly valuable for differentiating the degree of differentiation of pancreatic neuroendocrine tumors. The automatic segmentation of renal tumor, kidney and perirenal adipose tissue from computed tomography images and the extraction of radiomic features were performed to predict the pathological grade of clear cell renal cell carcinoma[22]. The results revealed that when the features of renal tumor, kidney and perirenal adipose tissue were incorporated, the combined model had the highest AUC value, and the AUC value of the internal validation group was 0.840. SHAP value analysis revealed that the first-order features of the tumor contributed the most to the pathological classification of clear cell renal cell carcinoma.
Different MRI sequences provided different information. FS-T2WI reflected tissue characteristics, and contrast-enhanced sequences (AP and PVP) reflected tissue blood supply status. Their contributions to the combined model were also different. Multisequence combined models have been shown to exhibit better prediction efficiency. In this study, the radiomic features of MR-T2WI and enhanced MR images in the AP and PVP, respectively, were acquired. The multisequence radiomics joint model achieved good predictive performance. The AUC values in the training group and the validation group were 0.917 and 0.901, respectively. The predictive performance of the combined radiomic-clinical model that included clinical factors was further improved, with AUC values in the training group and the validation group of 0.941 (95%CI: 0.857-0.945) and 0.932 (95%CI: 0.812-1.000), respectively. SHAP analysis revealed that the radiomic features with the top ranking in terms of contribution were texture features, and “FS-T2WI_wavelet-HLL_gldm_Large Dependence High Gray Level Emphasis” made the greatest positive contribution. These findings suggest that radiological texture features may play a key role in the prediction of HCC pathological grade, which is consistent with the results of Feng et al[24].
In addition, many studies have investigated the relationships between several clinical factors and the pathological classification of HCC. There is a consensus that the progression and outcome of HCC are associated with inflammation[25], and levels of the inflammatory markers neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, and lymphocyte-to-MONO ratio are significantly correlated with the pathological classification and prognosis of HCC[26-35]. Huang et al[36] reported that the AFP level, HBsAg titer, and neutrophil count were independent predictors of preoperative pathological classification of HCC; other studies also reported similar results[35-37]. Liu et al[18] and Yan et al[20] reported that tumor volume is an independent predictor of HCC pathological classification. These data revealed that the maximum tumor diameter, HBV titer and MONO count were independent risk factors associated with the pathological classification of HCC, which was consistent with the results of the aforementioned studies. In this set of data, no correlation between the AFP level and the pathological classification of HCC was observed; this may be related to the differences in the study populations and the small sample size, which remains to be studied in the future.
This study has the following limitations: (1) Because of the retrospective data employed for developing this ML model, this study may have some selection bias. Prospective studies should be conducted in the future to further validate the model; (2) This study used single-center data, and the sample size was small. In the future, multicenter data should be collected, and the sample size should be increased for further validation; and (3) Only the widely used logistic regression classifier was used for modeling. In the future, support vector machines, the K nearest neighbor algorithm, random forests[38] and other classifiers should be used for further study and comparison.
CONCLUSION
A radiomics model based on multisequence MRI features combined with clinical factors can better predict the pathological classification of HCC preoperatively. Combined with SHAP analysis, this model can explain the contribution of each feature contained within it, which will help clinicians better understand the model and develop personalized treatment plans.
Footnotes
Provenance and peer review: Invited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Radiology, nuclear medicine and medical imaging
Country of origin: China
Peer-review report’s classification
Scientific Quality: Grade A, Grade A, Grade A, Grade B
Novelty: Grade A, Grade A, Grade A, Grade B
Creativity or Innovation: Grade A, Grade A, Grade B, Grade B
Scientific Significance: Grade A, Grade A, Grade A, Grade B
P-Reviewer: Inam S, PhD, Assistant Professor, Researcher, Pakistan; Wen JW, PhD, China S-Editor: Bai Y L-Editor: A P-Editor: Lei YY
Xia TY, Zhou ZH, Meng XP, Zha JH, Yu Q, Wang WL, Song Y, Wang YC, Tang TY, Xu J, Zhang T, Long XY, Liang Y, Xiao WB, Ju SH. Predicting Microvascular Invasion in Hepatocellular Carcinoma Using CT-based Radiomics Model.Radiology. 2023;307:e222729.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 139][Reference Citation Analysis (1)]
Xu B, Dong SY, Bai XL, Song TQ, Zhang BH, Zhou LD, Chen YJ, Zeng ZM, Wang K, Zhao HT, Lu N, Zhang W, Li XB, Zheng SS, Long G, Yang YC, Huang HS, Huang LQ, Wang YC, Liang F, Zhu XD, Huang C, Shen YH, Zhou J, Zeng MS, Fan J, Rao SX, Sun HC. Tumor Radiomic Features on Pretreatment MRI to Predict Response to Lenvatinib plus an Anti-PD-1 Antibody in Advanced Hepatocellular Carcinoma: A Multicenter Study.Liver Cancer. 2023;12:262-276.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 4][Cited by in RCA: 36][Article Influence: 18.0][Reference Citation Analysis (0)]
Hu X, Li C, Wang Q, Wu X, Chen Z, Xia F, Cai P, Zhang L, Fan Y, Ma K. Development and External Validation of a Radiomics Model Derived from Preoperative Gadoxetic Acid-Enhanced MRI for Predicting Histopathologic Grade of Hepatocellular Carcinoma.Diagnostics (Basel). 2023;13:413.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 7][Reference Citation Analysis (0)]
Li S, Zhou Z, Gao M, Liao Z, He K, Qu W, Li J, Kamel IR, Chu Q, Zhang Q, Li Z. Incremental value of automatically segmented perirenal adipose tissue for pathological grading of clear cell renal cell carcinoma: a multicenter cohort study.Int J Surg. 2024;110:4221-4230.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 8][Reference Citation Analysis (0)]
Wenpei G, Yuan L, Liangbo L, Jingjun M, Bo W, Zhiqiang N, Yijie N, Lixin L. Predictive value of preoperative inflammatory indexes for postoperative early recurrence of hepatitis B-related hepatocellular carcinoma.Front Oncol. 2023;13:1142168.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 27][Reference Citation Analysis (0)]
Liao C, Li G, Bai Y, Zhou S, Huang L, Yan M, Qiu F, Chen J, Wang Y, Tian Y, Chen S. Prognostic value and association of sarcopenic obesity and systemic inflammatory indexes in patients with hepatocellular carcinoma following hepatectomy and the establishment of novel predictive nomograms.J Gastrointest Oncol. 2021;12:669-693.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 4][Cited by in RCA: 14][Article Influence: 3.5][Reference Citation Analysis (0)]
Gavriilidis P, Pawlik TM. Inflammatory indicators such as systemic immune inflammation index (SIII), systemic inflammatory response index (SIRI), neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) as prognostic factors of curative hepatic resections for hepatocellular carcinoma.Hepatobiliary Surg Nutr. 2024;13:509-511.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in RCA: 30][Reference Citation Analysis (0)]
