The present study aimed to develop and validate a preoperative model based on Gd-DTPA-enhanced magnetic resonance imaging (MRI) and clinical factors for predicting perineural invasion (PNI) in patients with intrahepatic cholangiocarcinoma (ICC), enabling clinicians to perform more accurate patient evaluation and and make individualized therapeutic decisions. Between July 2019 and February 2024, a total of 173 patients with pathologically confirmed ICC who underwent preoperative Gd‑DTPA‑enhanced MRI were retrospectively enrolled. These patients were randomly assigned to training and test cohorts at a 7:3 ratio. Multivariate logistic regression was used to identify independent predictors of PNI status and a predictive model was developed and presented as a nomogram. The model performance was assessed in terms of discrimination, calibration, and clinical utility. Peritumoral arterial hyper-enhancement (OR 14.99, 95

This study aims to develop a predictive model that combines MRI radiomics features with clinical factors to anticipate the preoperative presence of Vessels Encapsulating Tumor Clusters (VETC) and early recurrence in hepatocellular carcinoma (HCC) with tumors ≤ 3 cm. A total of 336 patients were included in the study, with 271 from Hospital 1, divided into a training group (n = 192) and a validation group (n = 79). An additional 65 patients from Hospital 2 were used as an external validation group. Key features were selected using Least Absolute Shrinkage and Selection Operator (LASSO) and logistic regression analyses. Clinical, radiomics, and combined clinical radiomics models were then constructed using machine learning algorithms. Model performance was evaluated using the Area Under the Roc Curve (AUC), calibration curves, and decision curves. Kaplan-Meier survival analysis assessed early recurrence in VETC positive and VETC negative HCC patients with tumors ≤ 3 cm. Both the clinical model utilizing logistic regression with GGT, AFP, non-smooth margin, and arterial peritumoral enhancement as independent risk factors, and the radiomics model incorporating 15 radiomics features with logistic regression, exhibited superior AUC values across the three groups (training, test, and external validation) when compared to other algorithms. Notably, the clinical radiomics model exhibited higher AUC values in the training group (0.885), validation group (0.867), and external validation group (0.849) compared to the standalone clinical and radiomics models. There was a significant difference (p < 0.05) in predicting early recurrence between VETC positive and VETC negative HCC patients (≤ 3 cm) when using the clinical radiomics models. The clinical radiomics model demonstrates robust performance in predicting VETC positivity in HCC patients (≤ 3 cm), thereby holding the potential to forecast early recurrence in these patients.

By generating an immune score reflecting the tumor immune microenvironment via Co-detection by Indexing (CODEX) Immunomics and integrating clinicoradiological features, we developed an interpretable machine learning model to predict postoperative survival in hepatocellular carcinoma (HCC) using SHapley Additive exPlanations (SHAP). We retrospectively enrolled 94 HCC patients who underwent the CODEX procedure and had preoperative magnetic resonance imaging. Patients were divided into a training set (n = 65) and a validation set (n = 29) in a 7:3 ratio. Univariate and multivariate Cox regression analyses identified clinicoradiological independent risk factors for 5-year survival to construct the Clinical model. For immunomics analysis, 36 immune-related molecules were evaluated using CODEX. Key features were selected through univariate Cox regression and Recursive Feature Elimination (RFE). The best-performing classifier among five machine learning algorithms was used to build the Immune model. The immune score from the Immune model and variables from the Clinical model were combined using multivariate Cox regression to identify independent risk factors, forming the Clinical-Immune model. Models were compared for discrimination, calibration, and clinical utility. SHAP was used to interpret the model’s predictions. Shape, arterial peritumoral enhancement, intratumoral necrosis constituted the Clinical model. Five immunomics features formed the Immune model using a survival decision algorithm. The Clinical-Immune model combined the immune score and arterial peritumoral enhancement. The concordance indexes (C-indexes) for the three models were 0.730, 0.832, and 0.852 in the training set, and 0.624, 0.815, and 0.870 in the validation set. Time-dependent area under the curve (timeAUC) values were 0.833, 0.907, and 0.969 in the training set, and 0.656, 0.919, and 1.000 in the validation set. The Clinical-Immune model, which demonstrated the best performance and offered superior predictive consistency and clinical utility, was selected as the final prediction model. We developed an interpretable machine learning model to predict postoperative survival in HCC patients using CODEX immunomics and clinicoradiological features. This robust model enhances survival prediction and supports clinical decision-making in HCC management.

