Cholangiocarcinoma identified in perforated choledochal cyst in a 3-year-old boy.

Cholangiocarcinoma in patients with Choledochal cysts is rare in childhood; however, it seriously affects the prognosis of the disease. The key to addressing this situation lies in completely removing the extrahepatic cyst. We herein present a case report of a 3-year-old boy with cholangiocarcinoma associated with a choledochal cyst (CDC). Preoperative 3D simulation, based on CT data, played an important role in the treatment of this patient.

Prognostic Impact of Visceral Adipose Tissue Imaging Parameters in Patients with Cholangiocarcinoma after Surgical Resection.

Visceral adiposity is known to be related to poor prognosis in patients with cholangiocarcinoma; however, the prognostic significance of the qualitative features of adipose tissue in cholangiocarcinoma has yet to be well defined. This study investigated the prognostic impact of adipose tissue imaging parameters reflecting the quantity and qualitative characteristics of subcutaneous (SAT) and visceral (VAT) adipose tissue on recurrence-free survival (RFS) and overall survival (OS) in 94 patients undergoing resection of cholangiocarcinoma. The area, mean computed tomography (CT) attenuation, and mean 2-deoxy-2-[18F]fluoro-D-glucose (FDG) uptake of SAT and VAT on positron emission tomography (PET)/CT for staging work-up were measured, and the relationship of these adipose tissue imaging parameters with clinicopathological factors and survival was assessed. TNM stage, histologic grade, lymphovascular invasion, and the size of cholangiocarcinoma showed positive correlations with adipose tissue imaging parameters. Multivariate survival analysis demonstrated that the visceral-to-subcutaneous adipose tissue area ratio (VSR) (p = 0.024; hazard ratio, 1.718) and mean FDG uptake of VAT (p = 0.033; hazard ratio, 9.781) were significant predictors for RFS, but all of the adipose tissue imaging parameters failed to show statistical significance for predicting OS. In addition to visceral adiposity, FDG uptake of VAT might be a promising prognostic parameter for predicting RFS in patients with cholangiocarcinoma.

68 Ga-FAPI-04 PET/CT for the Evaluation of Cholangiocarcinoma : Comparison With 18 F-FDG PET/CT and Abdominal 68 Ga-FAPI-04 PET/MR.

PURPOSE: In this study, we evaluated and compared the diagnostic performances of 68 Ga-FAPI-04 PET/CT and 18 F-FDG PET/CT for primary and metastatic cholangiocarcinoma (CCA) lesions. We also investigated the performance of PET/MR for visualizing and characterizing CCA and liver metastasis lesions. PATIENTS AND METHODS: Forty-four patients with suspected CCA were recruited and underwent 68 Ga-FAPI-04 and 18 F-FDG PET/CT within 1 week, including 30 patients who underwent simultaneous abdominal 68 Ga-FAPI-04 PET/MR scanning. The findings were confirmed by histopathology or radiographic follow-up. RESULTS: Compared with 18 F-FDG PET/CT, 68 Ga-FAPI-04 PET/CT showed higher sensitivity (94.3% vs 88.6%) and the same accuracy (86.4% vs 86.4%) in evaluating primary tumors. However, its specificity was lower (55.6% vs 77.8%). 68 Ga-FAPI-04 PET was superior to 18 F-FDG PET in both patient-based and lesion-based evaluations except for metastatic lesions in the liver and bone. For intrahepatic CCA, 68 Ga-FAPI-04 PET/CT and 18 F-FDG PET/CT (100% vs 100%) had similar detection rates, with similar uptake levels between tracers ( P > 0.05). However, for extrahepatic CCA, 68 Ga-FAPI-04 PET/CT had a higher detection rate (89.5% vs 78.9%), and 68 Ga-FAPI-04 had a higher uptake ( P < 0.05). PET/MR was more effective than PET/CT in terms of lesion conspicuity and diagnostic confidence for primary tumors and liver metastases. In addition, multisequence MRI identified more liver metastases than 68 Ga-FAPI-04 PET/CT and 18 F-FDG PET/CT. CONCLUSIONS: Compared with 18 F-FDG PET/CT, 68 Ga-FAPI-04 PET/CT showed a higher sensitivity in detecting primary CCA tumors, involved lymph nodes, and peritoneal metastases. Compared with 68 Ga-FAPI-04 PET/CT, PET/MR detected primary and liver metastatic lesions more accurately. For extrahepatic CCA, the combination of 68 Ga-FAPI-04 PET/CT and abdominal PET/MRI may replace 18 F-FDG PET/CT.

