User and system pitfalls in liver imaging with LI-RADS

Systematic training of LI-RADS CT v2018 improves interobserver agreements and performances in LR categorization for focal liver lesions

AIM

To retrospectively explored whether systematic training in the use of Liver Imaging Reporting and Data System (LI-RADS) v2018 on computed tomography (CT) can improve the interobserver agreements and performances in LR categorization for focal liver lesions (FLLs) among different radiologists.

MATERIALS AND METHODS

A total of 18 visiting radiologists and the liver multiphase CT images of 70 hepatic observations in 63 patients at high risk of HCC were included in this study. The LI-RADS v2018 training procedure included three thematic lectures, with an interval of 1 month. After each seminar, the radiologists had 1 month to adopt the algorithm into their daily work. The interobserver agreements and performances in LR categorization for FLLs among the radiologists before and after training were compared.

RESULTS

After training, the interobserver agreements in classifying the LR categories for all radiologists were significantly increased for most LR categories (P < 0.001), except for LR-1 (P = 0.053). After systematic training, the areas under the curve (AUCs) for LR categorization performance for all participants were significantly increased for most LR categories (P < 0.001), except for LR-1 (P = 0.062).

CONCLUSION

Systematic training in the use of the LI-RADS can improve the interobserver agreements and performances in LR categorization for FLLs among radiologists with different levels of experience.

LI-RADS CT and MRI Ancillary Feature Association with Hepatocellular Carcinoma and Malignancy: An Individual Participant Data Meta-Analysis

Background The independent contribution of each Liver Imaging Reporting and Data System (LI-RADS) CT or MRI ancillary feature (AF) has not been established. Purpose To evaluate the association of LI-RADS AFs with hepatocellular carcinoma (HCC) and malignancy while adjusting for LI-RADS major features through an individual participant data (IPD) meta-analysis. Materials and Methods Medline, Embase, Cochrane Central Register of Controlled Trials, and Scopus were searched from January 2014 to January 2022 for studies evaluating the diagnostic accuracy of CT and MRI for HCC using LI-RADS version 2014, 2017, or 2018. Using a one-step approach, IPD across studies were pooled. Adjusted odds ratios (ORs) and 95% CIs were derived from multivariable logistic regression models of each AF combined with major features except threshold growth (excluded because of infrequent reporting). Liver observation clustering was addressed at the study and participant levels through random intercepts. Risk of bias was assessed using a composite reference standard and Quality Assessment of Diagnostic Accuracy Studies 2. Results Twenty studies comprising 3091 observations (2456 adult participants; mean age, 59 years ± 11 [SD]; 1849 [75.3%] men) were included. In total, 89% (eight of nine) of AFs favoring malignancy were associated with malignancy and/or HCC, 80% (four of five) of AFs favoring HCC were associated with HCC, and 57% (four of seven) of AFs favoring benignity were negatively associated with HCC and/or malignancy. Nonenhancing capsule (OR = 3.50 [95% CI: 1.53, 8.01]) had the strongest association with HCC. Diffusion restriction (OR = 14.45 [95% CI: 9.82, 21.27]) and mild-moderate T2 hyperintensity (OR = 10.18 [95% CI: 7.17, 14.44]) had the strongest association with malignancy. The strongest negative associations with HCC were parallels blood pool enhancement (OR = 0.07 [95% CI: 0.01, 0.49]) and marked T2 hyperintensity (OR = 0.18 [95% CI: 0.07, 0.45]). Seventeen studies (85%) had a high risk of bias. Conclusion Most LI-RADS AFs were independently associated with HCC, malignancy, or benignity as intended when adjusting for major features. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Crivellaro in this issue.

Assessment of LI-RADS efficacy in classification of hepatocellular carcinoma and benign liver nodules using DCE-MRI features and machine learning

Purpose

The current study aimed to evaluate the efficiency of dynamic contrast-enhanced (DCE) MRI visual features in classifying benign liver nodules and hepatocellular carcinoma (HCC) using a machine learning model.

