To evaluate the prognostic significance of tumour mutation burden (TMB) in pancreatic ductal adenocarcinoma (PDAC) and explore the performance of dual-layer spectral CT (DLCT) for noninvasive TMB evaluation. This retrospective analysis enroled patients with histopathologically confirmed PDAC who underwent DLCT between June 2019 and December 2023. Clinical, qualitative radiological, and quantitative conventional CT and DLCT parameters were evaluated. Survival analysis evaluated TMB’s association with progression-free survival (PFS) and identified an optimal TMB cutoff. Independent TMB predictors were identified through univariable and LASSO regression. Predictive performance was quantified via receiver operating characteristic and precision-recall curve assessments. Among 75 patients (mean age 60.4 ± 11.2 years; 41 males, 34 females), median TMB was 2.13 mut/Mb (interquartile range: 1.00–4.26). A 5 mut/Mb cutoff revealed distinct prognostic groups, with high-TMB cases exhibiting better PFS (median PFS: 7 vs 5 months, p = 0.02). Normalised iodine concentration in the pancreatic phase (nICa) was the sole independent TMB predictor (area under the curve [AUC] = 0.901; cutoff = 0.089; accuracy = 89.3

Perineural invasion (PNI) is an independent predictive factor for pancreatic ductal adenocarcinoma (PDAC); however, preoperative prediction is difficult. This study aimed to evaluate the value of dual-layer spectral CT (DLCT) parameters in PNI in PDAC. This retrospective study included patients with pathologically confirmed PDAC who underwent DLCT between August 2019 and March 2024. Risk factors were identified using univariable and least absolute shrinkage and selection operator (LASSO) regression. Receiver operating characteristic (ROC), precision-recall (PR) curves, calibration curves and decision curve analysis (DCA) accessed diagnostic performance. We included 100 patients (mean age: 60.88 ± 11.39 years old, 59 males). There were 71 patients with PNI and 29 without PNI. Relevant factors for PNI included tumour diameter, normalised CT attenuation during the pancreatic parenchymal phase in conventional CT (nCTa), normalised CT attenuation during the pancreatic parenchymal phase in DLCT at 40 keV, and normalised iodine concentration during the pancreatic parenchymal phase (nDIa) (odds ratio [OR], 1.34; 95

Background: Postoperative pancreatic fistula (POPF) is a prevalent and severe complication of pancreaticoenteric anastomosis; however, its accurate preoperative prediction is challenging. Purpose: To investigate the utility of pancreatic stiffness and fluidity derived from tomoelastography and stratify the risk of POPF. Materials and methods: This prospective study included participants who underwent preoperative tomoelastography and pancreaticoenteric anastomosis between November 2021 and July 2024. Participants were divided into training and test sets in a ratio of 2:1. Stiffness and fluidity were quantified using maps of shear-wave speed ( c ) and phase angle (φ). A nomogram was constructed using independent predictive factors of POPF, which were determined using logistic regression analysis of the training set. Predictive performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration, and decision curve analysis (DCA) of both sets. Results: The POPF rate was 20.19% (21/104) and 24.52% (13/53) in the training and test sets, respectively. A moderate correlation was observed between c and fibrosis ( r = 0.66; P < 0.001) and between fat fraction and lipomatosis ( r = 0.55; P < 0.001) in the total set. Pancreatic c (odds ratio, [OR]: 0.27; P < 0.001), φ (OR: 0.17; P < 0.001), main pancreatic duct (MPD) (OR: 0.49; P = 0.002), and fat fraction (OR: 1.05; P = 0.028) in the resection margin were independent predictive factors for POPF in training set. The AUCs of the nomogram were higher than those of the conventional MRI model (fat fraction and MPD) in both the training (0.941 vs. 0.812, P = 0.002) and test sets (0.900 vs. 0.808, P = 0.046). The nomogram had a good calibration. DCA curves showed that the nomogram had better clinical applicability than the conventional MRI model. Conclusion: A nomogram constructed with pancreatic mechanical properties (stiffness and fluidity) quantified using tomoelastography can improve the predictive performance of conventional MRI for POPF risk stratification.