Predictive Factor of Tumor Aggressiveness in Patients with Extrahepatic Cholangiocarcinoma Based on Diffusion-weighted MRI.

BACKGROUND: Extrahepatic cholangiocarcinoma (EHCC), an exceedingly malignant neoplasm, often eludes early detection, culminating in a dire prognosis. Accurate cancer staging systems and pathological differentiation are designed to guide adjuvant interventions and predict postoperative prognoses. OBJECTIVE: This study sought to investigate the predictive capacity of DW-MRI in discerning T stages, lymph node metastasis, and pathological differentiation grades in patients with EHCC. METHODS: Eighty-five patients were pathologically diagnosed with EHCC and underwent abdominal MRI within two weeks before surgery at our hospital from Aug 2011 to Aug 2021. Tumor axial maximum area (AMA) and apparent diffusion coefficient (ADC) values for diverse T stages, N stages, and differentiation grades were retrospectively analyzed. RESULTS: The Mann-Whitney U test displayed significantly higher lesion AMA values (P =0.006) and lower tumor ADC values (P = 0.001) in the nodepositive group (median ADC and AMA value: 1.220×10-3 mm2/s, 82.231 mm2) than in the node-negative group (median ADC and AMA value: 1.316×10-3 mm2/s, 51.174 mm2). A tumor ADC value<1.249×10-3 mm2/s from the receiver operating characteristic curve (AUC=0.725, P=0.001) exhibited the capability to predict node-positive EHCC with a sensitivity of 64.29%, and specificity of 73.68%. Furthermore, a progressive decrease in the degree of EHCC differentiation was associated with a reduction in the tumor ADC value (P=0.000). CONCLUSION: The N stage and differentiation of EHCC can be evaluated non-invasively using diffusion-weighted MRI.

Component prediction in combined hepatocellular carcinoma-cholangiocarcinoma: habitat imaging and its biologic underpinnings.

PURPOSE: To construct an MRI-based habitat imaging model to help predict component percentage in combined hepatocellular carcinoma-cholangiocarcinoma (cHCC-CCA) preoperatively, and investigate the biologic underpinnings of habitat imaging in cHCC-CCA. METHODS: The study consisted of one retrospective model-building dataset and one prospective validation dataset from two hospitals. All voxels were assigned into different clusters according to the similarity of enhancement pattern by using K-means clustering method, and each habitat's volume fraction in each lesion was calculated. Least absolute shrinkage and selection operator (LASSO) regression analysis was performed to select optimal predictors, and then to establish an MRI-based habitat imaging model. R-squared was calculated to evaluate performance of the prediction models. Model performance was also verified in the prospective dataset with RNA sequencing data, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was then applied to investigate the biologic underpinnings of habitat imaging. RESULTS: A total of 129 patients were enrolled (mean age, 56.1 ± 10.4, 102 man), among which 104 patients were in the retrospective model-building set, while 25 patients in the prospective validation set. Three habitats, habitat1 (HCC-alike habitat), habitat2 (iCCA-alike habitat), and habitat3 (in-between habitat), were identified. Habitat 1's volume fraction, habitat 3's volume fraction, nonrim APHE, nonperipheral washout, and LI-RADS categorization were selected to develop an HCC percentage prediction model with R-squared of 0.611 in the model-building set and 0.541 in the validation set. Habitat 1's volume fraction was correlated with genes involved in regulation of actin cytoskeleton and Rap1 signaling pathway, which regulate cell migration and tumor metastasis. CONCLUSION: Preoperative prediction of HCC percentage in patients with cHCC-CCA was achieved using an MRI-based habitat imaging model, which may correlate with signaling pathways regulating cell migration and tumor metastasis.

