Inter-reader agreement of CEUS LI-RADS among radiologists with different levels of experience

Diagnostic Performance of Modified Contrast-Enhanced Ultrasound Liver Imaging Reporting and Data System in Patients Without Risk Factors for Hepatocellular Carcinoma: Comparison With World Federation for Ultrasound in Medicine and Biology Guideline

OBJECTIVE

The aim of this study was to assess the ability of the modified contrast-enhanced ultrasound (CEUS) Liver Imaging Reporting and Data System (LI-RADS) to distinguish malignancy in patients without known hepatocellular carcinoma (HCC) risk factors and compare diagnostic accuracy with that of the World Federation for Ultrasound in Medicine and Biology (WFUMB) guideline across radiologists with different levels of CEUS experience.

METHODS

A total of 848 individuals with no hepatitis infection presenting with 870 lesions in non-cirrhotic livers were included and divided into the Testing and Validation groups. The modified CEUS LI-RADS was proposed, including downgrading of focal nodular hyperplasia with typical features. Diagnostic performance of the modified CEUS LI-RADS was assessed in the Testing group. In the Validation group, two radiologists with more than 9 y of CEUS experience (Experts) and two radiologists with less than 6 mo of CEUS experience (Novices) used both the modified CEUS LI-RADS and the WFUMB guideline to evaluate performance in diagnosis of the lesions.

RESULTS

LR-5 + M (combination of modified LR-5 and modified LR-M) revealed optimal performance with a sensitivity, specificity and area under the curve (AUC) of 99.3%, 81.6% and 0.904, respectively. Novices using the modified CEUS LI-RADS outperformed those using the WFUMB guideline (AUC: 0.858 vs. 0.767, p = 0.005). Additionally, the sensitivity, specificity and AUC of Novices were comparable to those of Experts using the modified CEUS LI-RADS (94.1%, 77.6% and 0.858 vs. 96.1%, 77.6% and 0.868 for experts, respectively).

CONCLUSION

The modified CEUS LI-RADS is a valuable method for distinguishing hepatic malignancy in patients without HCC risk factors. This is particularly beneficial for radiologists with limited CEUS expertise.

Standardizing liver imaging reporting and interpretation: LI-RADS and beyond

Imaging plays a crucial role in diagnosis and post-treatment monitoring of primary liver cancers. Clear, consistent, and actionable communication of imaging results is crucial to avoid miscommunication and potential detrimental impact on patient care. In this review, we discuss the importance, advantages, and potential impact of universal adoption of standardized terminology and interpretive criteria for liver imaging, from the point of view of radiologists and clinicians.

Deep learning nomogram based on Gd-EOB-DTPA MRI for predicting early recurrence in hepatocellular carcinoma after hepatectomy

OBJECTIVES

The accurate prediction of post-hepatectomy early recurrence in patients with hepatocellular carcinoma (HCC) is crucial for decision-making regarding postoperative adjuvant treatment and monitoring. We aimed to explore the feasibility of deep learning (DL) features derived from gadoxetate disodium (Gd-EOB-DTPA) MRI, qualitative features, and clinical variables for predicting early recurrence.

METHODS

In this bicentric study, 285 patients with HCC who underwent Gd-EOB-DTPA MRI before resection were divided into training (n = 195) and validation (n = 90) sets. DL features were extracted from contrast-enhanced MRI images using VGGNet-19. Three feature selection methods and five classification methods were combined for DL signature construction. Subsequently, an mp-MR DL signature fused with multiphase DL signatures of contrast-enhanced images was constructed. Univariate and multivariate logistic regression analyses were used to identify early recurrence risk factors including mp-MR DL signature, microvascular invasion (MVI), and tumor number. A DL nomogram was built by incorporating deep features and significant clinical variables to achieve early recurrence prediction.

RESULTS

MVI (p = 0.039), tumor number (p = 0.001), and mp-MR DL signature (p < 0.001) were independent risk factors for early recurrence. The DL nomogram outperformed the clinical nomogram in the training set (AUC: 0.949 vs. 0.751; p < 0.001) and validation set (AUC: 0.909 vs. 0.715; p = 0.002). Excellent DL nomogram calibration was achieved in both training and validation sets. Decision curve analysis confirmed the clinical usefulness of DL nomogram.

CONCLUSION

The proposed DL nomogram was superior to the clinical nomogram in predicting early recurrence for HCC patients after hepatectomy.