Purpose:To develop and internally validate a multiphase contrast-enhanced computed tomography (CT) based radiomics-clinical nomogram for pre-operative prediction of inadequate future liver remnant (FLR) hypertrophy after portal vein embolization (PVE), with the aim of guiding individualized surgical planning. Patients and Methods:We retrospectively enrolled patients who underwent PVE at our centre. Contrast-enhanced CT and clinical data were collected before and 2-5 weeks after PVE for calculated FLR (cFLR≥30% defined as adequate hypertrophy). After liver segmentation on triphasic CT, radiomic features were extracted and reduced by LASSO to build a Rad-signature. This signature was integrated with significant clinical variables into an MLP fusion model, visualised as a nomogram, and evaluated using ROC, calibration and decision-curve analyses. Results:A total of 98 patients were included in this study. The fusion model achieved an area under the receiver-operating-characteristic curve (AUC) of 0.913 (95% CI 0.846-0.981) in the training cohort and 0.833 (95% CI 0.605-1.000) in the internal test cohort, outperforming the clinical-only model (AUC 0.714; DeLong P = 0.017). Sensitivity and specificity in the test set were 0.824 and 1.000, respectively. Pre-PVE FLR, cFLR, serum pre-albumin, prior chemotherapy, hepatitis status and underlying cirrhosis were independent predictors of adequate hypertrophy. Decision-curve analysis demonstrated net clinical benefit within a threshold probability range of 20-80%. Conclusion:The proposed multiphase CECT radiomics-clinical nomogram showed potential for identifying patients at high risk of insufficient FLR hypertrophy after PVE. This internally validated model holds promise for supporting personalised planning in hepatobiliary surgery.

Background and Objective:Lung cancer ranks as the most frequently diagnosed malignancy worldwide and the leading cause of cancer deaths. Pulmonary nodule growth serves as critical information for determining the probability of malignancy and guiding clinical decisions regarding surgical intervention and follow-up strategies. This review synthesizes current evidence on the multidimensional evaluation criteria for pulmonary nodule growth and the advancements in computed tomography (CT)-derived imaging prediction models, aiming to inform and optimize the clinical management of pulmonary nodules. Methods:A comprehensive literature search was performed across the Web of Science, PubMed, Cochrane Library, and EMBASE databases. The search strategy was designed to identify articles addressing pulmonary nodule growth and CT imaging features. The search was limited to original research articles and meta-analyses published in English between January 1, 2020 and February 3, 2026. Key Content and Findings:Current guidelines still exhibit subtle differences in defining pulmonary nodule growth, with the most commonly adopted criteria in contemporary research including an increase in mean diameter of 1.5 or 2 mm, a 25% volumetric increase, or a volume doubling time (VDT) of less than 400 days. In this context, CT imaging provides critical information associated with pulmonary nodule growth, such as nodule size, density, tumor characteristics, and peritumoral signs. Meanwhile, with the advancement of artificial intelligence (AI), radiomics approaches have further improved the accuracy of pulmonary nodule growth prediction, while deep learning has enabled the visualization of future nodule imaging. Additionally, natural language processing (NLP) models, as a rapidly developing technology, have exhibited considerable promise in predicting nodule growth and enhancing the efficiency of follow-up management protocols. Conclusions:This article provides a systematic overview of the evolution of assessment methods for pulmonary nodule growth and the latest breakthroughs in predictive modeling, and offers perspectives on future directions in this domain. With the continuous empowerment of AI in pulmonary nodule growth management, the corresponding clinical pathway stands to benefit from increasingly systematic optimization and evolution.

