Assessment of Hepatic Perfusion Using GRASP MRI: Bringing Liver MRI on a New Level

Improved Image Quality Through Deep Learning Acceleration of Gradient-Echo Acquisitions in Uterine MRI: First Application with the Female Pelvis

RATIONALE AND OBJECTIVES

The aim of this study was to compare the image quality of a deep learning (DL)-accelerated volumetric interpolated breath-hold examination (VIBE) sequence with a standard (ST) VIBE sequence in assessing the uterus.

MATERIALS AND METHODS

Between April and December 2023, a total of 61 female patients (aged 41 ± 14 years) who were referred for an magnetic resonance imaging (MRI) of the pelvis were included in this prospective study, after providing informed consent. All examinations were performed with a 1.5 T MRI scanner. The DL VIBE and ST VIBE were acquired before (noncontrast [NC]) and after (contrast-enhanced [CE]) contrast administration in the sagittal orientation. Three readers independently evaluated the following aspects of the images' quality using 4-point Likert scales (1 = nondiagnostic; 4 = excellent): global image quality, anatomy delineation, and lesion detection/demarcation. Motion artifacts and noise were also assessed (1 = no artifacts; 4 = severe artifacts). In addition, all three readers selected their preferred sequence and the sequence in which they had the highest diagnostic confidence.

RESULTS

After exclusions, the data for 54 patients were analyzed. The DL VIBE was preferred by all three readers in almost all cases (NC: 99%; CE: 96%) and rated highest for diagnostic confidence (NC: 98%; CE: 90%). The image quality of the DL VIBE was rated statistically significantly better than that of the ST VIBE, with simultaneously reduced noise and motion artifacts (p < 0.01). The image quality of the DL VIBE was predominantly rated with a score of 4 (NC: 54%; CE: 78%), while the image quality of the ST VIBE was mostly rated with a score of 3 (NC: 53%; CE: 80%). The anatomy of the female pelvis was significantly better delineated by the DL VIBE (p < 0.01; log[OR] = 5.3; 95% CI: 3.7-6.8), and lesions were more clearly demarcated (p < 0.01; log[OR] = 6.7; 95% CI: 4.5-8.8).

CONCLUSION

The DL VIBE sequence showed a significant overall improvement in all image quality characteristics for all readers and was preferred in most cases. The clinical implementation of DL VIBE in MRI of the female pelvis could improve the diagnostic value of the examination.

Tissue-matched analysis of MRI evaluating the tumor infiltrating lymphocytes in hepatocellular carcinoma

Tumor-infiltrating lymphocytes (TILs) play critical roles in the tumor microenvironment and immunotherapy response. This study aims to explore the feasibility of multi-parametric magnetic resonance imaging (MRI) in evaluating TILs and to develop an evaluation model that considers spatial heterogeneity. Multi-parametric MRI was performed on hepatocellular carcinoma (HCC) mice (N = 28). Three-dimensional (3D) printing was employed for tissue sampling, to match the multi-parametric MRI data with tumor tissues, followed by flow cytometry analysis and next-generation RNA-sequencing. Pearson's correlation, multivariate logistic regression, and receiver operating characteristic (ROC) curve analyses were utilized to model TIL-related MRI parameters. MRI quantitative parameters, including T1 relaxation times and perfusion, were correlated with the infiltration of leukocytes, T-cells, CD4+ T-cells, CD8+ T-cells, PD1 + CD8+ T-cells, B-cells, macrophages, and regulatory T-cells (correlation coefficients ranged from -0.656 to 0.482, p <.05) in tumor tissues. TILs were clustered into inflamed and non-inflamed subclasses, with the proportion of T-cells, CD8+ T-cells, and PD1 + CD8+ T-cells significantly higher in the inflamed group compared to the non-inflamed group (43.37% vs. 25.45%, 50.83% vs. 34.90%, 40.45% vs. 29.47%, respectively; p <.001). The TIL evaluation model, based on the Z-score combining Kep and T1post, was able to distinguish between these subgroups, yielding an area under the curve of 0.816 (95% confidence interval 0.721-0.910) and a cut-off value of -0.03 (sensitivity 68.4%, specificity 91.3%). Additionally, the Z-score was related to the gene expression of T-cell activation, chemokine production, and cell adhesion. The tissue-matched analysis of multi-parametric MRI offers a feasible method of regional evaluation and can distinguish between TIL subclasses.

Accelerated dynamic magnetic resonance imaging from Spatial-Subspace Reconstructions (SPARS)

Dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) ideally requires a high spatial and a high temporal resolution, but hardware limitations prevent acquisitions from achieving both simultaneously-either high temporal resolution is exchanged for spatial resolution, or vice versa. Even state-of-the-art image reconstruction techniques that infer missing data in a sparse acquisition space cannot recover the loss of spatial detail, especially at high temporal acceleration rates. The purpose of this paper is to introduce the concept of spatial subspace reconstructions (SPARS) and demonstrate its ability to reconstruct high spatial resolution dynamic images from as few as one acquired k-space spoke per time frame in a dynamic series. Briefly, a low-temporal-high-spatial resolution organization of the acquired raw data is used to estimate the basis vectors of the spatial subspace in which the high-temporal-high-spatial ground truth data resides. This subspace is then used to estimate entire images from single k-space spokes. In both simulated and human in-vivo data, the proposed SPARS reconstruction method outperformed standard GRASP and GRASP-Pro reconstruction, providing a shorter reconstruction time and yielding higher accuracy from both a spatial and temporal perspective.

Dynamic contrast-enhanced MRA of the aorta using a Golden-angle RAdial Sparse Parallel (GRASP) sequence: comparison with conventional time-resolved cartesian MRA (TWIST)

PURPOSE

To compare the application of two contrast-enhanced time-resolved magnetic resonance angiography sequences on an aortic disease patient cohort: the conventional Cartesian-sampling-based, Time-resolved angiography With Interleaved Stochastic Trajectories (TWIST) sequence, and the radial-sampling-based Golden-angle RAdial Sparse Parallel (GRASP) sequence. TWIST is highly sensitive to patient movement, which can lead to blurring and reduced sharpness of vascular structures, particularly in dynamic regions like the aorta. Such motion artifacts can compromise diagnostic accuracy. Radial-sampling-based techniques are less sensitive to motion than cartesian sampling and are expected to improve the image quality in body parts subjected to motion.