KEY POINTS

• Deep learning signature based on Gd-EOB-DTPA MRI was the predominant independent predictor of early recurrence for hepatocellular carcinoma (HCC) after hepatectomy. • Deep learning nomogram based on clinical factors and Gd-EOB-DTPA MRI features is promising for predicting early recurrence of HCC. • Deep learning nomogram outperformed the conventional clinical nomogram in predicting early recurrence.

Combining serum AFP and CEUS LI-RADS for better diagnostic performance in Chinese high-risk patients

PURPOSE

To evaluate and compare the diagnostic performance of revised contrast-enhanced ultrasound (CEUS) Liver Imaging Reporting and Data System version by combining LR-M category and serum alpha-fetoprotein (AFP) under different cut-off values.

MATERIAL AND METHODS

This retrospective study enrolled 152 high-risk patients with 152 histology-proven nodules. For revised LI-RADS, nodules in LR-M with different elevated AFP thresholds have been reclassified as the LR-5 category. The diagnostic performances of original and revised CEUS LI-RADS were evaluated and compared.

RESULTS

To compare with the original version, the sensitivity of revised LR-5 (adjusted with AFP value > 200 ng/ml or 400 ng/ml) for the diagnosis of hepatocellular carcinoma (HCC) improved from 52.5 to 69.2% or 65.0%, respectively (both p < 0.001) without compromising specificity (87.5% vs. 71.9% or 78.1%, respectively, both p > 0.05). For the diagnosis of non-HCC malignancy, the specificity of the LR-M after reclassification was improved (69.6% vs. 84.4% or 80.7%, respectively, both p < 0.001) with a non-significant sensitivity reduction (100.0 vs. 70.6% or 82.4%, respectively, both p > 0.05). After modification, the sensitivity of LR-5 also increased to 69.1% or 64.9% (both p < 0.001), while the specificity and PPV did not change (both p > 0.05) for larger nodules (> 20 mm).

CONCLUSION

The diagnostic performance of CEUS LI-RADS can be further improved by reclassifying LR-M nodules with elevated AFP thresholds to LR-5.

Ultrasound-based radiomics analysis for differentiating benign and malignant breast lesions: From static images to CEUS video analysis

Background

Continuous contrast-enhanced ultrasound (CEUS) video is a challenging direction for radiomics research. We aimed to evaluate machine learning (ML) approaches with radiomics combined with the XGBoost model and a convolutional neural network (CNN) for discriminating between benign and malignant lesions in CEUS videos with a duration of more than 1 min.

Methods

We gathered breast CEUS videos of 109 benign and 81 malignant tumors from two centers. Radiomics combined with the XGBoost model and a CNN was used to classify the breast lesions on the CEUS videos. The lesions were manually segmented by one radiologist. Radiomics combined with the XGBoost model was conducted with a variety of data sampling methods. The CNN used pretrained 3D residual network (ResNet) models with 18, 34, 50, and 101 layers. The machine interpretations were compared with prospective interpretations by two radiologists. Breast biopsies or pathological examinations were used as the reference standard. Areas under the receiver operating curves (AUCs) were used to compare the diagnostic performance of the models.

Results

The CNN model achieved the best AUC of 0.84 on the test cohort with the 3D-ResNet-50 model. The radiomics model obtained AUCs between 0.65 and 0.75. Radiologists 1 and 2 had AUCs of 0.75 and 0.70, respectively.

Conclusions

The 3D-ResNet-50 model was superior to the radiomics combined with the XGBoost model in classifying enhanced lesions as benign or malignant on CEUS videos. The CNN model was superior to the radiologists, and the radiomics model performance was close to the performance of the radiologists.

Contrast-enhanced ultrasound Liver Imaging Reporting and Data System: Lights and shadows in hepatocellular carcinoma and cholangiocellular carcinoma diagnosis

BACKGROUND

Contrast-enhanced ultrasound (CEUS) is considered a secondary examination compared to computed tomography (CT) and magnetic resonance imaging (MRI) in the diagnosis of hepatocellular carcinoma (HCC), due to the risk of misdiagnosing intrahepatic cholangiocarcinoma (ICC). The introduction of CEUS Liver Imaging Reporting and Data System (CEUS LI-RADS) might overcome this limitation. Even though data from the literature seems promising, its reliability in real-life context has not been well-established yet.

