Noninvasive Assessment of Portal Hypertension Using Spectral Computed Tomography

Spectral CT: Current Liver Applications

Using two different energy levels, dual-energy computed tomography (DECT) allows for material differentiation, improves image quality and iodine conspicuity, and allows researchers the opportunity to determine iodine contrast and radiation dose reduction. Several commercialized platforms with different acquisition techniques are constantly being improved. Furthermore, DECT clinical applications and advantages are continually being reported in a wide range of diseases. We aimed to review the current applications of and challenges in using DECT in the treatment of liver diseases. The greater contrast provided by low-energy reconstructed images and the capability of iodine quantification have been mostly valuable for lesion detection and characterization, accurate staging, treatment response assessment, and thrombi characterization. Material decomposition techniques allow for the non-invasive quantification of fat/iron deposition and fibrosis. Reduced image quality with larger body sizes, cross-vendor and scanner variability, and long reconstruction time are among the limitations of DECT. Promising techniques for improving image quality with lower radiation dose include the deep learning imaging reconstruction method and novel spectral photon-counting computed tomography.

Establishment of a non-invasive prediction model for the risk of oesophageal variceal bleeding using radiomics based on CT

AIM

To establish a non-invasive prediction model for the risk of oesophageal variceal bleeding (OVB) using radiomics based on computed tomography (CT).

MATERIALS AND METHODS

The study included 317 patients, 69 of whom were OVB-positive and 248 were OVB-negative. The OVB was caused by cirrhosis associated with hepatitis B. All patients underwent both oesophagogastroduodenoscopy (OGD) and triple-phase contrast-enhanced CT with spectral imaging mode within 14 days before OGD. The patients were divided chronologically into training (n=222) and validation (n=95) cohorts at a ratio of 7:3. The clinical and CT features were collected from a picture archiving and communication system, and radiomics features were extracted from the portal venous phase CT. Spearman's correlation, least absolute shrinkage, and selection operator regression analyses were used to select the most correlated features. Models were built using the selected features. The predictive performance of the models was evaluated using the area under the receiver operating characteristic curve (AUC).

RESULTS

One clinical feature, five CT features, and three radiomics features were selected, and three non-invasive models were built. Integration of the radiomics, CT, and clinical features model showed a better performance in predicting the risk of OVB, with an AUC of 0.89 (95% confidence interval [CI], 0.84-0.94) in the training dataset and 0.78 (95% CI, 0.68-0.87) in the validation dataset.

CONCLUSION

The combination of radiomics, CT, and clinical features may have added value in the non-invasive prediction of OVB, enabling early prevention and treatment.

Dual-energy computed tomography for non-invasive prediction of the risk of oesophageal variceal bleeding with hepatitis B cirrhosis

OBJECTIVE

Oesophageal variceal bleeding (OVB) is a fatal complication of cirrhosis and/or portal hypertension. We aimed to develop a non-invasive prediction model for the risk of OVB using dual-energy computed tomography (CT).

METHODS

317 oesophageal varices (OV) patients with hepatitis B virus-related cirrhosis were retrospectively assessed from January 2018 to December 2018. All patients underwent dual-energy CT scans within 14 days before endoscopy. 222 of 317 patients (174 OVB-negative patients and 48 OVB-positive patients) were included in the training cohort and 95 patients (74 OVB-negative patients and 21 OVB-positive patients) were included in the validation cohort chronologically. A model with the selected conventional CT features and a model with the conventional CT and dual-energy CT features were developed. The prediction accuracy was evaluated using the receiver operating characteristic (ROC) curve. The accuracy and reproducibility of the models for OVB risk prediction of cirrhosis were validated by the validation cohort. The areas under the curve (AUC) of the two models were compared with Delong test.

RESULTS

Diameter of oesophageal vein (OV(mm)), diameter of splenic vein (SPV(mm)), ascites (AS), iodine concentration in short gastric vein (SGV(HU)), iodine concentration in spleen (SP(HU)) were independent predictors of OVB risk (P < 0.05). Then, we developed a model with the selected conventional CT features [OV(mm), SPV(mm), AS] and a model with the conventional CT and dual-energy CT features [OV(mm), SPV(mm), AS, SGV(HU), SP(HU)]. The AUCs of the model built with the conventional CT and dual-energy CT features were higher than the model built only with the conventional CT features in the training (0.839 vs 0.809) and validation cohorts (0.798 vs 0.738).

CONCLUSION

The non-invasive prediction model developed with the conventional CT and dual-energy CT features may have added value in noninvasively predicting OVB than the model built only with the conventional CT features and may have significant clinical implications on early prevention and treatment of OVB.

ADVANCES IN KNOWLEDGE

Combination of dual-energy CT with conventional CT may have added value for non-invasive prediction of OVB compared to conventional CT.

Artificial Intelligence in Imaging of Chronic Liver Diseases: Current Update and Future Perspectives

Here we review artificial intelligence (AI) models which aim to assess various aspects of chronic liver disease. Despite the clinical importance of hepatocellular carcinoma in the setting of chronic liver disease, we focus this review on AI models which are not lesion-specific and instead review models developed for liver parenchyma segmentation, evaluation of portal circulation, assessment of hepatic fibrosis, and identification of hepatic steatosis. Optimization of these models offers the opportunity to potentially reduce the need for invasive procedures such as catheterization to measure hepatic venous pressure gradient or biopsy to assess fibrosis and steatosis. We compare the performance of these AI models amongst themselves as well as to radiomics approaches and alternate modality assessments. We conclude that these models show promising performance and merit larger-scale evaluation. We review artificial intelligence models that aim to assess various aspects of chronic liver disease aside from hepatocellular carcinoma. We focus this review on models for liver parenchyma segmentation, evaluation of portal circulation, assessment of hepatic fibrosis, and identification of hepatic steatosis. We conclude that these models show promising performance and merit a larger scale evaluation.

[Experimental studies for noninvasive assessment of portal vein pressure based on contrast enhanced subharmonic sonographic imaging]

Portal hypertension (PHT) is a common complication of liver cirrhosis, which could be measured by the means of portal vein pressure (PVP). However, there is no report about an effective and reliable way to achieve noninvasive assessment of PVP so far. In this study, firstly, we collected ultrasound images and echo signals of different ultrasound contrast agent (UCA) concentrations and different pressure ranges in a low-pressure environment based on an in vitro simulation device. Then, the amplitudes of the subharmonics in the echo signal were obtained by ultrasound grayscale image construction and fast Fourier transform (FFT). Finally, we analyzed the relationship between subharmonic amplitude (SA) and bionic portal vein pressure (BPVP) through linear regression. As a result, in the pressure range of 7.5-45 mm Hg and 8-20 mm Hg, the linear correlation coefficients (LCC) between SA and BPVP were 0.927 and 0.913 respectively when the UCA concentration was 1∶3 000, and LCC were 0.737 and 0.568 respectively when the UCA concentration was 1∶6 000. Particularly, LCC was increased to 0.968 and 0.916 respectively while the SAs of two UCA concentrations were used as the features of BPVP. Therefore, the results show a good performance on the linear relationship between SA and BPVP, and the LCC will be improved by using SAs obtained at different UCA concentrations as the features of BPVP. The proposed method provides reliable experimental verification for noninvasive evaluation of PVP through SA in clinical practice, which could be a guidance for improving the accuracy of PVP assessment.