Comparison of Radiologists and Deep Learning for US Grading of Hepatic Steatosis.

Background Screening for nonalcoholic fatty liver disease (NAFLD) is suboptimal due to the subjective interpretation of US images. Purpose To evaluate the agreement and diagnostic performance of radiologists and a deep learning model in grading hepatic steatosis in NAFLD at US, with biopsy as the reference standard. Materials and Methods This retrospective study included patients with NAFLD and control patients without hepatic steatosis who underwent abdominal US and contemporaneous liver biopsy from September 2010 to October 2019. Six readers visually graded steatosis on US images twice, 2 weeks apart. Reader agreement was assessed with use of κ statistics. Three deep learning techniques applied to B-mode US images were used to classify dichotomized steatosis grades. Classification performance of human radiologists and the deep learning model for dichotomized steatosis grades (S0, S1, S2, and S3) was assessed with area under the receiver operating characteristic curve (AUC) on a separate test set. Results The study included 199 patients (mean age, 53 years ± 13 [SD]; 101 men). On the test set (n = 52), radiologists had fair interreader agreement (0.34 [95% CI: 0.31, 0.37]) for classifying steatosis grades S0 versus S1 or higher, while AUCs were between 0.49 and 0.84 for radiologists and 0.85 (95% CI: 0.83, 0.87) for the deep learning model. For S0 or S1 versus S2 or S3, radiologists had fair interreader agreement (0.30 [95% CI: 0.27, 0.33]), while AUCs were between 0.57 and 0.76 for radiologists and 0.73 (95% CI: 0.71, 0.75) for the deep learning model. For S2 or lower versus S3, radiologists had fair interreader agreement (0.37 [95% CI: 0.33, 0.40]), while AUCs were between 0.52 and 0.81 for radiologists and 0.67 (95% CI: 0.64, 0.69) for the deep learning model. Conclusion Deep learning approaches applied to B-mode US images provided comparable performance with human readers for detection and grading of hepatic steatosis. Published under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Tuthill in this issue.

Prospective comparison of transient, point shear wave, and magnetic resonance elastography for staging liver fibrosis.

OBJECTIVES: To perform head-to-head comparisons of the feasibility and diagnostic performance of transient elastography (TE), point shear-wave elastography (pSWE), and magnetic resonance elastography (MRE). METHODS: This prospective, cross-sectional, dual-center imaging study included 100 patients with known or suspected chronic liver disease caused by hepatitis B or C virus, nonalcoholic fatty liver disease, or autoimmune hepatitis identified between 2014 and 2018. Liver stiffness measured with the three elastographic techniques was obtained within 6 weeks of a liver biopsy. Confounding effects of inflammation and steatosis on association between fibrosis and liver stiffness were assessed. Obuchowski scores and AUCs for staging fibrosis were evaluated and the latter were compared using the DeLong method. RESULTS: TE, pSWE, and MRE were technically feasible and reliable in 92%, 79%, and 91% subjects, respectively. At univariate analysis, liver stiffness measured by all techniques increased with fibrosis stages and inflammation and decreased with steatosis. For classification of dichotomized fibrosis stages, the AUCs were significantly higher for distinguishing stages F0 vs. ≥ F1 with MRE than with TE (0.88 vs. 0.71; p < 0.05) or pSWE (0.88 vs. 0.73; p < 0.05), and for distinguishing stages ≤ F1 vs. ≥ F2 with MRE than with TE (0.85 vs. 0.75; p < 0.05). TE, pSWE, and MRE Obuchowski scores for staging fibrosis stages were respectively 0.89 (95% CI 0.85-0.93), 0.90 (95% CI 0.85-0.94), and 0.94 (95% CI 0.91-0.96). CONCLUSION: MRE provided a higher diagnostic performance than TE and pSWE for staging early stages of liver fibrosis. TRIAL REGISTRATION: NCT02044523 KEY POINTS: • The technical failure rate was similar between MRE and US-based elastography techniques. • Liver stiffness measured by MRE and US-based elastography techniques increased with fibrosis stages and inflammation and decreased with steatosis. • MRE provided a diagnostic accuracy higher than US-based elastography techniques for staging of early stages of histology-determined liver fibrosis.