Methods

115 LI-RADS3, 137 LI-RADS4, and 140 LI-RADS5 nodules were included (392 nodules from 245 patients), which were evaluated by follow-up imaging for LR-3 and pathology results for LR-4 and LR-5 nodules. Data was collected retrospectively from 3 T and 1.5 T MRI scanners. All the lesions were categorized into 124 benign and 268 HCC lesions. Visual features included tumor size, arterial-phase hyper-enhancement (APHE), washout, lesion segment, mass/mass-like, and capsule presence. Gini-importance method extracted the most important features to prevent over-fitting. Final dataset was split into training(70%), validation(10%), and test dataset(20%). The SVM model was used to train the classifying algorithm. For model validation, 5-fold cross-validation was utilized, and the test data set was used to assess the final accuracy. The area under the curve and receiver operating characteristic curves were used to assess the performance of the classifier model.

Results

For test dataset, the accuracy, sensitivity, and specificity values for classifying benign and HCC lesions were 82%,84%, and 81%, respectively. APHE, washout, tumor size, and mass/mass-like features significantly differentiated benign and HCC lesions with p-value < .001.

Conclusions

The developed classification model employing DCE-MRI features showed significant performance of visual features in classifying benign and HCC lesions. Our study also highlighted the significance of mass and mass-like features in addition to LI-RADS categorization. For future work, this study suggests developing a deep-learning algorithm for automatic lesion segmentation and feature assessment to reduce lesion categorization errors.

Adherence to LI-RADS and EASL high-risk population criteria: A systematic review

BACKGROUND AND AIMS

The Liver Imaging Reporting and Data System (LI-RADS) and European Association for the Study of the Liver (EASL) diagnostic criteria for noninvasive diagnosis of HCC can only be applied to patients at a high risk of HCC. This systematic review assesses adherence to the LI-RADS and EASL high-risk population criteria in published studies.

APPROACH AND RESULTS

PubMed was searched for original research, published between January 2012 and December 2021, reporting LI-RADS and EASL diagnostic criteria on contrast-enhanced ultrasound, CT, or MRI. The algorithm version, publication year, risk status, and etiologies of chronic liver disease were recorded for each study. Adherence to high-risk population criteria was evaluated as optimal (unequivocal adherence), suboptimal (equivocal), or inadequate (clear violation). A total of 219 original studies were included, with 215 that used the LI-RADS criteria, 4 EASL only, and 15 that evaluated both LI-RADS and EASL criteria. Optimal, suboptimal, or inadequate adherence to high-risk population criteria was observed in 111/215 (51.6%), 86/215 (40.0%), and 18/215 (8.4%) LI-RADS and 6/19 (31.6%), 5/19 (26.3%), and 8/19 (42.1%) EASL studies ( p < 0.001) regardless of the imaging modality. Adherence to high-risk population criteria significantly improved according to the CT/MRI LI-RADS versions (optimal in v2018 in 64.5% of studies; v2017, 45.8%; v2014, 24.4%; v2013.1, 33.3%; p < 0.001) and the publication year (2020-2021, 62.5%; 2018-2019, 33.9%; 2014-2017, 39.3% of all LI-RADS studies; p = 0.002). No significant differences in adherence to high-risk population criteria were observed in the versions of contrast-enhanced ultrasound LI-RADS ( p = 0.388) or EASL ( p = 0.293).

CONCLUSION

Adherence to high-risk population criteria was optimal or suboptimal in about 90% and 60% of LI-RADS and EASL studies, respectively.

Strategy for Accurate Diagnosis by Contrast-Enhanced Ultrasound of Focal Liver Lesions in Patients Not at High Risk for Hepatocellular Carcinoma: A Preliminary Study

OBJECTIVE

To develop an effective strategy for accurate diagnosis of focal liver lesions (FLLs) in patients with non-high risk for hepatocellular carcinoma (HCC).

