Long-term outcomes of indeterminate focal hepatic observations less than 20 mm followed up with gadoxetic acid-enhanced magnetic resonance imaging (Gd-EOB-DTPA-MRI)

Introduction

Hepatocellular carcinoma (HCC), a primary malignant tumor, presents a significant health concern globally, with alarmingly high incidence and mortality rates (1,2). Although the clinical treatments for HCC have improved significantly, the 5-year survival rate of patients with HCC remains low because of the high likelihood of recurrence and metastasis (3,4). Therefore, early detection, diagnosis, and personalized management strategies are crucial for improving patient survival rates (5). Recently, gadoxetic acid-enhanced magnetic resonance imaging (Gd-EOB-MRI) has emerged as the preferred imaging technique for the early detection of HCC (6,7). This cutting-edge technology has demonstrated exceptional accuracy, sensitivity, and specificity, leading to a notably more timely and accurate diagnosis of liver cancer and thereby improving patient outcomes (8).

The Liver Imaging Reporting and Data System (LI-RADS), introduced by the American College of Radiology, was developed to establish a standardized imaging-based diagnosis of hepatic observations in at-risk individuals. The system categorizes these observations from LR-1 (“definitely benign”) to LR-5 (“definitely HCC”) (9). Among these categories, LR-3 indicates an intermediate probability for HCC that is associated with a baseline risk of 38% [95% confidence interval (CI): 31–45%] for HCC and 40% (95% CI: 31–50%) for malignancy (9-11). According to LI-RADS v2018, repeat or alternative diagnostic imaging within 3–6 months is recommended for patients with LR-3 observations (9,12). However, previous longitudinal studies have shown a wide variation in the progression rate from LR-3 to higher categories, ranging from 6% to 55.6% (13-23). This variability poses challenges to standardization of monitoring strategies for both radiologists and referring clinicians, as it creates uncertainty regarding the rate of malignancy associated with LR-3 observations and their progression. Moreover, to our knowledge, only a limited number of studies have assessed the potential risk factors for HCC progression using LR-3 observations (22,23).

This study aimed to retrospectively analyze the long-term outcomes of high-risk patients with LR-3 observations (based on LI-RADS v2018) sized at less than 20 mm via serial Gd-EOB-MRI scans, including cumulative incidence and significant predictors of progression. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-302/rc).

Methods

Management of the sample

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This single-center study was approved by the institutional ethics committee of the Third Affiliated Hospital of Nantong University (No. EK2024049). The requirement for individual consent was waived due to the retrospective nature of the analysis. The study analyzed the electronic medical records of 1,857 patients with high-risk factors for HCC as identified by histopathology, clinical history, or imaging findings, collected at the Third Affiliated Hospital of Nantong University between January 2016 and December 2022. These patients underwent Gd-EOB-MRI at both the initial diagnosis and follow-up between July 2015 and January 2022. The LI-RADS v2018 was used in this study, including all feature definitions and primary nodule classification schemes (9). An abdominal radiologist (F.X., with 10 years of experience in liver MRI) reviewed associated medical records and MRI reports to identify magnetic resonance (MR) images that corresponding with LR-3 observations. The inclusion criteria were as follows: (I) patients with cirrhosis with at least one LR-3 observation reported on index Gd-EOB-MRI; (II) index nodule size <20 mm and a maximum of three observations assessed in each patient to mitigate potential clustering bias; (III) at least one additional Gd-EOB-MRI follow-up with regular intervals of 6 months or less, and (IV) observations upgraded to LR-4 or LR-5 underwent histopathological confirmation of HCC. In contrast, non-progressive observations (LR remained stable or downgraded) were monitored for at least one year. The exclusion criteria were as follows: (I) Child-Pugh class C liver function, because hepatic dysfunction can impede adequate quality of HBP images; (II) local-regional therapy for hepatic observations between the baseline and follow-up examination; (III) a previous history of treatment for HCC before the index MRI examination, to minimize confounding effects on natural progression of disease; and (IV) unavailable or insufficient imaging quality. Because partial observations were identified using previous versions of LI-RADS (e.g., v2014 and v2017), the reviewer reinterpreted all index MRI examinations and recategorized all index observations according to LI-RADS v2018. A flowchart of the patient selection in this study is shown in Figure 1.