METHODS

30 patients (60.9 ± 16.1y.o.) with various aortic diseases underwent a 1.5T magnetic resonance angiography examination. Assessment of image quality in the ascending aorta (AA), descending aorta (DA), and abdominal aorta (AbA) on a 4-point Likert scale (1 = excellent, 4 = non-diagnostic) as well as max. aortic diameters (Dmax) were performed. T-test and multilevel mixed-effect proportional-odds models were used for the image analysis.

RESULTS

GRASP offered superior depiction of vascular structures in terms of vascular contrast for qualitative analysis (TWIST, reader 1: 1.6 ± 0.5; reader 2: 1.9 ± 0.4; reader 3: 1.1 ± 0.4; GRASP, reader 1: 1.5 ± 0.5; reader 2: 1.4 ± 0.5; reader 3: 1.0 ± 0.2) and vessel sharpness for qualitative (TWIST, reader 1: 1.9 ± 0.6; reader 2: 1.6 ± 0.6; reader 3: 2.0 ± 0.3; GRASP, reader 1: 1.4 ± 0.6; reader 2: 1.2 ± 0.4; reader 3: 1.3 ± 0.6) and quantitative analysis (TWIST, AA = 0.12 ± 0.04, DA = 0.12 ± 0.03, AbA = 0.11 ± 0.03; GRASP, AA = 0.20 ± 0.05, DA = 0.22 ± 0.06, AbA=0.20 ± 0.05). Streaking artefacts of GRASP were more visible compared to TWIST (TWIST, reader 1: 2.2 ± 0.6; reader 2: 1.9 ± 0.3; reader 3: 2.0 ± 0.5; GRASP, reader 1: 2.6 ± 0.6; reader 2: 2.3 ± 0.5; reader 3: 2.8 ± 0.6). Aortic Dmax comparison among the sequence showed no clinical relevance.

CONCLUSION

GRASP outperformed TWIST in SNR, vessel sharpness, and reduction in image blurring; streaking artefacts were stronger with GRASP, but did not affect diagnostic image quality.

Multi-average high-acceleration modified volumetric interpolated breath-hold examination (VIBE) for free-breathing multiphase contrast-enhanced liver MRI: a comparative study with breath-hold VIBE

BACKGROUND

Breath-hold volumetric interpolated breath-hold examination (BH-VIBE) of multiphase contrast-enhanced liver magnetic resonance imaging (MPCE-LMRI) requires good cooperative individuals to comply with multiple breath-holds.

PURPOSE

To develop a free-breathing modified VIBE (FB-mVIBE) as a substitute of BH-VIBE in MPCE-LMRI.

MATERIAL AND METHODS

We modified VIBE with a high acceleration factor (2 × 2) and four averages to produce the mVIBE scan. A total of 90 individuals (40 men; mean age = 54.6 ± 10.0 years) who had received MPCE-LMRI as part of a voluntary health check-up for oncology survey were enrolled. Each participant was scanned in four phases (pre-contrast, arterial phase, venous phase, and delay phase), and each phase had two sequential scans. To encounter the timing effect of contrast enhancement, three scan orders were designed: BH-VIBE and FB-mVIBE (group A, n = 30); BH-VIBE and FB-VIBE (group B, n = 30); and FB-mVIBE and BH-VIBE (group C, n = 30). The comparisons included the objective measurements and 25 visual-score by two abdominal radiologists independently.

RESULTS

Consistency between raters was observed for all three sequences (intraclass correlation coefficient [ICC] = 0.741-0.829). For rater 1, the mean scores of FB-mVIBE (23.67 ± 1.32) were equal to those of BH-VIBE (23.83 ± 1.98) in groups C and B (P = 0.852). The mean scores of FB-mVIBE (22.07 ± 3.02), but significantly higher than those of FB-VIBE (14.7 ± 3.41) in groups A and B (P <0.001). Similar scores were found for rater 2. The objective measurement of FB-mVIBE were equal to or higher than BH-VIBE and markedly superior to FB-VIBE.

CONCLUSION

FB-mVIBE is a practical alternative to BH-VIBE for individuals who cannot cooperate with multiple breath-holds for MPCE-LMRI.

Dynamic Radial MR Imaging for Endoleak Surveillance after Endovascular Repair of Abdominal Aortic Aneurysms with Inconclusive CT Angiography: A Prospective Study

Background/Objectives: To assess free-breathing, dynamic radial magnetic resonance angiography (MRA) for detecting endoleaks post-endovascular aortic repair (EVAR) in cases with inconclusive computed tomography angiography (CTA). Methods: This prospective single-center study included 17 participants (mean age, 70 ± 9 years; 13 males) who underwent dynamic radial MRI (Golden-angle RAdial Sparse Parallel-Volumetric Interpolated BrEath-hold, GRASP-VIBE) after inconclusive multiphasic CT for the presence of endoleaks during the follow-up of EVAR-treated abdominal aortic aneurysms. CT and MRI datasets were independently assessed by two radiologists for image quality, diagnostic confidence, and the presence/type of endoleak. Statistical analyses included interrater and intermethod agreement, and diagnostic performance (sensitivity, specificity, area under the curve (AUC)). Results: Subjective image analysis demonstrated good image quality and interrater agreement (k ≥ 0.6) for both modalities, while diagnostic confidence was significantly higher in MRA (p = 0.03). There was significantly improved accuracy for detecting type II endoleaks on MRA (AUC 0.97 [95% CI: 0.87, 1.0]) compared to CTA (AUC 0.66 [95% CI: 0.41, 0.91]; p = 0.03). Although MRA demonstrated higher values for sensitivity, specificity, AUC, and interrater agreement, none of the other types nor the overall detection rate for endoleaks showed differences in the diagnostic performance over CT (p ≥ 0.12). CTA and MRA revealed slight to moderate intermethod concordance in endoleak detection (k = 0.3-0.64). Conclusions: The GRASP-VIBE MRA characterized by high spatial and temporal resolution demonstrates clinical feasibility with good image quality and superior diagnostic confidence. It notably enhances diagnostic performance in detecting and classifying endoleaks, particularly type II, compared to traditional multiphase CTA with inconclusive findings.