Utility of modified pancreaticoduodenectomy (Hi-cut PD) for middle-third cholangiocarcinoma: an alternative to hepatopancreaticoduodenectomy.

BACKGROUND: The standard procedure for middle-third cholangiocarcinoma (MCC) is pancreaticoduodenectomy (PD); hepatopancreaticoduodenectomy (HPD) is often performed despite its high risk. There is no clear selection guidance for these procedures. METHODS: Patients with MCC who underwent HPD or PD were retrospectively evaluated. The conventional PD was modified (mPD) to transect the bile duct beyond or close to the cranial level of the portal bifurcation. RESULTS: The mPD group (n = 55) was characterized by older age, shorter operation time, less blood loss, and less frequent complications than were observed in the HPD group (n = 34). The median grossly tumor-free margin of the proximal bile duct (GM) was 13 mm vs 20 mm (P = 0.006). Overall survival did not differ significantly between groups (48% vs 53% at 5 years, P = 0.399). Multivariate analysis identified positive surgical margin as a sole independent prognostic factor (hazard ratio, 1.89; P = 0.043), which was statistically associated with GM length. Five-year survival for mPD patients with GM ≥15 mm was significantly better than that for those who had GM <15 mm (69% vs 33%, P = 0.011) and comparable to that of HPD patients (53%, P = 0.450). CONCLUSION: The mPD may be recommended in patients with MCC, provided that GM ≥15 mm is expected from the preoperative radiological imaging. Otherwise, HPD should be considered.

Survival analysis of patients with extrahepatic cholangiocarcinoma: a nomogram for clinical and MRI features.

BACKGROUND: This study aimed to establish a predictive model to estimate the postoperative prognosis of patients with extrahepatic cholangiocarcinoma (ECC) based on preoperative clinical and MRI features. METHODS: A total of 104 patients with ECC confirmed by surgery and pathology were enrolled from January 2013 to July 2021, whose preoperative clinical, laboratory, and MRI data were retrospectively collected and examined, and the effects of clinical and imaging characteristics on overall survival (OS) were analyzed by constructing Cox proportional hazard regression models. A nomogram was constructed to predict OS, and calibration curves and time-dependent receiver operating characteristic (ROC) curves were employed to assess OS accuracy. RESULTS: Multivariate regression analyses revealed that gender, DBIL, ALT, GGT, tumor size, lesion's position, the signal intensity ratio of liver to paraspinal muscle (SIRLiver/Muscle), and the signal intensity ratio of spleen to paraspinal muscle (SIRSpleen/Muscle) on T2WI sequences were significantly associated with OS, and these variables were included in a nomogram. The concordance index of nomogram for predicting OS was 0.766, and the AUC values of the nomogram predicting 1-year and 2-year OS rates were 0.838 and 0.863, respectively. The calibration curve demonstrated good agreement between predicted and observed OS. 5-fold and 10-fold cross-validation show good stability of nomogram predictions. CONCLUSIONS: Our nomogram based on clinical, laboratory, and MRI features well predicted OS of ECC patients, and could be considered as a convenient and personalized prediction tool for clinicians to make decisions.

Immunotherapy and targeted therapy for cholangiocarcinoma: Artificial intelligence research in imaging.

Cholangiocarcinoma (CCA) is a highly aggressive hepatobiliary malignancy, second only to hepatocellular carcinoma in prevalence. Despite surgical treatment being the recommended method to achieve a cure, it is not viable for patients with advanced CCA. Gene sequencing and artificial intelligence (AI) have recently opened up new possibilities in CCA diagnosis, treatment, and prognosis assessment. Basic research has furthered our understanding of the tumor-immunity microenvironment and revealed targeted molecular mechanisms, resulting in immunotherapy and targeted therapy being increasingly employed in the clinic. Yet, the application of these remedies in CCA is a challenging endeavor due to the varying pathological mechanisms of different CCA types and the lack of expressed immune proteins and molecular targets in some patients. AI in medical imaging has emerged as a powerful tool in this situation, as machine learning and deep learning are able to extract intricate data from CCA lesion images while assisting clinical decision making, and ultimately improving patient prognosis. This review summarized and discussed the current immunotherapy and targeted therapy related to CCA, and the research progress of AI in this field.