Introduction To develop radiomics and deep learning (DL) based interpretable models using MRI for preoperative prediction of perineural invasion (PNI) in intrahepatic cholangiocarcinoma (ICC). Materials and methods A total of 165 pathologically confirmed ICC patients with preoperative MRI were retrospectively enrolled from two centers (center1, training set, n = 115; validation set, n = 14; internal test set, n = 15; center 2, external test set, n = 21). Radiomics and DL models were constructed for single-phase (pre-contrast, arterial phase, portal venous phase, hepatobiliary phase [HBP]) and multi-phase MRI using the Shukun AI platform and PNI-MambaNet. Model performance was evaluated with the area under the receiver operating characteristic curve (AUC). Gradient-weighted class activation mapping (Grad-CAM) heatmaps visualized the regions prioritized by the DL models. Results The PNI positive rate was 42.4 % (61/144) and 28.6 % (6/21) in the two centers. Radiomics HBP models achieved the highest AUC in the internal test set, while multi-phase model performed best in the external test set (AUC: HBP, 0.778 and 0.733 for the internal and external test sets, respectively; multi-phase, 0.759 and 0.778). For DL models, multi-phase model achieved the highest AUC in the internal test set, while HBP model performed best in the external test set (AUC: HBP, 0.926 and 0.856; multi-phase, 0.944 and 0.844). DL models outperformed radiomics models in the external test set, with Grad-CAM visualizing tumor margin regions as the interest area. Conclusions DL models based on MRI effectively predict PNI in ICC, with visualizations enhancing clinical interpretability and potential application.

Background The mechanical properties of pancreatic diseases provide new insight for diagnosis and differentiation. Tomoelastography, a multifrequency MR elastography technique, provides high-resolution maps of shear wave speed (c in m/s) and phase angle (φ in rad) for evaluating pancreatic stiffness and fluidity. We explore the diagnostic performances of stiffness and fluidity quantified using tomoelastography for differentiating between non-functional pancreatic neuroendocrine neoplasms (NF-pNENs) and solid pseudopapillary neoplasms (SPNs). Methods This prospective study consecutive enrolled 92 untreated patients with pathologically confirmed NF-pNENs and SPNs who underwent tomoelastography examinations from September 2021 to September 2023. Two radiologists independently measured the stiffness and fluidity, and intra/interobserver agreements were assessed. Multivariate logistic regression analysis was performed to identify independent relevant clinical and radiological features to differentiate NF-pNENs and SPNs. The diagnostic performances of stiffness and fluidity and clinical features for tumor differentiation were evaluated using receiver operating characteristic curves. Results Thirty patients with SPNs; 62 with NF-pNENs. The radiologists showed substantial or near-perfect interobserver agreement in evaluating clinical and radiological features. SPNs had lower stiffness (1.87 vs 2.40 m/s, P < 0.001) and fluidity (0.97 vs 1.10 rad, P < 0.001) than NF-pNENs. In the multivariate analysis, the independent relevant factors for tumor differentiation were age (P = 0.002) and stiffness (P < 0.001). The areas under the curves (AUCs) of age and stiffness for tumor differentiation were 0.780 (cutoff, 47.5 years) and 0.876 (cutoff, 2.07 m/s), respectively. The differentiation performance of the combined model (c + age) was better than that of convention model (age + enhancement pattern) (AUC = 0.921 vs 0.813; P = 0.009), as well as compared to the age (AUC = 0.921 vs 0.780; P < 0.001), enhancement pattern (AUC = 0.921 vs 0.675; P < 0.001), and fluidity (φ) (AUC = 0.921 vs 0.788; P = 0.011) metrics, but the stiffness (c) metric alone had comparable differentiation performance (AUC = 0.921 vs 0.876; P = 0.115). Conclusions Tomoelastography quantified lesion stiffness values combined with clinical age metrics were effective in identifying NF-pNENs and SPNs, establishing the value of tomoelastography in the non-invasive preoperative quantitative identification of pancreas-associated neoplasms.