Feasibility and Implementation of a 4D Free-Breathing Variable Density Stack-of-Stars Functional Magnetic Resonance Urography in Young Children Without Sedation

BACKGROUND

Functional magnetic resonance urography (MRU) is well established in the diagnostic workup of urinary tract anomalies in children, providing comprehensive morphological and functional information. However, dynamic contrast-enhanced images acquired in the standard Cartesian k-space manner are prone to motion artifacts. A newly introduced 4D high spatiotemporal resolution dynamic contrast-enhanced magnetic resonance imaging based on variable density elliptical centric radial stack-of-stars sharing technique has shown improved image quality regarding motions under free breathing.

OBJECTIVE

The aims of this study were to implement this 4D free-breathing sequence for functional MRU and to compare its image quality and analyzability with standard breath-hold Cartesian MRU.

MATERIALS AND METHODS

We retrospectively evaluated all functional 4D MRU performed without general anesthesia between September 2021 and December 2022 and compared them with matched pairs (age, affected kidney, diagnosis) of standard Cartesian MRU between 2016 and 2022. Image analysis was performed by 2 radiologists independently regarding the following criteria using a 4-point Likert scale, with 4 being the best: overall image quality, diagnostic confidence, respiratory motion artifacts, as well as sharpness and contrast of aorta, kidneys, and ureters. We also measured vertical kidney motion due to respiratory motion and compared the variance for each kidney using F test. Finally, both radiologists calculated the volume, split renal volume (vDRF), split renal Patlak function (pDRF), and split renal function considering the volume and Patlak function (vpDRF) for each kidney. Values were compared using Bland-Altman plots and F test.

RESULTS

Forty children (20 for 4D free-breathing and standard breath-hold, respectively) were enrolled. Ten children of each group were examined using feed-and-sleep technique (median age: 4D, 3.3 months; standard, 4.2 months), 10 were awake (median age: 4D, 8.9 years; standard, 8.6 years). Overall image quality, diagnostic confidence, respiratory motion artifacts, as well as sharpness and contrast of the aorta, kidneys, and ureters were rated significantly better for 4D free-breathing compared with standard breath-hold by both readers ( P ranging from <0.0001 to 0.005). Vertical kidney motion was significantly reduced in 4D free-breathing for the right and the left kidney (both P < 0.001). There was a significantly smaller variance concerning the differences between the 2 readers for vpDRF in 4D MRU ( P = 0.0003). In contrast, no significant difference could be demonstrated for volume ( P = 0.05), vDRF ( P = 0.93), and pDRF ( P = 0.14).

CONCLUSIONS

We demonstrated the feasibility of applying a 4D free-breathing variable density stack-of-stars imaging for functional MRU in young pediatric patients with improved image quality, fewer motion artifacts, and improved functional analyzability.

Predicting intraoperative hemorrhage during curettage treatment of cesarean scar pregnancy using free-breathing GRASP DCE-MRI

OBJECTIVE

To explore the feasibility of the golden-angle radial sparse parallel (GRASP) dynamic magnetic resonance imaging (MRI) technique in predicting the intraoperative bleeding risk of scar pregnancy.

METHODS

A total of 49 patients with cesarean scar pregnancy (CSP) who underwent curettage and GRASP-MRI imaging were retrospectively selected between January 2021 and July 2022. The pharmacokinetic parameters, including Wash-in, Wash-out, time to peck (TTP), initial area under the curve (iAUC), the transfer rate constant (Ktrans), constant flow rate (Kep), and volume of extracellular space (Ve), were calculated. The amount of intraoperative bleeding was recorded by a gynecologist who performed surgery, after which patients were divided into non-hemorrhage (blood loss ≤ 200 mL) and hemorrhage (blood loss > 200 mL) groups. The measured pharmacokinetic parameters were statistically compared using the t-test or Mann-Whitney U test with a significant level set to be p < 0.05. The receiver operating characteristic (ROC) curve was constructed, and the area under the curve (AUC) was calculated to evaluate each parameter's capability in intraoperative hemorrhage subgroup classification.

RESULTS

Twenty patients had intraoperative hemorrhage (blood loss > 200 mL) during curettage. The hemorrhage group had larger Wash-in, iAUC, Ktrans, Ve, and shorter TTP than the non-hemorrhage group (all P > 0.05). Wash-in had the highest AUC value (0.90), while Ktrans had the lowest value (0.67). Wash-out and Kep were not significantly different between the two groups.

CONCLUSION

GRASP DCE-MRI has the potential to forecast intraoperative hemorrhage during curettage treatment of CSP, with Wash-in exhibiting the highest predictive performance. This data holds promise for advancing personalized treatment. However, further study is required to compare its effectiveness with other risk factors identified through anatomical MRI and ultrasound.

GRASPNET: Fast spatiotemporal deep learning reconstruction of golden-angle radial data for free-breathing dynamic contrast-enhanced magnetic resonance imaging

The purpose of the current study was to develop a deep learning technique called Golden-angle RAdial Sparse Parallel Network (GRASPnet) for fast reconstruction of dynamic contrast-enhanced 4D MRI acquired with golden-angle radial k-space trajectories. GRASPnet operates in the image-time space and does not use explicit data consistency to minimize the reconstruction time. Three different network architectures were developed: (1) GRASPnet-2D: 2D convolutional kernels (x,y) and coil and contrast dimensions collapsed into a single combined dimension; (2) GRASPnet-3D: 3D kernels (x,y,t); and (3) GRASPnet-2D + time: two 3D kernels to first exploit spatial correlations (x,y,1) followed by temporal correlations (1,1,t). The networks were trained using iterative GRASP reconstruction as the reference. Free-breathing 3D abdominal imaging with contrast injection was performed on 33 patients with liver lesions using a T1-weighted golden-angle stack-of-stars pulse sequence. Ten datasets were used for testing. The three GRASPnet architectures were compared with iterative GRASP results using quantitative and qualitative analysis, including impressions from two body radiologists. The three GRASPnet techniques reduced the reconstruction time to about 13 s with similar results with respect to iterative GRASP. Among the GRASPnet techniques, GRASPnet-2D + time compared favorably in the quantitative analysis. Spatiotemporal deep learning enables reconstruction of dynamic 4D contrast-enhanced images in a few seconds, which would facilitate translation to clinical practice of compressed sensing methods that are currently limited by long reconstruction times.