The value of an MRI-based radiomics model in predicting the survival and prognosis of patients with extrahepatic cholangiocarcinoma.

OBJECTIVES: The study aimed to establish radiomics models based on magnetic resonance imaging (MRI) multiparameter images to predict the survival and prognosis of patients with extrahepatic cholangiocarcinoma (ECC). METHODS: Seventy-eight patients with ECC confirmed by pathology were collected retrospectively. The radiomics model_a/b/c were constructed based on the 1/2/3-year survival of patients with ECC. The best texture features were selected according to postoperative survival time and ECC patient status to calculate the radiomics score (Rad-score). A cutoff value was selected, and patients were divided into high-risk and low-risk groups. RESULTS: Model_a, model_b, and model_c were used to predict 1-, 2-, and 3-year postoperative survival rates, respectively. The area under the curve values in the training and test groups were 1.000 and 0.933 for model_a, 0.909 and 0.907 for model_b, 1.000 and 0.975 for model_c, respectively. The survival prediction model based on the Rad-score showed that the postoperative mortality risk differed significantly between risk groups (p < 0.0001). CONCLUSIONS: The MRI radiomics model could be used to predict the survival and prognosis of patients with ECC.

Comparison of 3 Different Therapeutic Particles in Radioembolization of Locally Advanced Intrahepatic Cholangiocarcinoma.

Our objective was to compare 3 different therapeutic particles used for radioembolization in locally advanced intrahepatic cholangiocarcinoma. Methods: 90Y-glass, 90Y-resin, and 166Ho-labeled poly(l-lactic acid) microsphere prescribed activity was calculated as per manufacturer recommendations. Posttreatment quantitative 90Y PET/CT and quantitative 166Ho SPECT/CT were used to determine tumor-absorbed dose, whole-normal-liver-absorbed dose, treated-normal-liver-absorbed dose, tumor-to-nontumor ratio, lung-absorbed dose, and lung shunt fraction. Response was assessed using RECIST 1.1 and the [18F]FDG PET-based change in total lesion glycolysis. Hepatotoxicity was assessed using the radioembolization-induced liver disease classification. Results: Six 90Y-glass, 8 90Y-resin, and 7 166Ho microsphere patients were included for analysis. The mean administered activity was 2.6 GBq for 90Y-glass, 1.5 GBq for 90Y-resin, and 7.0 GBq for 166Ho microspheres. Tumor-absorbed dose and treated-normal-liver-absorbed dose were significantly higher for 90Y-glass than for 90Y-resin and 166Ho microspheres (mean tumor-absorbed dose, 197 Gy for 90Y-glass vs. 73 Gy for 90Y-resin and 50 Gy for 166Ho; mean treated-normal-liver-absorbed dose, 79 Gy for 90Y-glass vs. 37 Gy for 90Y-resin and 31 Gy for 166Ho). The whole-normal-liver-absorbed dose and tumor-to-nontumor ratio did not significantly differ between the particles. All patients had a lung-absorbed dose under 30 Gy and a lung shunt fraction under 20%. The 3 groups showed similar toxicity and response according to RECIST 1.1 and [18F]FDG PET-based total lesion glycolysis changes. Conclusion: The therapeutic particles used for radioembolization differed from each other and showed significant differences in absorbed dose, whereas toxicity and response were similar for all groups. This finding emphasizes the need for separate dose constraints and dose targets for each particle.

A multi-parameter intrahepatic cholangiocarcinoma scoring system based on modified contrast-enhanced ultrasound LI-RADS M criteria for differentiating intrahepatic cholangiocarcinoma from hepatocellular carcinoma.