Figure 1 Flowchart of the study population. The flowchart illustrates the study selection process on the basis of LI-RADS v2018 categorization of observations included in this study. There is an overlap of some patients who had multiple LR-3 observations during follow-up. Specifically, these patients had both progressing and non-progressing nodules, and thus were included in both the progression group and non-progression based on lesion-level analysis. Gd-EOB-MRI, gadoxetic acid-enhanced magnetic resonance imaging; HCC, hepatocellular carcinoma; LI-RADS, Liver Imaging Reporting and Data System; LR-3, Liver Imaging Reporting and Data System category 3; MRI, magnetic resonance imaging.

Demographic and clinical-related variables (age, sex, laboratory markers, and clinical risk stratification scores, including the Child-Pugh score and Model for End-Stage Liver Disease Score) for each patient were obtained through a review of the electronic medical records.

MRI procedure

All multiphase Gd-EOB-MRI examinations were performed in accordance with LI-RADS technical recommendations and standards. MRI data were acquired using a 3.0-T Intera Achieva or Ingenia MR scanner (Philips Healthcare, Best, The Netherlands) with a 16-channel phased-array torso coil. The index images included T1-weighted turbo field-echo in-phase and opposed-phase images, breath-hold multishot T2-weighted images, and respiratory-triggered T2-weighted images. Diffusion-weighted imaging (DWI) was performed using a respiratory-triggered single-shot echo-planar imaging sequence with b values of 0 and 800 s/mm². For the dynamic enhancement sequences, Gd-EOB-DTPA (Eovist, Bayer Healthcare Pharmaceuticals, Leverkusen, Germany) was administered intravenously with an automated contrast injector at a dose of 0.025 mmol/kg and a rate of 1.0 mL/s, followed by 20 mL of a 0.9% saline chaser at the same rate. A series of single-phase acquisitions were performed during each of the dynamic phases: unenhanced phase, late hepatic arterial phase (LAP; initiated at approximately 20–35 seconds post-injection), portal venous phase (PVP; 60 seconds), transitional phase (TP; 3 minutes), and hepatobiliary phase (HBP; 15–20 minutes), each acquired within approximately 14 seconds using a T1-weighted 3D turbo field-echo sequence [T1 high-resolution isotropic volume examination (THRIVE)]. The acquisition parameters are listed in Table S1.

Image analysis and LI-RADS categorization

Two abdominal radiologists (T.Z. and F.X., with 15 and 10 years of experience in liver MRI, respectively) independently and randomly reviewed and recorded the index Gd-EOB-MRI examinations. They assessed the major imaging features of LI-RADS, including size, non-rim arterial phase hyperenhancement (APHE), non-peripheral washout, and enhancing capsule. Additionally, they evaluated ancillary features (AFs) that suggested malignancy, such as restricted diffusion, mild-to-moderate T2 hyperintensity, TP hypointensity, and HBP hypointensity. The radiologists also assigned LI-RADS categorization based on their assessments. The last available follow-up imaging features were also reviewed, including all the aforementioned features as well as newly incorporated ones, specifically threshold growth (≥50% size increase of a mass in ≤6 months) as a major feature and subthreshold growth (unequivocal size increase of a mass, less than threshold growth) as an AF. Throughout the follow-up period, the reviewers documented the final LI-RADS category and category modifications, which included upgraded (recategorized as LR-4 or LR-5), remained stable (remained within the LR-3), or downgraded (recategorized as LR-2, LR-1, or resolved), and subdivided them into two groups: the progression group (category upgraded) and the non-progression group (category remained stable or downgraded). In addition, the interval between the initial and final examinations was recorded. Any discrepancies between the two readers were resolved by an additional abdominal radiologist (J.L., with 20 years of experience in liver MRI). The index and final observations were categorized based on major features according to the LI-RADS v2018 criteria.

