Search

The aim of this study was to use artificial intelligence (AI) to facilitate peer review for detection of missed suspicious liver lesions (SLLs) on CT pulmonary angiographic (CTPA) examinations.This retrospective study included 1 month of consecutive CTPA examinations from a multisite teleradiology practice. Visual classification (VC) software analyzed images for the presence (+) or absence (-) of SLLs (1 cm,20 Hounsfield units). Separately, a natural language processing (NLP) algorithm evaluated corresponding reports for description (+) of an SLL or lack thereof (-). Studies containing possible missed SLLs (VC+/NLP-) were reviewed by three abdominal radiologists in a two-step adjudication process to confirm if an SLL was missed by the interpreting radiologist. The number of VC+/NLP- cases, the number of images needing radiologist review, and the number of cases with confirmed missed SLLs were recorded. Interobserver agreement for SLLs was calculated for the radiologist readers.A total of 2,573 CTPA examinations were assessed, and 136 were classified as potentially containing missed SLLs (VC+/NLP-). After radiologist review, 13 cases with missed SLLs were confirmed, representing 0.5% of analyzed CT studies. Using AI, the ratio of CT studies requiring review to missed SLLs identified was 10:1; the ratio without the help of AI would be at least 66:1. Among the 136 cases reviewed by radiologists, interobserver agreement for SLLs was excellent (κ = 0.91).AI can accelerate meaningful peer review by rapidly assessing thousands of examinations to identify potentially clinically significant errors. Although radiologist involvement is necessary, the amount of effort required after initial AI screening is dramatically reduced.

Biliary malignancies include those arising from the intrahepatic and extrahepatic bile ducts as well as the gallbladder and hepatopancreatic ampulla of Vater. The majority of intrahepatic and extrahepatic malignancies are cholangiocarcinomas (CCAs). They arise owing to a complex interplay between the patient-specific genetic background and multiple risk factors and may occur in the liver (intrahepatic CCA), hilum (perihilar CCA), or extrahepatic bile ducts (distal CCA). Biliary-type adenocarcinoma constitutes the most common histologic type of ampullary and gallbladder malignancies. Its prognosis is poor and surgical resection is considered curative, so early detection is key, with multimodality imaging playing a central role in making the diagnosis. There are several risk factors for biliary malignancy as well as predisposing conditions that increase the risk; this review highlights the pertinent imaging features of these entities with histopathologic correlation. The predisposing factors are broken down into three major categories

Recognizing and organizing different series in an MRI examination is important both for clinical review and research, but it is poorly addressed by the current generation of picture archiving and communication systems (PACSs) and post-processing workstations. In this paper, we study the problem of using deep convolutional neural networks for automatic classification of abdominal MRI series to one of many series types. Our contributions are three-fold. First, we created a large abdominal MRI dataset containing 3717 MRI series including 188,665 individual images, derived from liver examinations. 30 different series types are represented in this dataset. The dataset was annotated by consensus readings from two radiologists. Both the MRIs and the annotations were made publicly available. Second, we proposed a 3D pyramid pooling network, which can elegantly handle abdominal MRI series with varied sizes of each dimension, and achieved state-of-the-art classification performance. Third, we performed the first ever comparison between the algorithm and the radiologists on an additional dataset and had several meaningful findings.

Exploring the genomic landscape of hepatocellular carcinoma (HCC) provides clues for therapeutic decision-making. Phosphatidylinositol-3 kinase (PI3K) signaling is one of the key pathways regulating HCC aggressiveness, and its genomic alterations have been correlated with sorafenib response. In this study, we aimed to predict somatic mutations of the PI3K signaling pathway in HCC samples through machine-learning-based radiomic analysis.HCC patients who underwent next-generation sequencing and preoperative contrast-enhanced CT were recruited from West China Hospital and The Cancer Genome Atlas for model training and validation, respectively. Radiomic features were extracted from volumes of interest (VOIs) covering the tumor (VOIA total of 132 HCC patients (mean age: 61.1 ± 14.7 years; 108 men) were enrolled. In the training set, the AUCs of radiomic signatures based on single CT phases were moderate (AUC 0.694-0.771). In the external validation set, the radiomic signature based on VOIMachine-learning-based radiomic analysis had potential for characterizing alterations of PI3K signaling in HCC and could help identify potential candidates for sorafenib treatment.