Background:The prognosis of a pancreatic neuroendocrine neoplasm (pNEN) is closely correlated with histological grade. While the role of tomoelastography in predicting tumor grades has been explored in various cancers, evidence regarding its association with the histological grade and clinical features of pNENs remains limited. This study aimed to investigate the association between tomoelastographic parameters and the histological grade and key clinical characteristics of pNENs. Methods:A retrospective study was conducted on 62 patients with pathologically confirmed pNENs, all of whom underwent tomoelastography prior to surgery without receiving neoadjuvant treatment. Patients were categorized into three groups: G1 (n = 28), G2 (n = 30), and G3/neuroendocrine carcinoma (NEC) (n = 4). The relationship between the tomoelastography parameters and clinicopathological characteristics was analysed by using the Kruskal-Wallis test, Spearman correlation, and ordinal logistic regression. Receiver-operating characteristic curves were used for evaluating the diagnostic performance of tomoelastography. Results:The shear wave speed (c), representing stiffness in tomoelastography, increased with tumor grade (1.63 m/s for G1 vs 2.23 m/s for G2 vs 2.53 m/s for G3&NEC, P < 0.001). Parameter c was positively correlated with the tumor size (r = 0.59, P < 0.001) and Ki67 index (r = 0.44, P < 0.001), and was notably higher in lesions with distant or regional lymph node metastases than in those without metastases. Identified as a hazardous factor for tumor grade (odds ratio = 3.92, 95% confidential interval [CI]: 1.88-8.16), c showed good performance in discriminating between G1 and G2 (area under the curve = 0.81, 95% CI: 0.70-0.93, P < 0.001). Conclusion:Tomoelastography offers a promising quantitative tool for assessing the histological grade of pNENs and identifying more aggressive tumor behavior via increased tissue stiffness.

To evaluate the value of dual-layer spectral detector CT (DLCT) for predicting Ki-67 proliferation status and p53 mutations in pancreatic ductal adenocarcinoma (PDAC). This retrospective study included untreated patients with pathologically confirmed PDAC who underwent DLCT between June 2019 and September 2023. Independent relevant clinical-radiological features and quantitative parameters for predicting Ki-67 proliferation status and p53 mutations were identified using multivariate logistic regression analysis. The diagnostic performances of independent variables were evaluated using receiver operating characteristic curves. We included 92 patients (60.19 ± 11.22 years old, 61 males). There were 40 patients with high Ki-67 expression (Ki-67 ≥ 25

Background:The evaluation of uncoupling protein 1 (UCP1) expression in brown adipose tissue (BAT) is critical for assessing the efficacy and prognosis of BAT-targeted therapies in metabolic diseases. This study aimed to explore the association between BAT UCP1 expression and hepatic inflammation in metabolic dysfunction-associated steatotic liver disease (MASLD) mice, and to verify the feasibility of predicting UCP1 expression non-invasively by quantifying hepatic inflammation using synthetic magnetic resonance imaging (SyMRI). Methods:In total, 80 SC57/BL6 and C57 db/db male mice with different diet modes were used for model construction. SyMRI was performed using a 3.0T magnetic resonance (MR) scanner. T1, T2, fat fraction (FF), and R2* values were obtained in the regions of interest (ROIs) delineated in the left and right liver lobes of each mouse. The liver T1 and T2 values were corrected by establishing a generalized linear model (GLM) to obtain fat- and iron-corrected T1 and T2 (cT1_A and cT2_A, respectively). The liver pathological scores were determined by two experienced pathologists using the clinical research network scoring standard for non-alcoholic steatosis hepatitis. BAT UCP1 expression was quantified as the percentage of the positively stained area in three representative regions. The association between the liver pathological score and BAT UCP1 expression was analyzed. The performance of the MRI parameters in evaluating liver inflammation was analyzed and compared. The efficacy of assessing BAT UCP1 expression using MRI parameters was also evaluated. Diagnostic thresholds were determined using Youden's J statistic. Pairwise comparisons of the area under the curve (AUC) values were performed using DeLong's test. Results:The mice models were divided into the normal control (NC; n=13) and MASLD (n=50) groups based on the liver pathological scores. There was a significant difference in BAT UCP1 expression between the NC and MASLD groups (P<0.001). UCP1 expression in the MASLD mice was positively correlated with liver inflammation activity (r=0.762, P<0.001). Among the MRI parameters, cT2_A was the best predictor of liver inflammation [AUC =0.717, 95% confidence interval (CI): 0.614-0.820]. K-means cluster analysis was used to divide the MASLD mice into high- and low-grade BAT UCP1 expression groups (F=370.404, P<0.001). The receiver operating characteristic (ROC) curve analysis showed that cT2_A demonstrated superior predictive value for UCP1 levels (AUC =0.741, 95% CI: 0.644-0.838). Conclusions:SyMRI-derived cT2_A values were used in the quantitative assessment of hepatic inflammation and to predict BAT UCP1 expression levels in MASLD mice. The results suggest that cT2_A could serve as a non-invasive biomarker in metabolic disease monitoring.