METHODS

From January 2012 to December 2015, consecutive patients with non-high risk for HCC who underwent contrast-enhanced ultrasound (CEUS) were included in this retrospective double-reader study. All patients were stratified into 2 different risks (intermediate, low-risk) groups according to criteria based on clinical characteristics, known as clinical risk stratification criteria. For the intermediate-risk group, the CEUS criteria for identifying benign lesions and HCCs were constructed based on selected CEUS features. The diagnostic performance of the clinical risk stratification criteria, and CEUS criteria for identifying benign lesions and HCCs was evaluated.

RESULTS

This study included 348 FLLs in 348 patients. The sensitivity and specificity of the clinical risk stratification criteria for malignancy was 97.8 and 69.8%. Patients were classified as intermediate risk if they were male, or older than 40 years of age, or HBcAb positive, or having positive tumor markers. Otherwise, patients were classified as low risk. Among the 348 patients, 327 were in the intermediate-risk group and 21 were in the low-risk group. In the intermediate-risk group, the CEUS criteria for identifying benign lesions were any of the following features: 1) hyper/isoenhancement in the arterial phase without washout, 2) nonenhancement in all phases, 3) peripheral discontinuous globular enhancement in the arterial phase, 4) centrifugal enhancement or peripheral enhancement followed by no central enhancement, or 5) enhanced septa. The accuracy, sensitivity, and specificity of the CEUS criteria for identifying benign lesions were 94.5, 83.0, and 99.6%, respectively. Arterial phase hyperenhancement followed by mild and late washout (>60 seconds) was more common in HCC patients than in non-HCC patients (P < .001). Using arterial phase hyperenhancement followed by mild and late washout as the CEUS criteria for identifying HCCs, the sensitivity and specificity were 52.6 and 95.3%, but unfortunately, the positive predictive value was only 82.0%. For the low-risk group, no further analysis was performed due to the small sample size.

CONCLUSIONS

Initial clinical risk stratification followed by assessment of certain CEUS features appears to be a promising strategy for the accurate diagnosis of FLLs in patients not at high risk for HCC.

Modifying LI-RADS on Gadoxetate Disodium-Enhanced MRI: A Secondary Analysis of a Prospective Observational Study

BACKGROUND

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.

PURPOSE

To modify LI-RADS (mLI-RADS) on EOB-MRI.

STUDY TYPE

Secondary analysis of a prospective observational study.

POPULATION

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.

FIELD STRENGTH/SEQUENCE

3.0 T T₂ -weighted fast spin-echo, diffusion-weighted spin-echo based echo-planar, and 3D T₁ -weighted gradient echo sequences.

ASSESSMENT

Three 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.

STATISTICAL TESTS

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.

RESULTS

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%).

DATA CONCLUSION

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.

LEVEL OF EVIDENCE

2 TECHNICAL EFFICACY: Stage 2.

Data-Driven Modification of the LI-RADS Major Feature System on Gadoxetate Disodium-Enhanced MRI: Toward Better Sensitivity and Simplicity

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 <0.05 was considered statistically significant.

RESULTS

The median (interquartile range) size of liver observations was 13 mm (7-27 mm). The diagnostic table for rLI-RADS encompassed 9 cells, as opposed to 16 cells for LI-RADS v2018. In the testing set, compared to LI-RADS v2018, rLI-RADS category 5 demonstrated a significantly superior sensitivity (76% vs. 61%) while maintaining comparably high PPV (92.5% vs. 94.1%, P = 0.126).

DATA CONCLUSION

Compared with LI-RADS v2018, rLI-RADS demonstrated improved simplicity and significantly superior diagnostic sensitivity for HCC in at-risk patients.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.

