Machine learning-based radiomics analysis of preoperative functional liver reserve with MRI and CT image.

OBJECTIVE: The indocyanine green retention rate at 15 min (ICG-R15) is a useful tool to evaluate the functional liver reserve before hepatectomy for liver cancer. Taking ICG-R15 as criteria, we investigated the ability of a machine learning (ML)-based radiomics model produced by Gd-EOB-DTPA-enhanced hepatic magnetic resonance imaging (MRI) or contrast-enhanced computed tomography (CT) image in evaluating functional liver reserve of hepatocellular carcinoma (HCC) patients. METHODS: A total of 190 HCC patients with CT, among whom 112 also with MR, were retrospectively enrolled and randomly classified into a training dataset (CT: n = 133, MR: n = 78) and a test dataset (CT: n = 57, MR: n = 34). Then, radiomics features from Gd-EOB-DTPA MRI and CT images were extracted. The features associated with the ICG-R15 classification were selected. Five ML classifiers were used for the ML-model investigation. The accuracy (ACC) and the area under curve (AUC) of receiver operating characteristic (ROC) with 95% confidence intervals (CI) were utilized for ML-model performance evaluation. RESULTS: A total of 107 different radiomics features were extracted from MRI and CT, respectively. The features related to ICG-R15 which was classified into 10%, 20% and 30% were selected. In MRI groups, classifier XGBoost performed best with its AUC = 0.917 and ACC = 0.882 when the threshold was set as ICG-R15 = 10%. When ICG-R15 = 20%, classifier Random Forest performed best with AUC = 0.979 and ACC = 0.882. When ICG-R15 = 30%, classifier XGBoost performed best with AUC = 0.961 and ACC = 0.941. For CT groups, the classifier XGBoost performed best when ICG-R15 = 10% with AUC = 0.822 and ACC = 0.842. When ICG-R15 = 20%, classifier SVM performed best with AUC = 0.860 and ACC = 0.842. When ICG-R15 = 30%, classifier XGBoost performed best with AUC = 0.938 and ACC = 0.965. CONCLUSIONS: Both the MRI- and CT-based machine learning models are proved to be valuable noninvasive methods for functional liver reserve evaluation.

A Bibliometric Analysis of the Top 100 Cited Articles on Hepatic Magnetic Resonance Imaging.

The purpose of this study is to guide the readers to the impact of the articles published on hepatic magnetic resonance imaging (MRI). We searched Scopus using 10 different search terms for hepatic MRI. The selected studies were thoroughly reviewed by two independent authors and any disagreement was sorted out by mutual consensus. The list of articles and journals was downloaded into an excel spreadsheet. Only the top 100 cited articles were selected by mutual consensus among all the authors. These articles were further read in the full-text form and were further categorized into subgroups. Three authors independently reviewed the top 100 selected articles, and subsequently data was extracted from them and analyzed. Our study showed that the highest number of top 100 cited articles on hepatic MRI were from Radiology (30 articles) followed by European Radiology (14 articles). The American Journal of Roentgenology, Radiographics, and Journal of Magnetic Resonance had seven articles each. The United States had the highest number of articles by region. Nineteen other journals contributed only one article each to the list of top 100 cited articles. The contribution of authors to the top 100 cited articles was reviewed; all the authors contributing with more than two articles to the highly cited articles are given in Table 3 in the supplementary material. The maximum number of articles were published during 2009 (14 articles), and for a five-year period, the maximum contribution was made during 2008-2013 (44 articles). Our analysis gives an insight on the frequency of citations of top articles on hepatic MRI, categorizes the subtopics, the timeline of the publications, and contributions from different geographic distributions.

Efficacy of the projection onto convex sets (POCS) algorithm at Gd-EOB-DTPA-enhanced hepatobiliary-phase hepatic MRI.

PURPOSE: To investigate the efficacy of the projection onto convex sets (POCS) algorithm at Gd-EOB-DTPA-enhanced hepatobiliary-phase MRI. METHODS: In phantom study, we scanned a phantom and obtained images by conventional means (P1 images), by partial-Fourier image reconstruction (PF, P2 images) and by PF with the POCS algorithm (P3 images). Then we acquired and compared subtraction images (P2-P1 images and P3-P1 images). In clinical study, 55 consecutive patients underwent Gd-EOB-DTPA (EOB)-enhanced 3D hepatobiliary-phase MRI on a 1.5T scanner. Images were obtained using conventional method (C1 images), PF (C2 images), and PF with POCS (C3 images). The acquisition time was 17-, 14-, and 14 s for protocols C1, C2 and C3, respectively. Two radiologists assigned grades for hepatic vessel sharpness and we compared the visual grading among the 3 protocols. And one radiologist compared signal-to-noise-ratio (SNR) of the hepatic parenchyma. RESULTS: In phantom study, there was no difference in signal intensity on a peripheral phantom column on P3-P1 images. In clinical study, there was no significant difference between C1 and C3 images (2.62 ± 0.49 vs. 2.58 ± 0.49, p = 0.70) in the score assigned for vessel sharpness nor in SNR (13.3 ± 2.67 vs. 13.1 ± 2.51, p = 0.18). CONCLUSION: The POCS algorithm makes it possible to reduce the scan time of hepatobiliary phase (from 17 to 14 s) without reducing SNR and without increasing artifacts.