Microvascular invasion (MVI) is an important risk factor in hepatocellular carcinoma (HCC), but its diagnosis mandates postoperative histopathologic analysis. We aimed to develop and externally validate a predictive scoring system for MVI.From July 2015 to November 2020, consecutive patients underwent surgery for HCC with preoperative gadoxetate disodium (EOB)-enhanced MRI was retrospectively enrolled. All MR images were reviewed independently by two radiologists who were blinded to the outcomes. In the training centre, a radio-clinical MVI score was developed via logistic regression analysis against pathology. In the testing centre, areas under the receiver operating curve (AUCs) of the MVI score and other previous MVI schemes were compared. Overall survival (OS) and recurrence-free survival (RFS) were analysed by the Kaplan-Meier method with the log-rank test.A total of 417 patients were included, 195 (47%) with pathologically-confirmed MVI. The MVI score included: non-smooth tumour margin (odds ratio [OR] = 4.4), marked diffusion restriction (OR = 3.0), internal artery (OR = 3.0), hepatobiliary phase peritumoral hypointensity (OR = 2.5), tumour multifocality (OR = 1.6), and serum alpha-fetoprotein400 ng/mL (OR = 2.5). AUCs for the MVI score were 0.879 (training) and 0.800 (testing), significantly higher than those for other MVI schemes (testing AUCs: 0.648-0.684). Patients with model-predicted MVI had significantly shorter OS (median 61.0 months vs not reached, P .001) and RFS (median 13.0 months vs. 42.0 months, P .001) than those without.A preoperative MVI score integrating five EOB-MRI features and serum alpha-fetoprotein level could accurately predict MVI and postoperative survival in HCC. Therefore, this score may aid in individualized treatment decision making.

Liver Imaging Reporting and Data System (LI-RADS) classifies liver nodules from LR-1 to LR-5 based on risk for hepatocellular carcinoma (HCC). It is challenging to know the nature of the LR-3 and LR-4 lesions.To test our hypothesis that in patients with a definite HCC (LR-5) or treated HCC (LR-TR), a coexisting LR-3 or LR-4 lesion is more likely to represent HCC compared to patients without LR-5 or LR-TR lesions.We conducted a retrospective study including all adult patients who received liver transplantation in our institution from 1/1/2014 to 3/3/2020 who had any LR-3 or LR-4 lesion on pre-transplant MRI.Seventy-eight patients were included in the final cohort (115 LR-3 and LR-4 lesions total). When accompanied by LR-5 or LR-TR lesions, 41% (28/69) of LR-3 lesions were HCC compared to 12% (3/25) when not accompanied by LR-5 LR-TR lesions. When accompanied by LR-5 or LR-TR lesions, 83% (10/12) of LR-4 lesions were HCC, versus 33% (3/9) when not accompanied by LR-5 or LR-TR lesions. In a multivariable analysis of all lesions, the presence of a LR-5 or LR-TR lesion was significantly associated with LR-3 or LR-4 lesions representing HCC (OR 6.4, p = 0.01).LR-3 and LR-4 lesions are more likely to be HCC in patients with LR-5 or LR-TR lesions. The presence of coexisting definite HCC may be a useful diagnostic feature to improve risk stratification of lesions without typical imaging features of HCC. This may also affect decision-making prior to liver transplant when HCC burden must be accurately determined.

The Liver Imaging Reporting and Data System (LI-RADS) is widely used for diagnosing hepatocellular carcinoma (HCC), however, with unsatisfactory sensitivity, complex ancillary features, and inadequate integration with gadoxetate disodium (EOB)-enhanced MRI.To modify LI-RADS (mLI-RADS) on EOB-MRI.Secondary analysis of a prospective observational study.Between July 2015 and September 2018, 224 consecutive high-risk patients (median age, 51 years; range, 26-83; 180 men; training/testing sets: 169/55 patients) with 742 (median size, 13 mm; interquartile range, 7-27; 498 HCCs) LR-3/4/5 observations.3.0 T TThree radiologists (with 5, 5, and 10 years of experience in liver MR imaging, respectively) blinded to the reference standard (histopathology or imaging follow-up) reviewed all MR images independently. In the training set, the optimal LI-RADS version 2018 (v2018) features selected by Random Forest analysis were used to develop mLI-RADS via decision tree analysis.In an independent testing set, diagnostic performances of mLI-RADS, LI-RADS v2018, and the Korean Liver Cancer Association (KLCA) guidelines were computed using a generalized estimating equation model and compared with McNemar's test. A two-tailed P 0.05 was statistically significant.Five features (nonperipheral "washout," restricted diffusion, nonrim arterial phase hyperenhancement [APHE], mild-moderate T2 hyperintensity, and transitional phase hypointensity) constituted mLI-RADS, and mLR-5 was nonperipheral washout coupled with either nonrim APHE or restricted diffusion. In the testing set, mLI-RADS was significantly more sensitive (72%) and accurate (80%) than LI-RADS v2018 (sensitivity, 61%; accuracy 74%; both P 0.001) and the KLCA guidelines (sensitivity, 64%; accuracy 74%; both P 0.001), without sacrificing positive predictive value (mLI-RADS, 94%; LI-RADS v2018, 94%; KLCA guidelines, 92%).In high-risk patients, the EOB-MRI-based mLI-RADS was simpler and more sensitive for HCC than LI-RADS v2018 while maintaining high positive predictive value.2 TECHNICAL EFFICACY: Stage 2.