Artificial intelligence in liver diseases: improving diagnostics, prognostics and response prediction

Clinical routine in hepatology involves the diagnosis and treatment of a wide spectrum of metabolic, infectious, autoimmune and neoplastic diseases. Clinicians integrate qualitative and quantitative information from multiple data sources to make a diagnosis, prognosticate the disease course, and recommend a treatment. In the last 5 years, advances in artificial intelligence (AI), particularly in deep learning, have made it possible to extract clinically relevant information from complex and diverse clinical datasets. In particular, histopathology and radiology image data contain diagnostic, prognostic and predictive information which AI can extract. Ultimately, such AI systems could be implemented in clinical routine as decision support tools. However, in the context of hepatology, this requires further large-scale clinical validation and regulatory approval. Herein, we summarise the state of the art in AI in hepatology with a particular focus on histopathology and radiology data. We present a roadmap for the further development of novel biomarkers in hepatology and outline critical obstacles which need to be overcome.

The role of dynamic and diffusion MR imaging in therapeutic response assessment after microwave ablation of hepatocellular carcinoma using LI-RADS v2018 treatment response algorithm

Background

Hepatocellular carcinoma (HCC) is considered as one of the major causes of morbidity and mortality worldwide. Microwave ablation (MWA) is a widely used treatment option having less morbidity and complications as compared with surgery and liver transplantation. MRI is the most widely used modality in the assessment of treatment response after MWA. Currently, LI-RADS v2018 algorithm is considered the cornerstone in daily clinical practice for assessment of the treatment response after locoregional therapy. The aim of the study was to assess the role of dynamic MRI and diffusion imaging in the assessment of treatment response and detection of tumor viability following microwave ablation therapy of HCC according to LI-RADS v2018 treatment response algorithm.

Results

This retrospective study was performed over 45 HCC lesions underwent MWA as the only therapeutic procedure and followed up by dynamic MRI with diffusion images and then classified according to the LI-RADS treatment response criteria into LR-TR viable and LR-TR nonviable groups. All the malignant lesions found in this study showed arterial phase hyperenhancment (APHE), whether in the early or late arterial phases. Delayed washout was found in all malignant lesions as well. In the diffusion analysis, the mean ADC value for the malignant lesions was 0.900 ± 0.126 × 10-3 mm2/s, while the mean ADC of the treatment-related specific benign parenchymal enhancement was 1.284 ± 0.129 × 10-3 mm2/s with a significant statistical difference in between (P = 0.0001) and a cutoff value of 1.11 × 10-3 mm2/s. Our findings showed that the dynamic MRI has 100% sensitivity, 93.5% specificity, 87.5% PPV, and 100% NPV in the detection of tumoral activity compared with 71.43% sensitivity, 93.55% specificity, 83.33% PPV, and 87.88% NPV for diffusion images.

Conclusion

LI-RADS 2018 provides a treatment response algorithm superior to the previously used assessment criteria. MRI with dynamic contrast-enhanced technique and diffusion imaging provide a powerful tool in the evaluation of treatment response after microwave ablation of hepatocellular carcinoma using the LI-RADS treatment response criteria and is considered a reliable method in differentiating between the recurrent or residual malignant lesions and the posttreatment benign liver changes.

Diagnosis of LI-RADS M lesions on gadoxetate-enhanced MRI: identifying cholangiocarcinoma-containing tumor with serum markers and imaging features

OBJECTIVES

The LI-RADS M (LR-M) category describes hepatic lesions probably or definitely malignant, but not specific for hepatocellular carcinoma in at-risk patients. Differentiation among LR-M entities, particularly detecting cholangiocarcinoma-containing tumors (M-CCs), is essential for treatment and prognosis. Thus, we aimed to develop diagnostic models on gadoxetate disodium-enhanced MRI comprising serum tumor markers and LI-RADS imaging features for M-CC.