Statistical analysis

All statistical analyses were performed using SPSS software version 27.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism software version 10.1.2 (Dotmatics, Boston, MA, USA). To compare the variables between the progression group and non-progression group in terms of patient demographics and the imaging features of LR-3 observations, we used the unpaired t-test or Mann-Whitney test for continuous variables and the Pearson chi-squared test or Fisher exact test for categoric variables. To identify potential predictors of progression (upgraded to LR-4 or LR-5), both univariate and multivariate Cox regression analyses were conducted. Variables with P<0.05 in the univariable analyses were included in the multivariable analysis. The overall cumulative incidence curve for progression for patients with LR-3 observations was estimated using the Kaplan-Meier method. The cumulative risk of progression and median interval to LR category upgrade were analyzed in the presence and absence of significant predictive risk factors, and the curve differences were compared using the log-rank test. Inter-reader agreement was determined via the intraclass correlation coefficient (ICC) for continuous variables (size on index and size change on follow-up) and via kappa coefficients for categorical variables (imaging features and LI-RADS categorization). Coefficients <0.21, 0.21–0.40, 0.41–0.60, 0.61–0.80, and 0.81–1.00 indicated poor, fair, moderate, good, and excellent agreement, respectively (24). Statistical significance was defined as two-sided P values less than 0.05 at a confidence level of 95%.

Results

Study cohort

Our final study population consisted of 125 high-risk patients with a total of 149 untreated LR-3 observations (<20 mm). The initial demographic data and LI-RADS imaging features on index MRI are summarized in Tables 1,2. Significant differences between the progression and non-progression groups were observed in the presence of non-rim APHE (P=0.02), mild-to-moderate T2 hyperintensity (P=0.01), and restricted diffusion (P=0.04). Inter-reader agreement for categorical variables, including imaging features (excluding size) and LI-RADS categorization, ranged from substantial to almost perfect (κ =0.73–1.00). For continuous variables, the agreement was excellent, with ICCs of 0.95 for initial observation size and 0.91 for follow-up size change.

Clinical follow-up data

The category transitions of the 149 index LR-3 observations on follow-up images are categorized and summarized in Figure 2. The median follow-up interval for all observations was 18.3 months (range, 2.7–78.5 months). Overall, 41.6% (62/149) of the patients were upgraded to LR-4 (n=15) or LR-5 (n=47) (Figures 3,4), 49.0% (73/149) remained stable, and 9.4% (14/149) were downgraded to LR-2 (n=9) or LR-1 (n=5). The median times of progression and non-progression were 13.8 months (range, 2.9–70.3 months) and 19.2 months (range, 12.0–78.5 months), respectively. In addition, 95.2% (59/62) of the patients showed progression with a median increase in size of 11 mm (range, 2–40 mm), with 16.9% (10/59) showing threshold growth (Figure 3) and 83.1% (49/59) showing subthreshold growth (Figure 4).

Figure 2 Diagram illustrating the changes in LI-RADS v2018 categories during follow-up (from index to final examination) for LR-3 observations included in this study. LI-RADS, Liver Imaging Reporting and Data System; LR-1, Liver Imaging Reporting and Data System category 1; LR-2, Liver Imaging Reporting and Data System category 2; LR-3, Liver Imaging Reporting and Data System category 3; LR-4, Liver Imaging Reporting and Data System category 4; LR-5, Liver Imaging Reporting and Data System category 5.

Figure 3 The LR-3 observation progressed to LR-5 based on LI-RADS v2018 during follow-up. (A-C) The index Gd-EOB-MRI showed an 8-mm non-rim APHE on the late arterial phase in hepatic segment II/III (A, arrow), without non-peripheral washout and enhancing capsule on the portal venous phase (B, arrow) or transitional phase (not shown), and hypointensity on HBP (C, arrow). The observation was initially categorized as LR-3. (D-F) After a 5-month Gd-EOB-MRI follow-up, the observation increased from 8 to 15 mm with non-rim APHE (D, arrow), with non-peripheral washout on the portal venous phase (E, arrow) and hypointensity on HBP (F, arrow). Due to the combination of imaging features (10–19 mm, non-rim APHE, non-peripheral washout, and threshold growth), the observation was consequently categorized as LR-5. APHE, arterial phase hyperenhancement; Gd-EOB-MRI, gadoxetic acid-enhanced magnetic resonance imaging; HBP, hepatobiliary phase; LI-RADS, Liver Imaging Reporting and Data System; LR-3, Liver Imaging Reporting and Data System category 3; LR-5, Liver Imaging Reporting and Data System category 5.

