Can Machine Learning Models Based on Computed Tomography Radiomics and Clinical Characteristics Provide Diagnostic Value for Epstein-Barr Virus-Associated Gastric Cancer?

OBJECTIVE: The aim of this study was to explore whether machine learning model based on computed tomography (CT) radiomics and clinical characteristics can differentiate Epstein-Barr virus-associated gastric cancer (EBVaGC) from non-EBVaGC. METHODS: Contrast-enhanced CT images were collected from 158 patients with GC (46 EBV-positive, 112 EBV-negative) between April 2018 and February 2023. Radiomics features were extracted from the volumes of interest. A radiomics signature was built based on radiomics features by the least absolute shrinkage and selection operator logistic regression algorithm. Multivariate analyses were used to identify significant clinicoradiological variables. We developed 6 ML models for EBVaGC, including logistic regression, Extreme Gradient Boosting, random forest (RF), support vector machine, Gaussian Naive Bayes, and K-nearest neighbor algorithm. The area under the receiver operating characteristic curve (AUC), the area under the precision-recall curves (AP), calibration plots, and decision curve analysis were applied to assess the effectiveness of each model. RESULTS: Six ML models achieved AUC of 0.706-0.854 and AP of 0.480-0.793 for predicting EBV status in GC. With an AUC of 0.854 and an AP of 0.793, the RF model performed the best. The forest plot of the AUC score revealed that the RF model had the most stable performance, with a standard deviation of 0.003 for AUC score. RF also performed well in the testing dataset, with an AUC of 0.832 (95% confidence interval: 0.679-0.951), accuracy of 0.833, sensitivity of 0.857, and specificity of 0.824, respectively. CONCLUSIONS: The RF model based on clinical variables and Rad_score can serve as a noninvasive tool to evaluate the EBV status of gastric cancer.

Radiomics nomogram based on optimal VOI of multi-sequence MRI for predicting microvascular invasion in intrahepatic cholangiocarcinoma.

OBJECTIVE: Microvascular invasion (MVI) is a significant adverse prognostic indicator of intrahepatic cholangiocarcinoma (ICC) and affects the selection of individualized treatment regimens. This study sought to establish a radiomics nomogram based on the optimal VOI of multi-sequence MRI for predicting MVI in ICC tumors. METHODS: 160 single ICC lesions with MRI scanning confirmed by postoperative pathology were randomly separated into training and validation cohorts (TC and VC). Multivariate analysis identified independent clinical and imaging MVI predictors. Radiomics features were obtained from images of 6 MRI sequences at 4 different VOIs. The least absolute shrinkage and selection operator algorithm was performed to enable the derivation of robust and effective radiomics features. Then, the best three sequences and the optimal VOI were obtained through comparison. The MVI prediction nomogram combined the independent predictors and optimal radiomics features, and its performance was evaluated via the receiver operating characteristics, calibration, and decision curves. RESULTS: Tumor size and intrahepatic ductal dilatation are independent MVI predictors. Radiomics features extracted from the best three sequences (T1WI-D, T1WI, DWI) with VOI10mm (including tumor and 10 mm peritumoral region) showed the best predictive performance, with AUCTC = 0.987 and AUCVC = 0.859. The MVI prediction nomogram obtained excellent prediction efficacy in both TC (AUC = 0.995, 95%CI 0.987-1.000) and VC (AUC = 0.867, 95%CI 0.798-0.921) and its clinical significance was further confirmed by the decision curves. CONCLUSION: A nomogram combining tumor size, intrahepatic ductal dilatation, and the radiomics model of MRI multi-sequence fusion at VOI10mm may be a predictor of preoperative MVI status in ICC patients.

Clinical-radiological predictive model in differential diagnosis of small (≤ 20 mm) solitary pulmonary nodules.

