MRI Extracellular Volume Fraction in Liver Fibrosis-A Comparison of Different Time Points and Blood Pool Measurements.

BACKGROUND: Extracellular volume (ECV) correlates with the degree of liver fibrosis. PURPOSE: To analyze the performance of liver MRI-based ECV evaluations with different blood pool measurements at different time points. STUDY TYPE: Prospective. SAMPLE: 73 consecutive patients (n = 31 females, mean age 56 years) with histopathology-proven liver fibrosis. FIELD STRENGTH/SEQUENCE: 3T acquisition within 90 days of biopsy, including shortened modified look-locker inversion recovery T1 mapping. ASSESSMENT: Polygonal regions of interest were manually drawn in the liver, aorta, vena cava, and in the main, left and right portal vein on four slices before and after Gd-DOTA administration at 5/10/15 minutes. ECV was calculated 1) on one single slice on portal bifurcation level, and 2) averaged over all four slices. STATISTICAL TESTS: Parameters were compared between patients with fibrosis grades F0-2 and F3-F4 with the Mann-Whitney U and fishers exact test. ROC analysis was used to assess the performance of the parameters to predict F3-4 fibrosis. A P-value <0.05 was considered statistically significant. RESULTS: ECV was significantly higher in F3-4 fibrosis (35.4% [33.1%-37.6%], 36.1% [34.2%-37.5%], and 37.0% [34.8%-39.2%] at 5/10/15 minutes) than in patients with F0-2 fibrosis (33.3% [30.8%-34.8%], 33.7% [31.6%-34.7%] and 34.9% [32.2%-36.0%]; AUC = 0.72-0.75). Blood pool T1 relaxation times in the aorta and vena cava were longer on the upper vs. lower slices at 5 minutes, but not at 10/15 minutes. AUC values were similar when measured on a single slice (AUC = 0.69-0.72) or based on blood pool measurements in the cava or portal vein (AUC = 0.63-0.67 and AUC = 0.65-0.70). DATA CONCLUSION: Liver ECV is significantly higher in F3-4 fibrosis compared to F0-2 fibrosis with blood pool measurements performed in the aorta, inferior vena cava, and portal vein at 5, 10, and 15 minutes. However, a smaller variability was observed for blood pool measurements between slices at 15 minutes. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY: Stage 3.

MSCs mediate long-term efficacy in a Crohn's disease model by sustained anti-inflammatory macrophage programming via efferocytosis.

Mesenchymal stem cells (MSCs) are novel therapeutics for the treatment of Crohn's disease. However, their mechanism of action is unclear, especially in disease-relevant chronic models of inflammation. Thus, we used SAMP-1/YitFc (SAMP), a chronic and spontaneous murine model of small intestinal inflammation, to study the therapeutic effects and mechanism of action of human bone marrow-derived MSCs (hMSC). hMSC dose-dependently inhibited naïve T lymphocyte proliferation via prostaglandin E2 (PGE2) secretion and reprogrammed macrophages to an anti-inflammatory phenotype. We found that the hMSCs promoted mucosal healing and immunologic response early after administration in SAMP when live hMSCs are present (until day 9) and resulted in a complete response characterized by mucosal, histological, immunologic, and radiological healing by day 28 when no live hMSCs are present. hMSCs mediate their effect via modulation of T cells and macrophages in the mesentery and mesenteric lymph nodes (mLN). Sc-RNAseq confirmed the anti-inflammatory phenotype of macrophages and identified macrophage efferocytosis of apoptotic hMSCs as a mechanism that explains their long-term efficacy. Taken together, our findings show that hMSCs result in healing and tissue regeneration in a chronic model of small intestinal inflammation and despite being short-lived, exert long-term effects via sustained anti-inflammatory programming of macrophages via efferocytosis.

Automated liver segmental volume ratio quantification on non-contrast T1-Vibe Dixon liver MRI using deep learning.