Background The Liver Imaging Reporting and Data System (LI-RADS) assigns a risk category for hepatocellular carcinoma (HCC) to imaging observations. Establishing the contributions of major features can inform the diagnostic algorithm. Purpose To perform a systematic review and individual patient data meta-analysis to establish the probability of HCC for each LI-RADS major feature using CT/MRI and contrast-enhanced US (CEUS) LI-RADS in patients at high risk for HCC. Materials and Methods Multiple databases (MEDLINE, Embase, Cochrane Central Register of Controlled Trials, and Scopus) were searched for studies from January 2014 to September 2019 that evaluated the accuracy of CT, MRI, and CEUS for HCC detection using LI-RADS (CT/MRI LI-RADS, versions 2014, 2017, and 2018; CEUS LI-RADS, versions 2016 and 2017). Data were centralized. Clustering was addressed at the study and patient levels using mixed models. Adjusted odds ratios (ORs) with 95% CIs were determined for each major feature using multivariable stepwise logistic regression. Risk of bias was assessed using Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) (PROSPERO protocol: CRD42020164486). Results A total of 32 studies were included, with 1170 CT observations, 3341 MRI observations, and 853 CEUS observations. At multivariable analysis of CT/MRI LI-RADS, all major features were associated with HCC, except threshold growth (OR, 1.6; 95% CI: 0.7, 3.6; P = .07). Nonperipheral washout (OR, 13.2; 95% CI: 9.0, 19.2; P = .01) and nonrim arterial phase hyperenhancement (APHE) (OR, 10.3; 95% CI: 6.7, 15.6; P = .01) had stronger associations with HCC than enhancing capsule (OR, 2.4; 95% CI: 1.7, 3.5; P = .03). On CEUS images, APHE (OR, 7.3; 95% CI: 4.6, 11.5; P = .01), late and mild washout (OR, 4.1; 95% CI: 2.6, 6.6; P = .01), and size of at least 20 mm (OR, 1.6; 95% CI: 1.04, 2.5; P = .04) were associated with HCC. Twenty-five studies (78%) had high risk of bias due to reporting ambiguity or study design flaws. Conclusion Most Liver Imaging Reporting and Data System major features had different independent associations with hepatocellular carcinoma; for CT/MRI, arterial phase hyperenhancement and washout had the strongest associations, whereas threshold growth had no association. © RSNA, 2021 Online supplemental material is available for this article.

Background Several early-phase clinical trials for the treatment of nonalcoholic steatohepatitis (NASH) use liver fat content as measured with the MRI-derived proton density fat fraction (PDFF) for a primary outcome. These trials have shown relative reductions in liver fat content with placebo treatment alone, a phenomenon termed "the placebo effect." This phenomenon confounds the results and limits generalizability to future trials. Purpose To quantify the effect of placebo treatment on change in the absolute PDFF value and to identify variables associated with this observed change. Materials and Methods This is a secondary analysis of prospectively collected data from seven early phase clinical trials that included participants with a diagnosis of NASH based on MRI and/or liver biopsy who received placebo treatment. The primary outcome was a greater than or equal to 30% relative reduction in PDFF after placebo treatment. Normalization of PDFF, relative change in alanine aminotransferase (ALT) level, and normalization of ALT level were also examined. An exploratory linear mixed-effects model was used to estimate an overall change in absolute PDFF and to explore parameters associated with this response. Results A total of 187 participants (median age, 52 years [IQR, 43-60 years]; 114 women) who received placebo treatment were evaluated. A greater than or equal to 30% relative reduction in baseline PDFF was seen in 20% of participants after 12 weeks of placebo treatment (10 of 49), 9% of participants after 16 weeks (two of 22), and 28% of participants after 24 weeks (34 of 122). A repeated-measures linear mixed-effects model estimated a decrease of 2.3 units (median relative reduction of 13%) in absolute PDFF values after 24 weeks of placebo treatment (95% CI: 3.2, 1.4

PURPOSE: To investigate the effect of training data type on generalizability of deep learning liver segmentation models. MATERIALS AND METHODS: This Health Insurance Portability and Accountability Act–compliant retrospective study included 860 MRI and CT abdominal scans obtained between February 2013 and March 2018 and 210 volumes from public datasets. Five single-source models were trained on 100 scans each of T1-weighted fat-suppressed portal venous (dynportal), T1-weighted fat-suppressed precontrast (dynpre), proton density opposed-phase (opposed), single-shot fast spin-echo (ssfse), and T1-weighted non–fat-suppressed (t1nfs) sequence types. A sixth multisource (DeepAll) model was trained on 100 scans consisting of 20 randomly selected scans from each of the five source domains. All models were tested against 18 target domains from unseen vendors, MRI types, and modality (CT). The Dice-Sørensen coefficient (DSC) was used to quantify similarity between manual and model segmentations. RESULTS: Single-source model performance did not degrade significantly against unseen vendor data. Models trained on T1-weighted dynamic data generally performed well on other T1-weighted dynamic data (DSC = 0.848 ± 0.183 [SD]). The opposed model generalized moderately well to all unseen MRI types (DSC = 0.703 ± 0.229). The ssfse model failed to generalize well to any other MRI type (DSC = 0.089 ± 0.153). Dynamic and opposed models generalized moderately well to CT data (DSC = 0.744 ± 0.206), whereas other single-source models performed poorly (DSC = 0.181 ± 0.192). The DeepAll model generalized well across vendor, modality, and MRI type and against externally sourced data. CONCLUSION: Domain shift in liver segmentation appears to be tied to variations in soft-tissue contrast and can be effectively bridged with diversification of soft-tissue representation in training data. Keywords: Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms, Supervised Learning, CT, MRI, Liver Segmentation Supplemental material is available for this article. © RSNA, 2023