BACKGROUND: There is a lack of clinical-radiological predictive models for the small (≤ 20 mm) solitary pulmonary nodules (SPNs). We aim to establish a clinical-radiological predictive model for differentiating malignant and benign small SPNs. MATERIALS AND METHODS: Between January 2013 and December 2018, a retrospective cohort of 250 patients with small SPNs was used to construct the predictive model. A second retrospective cohort of 101 patients treated between January 2019 and December 2020 was used to independently test the model. The model was also compared to two other models that had previously been identified. RESULTS: In the training group, 250 patients with small SPNs including 156 (62.4%) malignant SPNs and 94 (37.6%) benign SPNs patients were included. Multivariate logistic regression analysis indicated that older age, pleural retraction sign, CT bronchus sign, and higher CEA level were the risk factors of malignant small SPNs. The predictive model was established as: X = - 10.111 + [0.129 × age (y)] + [1.214 × pleural retraction sign (present = 1; no present = 0)] + [0.985 × CT bronchus sign (present = 1; no present = 0)] + [0.21 × CEA level (ug/L)]. Our model had a significantly higher region under the receiver operating characteristic (ROC) curve (0.870; 50% CI: 0.828-0.913) than the other two models. CONCLUSIONS: We established and validated a predictive model for estimating the pre-test probability of malignant small SPNs, that can help physicians to choose and interpret the outcomes of subsequent diagnostic tests.

Combined hepatocellular-cholangiocarcinoma: which preoperative clinical data and conventional MRI characteristics have value for the prediction of microvascular invasion and clinical significance?

OBJECTIVES: To explore which preoperative clinical data and conventional MRI findings may indicate microvascular invasion (MVI) of combined hepatocellular-cholangiocarcinoma (cHCC-CCA) and have clinical significance. METHODS: The study enrolled 113 patients with histopathologically confirmed cHCC-CCA (MVI-positive group [n = 56], MVI-negative group [n = 57]). Two radiologists retrospectively assessed the preoperative MRI features (qualitative analysis of morphology and dynamic enhancement features), and each lesion was assigned according to the LI-RADS. Preoperative clinical data were also evaluated. Logistic regression analyses were used to assess the relative value of these parameters as potential predictors of MVI. Recurrence-free survival (RFS) rates after hepatectomy in the two groups were estimated using Kaplan-Meier survival curves and compared using the log-rank test. RESULTS: The majority of cHCC-CCAs were categorized as LR-M. On multivariate analysis, a higher serum AFP level (OR, 0.523; 95% CI, 0.282-0.971; p = 0.040), intratumoral fat deposition (OR, 14.368; 95% CI, 2.749-75.098; p = 0.002), and irregular arterial peritumoral enhancement (OR, 0.322; 95% CI, 0.164-0.631; p = 0.001) were independent variables associated with the MVI of cHCC-CCA. After hepatectomy, patients with MVI of cHCC-CCA showed earlier recurrence than those without MVI (hazard ratio [HR], 0.402; 95% CI, 0.189-0.854, p = 0.013). CONCLUSION: A higher serum AFP level and irregular arterial peritumoral enhancement are potential predictive biomarkers for the MVI of cHCC-CCA, while intratumoral fat detected on MRI suggests a low risk of MVI. Furthermore, cHCC-CCAs with MVI may have worse surgical outcomes with regard to early recurrence than those without MVI. KEY POINTS: • Higher serum levels of AFP combined with irregular arterial peritumoral enhancement are independent risk factors for the MVI of cHCC-CCA, while fat deposition might be a protective factor. • cHCC-CCA with MVI may have a higher risk of early recurrence after surgery. • Most cHCC-CCAs were categorized as LR-M in this study, and no significant difference was found in MVI based on LI-RADS category.

The assessment of pathological response to neoadjuvant chemotherapy in muscle-invasive bladder cancer patients with DCE-MRI and DWI: a systematic review and meta-analysis.