PURPOSE: To evaluate the effectiveness of automated liver segmental volume quantification and calculation of the liver segmental volume ratio (LSVR) on a non-contrast T1-vibe Dixon liver MRI sequence using a deep learning segmentation pipeline. METHOD: A dataset of 200 liver MRI with a non-contrast 3 mm T1-vibe Dixon sequence was manually labeledslice-by-sliceby an expert for Couinaud liver segments, while portal and hepatic veins were labeled separately. A convolutional neural networkwas trainedusing 170 liver MRI for training and 30 for evaluation. Liver segmental volumes without liver vessels were retrieved and LSVR was calculated as the liver segmental volumes I-III divided by the liver segmental volumes IV-VIII. LSVR was compared with the expert manual LSVR calculation and the LSVR calculated on CT scans in 30 patients with CT and MRI within 6 months. RESULTS: Theconvolutional neural networkclassified the Couinaud segments I-VIII with an average Dice score of 0.770 ± 0.03, ranging between 0.726 ± 0.13 (segment IVb) and 0.810 ± 0.09 (segment V). The calculated mean LSVR with liver MRI unseen by the model was 0.32 ± 0.14, as compared with manually quantified LSVR of 0.33 ± 0.15, resulting in a mean absolute error (MAE) of 0.02. A comparable LSVR of 0.35 ± 0.14 with a MAE of 0.04 resulted with the LSRV retrieved from the CT scans. The automated LSVR showed significant correlation with the manual MRI LSVR (Spearman r = 0.97, p < 0.001) and CT LSVR (Spearman r = 0.95, p < 0.001). CONCLUSIONS: A convolutional neural network allowed for accurate automated liver segmental volume quantification and calculation of LSVR based on a non-contrast T1-vibe Dixon sequence.

Mesenchymal stem cells ameliorate inflammation in an experimental model of Crohn's disease via the mesentery.

Objective: Mesenchymal stem cells (MSCs) are novel therapeutics for treatment of Crohn's disease. However, their mechanism of action is unclear, especially in disease-relevant chronic models of inflammation. Thus, we used SAMP-1/YitFc, a chronic and spontaneous murine model of small intestinal inflammation, to study the therapeutic effect and mechanism of human bone marrow-derived MSCs (hMSC). Design: hMSC immunosuppressive potential was evaluated through in vitro mixed lymphocyte reaction, ELISA, macrophage co-culture, and RT-qPCR. Therapeutic efficacy and mechanism in SAMP were studied by stereomicroscopy, histopathology, MRI radiomics, flow cytometry, RT-qPCR, small animal imaging, and single-cell RNA sequencing (Sc-RNAseq). Results: hMSC dose-dependently inhibited naïve T lymphocyte proliferation in MLR via PGE 2 secretion and reprogrammed macrophages to an anti-inflammatory phenotype. hMSC promoted mucosal healing and immunologic response early after administration in SAMP model of chronic small intestinal inflammation when live hMSCs are present (until day 9) and resulted in complete response characterized by mucosal, histological, immunologic, and radiological healing by day 28 when no live hMSCs are present. hMSC mediate their effect via modulation of T cells and macrophages in the mesentery and mesenteric lymph nodes (mLN). Sc-RNAseq confirmed the anti-inflammatory phenotype of macrophages and identified macrophage efferocytosis of apoptotic hMSCs as a mechanism of action that explains their long-term efficacy. Conclusion: hMSCs result in healing and tissue regeneration in a chronic model of small intestinal inflammation. Despite being short-lived, exert long-term effects via macrophage reprogramming to an anti-inflammatory phenotype. Data Transparency Statement: Single-cell RNA transcriptome datasets are deposited in an online open access repository 'Figshare' (DOI: https://doi.org/10.6084/m9.figshare.21453936.v1 ).

Computer-aided diagnosis through medical image retrieval in radiology.

Currently, radiologists face an excessive workload, which leads to high levels of fatigue, and consequently, to undesired diagnosis mistakes. Decision support systems can be used to prioritize and help radiologists making quicker decisions. In this sense, medical content-based image retrieval systems can be of extreme utility by providing well-curated similar examples. Nonetheless, most medical content-based image retrieval systems work by finding the most similar image, which is not equivalent to finding the most similar image in terms of disease and its severity. Here, we propose an interpretability-driven and an attention-driven medical image retrieval system. We conducted experiments in a large and publicly available dataset of chest radiographs with structured labels derived from free-text radiology reports (MIMIC-CXR-JPG). We evaluated the methods on two common conditions: pleural effusion and (potential) pneumonia. As ground-truth to perform the evaluation, query/test and catalogue images were classified and ordered by an experienced board-certified radiologist. For a profound and complete evaluation, additional radiologists also provided their rankings, which allowed us to infer inter-rater variability, and yield qualitative performance levels. Based on our ground-truth ranking, we also quantitatively evaluated the proposed approaches by computing the normalized Discounted Cumulative Gain (nDCG). We found that the Interpretability-guided approach outperforms the other state-of-the-art approaches and shows the best agreement with the most experienced radiologist. Furthermore, its performance lies within the observed inter-rater variability.