AIM

To test the accuracy of CEUS LI-RADS for correctly diagnosing HCC and ICC in cirrhosis.

METHODS

CEUS LI-RADS class was retrospectively assigned to 511 nodules identified in 269 patients suffering from liver cirrhosis. The diagnostic standard for all nodules was either biopsy (102 nodules) or CT/MRI (409 nodules). Common diagnostic accuracy indexes such as sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were assessed for the following associations: CEUS LR-5 and HCC; CEUS LR-4 and 5 merged class and HCC; CEUS LR-M and ICC; and CEUS LR-3 and malignancy. The frequency of malignant lesions in CEUS LR-3 subgroups with different CEUS patterns was also determined. Inter-rater agreement for CEUS LI-RADS class assignment and for major CEUS pattern identification was evaluated.

RESULTS

CEUS LR-5 predicted HCC with a 67.6% sensitivity, 97.7% specificity, and 99.3% PPV (P < 0.001). The merging of LR-4 and 5 offered an improved 93.9% sensitivity in HCC diagnosis with a 94.3% specificity and 98.8% PPV (P < 0.001). CEUS LR-M predicted ICC with a 91.3% sensitivity, 96.7% specificity, and 99.6% NPV (P < 0.001). CEUS LR-3 predominantly included benign lesions (only 28.8% of malignancies). In this class, the hypo-hypo pattern showed a much higher rate of malignant lesions (73.3%) than the iso-iso pattern (2.6%). Inter-rater agreement between internal raters for CEUS-LR class assignment was almost perfect (n = 511, k = 0.94, P < 0.001), while the agreement among raters from separate centres was substantial (n = 50, k = 0.67, P < 0.001). Agreement was stronger for arterial phase hyperenhancement (internal k = 0.86, P < 2.7 × 10^-214; external k = 0.8, P < 0.001) than washout (internal k = 0.79, P < 1.6 × 10^-202; external k = 0.71, P < 0.001).

CONCLUSION

CEUS LI-RADS is effective but can be improved by merging LR-4 and 5 to diagnose HCC and by splitting LR-3 into two subgroups to differentiate iso-iso nodules from other patterns.

Diagnostic value of CEUS LI-RADS and serum tumor markers for combined hepatocellular-cholangiocarcinoma

PURPOSE

To analyze the contrast-enhanced ultrasound (CEUS) manifestations of combined hepatocellular-cholangiocarcinoma (CHC) and to investigate the diagnostic value of the 2017 version of the CEUS Liver Imaging Reporting and Data System (LI-RADS) and serum tumor markers in CHC.

METHODS

A total of 66 pathologically confirmed CHC nodules were retrospectively analyzed. We summarized the CEUS manifestations of CHC and analyzed the relationship between serum tumor markers and the enhancement pattern of CHC. We also classified CHC according to CEUS LI-RADS criteria. The Kappa test was used to assess the interreader agreement of CEUS LI-RADS between radiologists.

RESULT

According to the results, 52 of 62 (83.9%) patients had elevated alpha-fetoprotein (AFP), 19 of 61 (31.1%) had elevated carbohydrate antigen 199 (CA 199), and 13 of 61 (21.3%) had both elevated AFP and CA 199. Of the 66 CHC nodules, 64 (97.0%) were identified as malignant lesions by CEUS, 13 (19.7%) showed a hepatocellular carcinoma-like enhancement pattern, and 21 (31.8%) showed a cholangiocarcinoma-like enhancement pattern. For the CEUS LI-RADS categories, 39 of 53 (73.6%) CHC nodules were classified as LR-M, 12 (22.6%) were classified as LR-5, and 2 (3.8%) were classified as LR-4. The interreader agreement for the LI-RADS categories was 0.60.

CONCLUSIONS

Although CHC lacks specific CEUS features, CEUS LI-RADS and serum tumor markers can be useful tools for reducing the misdiagnosis of CHC. In addition, due to the relative complexity of the CEUS features involved in CHC, it is necessary for beginning radiologists to learn more about CEUS features.

Diagnostic performance and interreader agreement of CEUS LI-RADS in ≤ 30 mm liver nodules with different experienced radiologists

PURPOSE

To explore the diagnostic performance and interreader agreement of CEUS LI-RADS in diagnosing ≤ 30 mm liver nodules with different experienced radiologists.