Evaluation of Perfusion Change According to Pancreatic Cancer and Pancreatic Duct Dilatation Using Free-Breathing Golden-Angle Radial Sparse Parallel (GRASP) Magnetic Resonance Imaging

PURPOSE

To evaluate perfusion changes in the pancreas with pancreatic cancer and pancreatic duct dilatation using dynamic contrast-enhanced MRI (DCE-MRI).

METHOD

We evaluate the pancreas DCE-MRI of 75 patients. The qualitative analysis includes pancreas edge sharpness, motion artifacts, streak artifacts, noise, and overall image quality. The quantitative analysis includes measuring the pancreatic duct diameter and drawing six regions of interest (ROIs) in the three areas of the pancreas (head, body, and tail) and three vessels (aorta, celiac axis, and superior mesenteric artery) to measure the peak-enhancement time, delay time, and peak concentration. We evaluate the differences in three quantitative parameters among the ROIs and between patients with and without pancreatic cancer. The correlations between pancreatic duct diameter and delay time are also analyzed.

RESULTS

The pancreas DCE-MRI demonstrates good image quality, and respiratory motion artifacts show the highest score. The peak-enhancement time does not differ among the three vessels or among the three pancreas areas. The peak-enhancement time and concentrations in the pancreas body and tail and the delay time in the three pancreas areas are significantly longer (p < 0.05) in patients with pancreatic cancer than in those without pancreatic cancer. The delay time was significantly correlated with the pancreatic duct diameters in the head (p < 0.02) and body (p < 0.001).

CONCLUSION

DCE-MRI can display the perfusion change in the pancreas with pancreatic cancer. A perfusion parameter in the pancreas is correlated with the pancreatic duct diameter reflecting a morphological change in the pancreas.

Deep learning-based super-resolution gradient echo imaging of the pancreas: Improvement of image quality and reduction of acquisition time

PURPOSE

The purpose of this study was to evaluate the impact of a deep learning-based super-resolution technique on T1-weighted gradient-echo acquisitions (volumetric interpolated breath-hold examination; VIBE) on the assessment of pancreatic MRI at 1.5 T compared to standard VIBE imaging (VIBESTD).

MATERIALS AND METHODS

This retrospective single-center study was conducted between April 2021 and October 2021. Fifty patients with a total of 50 detectable pancreatic lesion entities were included in this study. There were 27 men and 23 women, with a mean age of 69 ± 13 (standard deviation [SD]) years (age range: 33-89 years). VIBESTD (precontrast, dynamic, postcontrast) was retrospectively processed with a deep learning-based super-resolution algorithm including a more aggressive partial Fourier setting leading to a simulated acquisition time reduction (VIBESR). Image analysis was performed by two radiologists regarding lesion detectability, noise levels, sharpness and contrast of pancreatic edges, as well as regarding diagnostic confidence using a 5-point Likert-scale with 5 being the best.

RESULTS

VIBESR was rated better than VIBESTD by both readers regarding lesion detectability (5 [IQR: 5, 5] vs. 5 [IQR: 4, 5], for reader 1; 5 [IQR: 5, 5] vs. 4 [IQR: 4, 5]) for reader 2; both P <0.001), noise levels (5 [IQR: 5, 5] vs. 5 [IQR: 4, 5] for reader 1; 5 [IQR: 5, 5] vs. 4 [IQR: 4, 5] for reader 2; both P <0.001), sharpness and contrast of pancreatic edges (5 [IQR: 5, 5] vs. 5 [IQR: 4, 5] for reader 1; 5 [IQR: 5, 5] vs. 4 [IQR: 4, 5] for reader 2; both P <0.001), as well as regarding diagnostic confidence (5 [IQR: 5, 5] vs. 5 [IQR: 4, 5] for reader 1; 5 [IQR: 5, 5] vs. 4 [IQR: 4, 5] for reader 2; both P <0.001). There were no significant differences between lesion sizes as measured by the two readers on VIBESR and VIBESTD images (P > 0.05). The mean acquisition time for VIBESTD (15 ± 1 [SD] s; range: 11-16 s) was longer than that for VIBESR (13 ± 1 [SD] s; range: 11-14 s) (P < 0.001).

CONCLUSION

Our results indicate that the newly developed deep learning-based super-resolution algorithm adapted to partial Fourier acquisitions has a positive influence not only on shortening the examination time but also on improvement of image quality in pancreatic MRI.

Highly time-resolved 4D MR angiography using golden-angle radial sparse parallel (GRASP) MRI

Current dynamic MRA techniques are limited by temporal resolution and signal-to-noise penalties. GRASP, a fast and flexible MRI technique combining compressed-sensing, parallel imaging, and golden-angle radial sampling, acquires volumetric data continuously and can be reconstructed post hoc for user-defined applications. We describe a custom pipeline to retrospectively reconstruct ultrahigh temporal resolution, dynamic MRA from GRASP imaging obtained in the course of routine practice. GRASP scans were reconstructed using a custom implementation of the GRASP algorithm and post-processed with MeVisLab (MeVis Medical Solutions AG, Germany). Twenty consecutive examinations were scored by three neuroradiologists for angiographic quality of specific vascular segments and imaging artifacts using a 4-point scale. Unsubtracted images, baseline-subtracted images, and a temporal gradient dataset were available in 2D and 3D reconstructions. Distinct arterial and capillary phases were identified in all reconstructions, with a median of 2 frames (IQR1-3 and 2-3, respectively). Median rating for vascular segments was 3 (excellent) in all reconstructions and for nearly all segments, with excellent intraclass correlation (range 0.91-1.00). No cases were degraded by artifacts. GRASP-MRI obtained in routine practice can be seamlessly repurposed to produce high quality 4D MRA with 1-2-s resolved isotropic cerebrovascular angiography. Further exploration into diagnostic accuracy in disease-specific applications is warranted.