The use of standardized terms in assessing and reporting disease processes has well-established benefits, such as clear communication between radiologists and other health care providers, improved diagnostic accuracy and reproducibility, and the enhancement and facilitation of research. Recently, the Liver Imaging Reporting and Data System (LI-RADS) Steering Committee released a universal liver imaging lexicon. The current version of the lexicon includes 81 vetted and precisely defined terms that are relevant to acquisition of images using all major liver imaging modalities and contrast agents, as well as lesion- and organ-level features. Most terms in the lexicon are applicable to all patients undergoing imaging of the liver, and only a minority of the terms are strictly intended to be used for patients with high risk factors for hepatocellular carcinoma. This pictorial atlas familiarizes readers with the liver imaging lexicon and includes discussion of general concepts, providing sample definitions, schematics, and clinical examples for a subset of the terms in the liver imaging lexicon. The authors discuss general, technical, and imaging feature terms used commonly in liver imaging, with the goal of illustrating their use for clinical and research applications.

Epigastric pain can have multiple etiologies including myocardial infarction, pancreatitis, acute aortic syndromes, gastroesophageal reflux disease, esophagitis, peptic ulcer disease, gastritis, duodenal ulcer disease, gastric cancer, and hiatal hernia. This document focuses on the scenarios in which epigastric pain is accompanied by symptoms such as heartburn, regurgitation, dysphagia, nausea, vomiting, and hematemesis, which raise suspicion for gastroesophageal reflux disease, esophagitis, peptic ulcer disease, gastritis, duodenal ulcer disease, gastric cancer, or hiatal hernia. Although endoscopy may be the test of choice for diagnosing these entities, patients may present with nonspecific or overlapping symptoms, necessitating the use of imaging prior to or instead of endoscopy. The utility of fluoroscopic imaging, CT, MRI, and FDG-PET for these indications are discussed. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

The TTo determine whether a denoised, single-average TRetrospective.A total of 45 males (age range 60-75 years) who underwent clinically indicated prostate MRI examinations, 21 of whom had pathologically proven prostate cancer.A 3 T; TNine readers randomly assessed complete exams including TGeneralized linear mixed effects models for differences in lesion detection, image quality features, and overall preference between TThere was no significant difference between sequences regarding identification of lesions with PI-RADS ≥3 (P = 0.10) or PI-RADS score (P = 0.77). Reader agreement was excellent for lesion identification (ICC = 0.84). There was no significant overall preference between the two sequences regarding image quality (P = 0.07, 95% CI: [-0.23, 0.01]). Reader agreement was good regarding sequence preference (ICC = 0.62).Use of single-average, denoised T3.Stage 3.

Radiomics is a high-throughput approach to image phenotyping. It uses computer algorithms to extract and analyze a large number of quantitative features from radiological images. These radiomic features collectively describe unique patterns that can serve as digital fingerprints of disease. They may also capture imaging characteristics that are difficult or impossible to characterize by the human eye. The rapid development of this field is motivated by systems biology, facilitated by data analytics, and powered by artificial intelligence. Here, as part of Abdominal Radiology's special issue on Quantitative Imaging, we provide an introduction to the field of radiomics. The technique is formally introduced as an advanced application of data analytics, with illustrating examples in abdominal radiology. Artificial intelligence is then presented as the main driving force of radiomics, and common techniques are defined and briefly compared. The complete step-by-step process of radiomic phenotyping is then broken down into five key phases. Potential pitfalls of each phase are highlighted, and recommendations are provided to reduce sources of variation, non-reproducibility, and error associated with radiomics.

The liver performs many vital functions for the human body. It stores essential vitamins and minerals, such as iron and vitamins A, D, K, and B12. It synthesizes proteins, such as blood clotting factors, albumin, and glycogen, as well as cholesterol, carbohydrates, and triglycerides. Additionally, it acts as a detoxifier, metabolizing and helping to clear alcohol, drugs, and ammonia. Typical MR imaging protocols for liver imaging include T2-weighted, chemical shift imaging, and precontrast and postcontrast T1-weighted sequences. This article discussed MR imaging of diffuse liver diseases and their typical imaging findings.

BACKGROUND The Liver Imaging Reporting and Data System (LI-RADS) is widely accepted as a reliable diagnostic scheme for hepatocellular carcinoma (HCC) in at-risk patients. However, its application is hampered by substantial complexity and suboptimal diagnostic sensitivity. PURPOSE To propose data-driven modifications to the LI-RADS version 2018 (v2018) major feature system (rLI-RADS) on gadoxetate disodium (EOB)-enhanced magnetic resonance imaging (MRI) to improve sensitivity and simplicity while maintaining high positive predictive value (PPV) for detecting HCC. STUDY TYPE Retrospective. POPULATION Two hundred and twenty-four consecutive at-risk patients (training dataset: 169, independent testing dataset: 55) with 742 LR-3 to LR-5 liver observations (HCC: N = 498 [67%]) were analyzed from a prospective observational registry collected between July 2015 and September 2018. FIELD STRENGTH/SEQUENCE 3.0 T/T2-weighted fast spin-echo, diffusion-weighted spin-echo based echo-planar and three-dimensional (3D) T1-weighted gradient echo sequences. ASSESSMENT All images were evaluated by three independent abdominal radiologists who were blinded to all clinical, pathological, and follow-up information. Composite reference standards of either histopathology or imaging follow-up were used. STATISTICAL TESTS In the training dataset, LI-RADS v2018 major features were used to develop rLI-RADS based on their associated PPV for HCC. In an independent testing set, diagnostic performances of LI-RADS v2018 and rLI-RADS were computed using a generalized estimating equation model and compared with McNemar's test. A P value