Convolutional neural network for automated segmentation of the liver and its vessels on non-contrast T1 vibe Dixon acquisitions.

We evaluated the effectiveness of automated segmentation of the liver and its vessels with a convolutional neural network on non-contrast T1 vibe Dixon acquisitions. A dataset of non-contrast T1 vibe Dixon liver magnetic resonance images was labelled slice-by-slice for the outer liver border, portal, and hepatic veins by an expert. A 3D U-Net convolutional neural network was trained with different combinations of Dixon in-phase, opposed-phase, water, and fat reconstructions. The neural network trained with the single-modal in-phase reconstructions achieved a high performance for liver parenchyma (Dice 0.936 ± 0.02), portal veins (0.634 ± 0.09), and hepatic veins (0.532 ± 0.12) segmentation. No benefit of using multi-modal input was observed (p = 1.0 for all experiments), combining in-phase, opposed-phase, fat, and water reconstruction. Accuracy for differentiation between portal and hepatic veins was 99% for portal veins and 97% for hepatic veins in the central region and slightly lower in the peripheral region (91% for portal veins, 80% for hepatic veins). In conclusion, deep learning-based automated segmentation of the liver and its vessels on non-contrast T1 vibe Dixon was highly effective. The single-modal in-phase input achieved the best performance in segmentation and differentiation between portal and hepatic veins.

T1 reduction rate with Gd-EOB-DTPA determines liver function on both 1.5 T and 3 T MRI.

Magnetic resonance T1 mapping before and after Gd-EOB-DTPA administration allows quantification of the T1 reduction rate as a non-invasive surrogate marker of liver function. A major limitation of T1 relaxation time measurement is its dependency on MRI field strengths. Since T1 reduction rate is calculated as the relative shortening of T1 relaxation time before and after contrast administration, we hypothesized that the T1 reduction rate is comparable between 1.5 and 3 T. We thus compared liver T1 relaxation times between 1.5 and 3 T in a total of 243 consecutive patients (124, 1.5 T and 119, 3 T) between 09/2018 and 07/2019. T1 reduction rates were compared between patients with no cirrhosis and patients with cirrhosis Child-Pugh A-C. There was no significant difference of T1 reduction rate between 1.5 and 3 T in any patient group (p-value 0.126-0.861). On both 1.5 T and 3 T, T1 reduction rate allowed to differentiate between patients with no cirrhosis and patients with liver cirrhosis Child A-C (p < 0.001). T1 reduction rate showed a good performance to predict liver cirrhosis Child A (AUC = 0.83, p < 0.001), Child B (AUC = 0.83, p < 0.001) and Child C (AUC = 0.92, p < 0.001). In conclusion, T1 reduction rate allows to determine liver function on Gd-EOB-DTPA MRI with comparable values on 1.5 T and 3 T.

Noninvasive assessment of clinically significant portal hypertension using ΔT1 of the liver and spleen and ECV of the spleen on routine Gd-EOB-DTPA liver MRI.