Figure 4 The LR-3 observation progressed to LR-5 based on LI-RADS v2018 during follow-up. (A-C) Baseline Gd-EOB-MRI showed a 15-mm observation without non-rim APHE on the late arterial phase (A, arrow), without non-peripheral washout on the portal venous phase (B, arrow) or transitional phase (not shown), with hypointensity on HBP in hepatic segment V/VI (C, arrow). The observation was initially categorized as LR-3. (D-F) After a 17-month Gd-EOB-MRI follow-up, the observation increased from 15 to 19 mm with non-rim APHE (D, arrow), with non-peripheral washout on the portal venous phase (E, arrow) and hypointensity on HBP (F, arrow). Due to the combination of imaging features (10–19 mm, non-rim APHE, non-peripheral washout, and subthreshold growth), the observation was consequently categorized as LR-5. APHE, arterial phase hyperenhancement; Gd-EOB-MRI, gadoxetic acid-enhanced magnetic resonance imaging; HBP, hepatobiliary phase; LI-RADS, Liver Imaging Reporting and Data System; LR-3, Liver Imaging Reporting and Data System category 3; LR-5, Liver Imaging Reporting and Data System category 5.

Risk factors predicting progression

The results of the univariate and multivariate Cox proportional hazards analyses are summarized in Table 3. In the univariate analysis, non-rim APHE (P<0.001), subthreshold growth (P=0.01), mild-to-moderate T2 hyperintensity (P<0.001), and restricted diffusion (P<0.001) were significant risk factors for progression. In the multivariate analysis, non-rim APHE [hazard ratio (HR) =2.19; 95% confidence interval (CI): 1.27–3.79; P=0.005], subthreshold growth (HR =2.78; 95% CI: 1.48–5.23; P=0.001), and mild-to-moderate T2 hyperintensity (HR =5.25; 95% CI: 2.05–13.43; P<0.001) showed significantly independent associations with progression.

Cumulative risk of progression and median time to upgrade

The results of the time to upgrade and cumulative risk analyses are summarized in Table 4, and the associated Kaplan-Meier curves are shown in Figure 5. The overall cumulative incidence of progression (upgraded to LR-4 or LR-5) for LR-3 observations was 41.6%, with 1.3%, 9.5%, 17.3%, and 37.3% at 3, 6, 12, and 24 months, respectively (Figure 5A). LR-3 observations with non-rim APHE (Figure 5B) or mild-to-moderate T2 hyperintensity (Figure 5C) demonstrated a significantly higher cumulative risk of progression (53.3% vs. 33.7% and 50.0% vs. 28.8%, respectively; both P<0.001) and a shorter median interval to LR category upgrade (14.7 vs. 18.9 months and 15.1 vs. 26.5 months, respectively; both P<0.001) compared to those without these features. In contrast, the presence of subthreshold growth did not significantly affect these outcomes (53.3% vs. 30.1%, P=0.056; 17.6 vs. 18.7 months, P=0.57) (Figure 5D).

Figure 5 Kaplan-Meier analysis of the cumulative incidence of progression. (A) Overall cumulative incidence of progression of LR-3 observations. (B) Comparison of observations with or without non-rim APHE. (C) Comparison of observations with or without mild-to-moderate T2 hyperintensity. (D) Comparison of observations with or without subthreshold growth. The cumulative risk rates were significantly higher in the presence of non-rim APHE (log-rank test P<0.001) or mild-to-moderate T2 hyperintensity (log-rank test P<0.001) but were not significantly affected by the presence of subthreshold growth (log-rank test P=0.056). APHE, arterial phase hyperenhancement; LR-3, Liver Imaging Reporting and Data System category 3; LR-4, Liver Imaging Reporting and Data System category 4; LR-5, Liver Imaging Reporting and Data System category 5.