Objectives To establish and externally validate a preoperative MRI-based risk scoring system for MVI prediction in IMCC≤5 cm, and to evaluate its prognostic value. Methods This multicenter retrospective study enrolled 171 patients with pathologically confirmed IMCC (≤5 cm) from four institutions who underwent curative surgical resection. Patients were divided into training (n = 122) and external validation (n = 49) cohorts. Univariable and multivariable logistic regression analyses were performed to identify independent predictors of MVI. A scoring system was established by incorporating the independent predictors based on logistic regression coefficients. According to this scoring system, patients were stratified into low- and high-risk MVI groups and disease-free survival (DFS) was analyzed by Kaplan-Meier survival analysis. Results In the multivariable analysis, non-smooth tumor margin (OR = 4.140; p = 0.005), arterial phase (AP) peritumoral enhancement (OR = 6.589; p < 0.001), and arterial edge enhancement ratio (AEER; OR = 0.916; p < 0.001) were included as independent predictors of MVI. The scoring system demonstrated high predictive accuracy, with AUCs of 0.837 in the training cohort and 0.813 in the external validation cohort. IMCC patients at high risk exhibited significantly shorter DFS compared to those at low risk for MVI in both cohorts (p < 0.05). Conclusion The preoperative MRI-based scoring system incorporating tumor margin, AP peritumoral enhancement, and AEER can effectively predict MVI in patients with IMCC ≤5 cm, providing a valuable predictive tool for risk stratification and prognosis assessment.

Objective:To develop and validate a non-invasive model for predicting early recurrence (ER) after microwave ablation (MWA) in patients with hepatocellular carcinoma (HCC) and comorbid type 2 diabetes (T2D). Methods:This retrospective study enrolled 186 HCC patients with T2D who underwent MWA and were divided into training (n=149) and validation (n=37) cohorts. Preoperative clinical parameters and portal-venous phase MRI radiomic features were extracted. A radiomics score (RAD -score) was constructed using the K-nearest neighbors (KNN) algorithm after feature selection. A clinical model was built using logistic regression, and a nomogram integrating clinical factors and the RAD -score was established. Model performance was evaluated using ROC curves, calibration curves, and decision curve analysis (DCA). Kaplan--Meier curves, pathological analysis, and immunohistochemistry validated prognostic stratification. Results:Four core radiomic features were identified. The combined nomogram achieved AUCs of 0.877 (training) and 0.846 (validation), outperforming the clinical (AUC 0.718) and radiomics (AUC 0.774) models alone. In the validation cohort, the sensitivity, specificity, and negative predictive value were 80.0%, 83.8%, and 90.1%, respectively. DCA showed a significant net benefit across threshold probabilities of 0.1-0.9. High-risk patients had a 5.8-fold (training) and 2.51-fold (validation) higher recurrence risk with significantly shorter progression-free survival. Pathological and immunohistochemical analyses confirmed more aggressive tumor features (higher rates of microvascular invasion, poor differentiation, thick-trabecular type, CD34 expression, and glypican-3 expression) in the high-risk group. Conclusion:The nomogram provides a satisfactory prediction of early recurrence within two years after MWA in HCC patients with T2D, enabling individualized postoperative surveillance and intervention.

Background:Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death globally, with a 5-year recurrence rate of 70% even after curative resection. Liver regeneration following hepatectomy shares molecular mechanisms with tumor recurrence, yet the relationship between early liver regeneration and HCC recurrence remains unclear. This study aimed to investigate the association between the daily increment of residual liver regeneration (DIRER) and post-hepatectomy outcomes. Methods:A single-center, retrospective cohort study was conducted at the Eastern Hepatobiliary Surgery Hospital, including 324 patients who underwent R0 hepatectomy between 2017 and 2019, with complete preoperative and postoperative three-dimensional (3D) visualization data. Of these, 128 patients (39.5%) had underlying liver cirrhosis. Liver resection volume was assessed intraoperatively using the drainage method. The residual liver volume (RLV) to body weight ratio (RLVw) was calculated as the ratio of RLV to preoperative body weight. DIRER was defined as the daily increase in RLV during regeneration. Prognostic risk factors were identified using Cox regression, and optimal DIRER cut-off points for recurrence-free survival (RFS) and overall survival (OS) were determined via X-tile software. Kaplan-Meier curves assessed clinical significance, with interaction and stratified analyses based on tumor size and surgical resection range. Logistic regression was employed to identify predictors of DIRER. Results:DIRER was found to be an independent risk factor for both RFS [hazard ratio (HR) =1.06, P=0.003] and OS (HR =1.06, P=0.004). DIRER was also associated with early recurrence (HR =1.05, P=0.02). Higher DIRER was linked to a significantly higher risk of multiple recurrences [relative risk (RR) =1.83, P=0.02]. For tumors ≤5 cm, a larger surgical resection range was associated with higher recurrence rates (HR =1.95, P=0.02). Additionally, a lower RLVw was correlated with higher DIRER and worse prognosis. Conclusions:Excessive hepatectomy may accelerate liver regeneration, leading to earlier recurrence and poorer prognosis. This underscores the importance of careful preoperative planning to balance the extent of radical resection with adequate RLV.