Metastatic liver tumor burden (LTB) is a prognostic factor affecting the survival of gastroenteropancreatic neuroendocrine tumors (GEP-NETs), but evaluation of the LTB usually depends on radiologic and functional imaging. This study aimed to develop a clinical model based on easily accessible clinicopathological markers to predict LTB level in GEP-NET patients. LTB was quantified based on 68Ga-DOTANOC PET/CT scan. The optimal cut-off value for high and low-LTB was determined based on our previous study. Serum levels of liver enzymes and tumor biomarkers were obtained within one week before PET/CT scan. The whole dataset was divided into training set and validation set. LASSO regression method was used to select predictors, and multivariate logistic regression was used to develop a clinical model which was further visualized by constructing a nomogram. Area under the curve (AUC) was applied to assess the accuracy of the constructed model. We retrospectively enrolled 200 patients with well-differentiated GEP-NETs. Ki-67 index, GGT (gamma-glutamyltransferase), LDH (lactate dehydrogenase), and NSE (neuron-specific enolase) were selected through the LASSO regression method, and a nomogram was built based on these variables. The predictive model yielded an AUC of 0.785 (95

The vanishing pancreas is a frequently overlooked condition which can result from partial or complete dorsal pancreatic agenesis, intra-pancreatic fat deposition (IPFD) and pancreatic atrophy caused by chronic pancreatitis. A variety of diseases, including cystic fibrosis, maturity-onset diabetes of the young type 8, Shwachman-Diamond syndrome, and Johanson-Blizzard syndrome, can manifest as IPFD. Dorsal pancreatic agenesis can, albeit rarely, coexist with abnormalities or tumors. This review aimed to summarize the various causes that may result in partial or complete vanishing pancreas on computed tomography/magnetic resonance imaging (CT/MRI). We provide a comprehensive review of these imaging findings and their corresponding clinical characteristics, which are crucial for ensuring an accurate diagnosis. By reviewing various causes of pancreatic vanishing, we summarize these imaging findings and their corresponding clinical characteristics, which is crucial for ensuring an accurate diagnosis and patient management.