Discussion

LR-3 observations correspond mainly to a heterogeneous group of hepatic lesions, including benign lesions, dysplastic nodules, early-stage HCCs, and hypervascularized HCCs (9). Therefore, relying solely on imaging features to accurately predict the natural outcomes of LR-3 observations can be challenging. Moreover, variations in methodologies across studies have led to heterogeneity in the findings, making it difficult to interpret pooled estimates of progression (25-28). For example, studies by Choi et al. and Tanabe et al. reported low progression rates (6–9%) from LR-3 to LR-4 or LR-5, with most observations either remaining stable or being downgraded during follow-up (15,16). In contrast, Vernuccio et al. focused on LR-3 observations exhibiting APHE⁺ ≥10 mm and reported a higher progression rate (55.6%) to LR-5 (20). Our study showed a consistent linear increase in the cumulative risk of LR-3 progression to LR-4 or higher with longer follow-up times. The risk was 1.3% at 3 months, increasing to 9.4% at 6 months, 17.5% at 12 months, and 37.3% at 24 months, aligning with recent studies (17). Compared to previous studies (16,17), our study had a higher reliability and diagnostic consistency for the LR category and imaging features based on Gd-EOB-DTPA, potentially offering more clinically accurate assessments of LR-3 progression and better guiding clinical decision-making. While many LR-3 observations warrant repeat or alternative imaging within the recommended 3- to 6-month surveillance interval (28), it is important to note that certain lesions, particularly hypervascular HCCs, may progress to LR-5 in a shorter timeframe (20). These findings underscore the need for improved risk stratification strategies, incorporating both imaging features and demographic characteristics, to better identify individuals at elevated risk of progression form LR-3 observations. These tools could enable more tailored surveillance strategies, improving the overall effectiveness of surveillance procedures and enhancing patient outcomes (17,29).

In this study, we identified certain features of LR-3 sized at less than 20 mm that were significantly associated with an increased risk of progression in high-risk patients. Specifically, the presence of mild-to-moderate T2 hyperintensity leading to categorization as LR-3 was associated with the highest risk among the features analyzed (HR =5.25). This finding is consistent with previous studies demonstrating that T2 hyperintensity levels correlate with factors such as HCC size, degree of hypervascularity, growth rate, and progression from dysplastic nodules to HCC (21). Although T2 hyperintensity is also a known risk factor for growth and subsequent hypervascularization in hypovascular nodules (30,31), it is not specific to HCC and can be observed in other malignancies. Therefore, it remains unclear whether T2 hyperintensities alone should prompt changes in management. Additionally, the presence of non-rim APHE was identified as another significant imaging feature. Observations with non-rim APHE showed a higher cumulative incidence of progression (50.0% vs. 28.8%) and a shorter median time to progression (15.1 vs. 26.5 months) compared to observations without non-rim APHE. These findings align with previous studies. For instance, Agnello et al. reported a progression rate of 31% from LR-3 without non-rim APHE to LR-5 during follow-up (13), while Vernuccio et al. observed a relatively higher progression rate to LR-5 (55.6%) in LR-3 observations with non-rim APHE and a size of ≥10 mm (20). This apparent discrepancy may be explained by differences in the underlying hepatocarcinogenic processes. Non-rim APHE observations likely indicate that hepatocarcinogenesis has already initiated (20). In contrast, non-hypervascular hypointense nodules (NHHNs) may represent precancerous lesions (e.g., high-grade dysplastic nodules or early-stage HCCs) with less aggressive biological features, requiring an additional window of hypervascularization before progressing to a higher stage (27,32).