Current Imaging Diagnosis of Hepatocellular Carcinoma

Simple Summary

The role of imaging in the management of hepatocellular carcinoma (HCC) has significantly evolved and expanded beyond the plain radiological confirmation of the tumor based on the typical appearance in a multiphase contrast-enhanced CT or MRI examination. The introduction of hepatobiliary contrast agents has enabled the diagnosis of hepatocarcinogenesis at earlier stages, while the application of ultrasound contrast agents has drastically upgraded the role of ultrasound in the diagnostic algorithms. Newer quantitative techniques assessing blood perfusion on CT and MRI not only allow earlier diagnosis and confident differentiation from other lesions, but they also provide biomarkers for the evaluation of treatment response. As distinct HCC subtypes are identified, their correlation with specific imaging features holds great promise for estimating tumor aggressiveness and prognosis. This review presents the current role of imaging and underlines its critical role in the successful management of patients with HCC.

Abstract

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer related death worldwide. Radiology has traditionally played a central role in HCC management, ranging from screening of high-risk patients to non-invasive diagnosis, as well as the evaluation of treatment response and post-treatment follow-up. From liver ultrasonography with or without contrast to dynamic multiple phased CT and dynamic MRI with diffusion protocols, great progress has been achieved in the last decade. Throughout the last few years, pathological, biological, genetic, and immune-chemical analyses have revealed several tumoral subtypes with diverse biological behavior, highlighting the need for the re-evaluation of established radiological methods. Considering these changes, novel methods that provide functional and quantitative parameters in addition to morphological information are increasingly incorporated into modern diagnostic protocols for HCC. In this way, differential diagnosis became even more challenging throughout the last few years. Use of liver specific contrast agents, as well as CT/MRI perfusion techniques, seem to not only allow earlier detection and more accurate characterization of HCC lesions, but also make it possible to predict response to treatment and survival. Nevertheless, several limitations and technical considerations still exist. This review will describe and discuss all these imaging modalities and their advances in the imaging of HCC lesions in cirrhotic and non-cirrhotic livers. Sensitivity and specificity rates, method limitations, and technical considerations will be discussed.

Dynamic Liver Magnetic Resonance Imaging During Free Breathing: A Feasibility Study With a Motion Compensated Variable Density Radial Acquisition and a Viewsharing High-Pass Filtering Reconstruction

OBJECTIVE

Robust dynamic contrast-enhanced T1-weighted images are crucial for accurate detection and categorization of focal liver lesions in liver/abdominal magnetic resonance imaging (MRI). As optimal dynamic imaging usually requires multiple breath-holds, its inherent susceptibility to motion artifacts frequently results in degraded image quality in incompliant patients. Because free-breathing imaging may overcome this drawback, the intention of this study was to evaluate a dynamic MRI sequence acquired during free breathing using the variable density, elliptical centric golden angle radial stack-of-stars radial sampling scheme, which so far has not been implemented in 4-dimensional applications.

MATERIALS AND METHODS

In a prospective pilot study, 27 patients received a routine abdominal MRI protocol including the prototype free-breathing sequence (4DFreeBreathing) for dynamic imaging. This enables more convenient and faster reconstruction through variable density, elliptical centric golden angle radial stack-of-stars without the use of additional reconstruction hardware, and even higher motion robustness through soft-gating. A standard breath-hold sequence performed subsequently served as reference standard. Of the continuous dynamic data sets, each dynamic phase was analyzed regarding image quality, motion artifacts and vessel conspicuity using 5-point Likert scales. Furthermore, correct timing of the late arterial phase was compared with the preexaminations.

RESULTS

4DFreeBreathing delivered motion-free dynamic images with high temporal resolution in each subject. Overall image quality scores were rated good or excellent for 4DFreeBreathing and the gold standard without significant differences (P = 0.34). There were significantly less motion artifacts in the 4DFreeBreathing sequence (P < 0.0001), whereas vessel conspicuity in each dynamic phase was comparable for both groups (P = 0.45, P > 0.99, P = 0.22, respectively). Correct timing of the late arterial phase could be achieved in 27 of 27 (100%) examinations using 4DFreeBreathing versus 35 of 53 (66%) preexaminations using gold standard (P < 0.001).

CONCLUSION

The benefit of convenient and fast image reconstruction combined with the superiority in motion robustness and timing compared with standard breath hold sequences renders 4DFreeBreathing an attractive alternative to existing free-breathing techniques in dynamic liver MRI.

Phase2Phase: Respiratory Motion-Resolved Reconstruction of Free-Breathing Magnetic Resonance Imaging Using Deep Learning Without a Ground Truth for Improved Liver Imaging

OBJECTIVES

Respiratory binning of free-breathing magnetic resonance imaging data reduces motion blurring; however, it exacerbates noise and introduces severe artifacts due to undersampling. Deep neural networks can remove artifacts and noise but usually require high-quality ground truth images for training. This study aimed to develop a network that can be trained without this requirement.

MATERIALS AND METHODS

This retrospective study was conducted on 33 participants enrolled between November 2016 and June 2019. Free-breathing magnetic resonance imaging was performed using a radial acquisition. Self-navigation was used to bin the k-space data into 10 respiratory phases. To simulate short acquisitions, subsets of radial spokes were used in reconstructing images with multicoil nonuniform fast Fourier transform (MCNUFFT), compressed sensing (CS), and 2 deep learning methods: UNet3DPhase and Phase2Phase (P2P). UNet3DPhase was trained using a high-quality ground truth, whereas P2P was trained using noisy images with streaking artifacts. Two radiologists blinded to the reconstruction methods independently reviewed the sharpness, contrast, and artifact-freeness of the end-expiration images reconstructed from data collected at 16% of the Nyquist sampling rate. The generalized estimating equation method was used for statistical comparison. Motion vector fields were derived to examine the respiratory motion range of 4-dimensional images reconstructed using different methods.

RESULTS

A total of 15 healthy participants and 18 patients with hepatic malignancy (50 ± 15 years, 6 women) were enrolled. Both reviewers found that the UNet3DPhase and P2P images had higher contrast (P < 0.01) and fewer artifacts (P < 0.01) than the CS images. The UNet3DPhase and P2P images were reported to be sharper than the CS images by 1 reviewer (P < 0.01) but not by the other reviewer (P = 0.22, P = 0.18). UNet3DPhase and P2P were similar in sharpness and contrast, whereas UNet3DPhase had fewer artifacts (P < 0.01). The motion vector lengths for the MCNUFFT800 and P2P800 images were comparable (10.5 ± 4.2 mm and 9.9 ± 4.0 mm, respectively), whereas both were significantly larger than CS2000 (7.0 ± 3.9 mm; P < 0.0001) and UNnet3DPhase800 (6.9 ± 3.2; P < 0.0001) images.