To evaluate the efficacy of thin-slice T2-weighted imaging (T2WI) and super-resolution reconstruction (SRR) for preoperative assessment of vascular invasion in pancreatic ductal adenocarcinoma (PDAC). Ninety-five PDACs with preoperative MRI were retrospectively enrolled as a training set, with non-reconstructed T2WI (NRT2) in different slice thicknesses (NRT2-3, 3 mm; NRT2-5, ≥ 5 mm). A prospective test set was collected with NRT2-5 (n = 125) only. A deep-learning network was employed to generate reconstructed super-resolution T2WI (SRT2) in different slice thicknesses (SRT2-3, 3 mm; SRT2-5, ≥ 5 mm). Image quality was assessed, including the signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and signal-intensity ratio (SIRt/p, tumor/pancreas; SIRt/b, tumor/background). Diagnostic efficacy for vascular invasion was evaluated using the area under the curve (AUC) and compared across different slice thicknesses before and after reconstruction. SRT2-5 demonstrated higher SNR and SIRt/p compared to NRT2-5 (74.18 vs 72.46; 1.42 vs 1.30; p < 0.05). SRT2-3 showed increased SIRt/p and SIRt/b over NRT2-3 (1.35 vs 1.31; 2.73 vs 2.58; p < 0.05). SRT2-5 showed higher CNR, SIRt/p and SIRt/b than NRT2-3 (p < 0.05). NRT2-3 outperformed NRT2-5 in evaluating venous invasion (AUC: 0.732 vs 0.597, p = 0.021). SRR improved venous assessment (AUC: NRT2-3, 0.927 vs 0.732; NRT2-5, 0.823 vs 0.597; p < 0.05), and SRT2-5 exhibits comparable efficacy to NRT2-3 in venous assessment (AUC: 0.823 vs 0.732, p = 0.162). Thin-slice T2WI and SRR effectively improve the image quality and diagnostic efficacy for assessing venous invasion in PDAC. Thick-slice T2WI with SRR is a potential alternative to thin-slice T2WI. Both thin-slice T2-WI and SRR effectively improve image quality and diagnostic performance, providing valuable options for optimizing preoperative vascular assessment in PDAC. Non-invasive and accurate assessment of vascular invasion supports treatment planning and avoids futile surgery.

Accurate prediction of pathological subtypes on radiological images is one of the most important deep learning (DL) tasks for the appropriate selection of clinical treatment. It is challenging for conventional DL models to obtain sufficient pathological labels for training because of the heavy workload, invasive surgery, and knowledge requirements in pathological analysis. However, existing methods based on limited annotations, such as active learning (AL) and semi-supervised learning (SSL), have difficulty in capturing lesion’s effective features because of the complicated semantic information of radiologic images. In this article, we introduce an efficient domain knowledge-guided semantic prediction framework that integrates domain knowledge-guided AL and SSL methods. This framework can effectively predict pathological subtypes on the basis of radiologic images with limited pathological annotations via three key modules: 1) the discriminative spatial-semantic feature extraction module captures the spatial-semantic features of lesions as semantic information that can better reflect the semantic relationship and effectively mitigate overfitting risk; 2) the explicit sign-guided anchor attention module measures the multimodal semantic distribution of samples under the guidance of clinical domain knowledge, thus selecting the most representative AL samples for pathological labeling; and 3) the implicit radiomics-guided dual-task entanglement module exploits the inherent constraint relationships between implicit radiomics features (IRFs) and pathological subtypes, facilitating the aggregation of unlabeled data. Experiments have been extensively conducted to evaluate our method in two clinical tasks: the pathological grading prediction in pancreatic neuroendocrine neoplasms (pNENs) and muscular invasiveness prediction in bladder cancer (BCa). The experimental results on both tasks demonstrate that the proposed method consistently outperforms the state-of-the-art approaches by a large margin.

Dual-layer spectral detector CT (DLCT) represents an advanced and emerging modality in CT imaging, offering multiparametric images that enhance the quantitative assessment of pancreatic diseases. Non-hypervascular non-functional pancreatic neuroendocrine neoplasms (NF-pNENs) and solid pseudopapillary neoplasms (SPNs) frequently exhibit overlapping clinical and imaging features, complicating their differentiation. This study aimed to investigate the valuable quantitative parameters of DLCT in preoperative differentiation between non-hypervascular NF-pNEN and SPN, as well as to analyze their diagnostic performance. This retrospective study included 52 patients with pathologically confirmed non-hypervascular NF-pNENs and SPNs who underwent DLCT examination before surgery between June 2019 and September 2025. To differentiate between non-hypervascular NF-pNENs and SPNs, independent relevant clinical-radiological features and quantitative parameters were identified using the Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis and multivariate logistic regression analysis. The diagnostic performances of independent variables were assessed through receiver operating characteristic curves. There were 34 patients with non-hypervascular NF-pNENs (46.7 ± 10.2 years, 19 females) and 18 patients with SPNs (32.8 ± 7.9 years, 14 females). Clinical, radiological features and parameters were evaluated with near-perfect agreements among two radiologists. Age and normalised iodine concentration of lesion in the arterial phase (nICa) were the independent factors for differentiating between non-hypervascular NF-pNENs and SPNs in multivariate logistic regression analysis. The areas under the receiver operating characteristic curves (AUCs) for age, nCTa and nICa in tumour differentiation were 0.855 (95