LI-RADS Version 2018 Ancillary Features at MRI.

The Liver Imaging Reporting and Data System (LI-RADS) standardizes performance of liver imaging in patients at risk for hepatocellular carcinoma (HCC) as well as interpretation and reporting of the results. Developed by experts in liver imaging and supported by the American College of Radiology, LI-RADS assigns to observations categories that reflect the relative probability of benignity, HCC, or other malignancy. While category assignment is based mainly on major imaging features, ancillary features may be applied to improve detection and characterization, increase confidence, or adjust LI-RADS categories. Ancillary features are classified as favoring malignancy in general, HCC in particular, or benignity. Those favoring malignancy in general or HCC in particular may be used to upgrade by a maximum of one category up to LR-4; those favoring benignity may be used to downgrade by a maximum of one category. If there are conflicting ancillary features (ie, one or more favoring malignancy and one or more favoring benignity), the category should not be adjusted. Ancillary features may be seen at diagnostic CT, MRI performed with extracellular agents, or MRI performed with hepatobiliary agents, with the exception of one ancillary feature assessed at US. This article focuses on LI-RADS version 2018 ancillary features seen at MRI. Specific topics include rules for ancillary feature application; definitions, rationale, and illustrations with clinical MRI examples; summary of evidence and diagnostic performance; pitfalls; and future directions. ©RSNA, 2018.

LI-RADS for MR Imaging Diagnosis of Hepatocellular Carcinoma: Performance of Major and Ancillary Features.

Purpose To evaluate the performance of major features, ancillary features, and categories of Liver Imaging Reporting and Data System (LI-RADS) version 2014 at magnetic resonance (MR) imaging for the diagnosis of hepatocellular carcinoma (HCC). Materials and Methods This retrospective institutional review board-approved study included patients with liver MR imaging and at least one pathologically proved lesion. Between 2004 and 2016, 102 patients (275 observations including 113 HCCs) met inclusion criteria. Two radiologists independently assessed major and ancillary imaging features for each liver observation and assigned a LI-RADS category. Per-lesion estimates of diagnostic performance of major features, ancillary features, and LI-RADS categories were assessed by using generalized estimating equation models. Results Major features (arterial phase hyperenhancement, washout, capsule, and threshold growth) had a sensitivity of 88.5%, 60.6%, 32.9%, and 41.6%, and a specificity of 18.6%, 84.8%, 98.8%, and 83.2% for HCC, respectively. Ancillary features (mild-moderate T2 hyperintensity, restricted diffusion, mosaic architecture, intralesional fat, lesional fat sparing, blood products, and subthreshold growth) had a sensitivity of 62.2%, 54.8%, 9.9%, 30.9%, 23.1%, 2.8%, and 48.3%, and a specificity of 79.4%, 90.6%, 99.4%, 94.2%, 83.1%, 99.3%, and 91.4% for HCC, respectively. The LR-5 or LR-5 V categories had a per-lesion sensitivity of 50.8% and a specificity of 95.8% for HCC, respectively. The LR-4, LR-5, or LR-5 V categories (determined by using major features only vs combination of major and ancillary features) had a per-lesion sensitivity of 75.9% and 87.9% and a per-lesion specificity of 87.5% and 86.2%, respectively. Conclusion The use of ancillary features in combination with major features increases the sensitivity while preserving a high specificity for the diagnosis of HCC.

Dilatation of the bile duct in patients after cholecystectomy: a retrospective study.

PURPOSE: Retrospective assessment of impact of cholecystectomy, age, and sex on bile duct (BD) diameter. MATERIALS AND METHODS: We retrospectively reviewed abdominal contrast-enhanced multidetector computed tomography and laboratory reports of 290 consecutive patients (119 men; mean age, 55.9 years) who presented without cholestasis to the emergency department of our institution between June 2009 and August 2010. BD diameters were measured in 3 locations, by 2 independent observers, twice, at 1-month intervals. Reproducibility and agreement were evaluated by intraclass correlation coefficients and Bland-Altman analyses. The effects of cholecystectomy, age, and sex on BD diameter were analysed with linear mixed models. RESULTS: BD diameter inter-reader reproducibility and agreement were excellent at the level of the right hepatic artery (intraclass correlation coefficient, 0.94). Sixty-one patients (21.0%) had a history of cholecystectomy. Among them, the 95th percentile of BD diameters at hepatic artery level was 7.9 mm (<50 years) and 12.3 mm (≥50 years). Among those without cholecystectomy, BD diameter was 6.2 mm (<50 years) and 7.7 mm (≥50 years). Cholecystectomy was associated with significantly larger BD diameters in both age groups (P < .001). Older age was associated with larger BD diameters (P = .004). Sex had no impact on BD diameter (P = .842). CONCLUSION: Patients after cholecystectomy may present with an enlarged BD unrelated to cholestasis. The BD diameter increases with age. Clinicians should rely on cholecystectomy status, age, and laboratory results to determine needs of further investigation.