METHODS

Between January 2018 and October 2020, 244 patients at high-risk for HCC who underwent CEUS were enrolled. Two novice radiologists and two expert radiologists independently evaluated LI-RADS categories and main features. Kappa (κ) and Kendall's tests were employed to evaluate the interreader agreement of CEUS LI-RADS. The diagnostic performance was determined based on sensitivity, specificity, accuracy, PPV and NPV.

RESULTS

The interreader agreement for arterial phase hyperenhancement, late and mild washout, early washout, and rim hyperenhancement was moderate to almost perfect (κ, 0.44-0.93) among the different levels of radiologists. The interreader agreement for the LI-RADS categories was substantial to almost perfect (κ, 0.78-0.88). However, the interreader agreement for marked washout was fair to moderate (κ, 0.28-0.50). When CEUS LR-5 was used as a diagnostic criterion for HCC, there were no statistical differences in sensitivity, specificity, accuracy, PPV and NPV among the radiologists (p > 0.05), except for the differences between Reader 4 and the remaining three radiologists in terms of accuracy and sensitivity (p < 0.05).

CONCLUSION

CEUS LI-RADS has good diagnostic agreement for ≤ 30 mm liver nodules among experienced radiologists.

Imaging Diagnosis of HCC: Future directions with special emphasis on hepatobiliary MRI and contrast-enhanced ultrasound

Hepatocellular carcinoma (HCC) is a unique cancer entity that can be noninvasively diagnosed using imaging modalities without pathologic confirmation. In 2018, several major guidelines for HCC were updated to include hepatobiliary contrast agent magnetic resonance imaging (HBA-MRI) and contrast-enhanced ultrasound (CEUS) as major imaging modalities for HCC diagnosis. HBA-MRI enables the achievement of high sensitivity in HCC detection using the hepatobiliary phase (HBP). CEUS is another imaging modality with real-time imaging capability, and it is reported to be useful as a second-line modality to increase sensitivity without losing specificity for HCC diagnosis. However, until now, there is an unsolved discrepancy among guidelines on whether to accept “HBP hypointensity” as a definite diagnostic criterion for HCC or include CEUS in the diagnostic algorithm for HCC diagnosis. Furthermore, there is variability in terminology and inconsistencies in the definition of imaging findings among guidelines; therefore, there is an unmet need for the development of a standardized lexicon. In this article, we review the performance and limitations of HBA-MRI and CEUS after guideline updates in 2018 and briefly introduce some future aspects of imaging-based HCC diagnosis.

Inter-reader reliability of contrast-enhanced ultrasound Liver Imaging Reporting and Data System: a meta-analysis

PURPOSE

To systematically determine the inter-reader reliability of the contrast-enhanced ultrasound (CEUS) Liver Imaging Reporting and Data System (LI-RADS), with emphasis on its major features for hepatocellular carcinoma (HCC) and LR-M (LI-RADS category M) features for non-HCC malignancy.

METHODS

MEDLINE, EMBASE, and Cochrane databases were searched from January 2016 to March 2021 to identify original articles reporting the inter-reader reliability of CEUS LI-RADS. Meta-analytic pooled kappa values (κ) were calculated for major features [nonrim arterial-phase hyperenhancement (APHE), mild and late washout], LR-M features (rim APHE, early washout), and LI-RADS categorization using the DerSimonian-Laird random-effects model. Meta-regression analysis was performed to explore any causes of study heterogeneity.

RESULTS

Twelve studies with a total of 2862 lesions were included. The meta-analytic pooled κ of nonrim APHE, mild and late washout, rim APHE, early washout, and LI-RADS categorization were 0.73 [95% confidence interval (CI), 0.67 - 0.79], 0.69 (95% CI, 0.54-0.84), 0.54 (95% CI, 0.37-0.71), 0.62 (95% CI, 0.45-0.79), and 0.75 (95% CI, 0.64-0.87), respectively. Compared with the major features, LR-M features had a lower meta-analytic pooled κ. Substantial study heterogeneity was noted in the LI-RADS categorization, and lesion size (p = 0.03) and the homogeneity in reader experience (p = 0.03) were significantly associated with study heterogeneity.

CONCLUSIONS

CEUS LI-RADS showed substantial inter-reader reliability for major features and LI-RADS categorization, but relatively lower reliability was found for LR-M features. In our opinion, the definitions of imaging features require further refinement to improve the inter-reader reliability of CEUS LI-RADS.