INTRODUCTION:Perineural invasion (PNI) is a significant factor associated with tumor recurrence and metastasis in intrahepatic cholangiocarcinoma (ICC). This study aims to develop predictive model for PNI in ICC using preoperative MRI features and explore its molecular basis. MATERIALS AND METHODS:We analyzed 165 ICC patients from two centers, dividing them into training (n = 120), internal validation (n = 24), and external validation (n = 21) cohorts. Clinical and MRI features associated with PNI were used to construct predictive models. The molecular mechanisms underlying PNI were analyzed using data from The Cancer Genome Atlas (TCGA) and were further validated with a subset of the training cohort. RESULTS:The study identified high CEA level, tumor morphology, location, intrahepatic bile duct dilatation, and tumor invasion into the portal vein as independent predictors of PNI. The clinical-radiological model showed an AUC of 0.839 in the training set, with AUCs of 0.754 and 0.872 in the internal and external validation cohorts, respectively. The COL1A1 gene was found to exhibit up-regulation in the PNI-positive group and was closely linked to poorer overall survival (OS). The arterial-phase enhancement pattern, tumor location, and tumor invasion into portal vein were significantly correlated with COL1A1 expression. CONCLUSIONS:This preoperative clinical-radiological model effectively predicts PNI in ICC, thereby aiding in the clinical management and prognosis stratification for patients. COL1A1 may serve as a significant biomarker for the development of PNI in ICC, which was significantly correlated with MRI features.

Background: Tumor fibrosis plays an important role in chemotherapy resistance in pancreatic ductal adenocarcinoma (PDAC); however, there remains a contradiction in the prognostic value of fibrosis. The authors aimed to investigate the relationship between tumor fibrosis and survival in patients with PDAC, classify patients into high- and low-fibrosis groups, and develop and validate a CT-based radiomics model to non-invasively predict fibrosis before treatment. Materials and methods: This retrospective, bicentric study included 295 patients with PDAC without any treatments before surgery. Tumor fibrosis was assessed using the collagen fraction (CF). Cox regression analysis was used to evaluate the associations of CF with overall survival (OS) and disease-free survival (DFS). Receiver operating characteristic (ROC) analyses were used to determine the rounded threshold of CF. An integrated model (IM) was developed by incorporating selected radiomic features and clinical-radiological characteristics. The predictive performance was validated in the test cohort (Center 2). Results: The CFs were 38.22±6.89% and 38.44±8.66% in center 1 (131 patients, 83 males) and center 2 (164 patients, 100 males), respectively (P=0.814). Multivariable Cox regression revealed that CF was an independent risk factor in the OS and DFS analyses at both centers. ROCs revealed that 40% was the rounded cut-off value of CF. IM predicted CF with areas under the curves (AUCs) of 0.829 (95% CI: 0.753-0.889) and 0.751 (95% CI: 0.677-0.815) in the training and test cohorts, respectively. Decision curve analyses revealed that IM outperformed radiomics model and clinical-radiological model for CF prediction in both cohorts. Conclusions: Tumor fibrosis was an independent risk factor for survival of patients with PDAC, and a rounded cut-off value of 40% provided a good differentiation of patient prognosis. The model combining CT-based radiomics and clinical-radiological features can satisfactorily predict survival-grade fibrosis in patients with PDAC.