The benefits of farnesoid X receptor (FXR) agonists in patients with non-alcoholic steatohepatitis (NASH) have been validated, although improvements in efficacy and/or tolerability remain elusive. Herein, we aimed to assess the performance of a structurally optimized FXR agonist in patients with NASH.In this 12-week, randomized, placebo-controlled study, we evaluated MET409 - a non-bile acid agonist with a unique chemical scaffold - in patients with NASH. Patients were randomized to receive either 80 mg (n = 20) or 50 mg (n = 19) of MET409, or placebo (n = 19).At Week 12, MET409 lowered liver fat content (LFC), with mean relative reductions of 55% (80 mg) and 38% (50 mg) vs. 6% in placebo (p0.001). MET409 achieved ≥30% relative LFC reduction in 93% (80 mg) and 75% (50 mg) of patients vs. 11% in placebo (p0.001) and normalized LFC (≤5%) in 29% (80 mg) and 31% (50 mg) of patients vs. 0% in placebo (p0.05). An increase in alanine aminotransferase (ALT) was observed with MET409, confounding Week 12 changes from baseline (-25% for 80 mg, 28% for 50 mg). Nonetheless, MET409 achieved ≥30% relative ALT reduction in 50% (80 mg) and 31% (50 mg) of patients vs. 17% in placebo. MET409 was associated with on-target high-density lipoprotein cholesterol decreases (mean changes of -23.4% for 80 mg and -20.3% for 50 mg vs. 2.6% in placebo) and low-density lipoprotein cholesterol (LDL-C) increases (mean changes of 23.7% for 80 mg and 6.8% for 50 mg vs. -1.5% in placebo). Pruritus (mild-moderate) occurred in 16% (50 mg) and 40% (80 mg) of MET409-treated patients.MET409 lowered LFC over 12 weeks in patients with NASH and delivered a differentiated pruritus and LDL-C profile at 50 mg, providing the first clinical evidence that the risk-benefit profile of FXR agonists can be enhanced through structural optimization.Activation of the farnesoid X receptor (FXR) is a clinically validated approach for treating non-alcoholic steatohepatitis (NASH), although side effects such as itching or increases in low-density lipoprotein cholesterol are frequently dose-limiting. MET409, an FXR agonist with a unique chemical structure, led to significant liver fat reduction and delivered a favorable side effect profile after 12 weeks of treatment in patients with NASH. These results provide the first clinical evidence that the risk-benefit profile of FXR agonists can be enhanced.

Purpose Incidental pulmonary embolism (IPE) can be found on body CT. The aim of this study was to evaluate the feasibility of using artificial intelligence to identify missed IPE on a large number of CT examinations. Methods This retrospective analysis included all single-phase chest, abdominal, and pelvic (CAP) and abdominal and pelvic (AP) CT examinations performed at a single center over 1 year, for indications other than identification of PE. Proprietary visual classification and natural language processing software was used to analyze images and reports from all CT examinations, followed by a two-step human adjudication process to classify cases as true positive, false positive, true negative, or false negative. Descriptive statistics were assessed for prevalence of IPE and features (subsegmental versus central, unifocal versus multifocal, right heart strain or not) of missed IPE. Interrater agreement for radiologist readers was also calculated. Results A total of 11,913 CT examinations (6,398 CAP, 5,515 AP) were included. Thirty false-negative examinations were identified on CAP (0.47%; 95% confidence interval [CI], 0.32%-0.67%) and nineteen false-negative studies on AP (0.34%; 95% CI, 0.21%-0.54%) studies. During manual review, readers showed substantial agreement for identification of IPE on CAP (κ = 0.76; 95% CI, 0.66-0.86) and nearly perfect agreement for identification of IPE on AP (κ = 0.86; 95% CI, 0.76-0.97). Forty-nine missed IPEs (0.41%; 95% CI, 0.30%-0.54%) were ultimately identified, compared with seventy-nine IPEs (0.66%; 95% CI, 0.53%-0.83%) identified at initial clinical interpretation. Conclusions Artificial intelligence can efficiently analyze CT examinations to identify potential missed IPE. These results can inform peer-review efforts and quality control and could potentially be implemented in a prospective fashion.