CONCLUSIONS

Without a ground truth, P2P can reconstruct sharp, artifact-free, and high-contrast respiratory motion-resolved images from highly undersampled data. Unlike the CS and UNet3DPhase methods, P2P did not artificially reduce the respiratory motion range.

Hepatectomy-Induced Alterations in Hepatic Perfusion and Function - Toward Multi-Scale Computational Modeling for a Better Prediction of Post-hepatectomy Liver Function

Liver resection causes marked perfusion alterations in the liver remnant both on the organ scale (vascular anatomy) and on the microscale (sinusoidal blood flow on tissue level). These changes in perfusion affect hepatic functions via direct alterations in blood supply and drainage, followed by indirect changes of biomechanical tissue properties and cellular function. Changes in blood flow impose compression, tension and shear forces on the liver tissue. These forces are perceived by mechanosensors on parenchymal and non-parenchymal cells of the liver and regulate cell-cell and cell-matrix interactions as well as cellular signaling and metabolism. These interactions are key players in tissue growth and remodeling, a prerequisite to restore tissue function after PHx. Their dysregulation is associated with metabolic impairment of the liver eventually leading to liver failure, a serious post-hepatectomy complication with high morbidity and mortality. Though certain links are known, the overall functional change after liver surgery is not understood due to complex feedback loops, non-linearities, spatial heterogeneities and different time-scales of events. Computational modeling is a unique approach to gain a better understanding of complex biomedical systems. This approach allows (i) integration of heterogeneous data and knowledge on multiple scales into a consistent view of how perfusion is related to hepatic function; (ii) testing and generating hypotheses based on predictive models, which must be validated experimentally and clinically. In the long term, computational modeling will (iii) support surgical planning by predicting surgery-induced perfusion perturbations and their functional (metabolic) consequences; and thereby (iv) allow minimizing surgical risks for the individual patient. Here, we review the alterations of hepatic perfusion, biomechanical properties and function associated with hepatectomy. Specifically, we provide an overview over the clinical problem, preoperative diagnostics, functional imaging approaches, experimental approaches in animal models, mechanoperception in the liver and impact on cellular metabolism, omics approaches with a focus on transcriptomics, data integration and uncertainty analysis, and computational modeling on multiple scales. Finally, we provide a perspective on how multi-scale computational models, which couple perfusion changes to hepatic function, could become part of clinical workflows to predict and optimize patient outcome after complex liver surgery.

Intraindividual Comparison of Compressed Sensing-Accelerated Cartesian and Radial Arterial Phase Imaging of the Liver in an Experimental Tumor Model

OBJECTIVES

The aim of this study was to intraindividually compare the performance of 2 compressed sensing (CS)-accelerated magnetic resonance imaging (MRI) sequences, 1 featuring Cartesian (compressed sensing volumetric interpolated breath-hold examination [CS-VIBE]) and the other radial (golden-angle radial sparse parallel [GRASP]) k-space sampling in continuous dynamic imaging during hepatic vascular phases, using extracellular and hepatocyte-specific contrast agents.

MATERIALS AND METHODS

Seven New Zealand white rabbits, with induced VX2 liver tumors (median number of lesions, 2 ± 0.83; range, 1-3), received 2 continuously acquired T1-weighted prototype CS-accelerated MRI sequences (CS-VIBE and GRASP) with high spatial (0.8 × 0.8 × 1.5 mm) and temporal resolution (3.5 seconds) in randomized order on 2 separate days using a 1.5-T scanner. In all animals, imaging was performed using first gadobutrol at a dose of 0.1 mmol/kg and, then 45 minutes later, gadoxetic acid at a dose of 0.025 mmol/kg.The following qualitative parameters were assessed using 3- and 5-point Likert scales (3 and 5 being the highest scores respectively): image quality (IQ), arterial and venous vessel delineation, tumor enhancement, motion artifacts, and sequence-specific artifacts. Furthermore, the following quantitative parameters were obtained: relative peak signal enhancement, time to peak, mean transit time, and plasma flow ratios. Paired sampled t tests and Wilcoxon signed rank tests were used for intraindividual comparison. Image analysis was performed by 2 radiologists.

RESULTS

Six of 7 animals underwent the full imaging protocol and obtained data were analyzed statistically. Overall IQ was rated moderate to excellent, not differing significantly between the 2 sequences.Gadobutrol-enhanced CS-VIBE examinations revealed the highest mean Likert scale values in terms of vessel delineation and tumor enhancement (arterial 4.4 [4-5], venous 4.3 [3-5], and tumor 2.9 [2-3]). Significantly, more sequence-specific artifacts were seen in GRASP examinations (P = 0.008-0.031). However, these artifacts did not impair IQ. Excellent Likert scale ratings were found for motion artifacts in both sequences. In both sequences, a maximum of 4 hepatic arterial dominant phases were obtained. Regarding the relative peak signal enhancement, CS-VIBE and GRASP showed similar results. The relative peak signal enhancement values did not differ significantly between the 2 sequences in the aorta, the hepatic artery, or the inferior vena cava (P = 0.063-0.536). However, significantly higher values were noted for CS-VIBE in gadoxetic acid-enhanced examinations in the portal vein (P = 0.031) and regarding the tumor enhancement (P = 0.005). Time to peak and mean transit time or plasma flow ratios did not differ significantly between the sequences.

CONCLUSIONS

Both CS-VIBE and GRASP provide excellent results in dynamic liver MRI using extracellular and hepatocyte-specific contrast agents, in terms of IQ, peak signal intensity, and presence of artifacts.

Simultaneous evaluation of perfusion and morphology using GRASP MRI in hepatic fibrosis

OBJECTIVES

To determine if golden-angle radial sparse parallel (GRASP) dynamic contrast-enhanced (DCE)-MRI allows simultaneous evaluation of perfusion and morphology in liver fibrosis.