OBJECTIVE: The purpose of this meta-analysis was to determine the value of dynamic contrast-enhanced-MRI (DCE-MRI) and diffusion-weighted imaging (DWI) in evaluating the pathological response of muscle invasive bladder cancer (MIBC) to neoadjuvant chemotherapy (NAC), and further indirectly compare the diagnostic performance of DCE-MRI and DWI. METHODS: Literatures associated to DCE-MRI and DWI in the evaluation of pathological response of MIBC to NAC were searched from PubMed, Cochrane Library, web of science, and EMBASE databases. The quality assessment of diagnostic accuracy studies 2 tool was used to assess the quality of studies. Pooled sensitivity (SE), specificity (SP), positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and the area under the receiver operating characteristic curves (AUC) with their 95% confidence intervals (CIs) were calculated to evaluate the diagnostic performance of DCE-MRI and DWI in predicting the pathological response to NAC in patients with MIBC. RESULTS: There were 11 studies involved, 6 of which only underwent DCE- MRI examination, 4 of which only underwent DWI examination, and 1 of which underwent both DCE- MRI and DWI examination. The pooled SE, SP, PLR, NLR, DOR of DCE-MRI were 0.88 (95% CI: 0.78-0.93), 0.88 (95% CI: 0.67-0.96), 7.4 (95% CI: 2.3-24.2), 0.14 (95% CI: 0.07-0.27), and 53 (95% CI: 10-288), respectively. The pooled SE, SP, PLR, NLR, DOR of DWI were 0.83 (95% CI: 0.75-0.88), 0.88 (95% CI: 0.81-0.93), 7.1 (95% CI: 4.3-11.7), 0.20 (95% CI: 0.14-0.28), and 36 (95% CI:18-73), respectively. The AUCs of SROC curve for DCE-MRI and DWI were 0.93 (95% CI: 0.91-0.95) and 0.92 (95% CI: 0.89-0.94), respectively. There were no significant differences between DWI and DCE-MRI for SE, SP, and AUC. CONCLUSION: This meta-analysis demonstrated high diagnostic performance of both DCE-MRI and DWI in predicting the pathological response to NAC in MIBC. DWI might be a potential substitute for DCE-MRI, with no significant difference in diagnostic performance between the two. However, caution should be taken when applying our results, as our results were based on indirect comparison. ADVANCES IN KNOWLEDGE: No previous studies have comprehensively analysed the value of DCE-MRI and DWI in evaluating the pathological response to NAC in MIBC. According to the current study, both DCE-MRI and DWI yielded high diagnostic performance, with the AUCs of 0.93 and 0.92, respectively. Indirect comparison no significant difference in the diagnostic performanceof DCE-MRI and DWI.

Clinical and magnetic resonance imaging features predict microvascular invasion in intrahepatic cholangiocarcinoma.

Introduction: Clinical features and magnetic resonance imaging (MRI)-related data are commonly employed in clinical settings and can be used to predict the microvascular invasion (MVI) status of intrahepatic cholangiocarcinoma (ICC) patients. Aim: To generate a clinical and MRI-based model capable of predicting the MVI status of ICC patients. Material and methods: Consecutive ICC patients evaluated from June 2015 to December 2018 were retrospectively enrolled in a training group to establish a predictive clinical MRI model. Consecutive ICC patients evaluated from January 2019 to June 2019 were prospectively enrolled in a validation group to test the reliability of this model. Results: In total, 143 patients were enrolled in the training group, of whom 46 (32.2%) and 96 (67.8%) were MVI-positive and MVI-negative, respectively. Logistics analyses revealed larger tumour size (p = 0.008) and intrahepatic duct dilatation (p = 0.01) to be predictive of MVI positivity, enabling the establishment of the following predictive model: -2.468 + 0.024 × tumour size + 1.094 × intrahepatic duct dilatation. The area under the receiver operating characteristic (ROC) curve (AUC) for this model was 0.738 (p < 0.001). An optimal cut-off value of -1.0184 was selected to maximize sensitivity (71.7%) and specificity (61.9%). When the data from the validation group were incorporated into the predictive model, the AUC value was 0.716 (p = 0.009). Conclusions: Both larger tumour size and intrahepatic duct dilatation were predictive of MVI positivity in patients diagnosed with ICC, and the predictive model developed based on these variables can offer quantitative guidance for assessing the risk of MVI.

A Multi-Parametric Radiomics Nomogram for Preoperative Prediction of Microvascular Invasion Status in Intrahepatic Cholangiocarcinoma.

Background: Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver cancer with increasing incidence in the last decades. Microvascular invasion (MVI) is a poor prognostic factor for patients with ICC, which correlates early recurrence and poor prognosis, and it can affect the selection of personalized therapeutic regime. Purpose: This study aimed to develop and validate a radiomics-based nomogram for predicting MVI in ICC patients preoperatively. Methods: A total of 163 pathologically confirmed ICC patients (training cohort: n = 130; validation cohort: n = 33) with postoperative Ga-DTPA-enhanced MR examination were enrolled, and a time-independent test cohort (n = 24) was collected for external validation. Univariate and multivariate analyses were used to determine the independent predictors of MVI status, which were then incorporated into the MVI prediction nomogram. Least absolute shrinkage and selection operator logistic regression was performed to select optimal features and construct radiomics models. The prediction performances of models were assessed by receiver operating characteristic (ROC) curve analysis. The performance of the MVI prediction nomogram was evaluated by its calibration, discrimination, and clinical utility. Results: Larger tumor size (p = 0.003) and intrahepatic duct dilatation (p = 0.002) are independent predictors of MVI. The final radiomics model shows desirable and stable prediction performance in the training cohort (AUC = 0.950), validation cohort (AUC = 0.883), and test cohort (AUC = 0.812). The MVI prediction nomogram incorporates tumor size, intrahepatic duct dilatation, and the final radiomics model and achieves excellent predictive efficacy in training cohort (AUC = 0.953), validation cohort (AUC = 0.861), and test cohort (AUC = 0.819), fitting well in calibration curves (p > 0.05). Decision curve and clinical impact curve further confirm the clinical usefulness of the nomogram. Conclusion: The nomogram incorporating tumor size, intrahepatic duct dilatation, and the final radiomics model is a potential biomarker for preoperative prediction of the MVI status in ICC patients.