Due to intrinsic differences in data formatting, data structure, and underlying semantic information, the integration of imaging data with clinical data can be non-trivial. Optimal integration requires robust data fusion, that is, the process of integrating multiple data sources to produce more useful information than captured by individual data sources. Here, we introduce the concept of fusion quality for deep learning problems involving imaging and clinical data. We first provide a general theoretical framework and numerical validation of our technique. To demonstrate real-world applicability, we then apply our technique to optimize the fusion of CT imaging and hepatic blood markers to estimate portal venous hypertension, which is linked to prognosis in patients with cirrhosis of the liver.To develop a measurement method of optimal data fusion quality deep learning problems utilizing both imaging data and clinical data.Our approach is based on modeling the fully connected layer (FCL) of a convolutional neural network (CNN) as a potential function, whose distribution takes the form of the classical Gibbs measure. The features of the FCL are then modeled as random variables governed by state functions, which are interpreted as the different data sources to be fused. The probability density of each source, relative to the probability density of the FCL, represents a quantitative measure of source-bias. To minimize this source-bias and optimize CNN performance, we implement a vector-growing encoding scheme called positional encoding, where low-dimensional clinical data are transcribed into a rich feature space that complements high-dimensional imaging features. We first provide a numerical validation of our approach based on simulated Gaussian processes. We then applied our approach to patient data, where we optimized the fusion of CT images with blood markers to predict portal venous hypertension in patients with cirrhosis of the liver. This patient study was based on a modified ResNet-152 model that incorporates both images and blood markers as input. These two data sources were processed in parallel, fused into a single FCL, and optimized based on our fusion quality framework.Numerical validation of our approach confirmed that the probability density function of a fused feature space converges to a source-specific probability density function when source data are improperly fused. Our numerical results demonstrate that this phenomenon can be quantified as a measure of fusion quality. On patient data, the fused model consisting of both imaging data and positionally encoded blood markers at the theoretically optimal fusion quality metric achieved an AUC of 0.74 and an accuracy of 0.71. This model was statistically better than the imaging-only model (AUC = 0.60; accuracy = 0.62), the blood marker-only model (AUC = 0.58; accuracy = 0.60), and a variety of purposely sub-optimized fusion models (AUC = 0.61-0.70; accuracy = 0.58-0.69).We introduced the concept of data fusion quality for multi-source deep learning problems involving both imaging and clinical data. We provided a theoretical framework, numerical validation, and real-world application in abdominal radiology. Our data suggests that CT imaging and hepatic blood markers provide complementary diagnostic information when appropriately fused.

PURPOSE: To perform a systematic review and meta-analysis using individual patient data to investigate the diagnostic performance of Liver Imaging Reporting and Data System (LI-RADS) Treatment Response (TR) algorithm for detecting incomplete necrosis on pathology. METHODS: PubMed and EMBASE were searched from Jan 1, 2017 until October 14, 2020. Studies reporting diagnostic accuracy of LI-RADS TR algorithm on CT or MRI for detecting incomplete necrosis on pathology as a reference standard were included. Sensitivity and specificity were pooled using random-effects model. Subgroup analyses were performed for LRT type, and imaging modality. RESULTS: Six studies (393 patients, 534 lesions) were included. Pooled sensitivity was 0.56 (95% confidence interval [CI] 0.43–0.69) and specificity was 0.91 (95%CI 0.84–0.96). Pooled sensitivity was highest using arterial phase hyperenhancement (APHE) (0.67 [95%CI 0.51–0.81]), followed by washout (0.43 [95%CI 0.26–0.62]) and enhancement similar to pretreatment (0.24 [95%CI 0.15–0.36]). Among lesions with incomplete necrosis, 2% (95%CI 0.00–0.05) manifested as washout but no APHE; 0% (95% CI 0.00–0.02) as enhancement similar to pretreatment without both APHE and washout. Pooled sensitivity was lower after ablation than embolization (0.42 [95%CI, 0.28–0.57] vs. 0.65 [95%CI, 0.53–0.77], p=0.033). MRI and CT were comparable (p=0.783 and 0.290 for sensitivity and specificity). CONCLUSIONS: LI-RADS TR algorithm shows moderate sensitivity and high specificity for detecting incomplete necrosis after LRT. APHE is the dominant criterion, a washout contributes to small but meaningful extent, while the contribution of enhancement similar to pretreatment may be negligible. LRT type may affect performance of the algorithm.

Locoregional therapies can be offered to hepatocellular carcinoma patients as a bridge to transplant, to downstage disease burden for transplant eligibility, or for disease control to prolong survival. Systemic therapies also play a large role in HCC treatment, occasionally in conjunction with other methods. This manuscript reviews the various treatment options for HCC with a historically noncurative intent.

Artificial Intelligence (AI) continues to shape the practice of radiology, with imaging of hepatocellular carcinoma (HCC) being of no exception. This article prepared by members of the LI-RADS Treatment Response (TR LI-RADS) work group and associates, presents recent trends in the utility of AI applications for the volumetric evaluation and assessment of HCC treatment response. Various topics including radiomics, prognostic imaging findings, and locoregional therapy (LRT) specific issues will be discussed in the framework of HCC treatment response classification systems with focus on the Liver Reporting and Data System treatment response algorithm (LI-RADS TRA).

