Evaluation of accuracy of automatic out-of-plane respiratory gating for DCEUS-based quantification using principal component analysis

Interpretable Machine Learning for Characterization of Focal Liver Lesions by Contrast-Enhanced Ultrasound

This work proposes an interpretable radiomics approach to differentiate between malignant and benign focal liver lesions (FLLs) on contrast-enhanced ultrasound (CEUS). Although CEUS has shown promise for differential FLLs diagnosis, current clinical assessment is performed only by qualitative analysis of the contrast enhancement patterns. Quantitative analysis is often hampered by the unavoidable presence of motion artifacts and by the complex, spatiotemporal nature of liver contrast enhancement, consisting of multiple, overlapping vascular phases. To fully exploit the wealth of information in CEUS, while coping with these challenges, here we propose combining features extracted by the temporal and spatiotemporal analysis in the arterial phase enhancement with spatial features extracted by texture analysis at different time points. Using the extracted features as input, several machine learning classifiers are optimized to achieve semiautomatic FLLs characterization, for which there is no need for motion compensation and the only manual input required is the location of a suspicious lesion. Clinical validation on 87 FLLs from 72 patients at risk for hepatocellular carcinoma (HCC) showed promising performance, achieving a balanced accuracy of 0.84 in the distinction between benign and malignant lesions. Analysis of feature relevance demonstrates that a combination of spatiotemporal and texture features is needed to achieve the best performance. Interpretation of the most relevant features suggests that aspects related to microvascular perfusion and the microvascular architecture, together with the spatial enhancement characteristics at wash-in and peak enhancement, are important to aid the accurate characterization of FLLs.

Automatic Respiratory Gating Hepatic DCEUS-based Dual-phase Multi-parametric Functional Perfusion Imaging using a Derivative Principal Component Analysis

Purpose: Angiogenesis in liver cancers can be characterized by hepatic functional perfusion imaging (FPI) on the basis of dynamic contrast-enhanced ultrasound (DCEUS). However, accuracy is limited by breathing motion which results in out-of-plane image artifacts. Current hepatic FPI studies do not correct for these artifacts and lack the evaluation of correction accuracy. Thus, a hepatic DCEUS-based dual-phase multi-parametric FPI (DM-FPI) scheme using a derivative principal component analysis (PCA) respiratory gating is proposed to overcome these limitations.

Materials and Methods: By considering severe 3D out-of-plane respiratory motions, the proposed scheme's accuracy was verified with in vitro DCEUS experiments in a flow model mimicking a hepatic vein. The feasibility was further demonstrated by considering in vivo DCEUS measurements in normal rabbit livers, and hepatic cavernous hemangioma and hepatocellular carcinoma in patients. After respiratory kinetics was extracted through PCA of DCEUS sequences under free-breathing condition, dual-phase respiratory gating microbubble kinetics was identified by using a derivative PCA zero-crossing dual-phase detection, respectively. Six dual-phase hemodynamic parameters were estimated from the dual-phase microbubble kinetics and DM-FPI was then reconstructed via color-coding to quantify 2.5D angiogenic hemodynamic distribution for live tumors.

Results: Compared with no respiratory gating, the mean square error of respiratory gating DM-FPI decreased by 1893.9 ± 965.4 (p < 0.05), and mean noise coefficients decreased by 17.5 ± 7.1 (p < 0.05), whereas correlation coefficients improved by 0.4 ± 0.2 (p < 0.01). DM-FPI observably removed severe respiratory motion artifacts on PFI and markedly enhanced the accuracy and robustness both in vitro and in vivo.

Conclusions: DM-FPI precisely characterized and distinguished the heterogeneous angiogenic hemodynamics about perfusion volume, blood flow and flow rate within two anatomical sections in the normal liver, and in benign and malignant hepatic tumors. DCEUS-based DM-FPI scheme might be a useful tool to help clinicians diagnose and provide suitable therapies for liver tumors.

A fast scheme for renal microvascular perfusion functional imaging: Assessed by an imaging quality evaluation model

PURPOSE

This study aimed to develop a fast scheme of multiparametric perfusion functional imaging (PFI) based on dynamic contrast-enhanced ultrasound (DCEUS) for assessing renal microvascular hemodynamics.

METHOD

The flow process in the DCEUS-based PFI was modified step-by-step to improve its operational efficiency, which was validated through in vivo renal perfusion experiments. A multiparametric model with a comprehensive coefficient of imaging quality (CIQ) was then built on four terms of the average information entropy, contrast, gray, and noise coefficient of PFIs to evaluate the sacrifice of imaging quality during modifications of DCEUS-based PFI.

RESULTS

The multiparametric model successfully evaluated modifications of DCEUS-based PFI from multiple perspectives (R²  = 0.73, P < 0.01). Compared with the raw scheme in the renal sagittal and coronal planes, the fast PFI scheme significantly improved its operational efficiency by 62.82 ± 1.07% (P < 0.01) and the nine PFIs simultaneously maintained a similar CIQ of 0.26 ± 0.06.

CONCLUSIONS

The inhomogeneous hemodynamic distributions with a ring-like feature in the renal microvasculature were accurately and efficiently characterized by the fast PFI scheme. The fast PFI scheme can be applied for early diagnosis, follow-up evaluation and monitoring treatment of chronic kidney disease.