Ablation versus laparoscopic adrenalectomy for the treatment of aldosterone­producing adenoma: a meta-analysis.

PURPOSE: To assess the relative clinical efficacy and safety of ablation and laparoscopic adrenalectomy (LA) for the treatment of adrenal aldosterone-producing adenoma (APA). MATERIALS AND METHODS: Pubmed, Embase, and Cochrane Library databases were searched for relevant studies, while the meta-analysis was performed with RevMan v5.3. RESULTS: After initially identifying 496 potentially relevant studies, five were ultimately included in the final meta-analysis. In total, these studies contained 128 patients that underwent LA and 91 patients that underwent ablation to treat APA. Clinical success rates were comparable between these two groups (OR: 0.55, P = 0.20), whereas the pooled decreases of systolic and diastolic blood pressure were significantly larger in the ablation group (P = 0.01 and 0.002, respectively). Pooled changes in the aldosterone-to-renin ratio (ARR), serum potassium levels, and medication use were similar in both groups (P = 0.62, 0.24, and 0.96, respectively). The average operative duration in the ablation group was somewhat shorter, but the difference was not significant (MD: - 57.99; P = 0.05), whereas the average blood loss and postoperative hospital stay duration of patients in the ablation group were decreased for patients in the ablation group compared to the LA group (P < 0.00001 and 0.00001, respectively). Major complication, minor complication, and hypertension crisis rates were comparable between these groups (P = 0.35, 0.69, and 0.09, respectively). CONCLUSIONS: Ablation offers comparable efficacy to LA when treating patients with APA, but is associated with a reduced operative duration, decreased intraoperative blood loss, and faster postoperative recovery.

[Clinical manifestations and neuroimaging characteristics of children with moyamoya disease].

OBJECTIVE: To study the clinical manifestations and neuroimaging characteristics of pediatric moyamoya disease. METHODS: The clinical data of 17 children with moyamoya disease were retrospectively studied. RESULTS: The onset age was between 3 and 14 years. The main clinical manifestations included motor weakness of extremities or hemiplegia, sensory disturbance and headache. Cranial CT or/and MRI examinations predominately showed cerebral infarct. Magnetic resonance angiography (MRA) and digital subtraction angiography (DSA) showed stenosis or occlusion at the terminus of the siphon portions of internal carotid arteries and proximal portions of anterior or middle cerebral arteries, and abnormal vascular networks at the base of brain. CONCLUSIONS: Cerebral ischemia is main clinical manifestations in children with moyamoya disease, presenting motor weakness of extremities or hemiplegia, sensory disturbance and headache. DSA is essential to the diagnosis of the disease.

Inferior vena cava obstruction in Budd-Chiari syndrome: a comparative study of rapid quantitative phase-contrast MRI and MRV.

PURPOSE: The purpose of this study was to analyze inferior vena cava (IVC) obstruction in Budd-Chiari syndrome (BCS) using rapid quantitative phase-contrast (PC) magnetic resonance (MR) imaging and compare the diagnostic efficacy of PC and MR venography (MRV) to explore the clinical prospects of applying PC quantitative diagnosis in IVC obstruction in BCS. MATERIALS AND METHODS: PC quantitative data and MRV qualitative data obtained in 35 DSA-recognized inferior vena cava obstruction in BCS patients and 35 patients with liver focal lesions found in physical examination were analyzed and compared. RESULTS: The areas under the ROC curve for diagnostic parameters related to PC quantitative parameters, including stroke volume, forward flow volume, absolute stroke volume, mean flux, stroke distance, mean velocity (MV), and peak velocity, were 0.7767, 0.7788, 0.7747, 0.7665, 0.9143, 0.9159, and 0.8947, respectively. The sensitivity, specificity, and accuracy of MV in the diagnosis of IVC obstruction were 88.57%, 85.71%, and 87.14%, respectively. For IVC obstruction in BCS, there was no significant difference between the diagnostic efficacy of MV (one of the quantitative parameters of PC) and that of MRV (P = 0.0768). CONCLUSION: PC can be used to diagnose IVC obstruction in BCS and improve the understanding of hemodynamics of IVC obstruction in BCS.