Purpose Chemical shift-encoded MRI (CSE-MRI) is well-established to quantify proton density fat fraction (PDFF) as a quantitative biomarker of hepatic steatosis. However, temperature is known to bias PDFF estimation in phantom studies. In this study, strategies were developed and evaluated to correct for the effects of temperature on PDFF estimation through simulations, temperature-controlled experiments, and a multi-center, multi-vendor phantom study. Theory and methods A technical solution that assumes and automatically estimates a uniform, global temperature throughout the phantom is proposed. Computer simulations modeled the effect of temperature on PDFF estimation using magnitude-, complex-, and hybrid-based CSE-MRI methods. Phantom experiments were performed to assess the temperature correction on PDFF estimation at controlled phantom temperatures. To assess the temperature correction method on a larger scale, the proposed method was applied to data acquired as part of a nine-site multi-vendor phantom study and compared to temperature-corrected PDFF estimation using an a priori guess for ambient room temperature. Results Simulations and temperature-controlled experiments show that as temperature deviates further from the assumed temperature, PDFF bias increases. Using the proposed correction method and a reasonable a priori guess for ambient temperature, PDFF bias and variability were reduced using magnitude-based CSE-MRI, across MRI systems, field strengths, protocols, and varying phantom temperature. Complex and hybrid methods showed little PDFF bias and variability both before and after correction. Conclusion Correction for temperature reduces temperature-related PDFF bias and variability in phantoms across MRI vendors, sites, field strengths, and protocols for magnitude-based CSE-MRI, even without a priori information about the temperature.

Resmetirom (MGL‐3196), a selective thyroid hormone receptor‐β agonist, was evaluated in a 36‐week paired liver biopsy study ({"type":"clinical-trial","attrs":{"text":"NCT02912260","term_id":"NCT02912260"}}NCT02912260) in adults with biopsy‐confirmed nonalcoholic steatohepatitis (NASH). The primary endpoint was relative liver fat reduction as assessed by MRI–proton density fat fraction (MRI‐PDFF), and secondary endpoints included histopathology. Subsequently, a 36‐week active treatment open‐label extension (OLE) study was conducted in 31 consenting patients (including 14 former placebo patients) with persistently mild to markedly elevated liver enzymes at the end of the main study. In patients treated with resmetirom (80 or 100 mg orally per day), MRI‐PDFF reduction at OLE week 36 was −11.1% (1.5%) mean reduction (standard error [SE]; P

There is a sparsity of data evaluating outcomes of patients with Liver Imaging Reporting and Data System (LI-RADS) (LR)-M lesions.To compare overall survival (OS) and progression free survival (PFS) between hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA) meeting LR-M criteria and to evaluate factors associated with prognosis.Retrospective.Patients at risk for HCC with at least one LR-M lesion with histologic diagnosis, from 8 academic centers, yielding 120 patients with 120 LR-M lesions (84 men [mean age 62 years] and 36 women [mean age 66 years]).A 1.5 and 3.0 T/3D TThe imaging categorization of each lesion as LR-M was made clinically by a single radiologist at each site and patient outcome measures were collected.OS, PFS, and potential independent predictors were evaluated by Kaplan-Meier method, log-rank test, and Cox proportional hazard model. A P value of0.05 was considered significant.A total of 120 patients with 120 LR-M lesions were included; on histology 65 were HCC and 55 were iCCA. There was similar median OS for patients with LR-M HCC compared to patients with iCCA (738 days vs. 769 days, P = 0.576). There were no significant differences between patients with HCC and iCCA in terms of sex (47:18 vs. 37:18, P = 0.549), age (63.0 ± 8.4 vs. 63.4 ± 7.8, P = 0.847), etiology of liver disease (P = 0.202), presence of cirrhosis (100% vs. 100%, P = 1.000), tumor size (4.73 ± 3.28 vs. 4.75 ± 2.58, P = 0.980), method of lesion histologic diagnosis (P = 0.646), and proportion of patients who underwent locoregional therapy (60.0% vs. 38.2%, P = 0.100) or surgery (134.8 ± 165.5 vs. 142.5 ± 205.6, P = 0.913). Using multivariable analysis, nonsurgical compared to surgical management (HR, 4.58), larger tumor size (HR, 1.19), and higher MELD score (HR, 1.12) were independently associated with worse OS.There was similar OS in patients with LR-M HCC and LR-M iCCA, suggesting that LR-M imaging features may more closely reflect patient outcomes than histology.3 TECHNICAL EFFICACY: Stage 5.

Incidental liver masses are commonly identified on imaging performed for other indications. Since the prevalence of benign focal liver lesions in adults is high, even in patients with primary malignancy, accurate characterization of incidentally detected lesions is of paramount clinical importance. This document reviews utilization of various imaging modalities for characterization of incidentally detected liver lesions, discussed in the context of several clinical scenarios. For each clinical scenario, a summary of current evidence supporting the use of a given diagnostic modality is reported. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