The purpose of this study is to mainly develop a predictive model based on clinicoradiological and radiomics features from preoperative gadobenate-enhanced (Gd-BOPTA) magnetic resonance imaging (MRI) using multilayer perceptron (MLP) deep learning to predict vessels encapsulating tumor clusters (VETC) in hepatocellular carcinoma (HCC) patients. A total of 230 patients with histopathologically confirmed HCC who underwent preoperative Gd-BOPTA MRI before hepatectomy were retrospectively enrolled from three hospitals (144, 54, and 32 in training, test, and validation set, respectively). Univariate and multivariate logistic regression analyses were used to determine independent clinicoradiological predictors significantly associated with VETC, which then constituted the clinicoradiological model. Regions of interest (ROIs) included four modes, intratumoral (Tumor), peritumoral area ≤ 2 mm (Peri2mm), intratumoral + peritumoral area ≤ 2 mm (Tumor + Peri2mm) and intratumoral integrated with peritumoral ≤ 2 mm as a whole (TumorPeri2mm). A total of 7322 radiomics features were extracted respectively for ROI(Tumor), ROI(Peri2mm), ROI(TumorPeri2mm) and 14644 radiomics features for ROI(Tumor + Peri2mm). Least absolute shrinkage and selection operator (LASSO) and univariate logistic regression analysis were used to select the important features. Seven different machine learning classifiers respectively combined the radiomics signatures selected from four ROIs to constitute different models, and compare the performance between them in three sets and then select the optimal combination to become the radiomics model we need. Then a radiomics score (rad-score) was generated, which combined significant clinicoradiological predictors to constituted the fusion model through multivariate logistic regression analysis. After comparing the performance of the three models using area under receiver operating characteristic curve (AUC), integrated discrimination index (IDI) and net reclassification index (NRI), choose the optimal predictive model for VETC prediction. Arterial peritumoral enhancement and peritumoral hypointensity on hepatobiliary phase (HBP) were independent risk factors for VETC, and constituted the Radiology model, without any clinical variables. Arterial peritumoral enhancement defined as the enhancement outside the tumor boundary in the late stage of arterial phase or early stage of portal phase, extensive contact with the tumor edge, which becomes isointense during the DP. MLP deep learning algorithm integrated radiomics features selected from ROI TumorPeri2mm was the best combination, which constituted the radiomics model (MLP model). A MLP score (MLP_score) was calculated then, which combining the two radiology features composed the fusion model (Radiology MLP model), with AUCs of 0.871, 0.894, 0.918 in the training, test and validation sets. Compared with the two models aforementioned, the Radiology MLP model demonstrated a 33.4