MRI features predict microvascular invasion in intrahepatic cholangiocarcinoma.

BACKGROUND: The presence of microvascular invasion (MVI) in intrahepatic cholangiocarcinoma (ICC) is a significant adverse prognostic factor. This study sought to investigate the correlation between preoperative imaging parameters and MVI in ICC. METHODS: A total of 108 patients with surgically resected single ICC tumors (34 MVI-positive and 74 MVI-negative lesions) who underwent MRI examination, including T1WI, T2WI, DWI, and dynamic enhancement imaging, were enrolled in this retrospective study. The following qualitative and quantitative characteristics were evaluated: tumor morphology, signal features on T1WI and T2WI, intrahepatic duct dilatation, hepatic capsule retraction, target sign on DWI, dynamic enhancement pattern, arterial phase enhancement pattern, dot-/band-like enhancement inside the tumor, visible vessel penetration inside the tumor (hepatic artery, portal vein, or hepatic vein), integrity of the enhancement edge of the arterial phase, peripheral hepatic enhancement, tumor size, maximum enhancement edge thickness, arterial edge enhancement ratio, and delayed phase enhancement ratio. Other clinicopathological features were also used to predict and evaluate MVI in ICC. Chi-square test, Fisher's exact test, and independent t-test were used for univariate analysis to determine the relationships among the presence of MVI and these MR parameters. Logistic regression analysis was used to identify predictors of MVI among these MR parameters. RESULTS: Among MRI characteristics, tumor morphology, intrahepatic duct dilatation, arterial phase enhancement pattern, visible hepatic artery penetration sign, maximum diameter of the tumor and the arterial phase edge enhancement ratio were correlated with MVI (P = 0.007, 0.003, 0.008, 0.000, 0.003, and 0.002, respectively). Furthermore, higher CA19-9 levels (≥37 U/ml) and pathological tumor grade III were also related to MVI (P = 0.014 and 0.004, respectively). However, multivariate logistic regression analysis demonstrated that none of the parameters were independent risk factors for the diagnosis of MVI in ICCs. CONCLUSION: For the preoperative prediction of MVI in ICC, six qualitative and quantitative data obtained on preoperative MRI, as well as one tumorigenic marker and the pathological tumor grade, were statistically significant. More research is needed to identify MR characteristics that can be used as independent risk factors.

HO-1 attenuates hippocampal neurons injury via the activation of BDNF-TrkB-PI3K/Akt signaling pathway in stroke.

Although recent studies have found that HO-1 plays an important role in neuronal survival, little is known about the precise mechanisms occurring during cerebral ischemia/reperfusion (I/R). Therefore, the aim of this study was to investigate the neuroprotective mechanisms of HO-1 against ischemic brain injury induced by cerebral I/R and to explore whether the BDNF-TrkB-PI3K/Akt signaling pathway contributed to the protection provided by HO-1. Over-expressed HO-1 plasmids were employed to induce the overexpression of HO-1 through hippocampi CA1 injection 5 days before the cerebral I/R animal model was induced by four-vessel occlusion for 15 min transient ischemia and followed by reperfusion in Sprague-Dawley rats. Immunoblotting was carried out to examine the expression of the related proteins, and HE-staining was used to detect the percentage of living neurons in the hippocampal CA1 region. The results showed that over-expressed HO-1 could significantly protect neurons against cerebral I/R. Furthermore, the protein expression of BDNF, TrkB and p-Akt also increased in the rats treated with over-expressed HO-1 plasmids. However, treatment with tropomyosin receptor kinase B (TrkB) receptor antagonist (K252a) reversed the HO-1-induced increase in BDNF and p-Akt protein levels and decreased the level of cleaved caspase-3 protein in I/R rats. In summary, our results imply that HO-1 can decrease cell apoptosis in the I/R rat brain and that the mechanism may be related to the activation of the BDNF-TrkB-PI3K/Akt signaling pathway.