To analyze abdominal manifestations in hereditary hemorrhagic telangiectasia (HHT). This retrospective study included 7 patients who were definitively diagnosed with HHT via the Curaçao criteria and underwent contrast-enhanced abdominal CT or MRI at our hospital from January 2016 to January 2024. We reviewed the literature and included patients with HHT who had abdominal CT and/or MRI data. A total of 7 patients were included in the study, 5 males and 2 females, ranging in age from 1 month to 79 years. All patients had at least 2 organs involved. Abdominal manifestations included liver involvement (telangiectasias, hepatic artery-hepatic vein fistula, hepatic artery-portal vein fistula, portal vein-hepatic vein fistula, or diffuse hyperplastic nodules) in 6 cases, splenic involvement (splenomegaly, dilated splenic capillaries, or aneurysm) in 4 cases, gastrointestinal involvement (collateral circulation establishment) in 5 cases, pancreatic involvement (increased and thickened pancreaticoduodenal artery) in 4 cases, and renal and perirenal involvement (no corticomedullary demarcation in the arterial phase, perinephric aneurysm) in 2 cases. Additionally, 139 cases of HHT with abdominal involvement were identified in the literature. Vascular fistula, aneurysm, capillary dilatation and the establishment of collateral circulation in abdominal organs are characteristic manifestations of HHT. MRI’s ability to visualize focal nodular hyperplasia-like nodules in the hepatobiliary phase and multiple arterial phases, together with CT-Angiography’s capacity to detect arteriovenous malformations across multiple organs, underscores the complementary strengths of these imaging modalities in the diagnosis of HHT.

To prospectively investigate the pancreatic stiffness (c) and fluidity (φ) of pancreatic neuroendocrine neoplasms (pNENs), measured using multifrequency magnetic resonance elastography (MRE), and evaluate their performance in predicting pNENs pathological grade. This study included 96 untreated patients with pathologically confirmed pNENs who underwent multifrequency MRE within 2 weeks before surgery between September 2021 and November 2023. Independent predictors of pathological grade were identified using multivariate regression analysis, and predictive performance was assessed using receiver operating characteristic curves. The study included 76 patients with low-grade pNENs (45 men; mean age: 48.7 ± 14.0 years; Grade 1: 34 patients, Grade 2: 42 patients) and 20 patients with high-grade pNENs (10 men; mean age: 54.4 ± 13.8 years; Grade 3: 15 patients, neuroendocrine carcinoma: 5 patients). The two radiologists showed substantial or near-perfect interobserver agreement in evaluating the quantitative parameters. The multivariate regression analysis identified c and relative enhancement in the portal venous phase (V) as independent predictors of pathological grade. The combined model (V + c) had the best predictive performance (area under the curve (AUC) = 0.930; sensitivity: 95.0

Objective: To determine the impact of trans-arterial embolization (TAE) on overall survival (OS) in patients with liver metastases from gastroenteropancreatic neuroendocrine tumors (LM-GEP-NETs) and to identify factors that may influence tumor response to TAE treatment. Methods: This study included patients with histologically and radiologically confirmed LM-GEP-NETs who received TAE treatment at The First Affiliated Hospital, Sun Yat-sen University, between November 2016 and January 2023. Imaging responses were assessed using Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 and modified RECIST (mRECIST) criteria. Tumor response was defined as complete or partial remission. Results: In total, 267 patients with LM-GEP-NETs were included. Patients with liver tumor burdens <25%, 25–50%, and ≥50% had progressively worse OS (p < 0.005). According to the RECIST criteria, 65.9% of patients exhibited tumor responses. Using the mRECIST criteria, 77.5% of patients showed tumor responses. Survival analyses with log-rank tests indicated that patients with tumor responses assessed using either the RECIST or mRECIST criteria had significantly better OS (p = 0.015 and p = 0.023, respectively). Further logistic regression analyses showed that early TAE (within 4 months after diagnosis of liver metastases) was associated with tumor responses assessed using RECIST or mRECIST. These results were further verified using propensity score matching and inverse probability treatment weighting adjusted datasets. Conclusions: A higher liver tumor burden was associated with poorer OS in patients with LM-GEP-NETs. Tumor response after TAE indicates survival benefits. Early TAE (within 4 months of diagnosis) was associated with better treatment responses.