METHODS

Consecutive at-risk patients with LR-M lesions exclusively (no co-existing LR-4 and/or LR-5 lesions) were retrieved retrospectively from a prospectively collected database spanning 3 years. Intrahepatic cholangiocarcinoma (ICC) and combined hepatocellular-cholangiocarcinoma (c-HCC-CCA) were classified together as M-CC. LI-RADS features determined by three independent radiologists and clinically relevant serum tumor markers were used to generate M-CC diagnostic models through logistic regression analysis against histology. Per-patient performance was evaluated using area under the receiver operating curve (AUC), sensitivity, and specificity.

RESULTS

Forty-five patients were included, 42.2% (19/45) with hepatocellular carcinoma, 33.3% (15/45) with ICC, 13.3% (6/45) with c-HCC-CCA, and 11.1% (5/45) with other hepatic lesions. Carbohydrate antigen (CA)19-9 > 38 U/mL, α-fetoprotein (AFP) > 4.8 ng/mL, and absence of the LI-RADS feature "blood products in mass" were significant predictors of M-CC. Combining three predictors demonstrated AUC of 0.862, sensitivity of 76%, and specificity of 88%. The risk of M-CC with all three criteria fulfilled was 98% (AUC, 0.690; sensitivity, 38%; specificity, 100%).

CONCLUSIONS

In at-risk patients with LR-M lesions, integrating CA19-9, AFP, and the LI-RADS feature "blood products in mass" achieved high diagnostic performance for M-CC. When all three criteria were fulfilled, the specificity for M-CC was 100%.

KEY POINTS

• In at-risk patients who had LR-M lesions exclusively (no concomitant LR-4/5 lesions), a model with carbohydrate antigen > 38 U/mL, α-fetoprotein > 4.8 ng/mL, and absence of the LI-RADS feature "blood products in mass" achieved high accuracy for diagnosing cholangiocarcinoma-containing tumors. • In patients of whom all three criteria were fulfilled, the specificity for M-CC was 100%, which might reduce or eliminate the need for biopsy confirmation.

LI-RADS v2018 major criteria: Do hepatocellular carcinomas in non-alcoholic steatohepatitis differ from those in virus-induced chronic liver disease on MRI?

PURPOSE

LI-RADS v2018 diagnostic system is used to diagnose hepatocellular carcinoma (HCC) in at risk patients. However, its applicability to HCC in non-alcoholic steatohepatitis (NASH) has not been specifically studied. The purpose of this study was to assess the applicability of LI-RADS v2018 diagnostic system for HCC in patients with NASH.

MATERIALS AND METHODS

The MRI examinations of 41 patients with HCC and NASH (NASH group) were reviewed and compared to those obtained in 41 patients with HCC and virus-induced chronic liver disease (Virus group). MRI examinations of the two groups were compared for imaging presentation, LI-RADS major criteria and LI-RADS categorization. Qualitative variables were compared using Fisher exact test and quantitative variables using Mann-Whitney U test Interreader agreement was assessed using kappa statistic.

RESULTS

No significant differences in qualitative and quantitative variables were observed between the two groups. Most common findings in the two groups were hyperenhancement during the arterial phase and visibility on T2-weighted images (93 % vs. 98 %, P = 0.616 and 85 % vs. 88 %, P = 1.000 for NASH group and Virus group, respectively). No differences in prevalence between the two groups were found for any major LI-RADS v2018 criterion. Interreader agreement for LI-RADS categorization was strong for the NASH group (kappa = 0.802) and moderate for the virus group (kappa = 0.720). No differences were found between the two groups for LI-RADS categories (P =  0.303).

CONCLUSIONS

The LI-RADS v2018 diagnostic algorithm can be applied in patients with NASH.