PURPOSE: To develop a multi-parameter intrahepatic cholangiocarcinoma (ICC) scoring system and compare its diagnostic performance with contrast-enhanced ultrasound (CEUS) liver imaging reporting and data system M (LR-M) criteria for differentiating ICC from hepatocellular carcinoma (HCC). METHODS: This retrospective study enrolled 62 high-risk patients with ICCs and 62 high-risk patients with matched HCCs between January 2022 and December 2022 from two institutions. The CEUS LR-M criteria was modified by adjusting the early wash-out onset (within 45 s) and the marked wash-out (within 3 min). Then, a multi-parameter ICC scoring system was established based on clinical features, B-mode ultrasound features, and modified LR-M criteria. RESULT: We found that elevated CA 19-9 (OR=12.647), lesion boundary (OR=11.601), peripheral rim-like arterial phase hyperenhancement (OR=23.654), early wash-out onset (OR=7.211), and marked wash-out (OR=19.605) were positive predictors of ICC, whereas elevated alpha-fetoprotein (OR=0.078) was a negative predictor. Based on these findings, an ICC scoring system was established. Compared with the modified LR-M and LR-M criteria, the ICC scoring system showed the highest area under the curve (0.911 vs. 0.831 and 0.750, both p<0.05) and specificity (0.935 vs. 0.774 and 0.565, both p<0.05). Moreover, the numbers of HCCs categorized as LR-M decreased from 27 (43.5%) to 14 (22.6%) and 4 (6.5%) using the modified LR-M criteria and ICC scoring system, respectively. CONCLUSION: The modified LR-M criteria-based multi-parameter ICC scoring system had the highest specificity for diagnosing ICC and reduced the number of HCC cases diagnosed as LR-M category.

Nomogram based on preoperative clinical and MRI features to estimate the microvascular invasion status and the prognosis of solitary intrahepatic mass-forming cholangiocarcinoma.

PURPOSE: To develop a nomogram based on preoperative clinical and magnetic resonance imaging (MRI) features for the microvascular invasion (MVI) status in solitary intrahepatic mass-forming cholangiocarcinoma (sIMCC) and to evaluate whether it could predict recurrence-free survival (RFS). METHODS: We included 115 cases who experienced MRI examinations for sIMCC with R0 resection. The preoperative clinical and MRI features were extracted. Independent predictors related to MVI+ were evaluated by stepwise multivariate logistic regression, and a nomogram was constructed. A receiver operating characteristic (ROC) curve was used to assess the predictive ability. All patients were classified into high- and low-risk groups of MVI. Then, the correlations of the nomogram with RFS in patents with sIMCC were analyzed by Kaplan-Meier method. RESULTS: The occurrence rate of MVI+ was 38.3% (44/115). The preoperative independent predictors of MVI+ were carbohydrate antigen 19-9 > 37 U/ml, tumor size > 5 cm, and an ill-defined tumor boundary. Integrating these predictors, the nomogram exerted a favorable diagnostic performance with areas under the ROC curve of 0.767 (95% confidence interval [CI] 0.654-0.881) in the development cohort, and 0.760 (95% CI 0.591-0.929) in the validation cohort. In the RFS analysis, significant differences were observed between the high- and low-risk MVI groups (6-month RFS rates: 64.5% vs. 78.8% and 46.7% vs. 82.4% in the development and validation cohorts, respectively) (P < 0.05). CONCLUSIONS: A nomogram based on clinical and MRI features is a potential biomarker of MVI and may be a potent method to classify the risk of recurrence in patients with sIMCC.

Incremental value of volumetric multiparametric MRI over Fudan score for prognosis of unresectable intrahepatic cholangiocarcinoma treated with systemic chemotherapy.