PURPOSE: To analyze the predictive value of ΔT1 of the liver and spleen as well as the extracellular volume fraction (ECV) of the spleen as noninvasive biomarkers for the determination of clinically significant portal hypertension (CSPH) on routine Gd-EOB-DTPA liver MRI. METHOD: 195 consecutive patients with known or suspected chronic liver disease from 9/2018 to 7/2019 with Gd-EOB-DTPA liver MRI and abdominal T1 mapping were retrospectively included. Based on the presence of splenomegaly with thrombocytopenia, ascites and portosystemic collaterals, the patients were divided into noCSPH (n = 113), compensated CSPH (cCSPH, ≥1 finding without ascites; n = 55) and decompensated CSPH (dCSPH, ascites ± other findings; n = 27). T1 times were measured in the liver, spleen and abdominal aorta in the unenhanced and contrast-enhanced T1 maps. Native T1 times and ΔT1 of the liver and spleen as well as ECV of the spleen were compared between groups using the Kruskal-Wallis test with Dunn's post hoc test. Furthermore, cutoff values for group differentiation were calculated using ROC analysis with Youden's index. RESULTS: ΔT1 of the liver was significantly lower in patients with cCSPH and dCSPH (p < 0.001) compared to patients with noCSPH. In the ROC analyses for differentiation between noCSPH and CSPH (cCSPH + dCSPH), a cutoff of < 0.67 for ΔT1 of the liver (AUC = 0.79) performed better than ΔT1 (AUC = 0.69) and ECV (AUC = 0.63) of the spleen with cutoffs of > 0.29 and > 41.9, respectively. CONCLUSION: ΔT1 of the liver and spleen in addition to ECV of the spleen allow for determination of CSPH on routine Gd-EOB-DTPA liver MRI.

Cystic Hepatic Neuroendocrine Tumor: A Rare Entity.

ABSTRACT: A 70-year-old man presented with unspecific abdominal symptoms and weight loss was referred for a sonographic examination. Sonography revealed 3 cystic hepatic masses in an otherwise unremarkable liver. Contrast-enhanced MRI of the liver was performed to characterize the hepatic lesions and elucidate their etiology. The differential diagnosis was primarily parasitic disease or metastases with cystic transformations. 68Ga-DOTATOC PET/CT revealed the neuroendocrine origin of these lesions, confirmed by biopsy. However, the primary site of the neuroendocrine tumor remained unclear, leaving primary hepatic neuroendocrine tumor and neuroendocrine cancer of unknown primary as possible diagnostic options.

Correlation of gastrointestinal perforation location and amount of free air and ascites on CT imaging.

PURPOSE: To analyze the amount of free abdominal gas and ascites on computed tomography (CT) images relative to the location of a perforation. METHODS: We retrospectively included 172 consecutive patients (93:79 = m:f) with GIT perforation, who underwent abdominal surgery (ground truth for perforation location). The volume of free air and ascites were quantified on CT images by 4 radiologists and a semiautomated software. The relation of the perforation location (upper/lower GIT) and amount of free air and ascites was analyzed by the Mann-Whitney test. Furthermore, best volume cutoff for upper and lower GIT perforation, areas under the curve (AUC), and interreader volume agreement were assessed. RESULTS: There was significantly more abdominal ascites with upper GIT perforation (333 ml, range 5 to 2000 ml) than with lower GIT perforation (100 ml, range 5 to 2000 ml, p = 0.022). The highest volume of free air was found with perforations of the stomach, descending colon and sigmoid colon. Significantly less free air was found with perforations of the small bowel and ascending colon compared to the aforementioned. An ascites volume > 333 ml was associated with an upper GIT perforation demonstrating an AUC of 0.63 ± 0.04. CONCLUSION: Using a two-step process based on the volumes of free air and free fluid can help localizing the site of perforation to the upper, middle or lower GI tract.

Diagnosis and staging of hepatobiliary malignancies: Potential incremental value of (18)F-FDG-PET/MRI compared to MRI of the liver.

OBJECTIVE: The purpose of the study was to investigate the potential added value of 18F-FDG-PET/MRI (functional information derived from PET) over standard diagnostic liver MRI (excellent soft tissue characterization) in diagnosing and staging suspected primary hepatobiliary malignancies including extrahepatic cholangiocarcinoma (ECC), intrahepatic cholangiocellular carcinoma (ICC) and gallbladder cancer (GBCA). METHODS: Twenty consecutive patients with suspected hepatobiliary malignancy were included in this retrospective study. All patients underwent combined whole-body (WB) 18F-FDG-PET/MRI including contrast-enhanced MRI of the liver, contrast-enhanced WB-MRI and WB 18F-FDG-PET. Two experienced readers staged hepatobiliary disease using TNM criteria: first based on MRI alone and then based on combined 18F-FDG-PET/MRI. Subsequently, the impact of FDG-PET/MRI on clinical management compared to MRI alone was recorded. Histopathologic proof served as the reference standard. RESULTS: Hepatobiliary neoplasms were present in 16/20 patients (ECC n = 3, ICC n = 8, GBCA n = 5), two patients revealed benign disease, two were excluded. TNM staging with 18F-FDG-PET/MRI was identical to MRI alone in 11/18 (61.1 %) patients and correctly changed the stage in 4/18 (22.2 %), resulting in a change in management for 2/4 patients (11.1 %). 18F-FDG-PET/MRI was false-positive in 3/18 cases (16.7 %). Both MRI and 18F-FDG-PET/MRI were falsely positive in 1 case without malignancy. CONCLUSIONS: A small incremental benefit of 18F-FDG-PET/MRI over standard MRI of the liver was observed. However, in some cases 18F-FDG-PET/MRI may lead to false-positive findings. Overall there is seemingly limited role of 18F-FDG-PET/MRI in patients with suspected hepatobiliary malignancy.