RATIONALE AND OBJECTIVES:This study aimed to develop an interpretable machine learning model using magnetic resonance imaging (MRI) radiomics features to predict preoperative microscopic peritumoral small cancer foci (MSF) and explore its relationship with early recurrence in hepatocellular carcinoma (HCC) patients. METHODS:A total of 1049 patients from three hospitals were divided into a training set (Hospital 1: 614 cases), a test set (Hospital 2: 248 cases), and a validation set (Hospital 3: 187 cases). Independent risk factors from clinical and MRI features were identified using univariate and multivariate logistic regression to build a clinicoradiological model. MRI radiomics features were then selected using methods like least absolute shrinkage and selection operator (LassoCV) and modeled with various machine learning algorithms, choosing the best-performing model as the radiomics model. The clinical and radiomics features were combined to form a fusion model. Model performance was evaluated by comparing receiver operating characteristic (ROC) curves, area under the curve (AUC) values, calibration curves, and decision curve analysis (DCA) curves. Net reclassification improvement (NRI) and integrated discrimination improvement (IDI) values assessed improvements in predictive efficacy. The model's prognostic value was verified using Kaplan-Meier analysis. SHapley Additive exPlanations (SHAP) was used to interpret how the model makes predictions. RESULTS:Three models were developed as follows: Clinical Radiology, XGBoost, and Clinical XGBoost. XGBoost was selected as the final model for predicting MSF, with AUCs of 0.841, 0.835, and 0.817 in the training, test, and validation sets, respectively. These results were comparable to the Clinical XGBoost model (0.856, 0.826, 0.837) and significantly better than the Clinical Radiology model (0.688, 0.561, 0.613). Additionally, the XGBoost model effectively predicted early recurrence in HCC patients. CONCLUSION:This study successfully developed an interpretable XGBoost machine learning model based on MRI radiomics features to predict preoperative MSF and early recurrence in HCC patients.

Primary liver carcinosarcoma (CS) and sarcomatoid carcinoma (SC) are rare malignant tumors of the liver. Although the two tumors often overlap in clinical and imaging manifestations, there are currently no reports comparing the imaging features of these two tumors. Our study aims to compare the clinical characteristics and imaging features of these two tumors to further describe their distinct features, thereby enhancing understanding and diagnostic accuracy. A retrospective analysis was conducted on the clinical and imaging data of 17 patients with CS and 27 patients with SC diagnosed by surgical or needle biopsy between September 2010 and December 2024 at our hospital. The data were summarized and statistically analyzed. Both groups were predominantly male, with a lower mean age (56.65 ± 11.82) in the CS group compared to the SC group (64.93 ± 8.15) (P = 0.01). Compared to the SC group, the CS group more commonly presented with hepatitis B, cirrhosis, and elevated AFP levels. Both groups were more commonly located in the right hepatic lobe, with larger tumors that were often solitary, irregularly shaped, and lobulated. Most tumors exhibited necrosis and hemorrhage. Calcification was observed in two cases in the CS group on CT scans. The tumor margins were predominantly indistinct, and the majority of tumors did not show a capsule. Approximately half of the patients in the SC group had lymph node involvement, which was significantly higher than in the CS group (P = 0.023). After contrast enhancement, all cases in both groups showed heterogeneous enhancement in the arterial phase. Regarding enhancement distribution, the CS group more commonly exhibited enhancement at the margins and in the solid components, while most cases in the SC group showed enhancement at the margins and in the septa. In terms of dynamic enhancement patterns, the CS group more commonly exhibited partial or complete regression in the delayed phase, while the SC group more commonly exhibited progressive or persistent enhancement in the delayed phase, with statistical significance (P = 0.042). Patients in the SC group had significantly higher age and lymph node involvement than those in the CS group. In terms of tumor enhancement patterns, the CS group primarily exhibited delayed-phase regression or partial regression, while the SC group primarily exhibited delayed-phase persistent or progressive enhancement.

Purpose:To explore the application value of clinical indicators, radiological features, and magnetic resonance imaging (MRI) radiomics to predict the grading of MVI in nodular hepatocellular carcinoma (≤3cm). Methods:A total of 131 patients with hepatocellular carcinoma (HCC) and confirmed microvascular invasion (MVI) who underwent surgical resection between January 2016 and December 2022 were retrospectively analyzed. A clinical-radiological (CR) model was constructed using independent risk factors identified by logistic regression. Radiomics models based on MRI (arterial phase, portal venous phase, delayed phase) across various regions (AVDPintra, AVDPintra+peri3mm, AVDPintra+peri5mm, AVDPintra+peri10mm) were developed using the Logistic Regression (LR) classifiers. The optimal radiomics model was subsequently integrated with the CR model to construct a combined clinical-radiological-radiomics (CRR) model. Model performance was assessed using the area under the curve (AUC). Results:Non-smooth margin and intratumoral artery were risk factors for MVI grading. The combined CRR model demonstrated the best predictive performance, with AUCs of 0.907 and 0.917 in the training and testing sets, respectively. Compared with the CR model alone, the CRR model showed a statistically significant improvement (p = 0.008, DeLong test). Conclusion:The AVDPintra+peri3mm model based on MRI radiomics demonstrates good predictive performance in predicting MVI grading in HCC (≤3cm). Combining features from the CR model with those of the AVDPintra+peri3mm model to construct the CRR model further enhances the prediction of MVI grading.