BACKGROUND: Individualized patient care requires prognostic models customized to a tumor and an individual's disease profile for reliable survival prediction. MRI has prognostic value for intrahepatic cholangiocarcinoma (ICCA). Existing prognostic models for ICCA exclude imaging-based information about an individual's tumor that may reflect important aspects of tumor's biology. Fudan score, a prognostic model applicable to unresectable ICCA, is limited by subjective morphologic imaging parameters. OBJECTIVES: To assess the prognostic value of baseline volumetric multiparametric MRI in unresectable intrahepatic cholangiocarcinoma (ICCA) treated with systemic chemotherapy and the incremental value of MRI over the Fudan score. METHODS: This retrospective study included 114 ICCA patients treated with systemic chemotherapy between 2007 and 2021 after a baseline MRI. The single largest tumor was volumetrically assessed for anatomic (total tumor volume and diameter) and functional parameters (viable tumor volume, percentage-viable tumor volume, viable tumor burden, and ADC). A derivation cohort of 30 patients was utilized to identify MRI parameters associated with overall survival (OS) using Cox regression analysis. The incremental value of MRI over Fudan score was assessed on an independent sub-cohort of 84 patients using Kaplan-Meier analysis and C-index. RESULTS: 114 patients (64 years +/- 11; 61 women) were evaluated. Pre-treatment high (>1350x10-6 mm2/sec) ADC was the only independent predictor of OS (HR, 8.07; P < 0.001). Replacing subjective tumor boundary with objective ADC value, and using modified biochemical thresholds increased the prognostic stratification for the risk groups in the modified ADC-Fudan model compared to the original Fudan model (median survival 12 and 4.5 months; P = 0.055; vs. 11 and 3 months; P < 0.001). The modified ADC-Fudan model demonstrated an 11 % improvement over the original Fudan model (c-index: 0.80 vs. 0.69; P = 0.044) for survival prediction. CONCLUSIONS: High pre-treatment volumetric ADC was associated with unfavorable prognosis in patients with unresectable intrahepatic cholangiocarcinoma treated with systemic chemotherapy. Supplementing the original Fudan model with ADC and modified serum marker thresholds improved the survival prediction performance by 11% in the resulting modified ADC-Fudan model. CLINICAL IMPACT: Volumetric MRI could improve the survival prediction among ICCA patients prior to receiving potentially toxic and expensive palliative chemotherapies. This could potentially guide individualized therapy for this patient cohort.

Systemic Chemotherapy with or without Hepatic Arterial Infusion Chemotherapy for Intrahepatic Cholangiocarcinoma with Extrahepatic Oligometastasis: A Propensity Score-Matched Analysis.

PURPOSE: To compare the effectiveness of hepatic arterial infusion chemotherapy (HAIC) plus systemic chemotherapy (SYS) with that of SYS alone in patients with intrahepatic cholangiocarcinoma (ICC) with extrahepatic oligometastasis in terms of overall survival (OS) and mortality related to liver failure. MATERIALS AND METHODS: Consecutive patients diagnosed with ICC with extrahepatic oligometastasis who received either HAIC plus SYS or SYS alone between January 2019 and January 2021 were included in this retrospective cohort study. Propensity score matching (PSM) analysis was performed to address potential confounding factors. OS, progression-free survival (PFS), and intrahepatic progression-free survival (IPFS) were analyzed. The occurrence of death due to liver failure was also assessed. RESULTS: The study included a total of 179 patients, with 96 receiving SYS alone and 83 receiving HAIC plus SYS. After PSM, 83 pairs were included for further analysis. The median OS and IPFS were significantly longer in the HAIC plus SYS group compared to the SYS alone group (OS: 15.8 months vs 12.7 months; P = .023; IPFS: 9.7 vs 6.1 months; P < .001). No difference was found in PFS between the 2 groups. The HAIC plus SYS group had a significantly lower rate of mortality due to liver failure compared to the SYS alone group (42% vs 72%; P = .002). CONCLUSIONS: HAIC plus SYS is a promising treatment approach for patients with ICC and extrahepatic oligometastasis with improved OS, IPFS, and freedom from liver failure mortality compared with SYS alone.

Noninvasive prediction of perineural invasion in intrahepatic cholangiocarcinoma by clinicoradiological features and computed tomography radiomics based on interpretable machine learning: a multicenter cohort study.