The Canadian Association of Radiologists guidelines for the prevention of contrast-induced nephropathy: a critical appraisal.

OBJECTIVE: The purpose of this study is to critically appraise the Canadian Association of Radiologists (CAR) guidelines on the prevention of contrast-induced nephropathy (CIN). METHODS: The Appraisal of Guidelines Research and Evaluation (AGREE) tool is a questionnaire that consists of 23 key items organized in 6 domains (scope and purpose, stakeholder involvement, rigor of development, clarity and presentation, applicability, editorial independence). Four radiologists read the guidelines and completed the questionnaire independently. To assess the quality of the evidence, the articles listed in the bibliography were reviewed, and the following data were collected and tabulated: the type of contrast, the administration route, and the level of evidence (Centre for Evidence Based Medicine, University of Oxford). RESULTS: The highest score was for scope and purpose (92%), whereas the lowest scores were for rigor of development (50%) and applicability (40%). The score for the remaining domains were the following: stakeholder involvement (59%), clarity and presentation (69%), and editorial independence (63%). Although the bibliography of the CAR guidelines list 46 articles, only 33 were deemed pertinent to support the recommendations related to risk stratification or risk reduction of CIN. Only 3 articles dealt specifically with intravenous injection of iodinated contrast. Four articles dealt with ionic contrast, and, in 17 references, the contrast type was not specified. The best evidence (level 1) was in support of risk-reduction recommendations, but, in 8 of the 9 articles cited, the route of administration studied was exclusively intra-arterial. CONCLUSION: It would be appropriate to revisit the topic of CIN and formulate new guidelines. A formal systematic review of the literature should be undertaken and the data extraction should specifically address contrast type and route of administration, as well as the applicability of any recommendations.

Fatty liver deposition and sparing: a pictorial review.

OBJECTIVE: Fatty liver deposition is a very common finding, but it has many atypical patterns of distribution that can represent diagnostic pitfalls. The purpose of this pictorial essay is to review different patterns of fatty liver deposition and sparing. METHODS: We searched our archive retrospectively, reviewed the literature, and identified six patterns of liver steatosis. RESULTS: Steatosis may be diffuse, geographic, focal, subcapsular, multifocal or perivascular. CONCLUSIONS: Previous knowledge of atypical patterns of steatosis distribution may prevent misdiagnosis of infiltrative disease or focal liver lesions. When an unusual form of fatty liver deposition is suspected on ultrasound or computed tomography, magnetic resonance imaging may be used to confirm the diagnosis.

A primer to common major gastrointestinal post-surgical anatomy on CT-a pictorial review.

The post-operative abdomen can be challenging and knowledge of normal post-operative anatomy is important for diagnosing complications. The aim of this pictorial essay is to describe a few selected common, major gastrointestinal surgeries, their clinical indications and depict their normal post-operative computed tomography (CT) appearance. This essay provides some clues to identify the surgeries, which can be helpful especially when surgical history is lacking: recognition of the organ(s) involved, determination of what was resected and familiarity with the type of anastomoses used.

Small and large bowel volvulus: Clues to early recognition and complications.

Small and large bowel volvulus are uncommon causes of bowel obstruction with nonspecific clinical manifestations which may delay the diagnosis and increase morbidity. Therefore, radiologists play an important role in promptly establishing the diagnosis, recognizing underlying congenital or acquired risk factors and detecting potentially life-threatening complications. Multidetector CT performed with intravenous contrast is currently the preferred modality for the evaluation of volvulus, which is best appreciated when imaging is perpendicular to the axis of bowel rotation, hence the benefit of multiplanar reformations. In this pictorial essay we review the pathophysiology of the different types of intestinal volvulus, discuss diagnostic criteria for prompt diagnosis of volvulus and emphasize early recognition of the complications.