Objective:This study aims to evaluate the prognostic predictive efficacy of Gadobenate dimeglumine (Gd-BOPTA)-enhanced magnetic resonance imaging (MRI) in patients with solitary hepatocellular carcinoma (HCC) without microvascular invasion (MVI) and to investigate the potential clinical and imaging parameters for stratifying the risk of recurrence following hepatectomy. Methods:This retrospective study included 134 patients with histopathologically confirmed solitary HCC without microvascular invasion (MVI) from two hospital districts, which divided into the training cohort and validation cohort. MRI features were independently assessed by two radiologists. Univariate and multivariate Cox regression analyses were conducted to identify independent risk factors associated with recurrence-free survival (RFS). A nomogram was developed based on these factors, and its performance was validated in the validation cohort. RFS was analyzed using Kaplan-Meier curves and the Log rank test. Results:The median RFS for the 134 patients was 45.7 months, with 41.8% of patients experiencing tumor recurrence after hepatectomy. Univariate Cox regression analysis identified hepatitis Be antigen (HBeAg) positivity, tumor size, tumor growth subtype, non-peripheral washout, nodule-in-nodule architecture, mosaic architecture, and intratumoral arteries as significant risk factors for RFS. Multivariate Cox regression analysis revealed that HBeAg positive, tumor growth subtype, non-peripheral washout, mosaic architecture, and internal arteries were independent prognostic factors for RFS in patients with solitary HCC without MVI. The nomogram based on these variables demonstrated good predictive accuracy, with concordance indices (C-index) of 0.740 and 0.701 in the training and validation cohorts, respectively. Additionally, patients in the high-risk group exhibited significantly lower RFS compared to those in the low-risk group. Conclusion:A model incorporating Gd-BOPTA-enhanced MRI and clinical features can effectively predict RFS in solitary HCC patients without MVI and assist in risk stratification for recurrence after hepatectomy.

Purpose:This study aimed to develop a predictive model for the prognosis of patients with hepatocellular carcinoma (HCC) after resection. Methods:Eighty-two HCC patients were randomly divided into a training cohort (n = 62) and a validation cohort (n = 20). Clinicopathological, multiproteomics features based on CO-Detection by Indexing (Codex), and radiomics features extracted from magnetic resonance imaging (MRI) were used to construct four models: clinicopathological model, radiomics model, proteomics model, and combined model. Model performance was evaluated using the C-index, calibration curves, receiver operating characteristic (ROC) curves, survival curves, and decision curve analysis (DCA). Results:The combined model, integrating clinicopathological, radiomics, and multi-proteomic features, demonstrated the best performance of overall survival (OS) prediction in both the training cohort (C-index = 0.821, 95% CI: 0.745-0.897) and validation cohort (C-index = 0.791, 95% CI: 0.628-0.954). The calibration curve showed high accuracy of the combined nomogram in predicting OS. Conclusion:This study innovatively integrates CODEX-based multiproteomics, radiomics, and clinicopathological features to construct a prognostic prediction model for HCC. The combined model demonstrates improved prognostic predictive efficacy compared with single-modality models. This approach establishes a theoretical foundation for personalized diagnosis and treatment. However, its clinical utility requires further validation through large-scale, multi-center studies.