CT artifacts after contrast media injection in chest imaging: evaluation of post-processing algorithms, virtual monoenergetic images and their combination for artifact reduction.

BACKGROUND: After injection into a brachial vein, high contrast media concentration in axillary and subclavian veins can cause artifacts that impair diagnostic utility. This study assessed artifact reduction by artifact-reduction-algorithms (ARA) and virtual-monoenergetic-images (VMI), as well as their combination (VMIARA) compared to conventional CT-images (CI). METHODS: Forty-six spectral-detector-CT (SDCT) examinations of patients that received ARA-reconstructions due to perivenous-artifacts were included in this retrospective study. CI, ARA, VMI, and VMIARA (range: 70-200 keV) were reconstructed. Objective analysis was performed with ROI-based assessment of mean and standard deviation of attenuation (HU) in hypo- and hyperdense artifacts and impaired muscle and arteries as well as artifact-free reference-tissue. Extent of artifact reduction, assessment of surrounding soft tissue and vessels, and appearance of new artifacts were rated visually by two radiologists. RESULTS: Hypo- and hyperdense artifacts showed significant improvement as evidenced by decreasing attenuation differences between artifact impaired and artifact-free reference tissue in ARA, VMI ≥80 keV, and VMIARA between 70-200 keV (e.g., CI/ARA/VMI100keV/VMIARA100keV: hypodense artifacts, (-)264.8±150.9/(-)87.1±78.9/(-)48.6±64.6/9.9±63.9 HU; P<0.001); hyperdense artifacts, 164.2±51.1/82.1±73.2/7.9±34.7/(-)17.3±50.7 HU; P<0.001). Artifacts impairing surrounding muscle and arteries were also reduced by all three approaches. In visual assessment, ARA, VMI ≥100 keV, and VMIARA between 70-200 keV also showed significant artifact reduction and improved assessment; however, for assessment of arteries improvement was not significant using ARA alone. New artifacts were reported, particularly at higher keV-values. CONCLUSIONS: In presence of perivenous-artifacts, ARA, VMI and their combination allow for significant artifact reduction; however, their combination and VMI as a standalone approach yielded best results and should therefore be used, if available.

T1 mapping of the liver and the spleen in patients with liver fibrosis-does normalization to the blood pool increase the predictive value?

PURPOSE: To analyze whether the T1 relaxation time of the liver is a good predictor of significant liver fibrosis and whether normalization to the blood pool improves the predictive value. METHODS: This prospective study was conducted between 03/2016 and 02/2018. One hundred seventy-three patients underwent multiparametric liver MRI at 3 T. The T1 relaxation time was measured in the liver and the spleen, in the aorta, the portal vein, and the inferior vena cava (IVC). T1 relaxation times with and without normalization to the blood pool were compared between patients with (n = 26) and without (n = 141) significant liver fibrosis, based on a cutoff value of 3.5 kPa in MRE as the noninvasive reference standard. For statistics, Student's t test, receiver operating characteristic (ROC) curve analysis, and Pearson's correlation were used. RESULTS: The T1 relaxation time of the liver was significantly longer in patients with liver fibrosis, both with and without blood pool normalization (p < 0.001). T1 relaxation time of the liver allowed prediction of significant liver fibrosis (AUC = 0.88), while normalization to the IVC resulted in a slightly lower performance (AUC = 0.82). The lowest performance was achieved when the T1 relaxation times of the liver were normalized to the aorta (AUC = 0.66) and to the portal vein (AUC = 0.62). The T1 relaxation time of the spleen detected significant liver fibrosis with an AUC of 0.68, and 0.51-0.64 with normalization to the blood pool. CONCLUSION: The T1 relaxation time of the liver is a good predictor of significant liver fibrosis. However, normalization of the blood pool did not improve the predictive value. KEY POINTS: • The T1 relaxation time of the liver is a good predictor of significant liver fibrosis. • Normalization to the blood pool did not improve the predictive value of T1 mapping. • If the blood pool normalization was weighted 30% to the aorta and 70% to the portal vein, the performance was better than normalization to the aorta alone but still lower than normalization to the IVC.