Diagnostic Performance of LI-RADS Version 2018 for Evaluation of Pediatric Hepatocellular Carcinoma

Background The Liver Imaging Reporting and Data System (LI-RADS) has standardized the evaluation of adult but not pediatric hepatocellular carcinoma (HCC). Purpose To evaluate the performance of LI-RADS version 2018 for diagnosis of pediatric HCC. Materials and Methods This multi-institution retrospective study evaluated all available dynamic CT and/or MRI scans of pediatric (≤18 years) HCC from five institutions between July 2009 and April 2019. The control group included an equal number of other enhancing hepatic lesions. Blinded to final diagnosis, three radiologists independently applied LI-RADS version 2018 criteria. The reference standard was pathologic examination or more than 1 year follow-up. Sensitivity and specificity of LI-RADS were computed using a dichotomous classification of LR-1, LR-2, or LR-3 versus LR-4, LR-5, LR-TIV (tumor in vein), or LR-M (probably or definitely malignant but not HCC-specific) for predicting hepatic malignancy in the entire cohort and in patients at risk for HCC. Results The cohort consisted of 116 children: 58 with HCC (mean age, 12 years ± 5; 31 girls) and 58 with other enhancing hepatic masses (mean age, 12 years ± 5; 42 girls). Frequencies of major criteria in classic HCC for the three readers were as follows: nonrim arterial phase hyperenhancement, 49%-62% (19-24 of 39 patients); nonperipheral "washout," 36%-59% (14-23 of 39 patients); and enhancing "capsule," 28%-38% (11-15 of 39 patients). For the full cohort, the sensitivity of LR-4, LR-5, LR-TIV, or LR-M for malignancy among the three readers ranged from 85% (95% CI: 76, 94) to 88% (95% CI: 80, 96); specificity of LR-1, LR-2, or LR-3 for benignity ranged from 54% (95% CI: 40, 68) to 70% (95% CI: 57, 83). In the at-risk subgroup, sensitivity ranged from 58% (95% CI: 36, 80) to 68% (95% CI: 48, 89); specificity ranged from 56% (95% CI: 37, 74) to 63% (95% CI: 45, 81). All lesions categorized as LR-TIV (n = 10-13) were HCCs. Conclusion Liver Imaging Reporting and Data System version 2018 had moderate sensitivity but low specificity for the diagnosis of pediatric hepatocellular carcinoma (HCC), which had low frequencies of the major criteria used for adult HCC diagnosis. © RSNA, 2021 See also the editorial by Paltiel in this issue.

Diagnostic Performance of LI-RADS Version 2018, LI-RADS Version 2017, and OPTN Criteria for Hepatocellular Carcinoma

OBJECTIVE. Liver Imaging Reporting and Data System (LI-RADS) was updated in 2018 (LI-RADS version 2018 [LI-RADSv2018]) to facilitate integration into the American Association for the Study of Liver Diseases 2018 clinical practice guidelines and involved changes in LR-5 categorization and threshold growth definitions. There are also differences between the criteria for LI-RADSv2018 LR-5 category and the criteria for Organ Procurement and Transplantation Network (OPTN) class 5. The objective of our study was to compare the diagnostic performances of LI-RADSv2018, LI-RADS version 2017 (LI-RADSv2017), and OPTN criteria for diagnosing hepatocellular carcinoma (HCC) on MRI. MATERIALS AND METHODS. In this retrospective study, 122 patients with 159 observations were included who met LI-RADS criteria for at risk for HCC and had at least one hepatic observation on MRI performed between January 1, 2015, and January 1, 2018 and who had histopathology results (n = 104) or follow-up imaging (n = 55) as reference standards. Three abdominal radiologists assigned categories independently and in consensus using LI-RADSv2017, LI-RADSv2018, and OPTN criteria. Diagnostic performance was compared among the guidelines with a generalized estimating equation. RESULTS. Fourteen of 159 (8.8%) observations were assigned a different category according to LI-RADSv2018 compared with LI-RADSv2017. Eight of 31 (25.8%) LR-4 observations using v2017 were recategorized as LR-5 using v2018, and all eight were HCC. Six of 31 (19.4%) LR-4 observations based on v2017 were recategorized as LR-3 using v2018, and all six were non-HCCs. Seven of 114 (6.1%) observations not meeting OPTN class 5 criteria were LR-5 using v2018, and all seven were HCC. Sensitivity for HCC of LR-5 and LR-TIV+5 (i.e., LR-TIV [tumor in vein] definitely due to HCC) categories based on v2018 was significantly higher than that based on v2017 (63.9% vs 55.2%, respectively; p = 0.008) without a difference in specificity (97.3% vs 97.3%; p = 1.00). Sensitivity of LR-5 and LR-TIV+5 in LI-RADSv2018 was significantly higher than the sensitivity of class 5 in OPTN criteria (63.9% vs 53.6%; p = 0.004) without a difference in specificity (97.3% vs 97.3%; p = 1.00). Reader agreement was moderate for overall LIRADSv2017 and LI-RADSv2018 categories (κ = 0.504 and 0.561, respectively); substantial for LR-5 and LR-TIV+5 categories as diagnostic of HCC versus other categories for both v2017 and v2018 (κ = 0.758 and 0.802, respectively); and substantial for OPTN class 5 criteria (κ = 0.756). CONCLUSION. The diagnostic performance of LI-RADSv2018 is higher, with higher sensitivity and similar specificity, than the diagnostic performance of LI-RADSv2017 and OPTN criteria for HCC.