To determine the rate of development of clinically significant liver nodules (LR-4, LR-5, LR-M) after a negative MRI in an HCC screening population. This retrospective study included patients at risk of developing HCC requiring imaging surveillance who had undergone multiphase Gadobenate dimeglumine-enhanced MRI that was negative and had follow up LI-RADS compliant multiphase CTs or MRIs for at least 12 months or positive follow-up within 12 months. Follow-up examinations were classified as negative (no nodules or only LR-1 nodules) or positive (nodule other than LR-1). Time-to-first positive examination, types of nodules, and cumulative incidence of nodule development were recorded. 204 patients (mean age 58.9 ± 10.2 years, 128 women), including 172 with cirrhosis, were included. Based CT/MRI follow-up (median 35 months, range 12–80 months), the overall cumulative incidence of developing a nodule was 10.5%. Cumulative incidence of nodule development was: 0.5% at 6–9 months and 2.1% at 12 ± 3 months, including one LR-4 nodule, one LR-M nodule, and two LR-3 nodules. The cumulative incidence of clinically significant nodule development was 1.1% at 9–15 months. 70% (143/204) of patients also underwent at least one US follow-up, and no patient developed a positive US examination following index negative MRI. Clinically significant liver nodules develop in 1.1% of at-risk patients in the first year following negative MRI. While ongoing surveillance is necessary for at-risk patients, our study suggests than longer surveillance intervals after a negative MRI may be reasonable and that further research is needed to explore this possibility.

The Project Baseline Health Study (PBHS) was launched to map human health through a comprehensive understanding of both the health of an individual and how it relates to the broader population. The study will contribute to the creation of a biomedical information system that accounts for the highly complex interplay of biological, behavioral, environmental, and social systems. The PBHS is a prospective, multicenter, longitudinal cohort study that aims to enroll thousands of participants with diverse backgrounds who are representative of the entire health spectrum. Enrolled participants will be evaluated serially using clinical, molecular, imaging, sensor, self-reported, behavioral, psychological, environmental, and other health-related measurements. An initial deeply phenotyped cohort will inform the development of a large, expanded virtual cohort. The PBHS will contribute to precision health and medicine by integrating state of the art testing, longitudinal monitoring and participant engagement, and by contributing to the development of an improved platform for data sharing and analysis.

Incidental pancreatic cysts are increasingly detected on imaging studies performed for unrelated indications and may be incompletely characterized on these studies. Adequate morphological characterization is critical due to the small risk of malignant degeneration associated with neoplastic pancreatic cysts, as well as the risk of associated pancreatic adenocarcinoma. For all pancreatic cysts, both size and morphology determine management. Specifically, imaging detection of features, such as pancreatic ductal communication and presence or absence of worrisome features or high-risk stigmata, have important management implications. The recommendations in this publication determine the appropriate initial imaging study to further evaluate a pancreatic cyst that was incidentally detected on a nondedicated imaging study. The recommendations are designed to maximize the yield of diagnostic information in order to better risk-stratify pancreatic cysts and assist in guiding future management. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

To determine whether LI-RADS ancillary features predict longitudinal LR-3 observation category changes. This exploratory, retrospective, single-center study with an independent reading center included patients who underwent two or more multiphase CT or MRI examinations for hepatocellular carcinoma assessment between 2011 and 2015. Three readers independently evaluated each observation using CT/MRI LI-RADS v2017, and observations categorized LR-3 using major features only were included in the analysis. Prevalence of major and ancillary features was calculated. After excluding low-frequency (

Background The Liver Imaging Reporting and Data System (LI-RADS) treatment response algorithm (TRA) is used to assess presumed hepatocellular carcinoma (HCC) after local-regional therapy, but its performance has not been extensively assessed. Purpose To assess the performance of LI-RADS version 2018 TRA in the evaluation of HCC after ablation. Materials and Methods In this retrospective study, patients who underwent ablation therapy for presumed HCC followed by liver transplantation between January 2011 and December 2015 at a single tertiary care center were identified. Lesions were categorized as completely (100%) or incompletely (≤99%) necrotic based on transplant histology. Three radiologists assessed pre- and posttreatment MRI findings using LI-RADS version 2018 and the TRA, respectively. Interreader agreement was assessed by using the Fleiss κ test. Performance characteristics for predicting necrosis category based on LI-RADS treatment response (LR-TR) category (viable or nonviable) were calculated by using generalized mixed-effects models to account for clustering by subject. Results A total of 36 patients (mean age, 58 years ± 5 [standard deviation]; 32 men) with 53 lesions was included. Interreader agreement for pretreatment LI-RADS category was 0.40 (95% confidence interval [CI]: 0.15, 0.67; P < .01) and was lower than the interreader agreement for TRA category (κ = 0.71; 95% CI: 0.59, 0.84; P < .01). After accounting for clustering by subject, sensitivity of tumor necrosis across readers ranged from 40% to 77%, and specificity ranged from 85% to 97% when LR-TR equivocal assessments were treated as nonviable. When LR-TR equivocal assessments were treated as viable, sensitivity of tumor necrosis across readers ranged from 81% to 87%, and specificity ranged from 81% to 85% across readers. Six (11%) of 53 treated lesions were LR-TR equivocal by consensus, with most (five of six) incompletely necrotic at histopathology. Conclusion The Liver Imaging Reporting and Data System treatment response algorithm can be used to predict viable or nonviable hepatocellular carcinoma after ablation. Most ablated lesions rated as treatment response equivocal were incompletely necrotic at histopathology. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Do and Mendiratta-Lala in this issue.