Metal artifacts from sternal wires: evaluation of virtual monoenergetic images from spectral-detector CT for artifact reduction.

PURPOSE: In chest imaging, sternal wires can cause metal artifacts that hamper depiction of surrounding soft tissue and bone. This study investigated if these artifacts may be reduced by means of virtual monoenergetic images (VMI) obtained from a novel detector-based spectral CT scanner (SDCT) in comparison to conventional CT images (CI). MATERIALS AND METHODS: 30 patients with clinically indicated SDCT scans of the chest exhibiting artifacts due to sternal wires were included in this IRB-approved study. CI and VMI (40-200 keV, 10 keV increment) were reconstructed. Quantitative image analysis was conducted by ROI-based measurement of attenuation (HU) and standard deviation within the most pronounced hypo- and hyperdense artifacts. Visually, artifact reduction and diagnostic assessment of surrounding soft tissue as well as sternal bone were independently rated by two radiologists on 5-point Likert-scales. RESULTS: In comparison to CI, high-keV VMI showed an effective reduction of hypo- and hyperattenuating parasternal artifacts, as corrected HU-values approximated their true expected values (hypodense: CI -66.2 ± 70.8; VMI200keV 2.4 ± 29.2; hyperdense: CI 156.7 ± 70.8HU; VMI200keV 76.9 ± 45.4, both p < 0.05). In addition, image noise was significantly lower in high-keV VMI compared to CI. Subjective analysis confirmed that VMI of ≥100 keV significantly reduced artifacts and improved diagnostic assessment of surrounding soft tissue and bone. Interrater-agreement was excellent (intraclass-correlation-coefficient = 0.83). CONCLUSIONS: High-keV VMI yielded a significant reduction of artifacts from sternal wires and improved assessment of surrounding structures.

Virtual non-contrast for evaluation of liver parenchyma and vessels: results from 25 patients using multi-phase spectral-detector CT.

BACKGROUND: In abdominal imaging, contrast-enhanced computed tomography (CT) examinations are most commonly applied; however, unenhanced examinations are still needed for several clinical questions but require additional scanning and radiation exposure. PURPOSE: To evaluate accuracy of virtual non-contrast (VNC) from arterial and venous phase spectral-detector CT (SDCT) scans compared to true-unenhanced (TNC) images for the evaluation of liver parenchyma and vessels. MATERIAL AND METHODS: A total of 25 patients undergoing triphasic SDCT examinations were included. VNC was reconstructed from arterial and venous phases and compared to TNC images. Quantitative image analysis was performed by region of interest (ROI)-based assessment of mean and SD of attenuation (HU) in each liver segment, spleen, portal vein, common hepatic artery, and abdominal aorta. Subjectively, iodine subtraction and diagnostic assessment were rated on 5-point Likert scales. RESULTS: Attenuation and image noise measured in the liver from VNC were not significantly different from TNC (TNC: 54.6 ± 10.8 HU, VNC arterial phase: 55.7 ± 10.8 HU; VNC venous phase: 58.3 ± 10.0 HU; P > 0.05). VNC also showed accurate results regarding attenuation and image noise for spleen, portal vein, and abdominal aorta. Only iodine subtraction in the common hepatic artery in the arterial phase was insufficient which was confirmed by the subjective reading. Apart from that, subjective reading showed accurate iodine subtraction and comparable diagnostic assessment. CONCLUSION: VNC from the arterial and venous phases were very similar to TNC yielding mostly negligible differences in attenuation, image noise, and diagnostic utility. Inadequate iodine subtraction occurred in hepatic arteries in the arterial phase.