Deep learning assisted differentiation of hepatocellular carcinoma from focal liver lesions: choice of four-phase and three-phase CT imaging protocol

PURPOSE

To evaluate whether a three-phase dynamic contrast-enhanced CT protocol, when combined with a deep learning model, has similar accuracy in differentiating hepatocellular carcinoma (HCC) from other focal liver lesions (FLLs) compared with a four-phase protocol.

METHODS

Three hundred and forty-two patients (mean age 49.1 ± 10.5 years, range 19-86 years, 65.8% male) scanned with a four-phase CT protocol (precontrast, arterial, portal-venous and delayed phases) were retrospectively enrolled. A total of 449 FLLs were categorized into HCC and non-HCC groups based on the best available reference standard. Three convolutional dense networks (CDNs) with the input of four-phase CT images (model A), three-phase images without portal-venous phase (model B) and three-phase images without precontrast phase (model C) were trained on 80% of lesions and evaluated in the other 20% by receiver operating characteristics (ROC) and confusion matrix analysis. The DeLong test was performed to compare the areas under the ROC curves (AUCs) of A with B, B with C, and A with C.

RESULTS

The diagnostic accuracy in differentiating HCC from other FLLs on test sets was 83.3% for model A, 81.1% for model B and 85.6% for model C, and the AUCs were 0.925, 0.862 and 0.920, respectively. The AUCs of models A and C did not differ significantly (p = 0.765), but the AUCs of models A and B (p = 0.038) and of models B and C (p = 0.028) did.

CONCLUSIONS

When combined with a CDN, a three-phase CT protocol without precontrast showed similar diagnostic accuracy as a four-phase protocol in differentiating HCC from other FLLs, suggesting that the multiphase CT protocol for HCC diagnosis might be optimized by removing the precontrast phase to reduce radiation dose.

Liver Imaging Reporting and Data System Version 2018: What Radiologists Need to Know

In this article, we aim to review Liver Imaging Reporting and Data System version 18 (LI-RADS v2018). Hepatocellular carcinoma (HCC) is the most common primary hepatic malignancy. Liver Imaging Reporting and Data System developed for standardizing interpreting, reporting, and data collection of HCC describes 5 major features for accurate HCC diagnosis and several ancillary features, some favoring HCC in particular or malignancy in general and others favoring benignity. Untreated hepatic lesions LI-RADS affords 8 unique categories based on imaging appearance on computed tomography and magnetic resonance imaging, which indicate the possibility of HCC or malignancy with or without tumor in vein. Furthermore, LI-RADS defines 4 treatment response categories for treated HCCs after different locoregional therapy. These continuous recent updates on LI-RADS improve the communication between the radiologists and the clinicians for better management and patient outcome.