[Diagnosis of an increased serum level of ferritin]. / Démarche diagnostique devant une hyperferritinémie.

The discovery of a hyperferritinemia is most of the time fortuitous. The diagnostic approach aims at looking for the responsible etiology and at verifying if an iron hepatic overload is present or not. Three diagnostic steps are proposed. The clinical elements and a few straightforward biological tests are sufficient at first to identify one of the four main causes: alcoholism, inflammatory syndrome, cytolysis, and metabolic syndrome. None of these causes is associated with a significant iron hepatic overload. If the transferring saturation coefficient is raised (>50%) a hereditary hemochromatosis should be discussed. Secondly, less common disorders will be discussed. Among these, only the chronic hematological disorders either acquired or congenital are at risk of iron hepatic overload. Thirdly, if a doubt persists in the etiologic research, and the serum ferritin level is very high or continues to rise, it is essential to verify that there is no iron hepatic overload. For that purpose, the MRI with study of the iron overload is the main test, which will guide the therapeutic attitude. Identification of more than a single etiology occurs in more than 40% of the cases.

An unusual case of primary hepatic lymphoma mimicking sarcoidosis in MRI.

Sarcoidosis is a granulomatous disease, in which liver affection is common, contrary to a primary hepatic lymphoma that is very rarely seen. On MRI both present with almost the same imaging features: hypointense in T1-weighted and hyperintense in T2-weighted sequences. Our patient with a histologically confirmed sarcoidosis in the lungs showed liver lesions that were similar to sarcoidosis manifestations of the liver. Due to size, progression and overlapping features with secondary malignant liver lesions within an interval of 5 months, a biopsy was conducted and confirmed a primary hepatic lymphoma with diffuse large b-cells. Thus, we would recommend performing a biopsy in ambiguous lesions with indistinguishable characteristics and progression within a short follow-up interval.

Hepatic MR imaging techniques, optimization, and artifacts.

This article describes a basic 1.5-T hepatic magnetic resonance (MR) imaging protocol, strategies for optimizing pulse sequences while managing artifacts, the proper timing of postgadolinium 3-dimensional gradient echo sequences, and an effective order of performing pulse sequences with the goal of creating an efficient and high-quality hepatic MR imaging examination. The authors have implemented this general approach on General Electric, Philips, and Siemens clinical scanners.

MRI monitoring of myocardial iron overload: use of cardiac MRI combined with hepatic MRI in a cohort of multi-transfused patients with thalassaemia.

PURPOSE: We report the results of combining cardiac and hepatic MRI in the same examination to monitor 48 multi-transfused patients presenting iron overload secondary to their transfusions. This cardiac MRI technique uses acquisition sequences and calculation software that are readily available for 1.5 T systems, and it has been validated to screen for patients at risk of cardiac complications who present myocardial iron overload (T2*<20milliseconds). PATIENTS AND METHODS: A total of 176 combined MRI examinations were performed between May 2006 and January 2012 in 48 patients who had received transfusions due to thalassaemia. This monocentric retrospective study brings together all of the imaging examinations carried out. RESULTS: There was a positive correlation between the cardiac T2* values and left ventricular ejection fraction, which were measured in the same examination. At the first assessment 23/48 patients had a T2*<20ms. These patients showed a significant improvement in cardiac T2* over time while their iron chelation therapy was being intensified. CONCLUSION: This study validates the application of the cardiac MRI technique used to monitor cardiac iron overload in patients who have undergone multiple transfusions.

Morphine three-dimensional T1 gadoxetate MR cholangiography of potential living related liver donors.

PURPOSE: To assess low-dose morphine for distension and improved visualization of intrahepatic bile ducts on T1 MR cholangiography (MRC) in preoperative imaging of potential liver donors. MATERIALS AND METHODS: Sixty-nine consecutive potential living related liver donors (mean age, 39 years; age range, 20 to 59 years) referred for pre-transplant MRI evaluation were evaluated without (n=30) or with (n=39) intravenous morphine injection (0.04 mg/kg). Morphine was injected pre-MRI while establishing intravenous access to allow ∼1 h for biliary distension before T1 MRC. Three radiologists reviewed intrahepatic biliary branch order visualization, common bile duct (CBD) diameter, and overall image quality. In 25 patients undergoing liver donation surgery, T1 MRC findings were correlated with intraoperative findings. This retrospective study was approved by the institutional review board. RESULTS: Biliary visualization was improved post-morphine administration with biliary duct branch order visualization score of 3.2 and 3.3 at 45 and 60 min, respectively, compared with 2.7 without morphine (P<0.002); CBD diameter measured 5.3 and 5.5 versus 4.1 mm (P<0.005), and overall image quality score was 2.4 and 2.6 versus 1.8 (P<0.0006). Operative notes confirmed T1 MRC findings in 6/11 donors without morphine and 14/14 donors with morphine. CONCLUSION: Intravenous low-dose morphine distends and improves visualization of bile ducts on T1 gadoxetate MRC.