METHODS

Participants who were scheduled for liver biopsy or resection were enrolled (NCT02480972). Images were reconstructed at 12-s temporal resolution for morphologic assessment and at 3.3-s temporal resolution for quantitative evaluation. The image quality of the morphologic images was assessed on a four-point scale, and the Liver Imaging Reporting and Data System score was recorded for hepatic observations. Comparisons were made between quantitative parameters of DCE-MRI for the different fibrosis stages, and for hepatocellular carcinoma (HCCs) with different LR features.

RESULTS

DCE-MRI of 64 participants (male = 48) were analyzed. The overall image quality consistently stood at 3.5 ± 0.4 to 3.7 ± 0.4 throughout the exam. Portal blood flow significantly decreased in participants with F2-F3 (n = 18, 175 ± 110 mL/100 mL/min) and F4 (n = 12, 98 ± 47 mL/100 mL/min) compared with those in participants with F0-F1 (n = 34, 283 ± 178 mL/100 mL/min, p < 0.05 for all). In participants with F4, the arterial fraction and extracellular volume were significantly higher than those in participants with F0-F1 and F2-F3 (p < 0.05). Compared with HCCs showing non-LR-M features (n = 16), HCCs with LR-M (n = 5) had a significantly prolonged mean transit time and lower arterial blood flow (p < 0.05).

CONCLUSIONS

Liver MRI using GRASP obtains both sufficient spatial resolution for confident diagnosis and high temporal resolution for pharmacokinetic modeling. Significant differences were found between the MRI-derived portal blood flow at different hepatic fibrosis stages.

KEY POINTS

A single MRI examination is able to provide both images with sufficient spatial resolution for anatomic evaluation and those with high temporal resolution for pharmacokinetic modeling. Portal blood flow was significantly lower in clinically significant hepatic fibrosis and mean transit time and extracellular volume increased in cirrhosis, compared with those in no or mild hepatic fibrosis. HCCs with different LR features showed different quantitative parameters of DCE-MRI: longer mean transit time and lower arterial flow were observed in HCCs with LR-M features.

Transarterial Chemoembolization of Hepatocellular Carcinoma Using Radiopaque Drug-Eluting Embolics: Impact of Embolic Density and Residual Tumor Perfusion on Tumor Recurrence and Survival

Purpose

To evaluate the value of dual-phase parenchymal blood volume (PBV) C-arm mounted cone-beam-CT (CBCT) to enable assessment of radiopaque, doxorubicin-loaded drug-eluting embolics (rDEE) based on the visual degree of embolization, embolic density and residual tumor perfusion as early predictors for tumor recurrence after transarterial chemoembolization (TACE) of hepatocellular carcinoma (HCC).

Material and Methods

Thirty patients (50 HCCs) were prospectively enrolled, underwent cross-sectional imaging before and after TACE using 100–300 µm rDEE and had regular follow-up examinations. Directly before and after the TACE procedure, PBV-CBCT was acquired. The response was evaluated and compared to visual degree of embolization (DE) and embolic density (ED) of rDEE deposits, as well as the presence of residual tumor perfusion (RTP) derived from PBV-CBCT. Outcome was assessed by mid-term tumor response applying mRECIST and patient survival after 12 months.

Results

RTP was detected in 16 HCCs and correlated negatively with DE (p = .03*) and ED (p = .0009*). The absence of RTP significantly improved lesion-based mid-term response rates regarding complete response (CR, 30/34 (88%) vs 2/16 (12.5%), p = .0002*), lesion-based complete response rate was 75% (21/28) for DE ≥ 50% vs. 50% (11/22) for DE < 50% (p =  .08) and 82% (27/33) for ED ≥ 2 vs. 29% for ED < 2 (5/17), p =  .005*). Thirteen patients were treated with re-TACE within 12 months, 11 of which had shown RTP. 12-month survival rate was 93%.

Conclusion

Residual tumor perfusions as assessed by PBV-CBCT during rDEE-TACE proved to be the best parameter to predict mid-term response.

“Level of Evidence: Level 3”

Supplementary Information

The online version contains supplementary material available at 10.1007/s00270-021-02858-6.

Spiral 3-Dimensional T1-Weighted Turbo Field Echo: Increased Speed for Magnetization-Prepared Gradient Echo Brain Magnetic Resonance Imaging

OBJECTIVES

Spiral magnetic resonance imaging acquisition may enable improved image quality and higher scan speeds than Cartesian trajectories. We tested the performance of four 3D T1-weighted (T1w) TFE sequences (magnetization-prepared gradient echo magnetic resonance sequence) with isotropic spatial resolution for brain imaging at 1.5 T in a clinical patient cohort based on qualitative and quantitative image quality metrics. Two prototypical spiral TFE sequences (spiral 1.0 and spiral 0.85) and a Cartesian compressed sensing technology accelerated TFE sequence (CS 2.5; acceleration factor of 2.5) were compared with a conventional (reference standard) Cartesian parallel imaging accelerated TFE sequence (SENSE; acceleration factor of 1.8).

MATERIALS AND METHODS

The SENSE (5:52 minutes), CS 2.5 (3:17 minutes), and spiral 1.0 (2:16 minutes) sequences all had identical spatial resolutions (1.0 mm). The spiral 0.85 (3:47 minutes) had a higher spatial resolution (0.85 mm). The 4 TFE sequences were acquired in 41 patients (20 with and 21 without contrast media). Three readers rated qualitative image quality (12 categories) and selected their preferred sequence for each patient. Two readers performed quantitative analysis whereby 6 metrics were derived: contrast-to-noise ratio for white and gray matter (CNRWM/GM), contrast ratio for gray matter-CSF (CRGM/CSF), and white matter-CSF (CRWM/CSF); and coefficient of variations for gray matter (CVGM), white matter (CVWM), and CSF (CVCSF). Friedman tests with post hoc Nemenyi tests, exact binomial tests, analysis of variance with post hoc Dunnett tests, and Krippendorff alphas were computed.

RESULTS

Concerning qualitative analysis, the CS 2.5 sequence significantly outperformed the SENSE in 4/1 (with/without contrast) categories, whereas the spiral 1.0 and spiral 0.85 showed significantly improved scores in 10/9and 7/7 categories, respectively (P's < 0.001-0.039). The spiral 1.0 was most frequently selected as the preferred sequence (reader 1, 10/15 times; reader 2, 9/12 times; reader 3, 11/13times [with/without contrast]). Interreader agreement ranged from substantial to almost perfect (alpha = 0.615-0.997). Concerning quantitative analysis, compared with the SENSE, the CS 2.5 had significantly better scores in 2 categories (CVWM, CVCSF) and worse scores in 2 categories (CRGM/CSF, CRWM/CSF), the spiral 1.0 had significantly improved scores in 4 categories (CNRWM/GM, CRGM/CSF, CRWM/CSF, CVWM), and the spiral 0.85 had significantly better scores in 2 categories (CRGM/CSF, CRWM/CSF).