Background: Microvascular invasion (MVI) is a critical prognostic factor in hepatocellular carcinoma (HCC), but preoperative three-class prediction remains challenging. Radiomics and clinical biomarkers may enable more accurate and individualized assessment. Aim: The aim of this study was to develop and validate a Transformer-based deep learning framework that integrates radiomic and clinical features for direct three-class MVI classification in HCC patients. Methods: This retrospective study included 437 patients with pathologically confirmed hepatocellular carcinoma (HCC) and microvascular invasion (MVI) status from two campuses of a single institution. Patients from Hospital A (n = 305) were randomly divided into training and internal test cohorts, while patients from Hospital B (n = 132) were used as an independent external validation cohort. Radiomic features were extracted from preoperative Gd-BOPTA-enhanced MRI, and clinical laboratory data were collected. A two-stage feature selection strategy, combining univariate statistical testing and recursive feature elimination, was applied. A Transformer-based model was built to classify three MVI categories (M0, M1, M2), and its performance was evaluated in both the internal test cohort and the external validation cohort. Results were compared with those from traditional machine learning models, including Random Forest, Logistic Regression, XGBoost, and LightGBM. Results: On the internal test set (n = 76, Hospital A), the model achieved an accuracy of 0.733 (95% CI: 0.64-0.83), a weighted F1-score of 0.733, and a macro-average AUC of 0.880 (95% CI: 0.807-0.953). The sensitivity and specificity for M1 were 0.56 (95% CI: 0.31-0.78) and 0.86 (95% CI: 0.74-0.94), respectively; for high-risk M2 cases, the sensitivity was 0.73 (95% CI: 0.64-0.81) and the specificity was 0.91 (95% CI: 0.85-0.96). On the external validation set (n = 132, Hospital B), performance remained stable with an accuracy of 0.758, a weighted F1-score of 0.768, and a macro-average AUC of 0.886 (95% CI: 0.833-0.940). Conclusions: This Transformer-based model enables accurate and objective three-class MVI prediction using multi-modal features, supporting individualized surgical planning and improved clinical outcomes. In particular, the ability to preoperatively identify high-risk M2 patients may inform surgical margin design, guide adjuvant therapy strategies, and influence liver transplantation eligibility.

Objective:To investigate the role of MRI peritumoral imaging in predicting microvascular invasion (MVI) status in patients with combined hepatocellular carcinoma and cholangiocarcinoma (cHCC-CCA). Methods:Clinical and pathological data and MRI images of 118 patients with surgically resected and pathologically confirmed cHCC-CCA were retrospectively collected. The tumor in MRI images was segmented by ITK-SNAP software in three dimensions and extended 1 centimeter(cm) towards the tumor periphery. Then, the Python open-source platform was used for radiomics analysis. Mutual information and recursive elimination methods were used to select the optimal features. Clinical models and radiomics models were constructed based on six classifiers. The model's effectiveness was comprehensively evaluated using receiver operating characteristic (ROC), area under curve (AUC), and decision curve analysis (DCA), and the model results were output using Shapley Additive exPlans (SHAP). Results:The differences in HBeAg, capsule, target sign, and lymph node metastasis between MVI negative and positive groups were statistically significant (p < 0.05). Based on peritumoral, 1cm fusion model (in arterial phase) has an AUC of 0.940 (95% CI: 0.801-0.947) and 0.825 (95% CI: 0.633-0.917) in the training/testing set when identifying the MVI status of cHCC-CCA. The accuracy, sensitivity, and specificity in the testing set are 0.778, 0.800, and 0.726, respectively. The DCA shows that when the threshold is approximately 11.08%-66.47%, the net return of the fusion model is higher than that of the clinical and radiomics models under the same conditions. Conclusion:Radiomics with a 1cm extension around the tumor can improve the performance of machine-learning models in predicting MVI labels.