Focal nodular hyperplasia and hepatic adenoma: differentiation with low-mechanical-index contrast-enhanced sonography.

OBJECTIVE: The purpose of our study was to determine the differentiating features of focal nodular hyperplasia (FNH) and hepatic adenoma on contrast-enhanced sonography. MATERIALS AND METHODS: Sixty-two patients who underwent contrast-enhanced sonography and were confirmed to have FNH (n = 43) or hepatic adenoma (n = 19) were assessed retrospectively for arterial phase enhancement, filling direction, stellate arteries, and portal phase enhancement. An algorithm was applied to these interpreted features to determine the contrast-enhanced sonography diagnosis. RESULTS: All lesions were hypervascular in the arterial phase. Centrifugal filling was more common in FNH (39 and 32 of 43, 91% and 74% [reader 1 and reader 2]) than in adenoma (3 and 3 of 19, 16%). Centripetal or mixed filling was more common in adenoma (16 and 16 of 19, 84%) than in FNH (4 and 11 of 43, 9% and 26%) (p < 0.001, kappa = 0.61). Stellate arteries characterized FNH (29 and 26 of 43, 67% and 60%) but not adenoma (3 and 2 of 19, 16% and 11%) (p < 0.001, kappa = 0.36). Sustained portal phase enhancement was more common in FNH (37 and 39 of 43, 86% and 91%) than in adenoma (9 and 12 of 19, 47% and 63%) (p < 0.02, kappa = 0.79). The sensitivity, specificity, positive predictive value, and negative predictive value of sonography for diagnosing FNH were 95% and 86%, 74% and 79%, 89% and 90%, and 88% and 71%, (reader 1 and reader 2, respectively). CONCLUSION: FNH is predicted on the basis of arterial phase centrifugal filling and stellate vascularity on contrast-enhanced sonography. Adenoma is less reliably predicted on the basis of centripetal or mixed filling without stellate vascularity. Sustained portal phase enhancement is more common in FNH than in adenoma but contributes less to the differentiation of these lesions.

Are metastases really hypovascular in the arterial phase? The perspective based on contrast-enhanced ultrasonography.

OBJECTIVE: The purpose of this study was to describe enhancement and vascularity characteristics of liver metastases on real-time low-mechanical index contrast-enhanced ultrasonography. METHODS: This retrospective study was approved for chart review by our Research Ethics Board. Informed consent was waived. Fifty metastases (colorectal [n = 28], neuroendocrine [n = 6], pancreatic [n = 6], melanoma [n = 3], and other [n = 7]) in 50 patients (38-84 years, 24 male and 26 female) were analyzed. Contrast-enhanced ultrasonography was performed after intravenous injection of a microbubble contrast agent. Two radiologists independently reviewed digital cine clips and static images for the arterial phase intensity and pattern of enhancement and the presence of dysmorphic vessels. Observations on wash-out included its presence and completeness. Disagreement was resolved by consensus. The interval to peak arterial enhancement and beginning of wash-out were determined. Reader agreement was estimated with the kappa statistic. RESULTS: All but 6 metastases (44/50 [88%]) showed arterial hypervascularity, with dysmorphic vessels in 21 (42%) of 50. The pattern of enhancement was rim in 21 (42%) of 50 and diffuse in 29 (58%) of 50. The time to peak arterial enhancement ranged from 8 to 27 seconds (mean, 15.1 seconds), and the beginning of wash-out ranged from 13 to 50 seconds (mean, 25.2 seconds). Although a thin margin of residual enhancement was seen in 27 (54%) of 50 lesions in the early wash-out phase, all lesions (50/50) showed uniform complete wash-out in the portal phase. CONCLUSIONS: Contrary to popular belief based on computed tomography and magnetic resonance imaging studies, most hepatic metastases, including those thought to be hypovascular, show transient arterial hypervascularity on contrast-enhanced ultrasonography, followed by rapid and complete wash-out initiated within the conventional arterial phase.