In our study, subthreshold growth emerged as a significant predictor of an increased risk of upgrade (HR =2.78). However, it may not necessarily have a major impact on the cumulative risk of progression or the median time to upgrade. Subthreshold growth refers to the observable growth of a mass that falls short of meeting the threshold for significant growth. In LI-RADS v2018, the definition of threshold growth has been simplified to include only a ≥50% increase in size of a mass within ≤6 months. Previous definitions, such as a ≥100% size increase in >6 months or a new observation of ≥10 mm within 24 months, are now classified as subthreshold growth. According to one study (33), subthreshold growth as a stand-alone feature exhibited a sensitivity of 48% and a specificity of 91% for diagnosing HCC. In our study, the majority (83.1%, 49/59) of LR-3 observations less than 20 mm that were upgraded to a higher category displayed subthreshold growth. Although retrospective and prospective studies are lacking, indirect evidence and biological plausibility suggest that subthreshold growth increases the likelihood of malignancy (34). Previous studies have suggested that hypointensity on the HBP may predict progression to HCC in LR-3 observations (20,35). However, in our study, HBP hypointensity was not a significant predictor of progression. This may be due to the fact that the majority of LR-3 observations (98.7%, 147/149) in our study were already hypointense on HBP images. This likely improved the detection rate of non-hypervascular hepatocellular neoplasms and effectively minimized false positives from hyperenhancing pseudolesions or perfusion alterations (31,36). Our findings suggest that the identified imaging features and clinical factors may offer greater clinical precision in understanding the natural history of LR-3 observations and predicting the risk of malignant progression.

This study has several limitations. First, the analysis was limited to patients who underwent follow-up imaging, potentially introducing selection bias, as treated or lost to follow up cases were excluded. Nonetheless, the focus on untreated patients still provides meaningful insights for guiding surveillance strategies. Second, the retrospective design resulted in heterogeneity in follow-up intervals and durations, complicating the determination of optimal monitoring timelines. Third, histopathological confirmation was not performed, consistent with LI-RADS guidelines for categorizing indeterminate observations based on imaging features alone. Fourth, only a subset of AF was assessed, constrained by protocol availability and prior evidence, which may have limited the identification of additional predictors. Finally, the single-center cohort consisted exclusively of Chinese patients with hepatitis B, potentially limiting generalizability to other populations. Further prospective, multicenter studies with standardized follow-up protocols and broader patient cohorts are needed to validate these findings and establish evidence-based recommendations.

Conclusions

Our findings suggest that several factors are associated with an increased risk of LR-3 observations (<20 mm) progressing to higher categories. Notably, the presence of non-rim APHE, subthreshold growth, and mild-to-moderate T2 hyperintensity emerged as significant predictors of progression. LR-3 observations exhibiting non-rim APHE or mild-to-moderate T2 hyperintensity were particularly associated with a higher cumulative risk of progression and a shorter median time for upgrading. These results underscore the importance of considering these imaging features when assessing the likelihood of LR-3 observations progressing to more advanced categories.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-302/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-302/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-302/prf