CONCLUSIONS

Spiral T1w TFE sequences may deliver high-quality clinical brain imaging, thus matching the performance of conventional parallel imaging accelerated T1w TFEs. Imaging can be performed at scan times as short as 2:16 minutes per sequence (61.4% scan time reduction compared with SENSE). Optionally, spiral imaging enables increased spatial resolution while maintaining the scan time of a Cartesian-based acquisition schema.

Quantitative magnetic resonance imaging for focal liver lesions: bridging the gap between research and clinical practice

Magnetic resonance imaging (MRI) is highly important for the detection, characterization, and follow-up of focal liver lesions. Several quantitative MRI-based methods have been proposed in addition to qualitative imaging interpretation to improve the diagnostic work-up and prognostics in patients with focal liver lesions. This includes DWI with apparent diffusion coefficient measurements, intravoxel incoherent motion, perfusion imaging, MR elastography, and radiomics. Multiple research studies have reported promising results with quantitative MRI methods in various clinical settings. Nevertheless, applications in everyday clinical practice are limited. This review describes the basic principles of quantitative MRI-based techniques and discusses the main current applications and limitations for the assessment of focal liver lesions.

Fusion of Preinterventional MR Imaging With Liver Perfusion CT After RFA of Hepatocellular Carcinoma: Early Quantitative Prediction of Local Recurrence

OBJECTIVES

The aim of this study was to evaluate the ability of fusion of pretreatment magnetic resonance (MR) imaging with posttreatment perfusion-CT (P-CT) after radiofrequency ablation (RFA) of hepatocellular carcinomas (HCCs) and to determine treatment success in an objective, quantitative way.

MATERIALS AND METHODS

In this institutional review board-approved study, 39 patients (78.4% male; mean age 68.2 ± 8.5 years) with a total of 43 HCCs, who underwent RFA at our institution and had diagnostic pre-RFA MR imaging and post-RFA P-CT, were included in the study. Post-RFA P-CT was performed within 24 hours after RFA. In a first step, the pre-RFA MR imaging, depicting the HCC, was registered onto the post-RFA P-CT using nonrigid image registration. After image registration, the MR data were reloaded jointly with the calculated perfusion parameter volumes into the perfusion application for quantitative analysis. A 3-dimensional volume of interest was drawn around the HCC and the ablation zone; both outlines were automatically projected onto all perfusion maps. Resulting perfusion values (normalized peak enhancement [NPE, %]; arterial liver perfusion [ALP, in mL/min/100 mL]; BF [blood flow, mL/100 mL/min]; and blood volume [BV, mL/100 mL]) and histogram data were recorded. Local tumor recurrence was defined in follow-up imaging according to the EASL guidelines.

RESULTS

Image registration of MR imaging and CT data was successful in 37 patients (94.9%). Local tumor recurrence was observed in 5 HCCs (12%). In the local tumor recurrence group (LTR-group), HCC size was significantly larger (22.7 ± 3.9 cm vs 17.8 ± 5.3 cm, P = 0.035) and the ablation zone was significantly smaller (29.8 ± 6.9 cm vs 39.3 ± 6.8 cm, P = 0.014) compared with the no-local tumor recurrence group (no-LTR group). The differences (ablation zone - tumor) of the perfusion parameters NPE, ALP, BF, and BV significantly differed between the 2 groups (all P's < 0.005). Especially, the difference (ablation zone - tumor) of NPE and ALP, with a cutoff value of zero, accurately differentiated between LTR or no-LTR in all cases. A negative difference of these perfusion parameters identified local tumor recurrence in all cases.

CONCLUSIONS

Image registration of pre-RFA MR imaging onto post-RFA P-CT is feasible and allows to predict local tumor recurrence within 24 hours after RFA in an objective, quantitative manner and with excellent accuracy.

Noninvasive imaging assessment of portal hypertension

Portal hypertension (PH) is a spectrum of complications of chronic liver disease (CLD) and cirrhosis, with manifestations including ascites, gastroesophageal varices, splenomegaly, hypersplenism, hepatic hydrothorax, hepatorenal syndrome, hepatopulmonary syndrome and portopulmonary hypertension. PH can vary in severity and is diagnosed via invasive hepatic venous pressure gradient measurement (HVPG), which is considered the reference standard. Accurate diagnosis of PH and assessment of severity are highly relevant as patients with clinically significant portal hypertension (CSPH) are at higher risk for developing acute variceal bleeding and mortality. In this review, we discuss current and upcoming noninvasive imaging methods for diagnosis and assessment of severity of PH.

Imaging biomarkers of diffuse liver disease: current status

We are happy to introduce this special issue of Abdominal Radiology on "diffuse liver disease". We have invited imaging experts to discuss various topics pertaining to diffuse liver disease, covering a vast array of imaging techniques including ultrasound (US), CT, MRI and new molecular imaging agents. Below, we briefly discussed the current status, limitations, and future directions of imaging biomarkers of diffuse liver disease.

CT and MR perfusion techniques to assess diffuse liver disease

Perfusion imaging allows for the quantitative extraction of physiological perfusion parameters of the liver microcirculation at levels far below the spatial the resolution of CT and MR imaging. Because of its peculiar structure and architecture, perfusion imaging is more challenging in the liver than in other organs. Indeed, the liver is a mobile organ and significantly deforms with respiratory motion. Moreover, it has a dual vascular supply and the sinusoidal capillaries are fenestrated in the normal liver. Using extracellular contrast agents, perfusion imaging has shown its ability to discriminate patients with various stages of liver fibrosis. The recent introduction of hepatobiliary contrast agents enables quantification of both the liver perfusion and the hepatocyte transport function using advanced perfusion models. The purpose of this review article is to describe the characteristics of liver perfusion imaging to assess chronic liver disease, with a special focus on CT and MR imaging.