BACKGROUND: Perineural invasion (PNI) of intrahepatic cholangiocarcinoma (ICC) is a strong independent risk factor for tumour recurrence and long-term patient survival. However, there is a lack of noninvasive tools for accurately predicting the PNI status. The authors develop and validate a combined model incorporating radiomics signature and clinicoradiological features based on machine learning for predicting PNI in ICC, and used the Shapley Additive explanation (SHAP) to visualize the prediction process for clinical application. METHODS: This retrospective and prospective study included 243 patients with pathologically diagnosed ICC (training, n =136; external validation, n =81; prospective, n =26, respectively) who underwent preoperative contrast-enhanced computed tomography between January 2012 and May 2023 at three institutions (three tertiary referral centres in Guangdong Province, China). The ElasticNet was applied to select radiomics features and construct signature derived from computed tomography images, and univariate and multivariate analyses by logistic regression were used to identify the significant clinical and radiological variables with PNI. A robust combined model incorporating radiomics signature and clinicoradiological features based on machine learning was developed and the SHAP was used to visualize the prediction process. A Kaplan-Meier survival analysis was performed to compare prognostic differences between PNI-positive and PNI-negative groups and was conducted to explore the prognostic information of the combined model. RESULTS: Among 243 patients (mean age, 61.2 years ± 11.0 (SD); 152 men and 91 women), 108 (44.4%) were diagnosed as PNI-positive. The radiomics signature was constructed by seven radiomics features, with areas under the curves of 0.792, 0.748, and 0.729 in the training, external validation, and prospective cohorts, respectively. Three significant clinicoradiological features were selected and combined with radiomics signature to construct a combined model using machine learning. The eXtreme Gradient Boosting exhibited improved accuracy and robustness (areas under the curves of 0.884, 0.831, and 0.831, respectively). Survival analysis showed the construction combined model could be used to stratify relapse-free survival (hazard ratio, 1.933; 95% CI: 1.093-3.418; P =0.021). CONCLUSIONS: We developed and validated a robust combined model incorporating radiomics signature and clinicoradiological features based on machine learning to accurately identify the PNI statuses of ICC, and visualize the prediction process through SHAP for clinical application.

A multi-center diagnostic system for intrahepatic mass-forming cholangiocarcinoma based on preoperative MRI and clinical features.

OBJECTIVES: To establish a non-invasive diagnostic system for intrahepatic mass-forming cholangiocarcinoma (IMCC) via decision tree analysis. METHODS: Totally 1008 patients with 504 pathologically confirmed IMCCs and proportional hepatocellular carcinomas (HCC) and combined hepatocellular cholangiocarcinomas (cHCC-CC) from multi-centers were retrospectively included (internal cohort n = 700, external cohort n = 308). Univariate and multivariate logistic regression analyses were applied to evaluate the independent clinical and MRI predictors for IMCC, and the selected features were used to develop a decision tree-based diagnostic system. Diagnostic efficacy of the established system was calculated by the receiver operating characteristic curve analysis in the internal training-testing and external validation cohorts, and also in small lesions ≤ 3 cm. RESULTS: Multivariate analysis revealed that female, no chronic liver disease or cirrhosis, elevated carbohydrate antigen 19-9 (CA19-9) level, normal alpha-fetoprotein (AFP) level, lobulated tumor shape, progressive or persistent enhancement pattern, no enhancing tumor capsule, targetoid appearance, and liver surface retraction were independent characteristics favoring the diagnosis of IMCC over HCC or cHCC-CC (odds ratio = 3.273-25.00, p < 0.001 to p = 0.021). Among which enhancement pattern had the highest weight of 0.816. The diagnostic system incorporating significant characteristics above showed excellent performance in the internal training (area under the curve (AUC) 0.971), internal testing (AUC 0.956), and external validation (AUC 0.945) cohorts, as well as in small lesions ≤ 3 cm (AUC 0.956). CONCLUSIONS: In consideration of the great generalizability and clinical efficacy in multi-centers, the proposed diagnostic system may serve as a non-invasive, reliable, and easy-to-operate tool in IMCC diagnosis, providing an efficient approach to discriminate IMCC from other HCC-containing primary liver cancers. CLINICAL RELEVANCE STATEMENT: This study established a non-invasive, easy-to-operate, and explainable decision tree-based diagnostic system for intrahepatic mass-forming cholangiocarcinoma, which may provide essential information for clinical decision-making. KEY POINTS: • Distinguishing intrahepatic mass-forming cholangiocarcinoma (IMCC) from other primary liver cancers is important for both treatment planning and outcome prediction. • The MRI-based diagnostic system showed great performance with satisfying generalization ability in the diagnosis and discrimination of IMCC. • The diagnostic system may serve as a non-invasive, easy-to-operate, and explainable tool in the diagnosis and risk stratification for IMCC.