Funding: This work was supported by the Clinical Medicine Special Research Fund Project of Nantong University (grant No. 2023JQ016 to T.Z.), the Nantong Health Commission Project (grant No. MS2023078 to T.Z.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-302/coif). F.X. reports funding support from the Nantong University Special Research Fund for Clinical Medicine (No. EK2021017), outside the submitted work. S.M. reports funding support from Funding of University “Qinglan Project” in Jiangsu Province and Teaching Reform Research Project of Nantong University in 2024 (No. 2024E05), outside the submitted work. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Institutional Ethics Committees of Third Affiliated Hospital of Nantong University (No. EK2024049). The requirement for individual consent was waived due to the retrospective nature of the analysis.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Xia C, Chen Y, Zhu Y, et al. Identification of DLAT as a potential therapeutic target via a novel cuproptosis-related gene signature for the prediction of liver cancer prognosis. J Gastrointest Oncol 2024;15:2230-51. [Crossref] [PubMed]
3. Maluccio M, Covey A. Recent progress in understanding, diagnosing, and treating hepatocellular carcinoma. CA Cancer J Clin 2012;62:394-9. [Crossref] [PubMed]
4. Xu XF, Xing H, Han J, et al. Risk Factors, Patterns, and Outcomes of Late Recurrence After Liver Resection for Hepatocellular Carcinoma: A Multicenter Study From China. JAMA Surg 2019;154:209-17. [Crossref] [PubMed]
5. Koza A, Bhogal RH, Fotiadis N, et al. The Role of Ablative Techniques in the Management of Hepatocellular Carcinoma: Indications and Outcomes. Biomedicines 2023;11:1062. [Crossref] [PubMed]
6. Murakami T, Sofue K, Hori M. Diagnosis of Hepatocellular Carcinoma Using Gd-EOB-DTPA MR Imaging. Magn Reson Med Sci 2022;21:168-81. [Crossref] [PubMed]
7. Singal AG, Zhang E, Narasimman M, et al. HCC surveillance improves early detection, curative treatment receipt, and survival in patients with cirrhosis: A meta-analysis. J Hepatol 2022;77:128-39. [Crossref] [PubMed]
8. Huang P, Wu F, Hou K, et al. Diagnostic algorithm for subcentimeter hepatocellular carcinoma using alpha-fetoprotein and imaging features on gadoxetic acid-enhanced MRI. Eur Radiol 2024;34:2271-82. [Crossref] [PubMed]
9. American College of Radiology. CT/MRI LI-RADS v2018 CORE. 2018. Available online:
10. van der Pol CB, Lim CS, Sirlin CB, et al. Accuracy of the Liver Imaging Reporting and Data System in Computed Tomography and Magnetic Resonance Image Analysis of Hepatocellular Carcinoma or Overall Malignancy-A Systematic Review. Gastroenterology 2019;156:976-86. [Crossref] [PubMed]
11. van der Pol CB, McInnes MDF, Salameh JP, et al. CT/MRI and CEUS LI-RADS Major Features Association with Hepatocellular Carcinoma: Individual Patient Data Meta-Analysis. Radiology 2022;302:326-35. [Crossref] [PubMed]
12. Mitchell DG, Bashir MR, Sirlin CB. Management implications and outcomes of LI-RADS-2, -3, -4, and -M category observations. Abdom Radiol (NY) 2018;43:143-8. [Crossref] [PubMed]
13. Agnello F, Albano D, Sparacia G, et al. Outcome of LR-3 and LR-4 observations without arterial phase hyperenhancement at Gd-EOB-DTPA-enhanced MRI follow-up. Clin Imaging 2020;68:169-74. [Crossref] [PubMed]
14. Ranathunga D, Osman H, Islam N, et al. Progression Rates of LR-2 and LR-3 Observations on MRI to Higher LI-RADS Categories in Patients at High Risk of Hepatocellular Carcinoma: A Retrospective Study. AJR Am J Roentgenol 2022;218:462-70. [Crossref] [PubMed]
15. Choi JY, Cho HC, Sun M, et al. Indeterminate observations (liver imaging reporting and data system category 3) on MRI in the cirrhotic liver: fate and clinical implications. AJR Am J Roentgenol 2013;201:993-1001. [Crossref] [PubMed]
16. Tanabe M, Kanki A, Wolfson T, et al. Imaging Outcomes of Liver Imaging Reporting and Data System Version 2014 Category 2, 3, and 4 Observations Detected at CT and MR Imaging. Radiology 2016;281:129-39. [Crossref] [PubMed]
17. Xing F, Zhang T, Miao X, et al. Long-term evolution of LR-2, LR-3 and LR-4 observations in HBV-related cirrhosis based on LI-RADS v2018 using gadoxetic acid-enhanced MRI. Abdom Radiol (NY) 2023;48:3703-13. [Crossref] [PubMed]