OBJECTIVES:To evaluate the value of contrast-enhanced ultrasound (CEUS) combined with two-dimensional shear wave elastography (2D-SWE) in the differential diagnosis of xanthogranulomatous cholecystitis (XGC) and gallbladder carcinoma (GBC). METHODS:This study included 90 patients who underwent surgical treatment from December 2021 to September 2023. Pathological results divided them into XGC (22 cases) and GBC (68 cases) groups. Both groups underwent 2D, 2D-SWE, and CEUS for the differential diagnosis of XGC and GBC. 2D-SWE assessed the uniformity pattern, margin, and pericholecystic enlargement, while the CEUS analyzed the GB mucosal line, hypo-echoic nodules, vascular imaging types, and enhancement patterns. RESULTS:The 2D-SWE assessment revealed that heterogeneity, irregularity, and pericholecystic enlargement were more common in the GBC group than in the XGC group (p < 0.001). Quantitative analysis revealed a significantly higher GB/liver stiffness ratio in GBC than in XGC (p < 0.001). As assessed through CEUS, the XGC group was characterised by synchronised wash-in and slow wash-out (p < 0.001). The GBC group showed 'rapid wash-in and rapid wash-out' mode (p < 0.001). The diagnostic sensitivity, specificity, and accuracy of differentiating XGC lesions from those of GBC using CEUS combined with 2D-SWE were higher than those of either method alone, which were 91.0 %, 90.9 %, and 90.9 %, respectively. CONCLUSIONS:CEUS combined with 2D-SWE improved the diagnostic efficiency of XGC and GBC, with high sensitivity, specificity, and accuracy. CLINICAL RELEVANCE STATEMENT:CEUS combined with 2D-SWE improved the diagnostic efficiency of XGC and GBC with higher sensitivity, specificity, and accuracy.

BACKGROUND:Vessels Encapsulating Tumor Clusters (VETC) are now recognized as independent indicators of recurrence and overall survival in hepatocellular carcinoma (HCC) patients. However, there has been limited investigation into predicting the VETC pattern using hepatobiliary phase (HBP) features from preoperative gadobenate-enhanced MRI. METHODS:This study involved 252 HCC patients with confirmed VETC status from three different hospitals (Hospital 1: training set with 142 patients; Hospital 2: test set with 64 patients; Hospital 3: validation set with 46 patients). Independent predictive factors for VETC status were determined through univariate and multivariate logistic analyses. Subsequently, these factors were used to construct two distinct VETC prediction models. Model 1 included all independent predictive factors, while Model 2 excluded HBP features. The performance of both models was assessed using the Area Under the Curve (AUC), Decision Curve Analysis, and Calibration Curve. Prediction accuracy between the two models was compared using Net Reclassification Improvement (NRI) and Integrated Discriminant Improvement (IDI). RESULTS:CA199, IBIL, shape, peritumoral hyperintensity on HBP, and arterial peritumoral enhancement were independent predictors of VETC. Model 1 showed robust predictive performance, with AUCs of 0.836 (training), 0.811 (test), and 0.802 (validation). Model 2 exhibited moderate performance, with AUCs of 0.813, 0.773, and 0.783 in the respective sets. Calibration and decision curves for both models indicated consistent predictions between predicted and actual VETC, benefiting HCC patients. NRI showed Model 1 increased by 0.326, 0.389, and 0.478 in the training, test, and validation sets compared to Model 2. IDI indicated Model 1 increased by 0.036, 0.028, and 0.025 in the training, test, and validation sets compared to Model 2. CONCLUSION:HBP features from preoperative gadobenate-enhanced MRI can enhance the predictive performance of VETC in HCC.

?????????????????? ?????????????, ????????(HCC)?????????????. ?????????????????. ????????????????(DL)???????, ????????????????????????. ?????????????????492?(???+?????)??????173???(??LT???), ????????????161?(?????), ?????????????????????(CT)??????????; ????????????7:3?????????????(???345?, ???147?), ????????????????????, ?????????, ??????????????????????; ??????CT???3???(L3)??????(SM)????(PM)????????; ??SM?PM??????, ????????(AutoEncoder)????, TFDeepSurv????????DL??????, ??HCC???????(RFS)???????(OS); ??????????????????(ROC)??????(AUC)??????(C-index), ?????????????(NRI)???????(DCA)????????. ??????CT??L3???SM?PM??????1343???????. AutoEncoder?????????????100???. ??TFDeepSurv????????DL???????, ?HCC??????????????????, ???HCC??????????????????????. ??, ??Kaplan-Meier???????????DL??????????????LT?????????????????????, ????????0.775?0.613. NRI?DCA????DL???????????. ????????????????DL???????????; ????????????????????????????????, ?????????OS?????, ???????????????????.