18. Cannella R, Vernuccio F, Celsa C, et al. Long-term evolution of LI-RADS observations in HCV-related cirrhosis treated with direct-acting antivirals. Liver Int 2021;41:2179-88. [Crossref] [PubMed]
19. Kim BJ, Choi SH, Kim SY, et al. Liver Imaging Reporting and Data System categories: Long-term imaging outcomes in a prospective surveillance cohort. Liver Int 2022;42:1648-57. [Crossref] [PubMed]
20. Vernuccio F, Cannella R, Choudhury KR, et al. Hepatobiliary phase hypointensity predicts progression to hepatocellular carcinoma for intermediate-high risk observations, but not time to progression. Eur J Radiol 2020;128:109018. [Crossref] [PubMed]
21. Cerny M, Chernyak V, Olivié D, et al. LI-RADS Version 2018 Ancillary Features at MRI. Radiographics 2018;38:1973-2001. [Crossref] [PubMed]
22. Arvind A, Joshi S, Zaki T, et al. Risk of Hepatocellular Carcinoma in Patients With Indeterminate (LI-RADS 3) Liver Observations. Clin Gastroenterol Hepatol 2023;21:1091-1093.e3. [Crossref] [PubMed]
23. Shropshire E, Mamidipalli A, Wolfson T, et al. LI-RADS ancillary feature prediction of longitudinal category changes in LR-3 observations: an exploratory study. Abdom Radiol (NY) 2020;45:3092-102. [Crossref] [PubMed]
24. Kundel HL, Polansky M. Measurement of observer agreement. Radiology 2003;228:303-8. [Crossref] [PubMed]
25. Kim TH, Kim SY, Tang A, et al. Comparison of international guidelines for noninvasive diagnosis of hepatocellular carcinoma: 2018 update. Clin Mol Hepatol 2019;25:245-63. [Crossref] [PubMed]
26. Kim BR, Lee JM, Lee DH, et al. Diagnostic Performance of Gadoxetic Acid-enhanced Liver MR Imaging versus Multidetector CT in the Detection of Dysplastic Nodules and Early Hepatocellular Carcinoma. Radiology 2017;285:134-46. [Crossref] [PubMed]
27. Kim TH, Woo S, Han S, et al. Risk Factors for Hypervascularization in Hepatobiliary Phase Hypointense Nodules without Arterial Phase Hyperenhancement: A Systematic Review and Meta-analysis. Acad Radiol 2022;29:198-210. [Crossref] [PubMed]
28. Hussain HK. Hepatobiliary-Phase Hypointense Nodules without Arterial Phase Hyperenhancement: Time to Act. Acad Radiol 2022;29:211-2. [Crossref] [PubMed]
29. Mitropoulou G, Kompoura V, Saffioti F, et al. The Role of Matrix Metalloproteinases in Liver Function and Disease. Front Biosci (Landmark Ed) 2025;30:27127. [Crossref] [PubMed]
30. Öcal O, Zech CJ, Fabritius MP, et al. Non-hypervascular hepatobiliary phase hypointense lesions detected in patients with hepatocellular carcinoma: a post hoc analysis of SORAMIC trial to identify risk factors for progression. Eur Radiol 2023;33:493-500. [Crossref] [PubMed]
31. Bae JS, Lee JY, Suh KS, et al. Characterization of hepatobiliary phase hypointense nodules without arterial phase hyperenhancement on gadoxetic acid-enhanced MRI via contrast-enhanced ultrasound using perfluorobutane. Abdom Radiol (NY) 2023;48:2321-30. [Crossref] [PubMed]
32. Brandi N, Renzulli M. Liver Lesions at Risk of Transformation into Hepatocellular Carcinoma in Cirrhotic Patients: Hepatobiliary Phase Hypointense Nodules without Arterial Phase Hyperenhancement. J Clin Transl Hepatol 2024;12:100-12. [Crossref] [PubMed]
33. Joo I, Kim SY, Kang TW, et al. Radiologic-Pathologic Correlation of Hepatobiliary Phase Hypointense Nodules without Arterial Phase Hyperenhancement at Gadoxetic Acid-enhanced MRI: A Multicenter Study. Radiology 2020;296:335-45. [Crossref] [PubMed]
34. Cerny M, Bergeron C, Billiard JS, et al. LI-RADS for MR Imaging Diagnosis of Hepatocellular Carcinoma: Performance of Major and Ancillary Features. Radiology 2018;288:118-28. [Crossref] [PubMed]
35. Cannella R, Vernuccio F, Sagreiya H, et al. Liver Imaging Reporting and Data System (LI-RADS) v2018: diagnostic value of ancillary features favoring malignancy in hypervascular observations ≥ 10 mm at intermediate (LR-3) and high probability (LR-4) for hepatocellular carcinoma. Eur Radiol 2020;30:3770-81. [Crossref] [PubMed]
36. Li N, Wang Y, Yang L, et al. Hepatobiliary-Phase Hypointense Nodules Without Arterial-Phase Hyperenhancement: Developing a Risk Stratification for Hypervascular Transformation Based on a Real-World Observational Cohort Study. Technol Cancer Res Treat 2024;23:15330338241299003. [Crossref] [PubMed]

(English Language Editor: J. Gray)