Objectives To compare the performance of gadoxetic acid-enhanced MRI (EOB-MRI) and extracellular contrast agent-enhanced MRI (ECA-MRI) for diagnosing subcentimeter recurrent HCC. Materials and methods Between January 2017 and December 2023, patients with newly detected suspicious subcentimeter hepatic lesions who underwent both EOB-MRI and ECA-MRI after HCC hepatectomy were retrospectively and consecutively included and divided into training and time-independent test sets. The generalized estimating equation model identified significant MR characteristics in the training set. Diagnostic performances of typical vascular pattern and modified diagnostic algorithms were calculated and compared using the McNemar test in both sets. Results A total of 153 patients (mean age, 55.30 ± 11.09; 118 men) with 185 subcentimeter lesions were included. Typical vascular pattern exhibited comparable but unsatisfactory performance across modalities (both sensitivity < 0.50). In the training set, the optimal modified algorithms for EOB-MRI and ECA-MRI combined all identified significant MR characteristics: T2WI mild-to-moderate hyperintensity, nonperipheral washout (on PVP/TP for EOB-MRI; on PVP/DP for ECA-MRI), and restricted diffusion. Both modified algorithms maintained a comparable high specificity compared to the typical vascular pattern. However, the EOB-MRI-based algorithm demonstrated significantly higher sensitivity than the typical vascular pattern (training set, 0.812 vs. 0.400, p < 0.001; time-independent test set, 0.792 vs. 0.375, p = 0.006), an improvement not observed with the ECA-MRI-based algorithm. Furthermore, EOB-MRI-based modified algorithm demonstrated higher sensitivity than its ECA-MRI counterpart (training set, 0.812 vs. 0.425, p < 0.001; time-independent test set, 0.792 vs. 0.417, p = 0.004), with comparable specificity. Conclusion EOB-MRI with a modified algorithm exhibited superior diagnostic performance for subcentimeter recurrent HCC when compared with ECA-MRI. Critical relevance statement This study provides evidence comparing gadoxetic and extracellular MRI in diagnosing subcentimeter recurrent HCC, highlighting the superior diagnostic performance of EOB-MRI-based modified algorithm and its potential to optimize surveillance strategies and clinical decision-making. Key Points <UnorderedList Mark="Bullet"> <ItemContent> Early detection of subcentimeter recurrent HCC optimizes treatment windows. </ItemContent> <ItemContent> EOB-MRI-based modified algorithm improves sensitivity while maintaining high specificity. </ItemContent> <ItemContent> EOB-MRI-based modified algorithm expedites diagnosing subcentimeter recurrences, ultimately facilitating salvage treatment. </ItemContent> </UnorderedList> Graphical Abstract

Performance comparison of gadoxetic and extracellular MRI in diagnosing subcentimeter recurrent HCC after hepatectomy: a modified algorithm

期刊论文
出版模式：

开放获取
发表日期：

2026年4月8日

查看全文

Yuyao Xiao,
Fei Wu,
Changwu Zhou,
Peng Huang,
Cheng Wang,
Haoran Dai,
Xinyue Liang,
Xi Jia,
Mengsu Zeng

摘要

Objectives

To compare the performance of gadoxetic acid-enhanced MRI (EOB-MRI) and extracellular contrast agent-enhanced MRI (ECA-MRI) for diagnosing subcentimeter recurrent HCC.

Materials and methods

Between January 2017 and December 2023, patients with newly detected suspicious subcentimeter hepatic lesions who underwent both EOB-MRI and ECA-MRI after HCC hepatectomy were retrospectively and consecutively included and divided into training and time-independent test sets. The generalized estimating equation model identified significant MR characteristics in the training set. Diagnostic performances of typical vascular pattern and modified diagnostic algorithms were calculated and compared using the McNemar test in both sets.

Results

A total of 153 patients (mean age, 55.30 ± 11.09; 118 men) with 185 subcentimeter lesions were included. Typical vascular pattern exhibited comparable but unsatisfactory performance across modalities (both sensitivity < 0.50). In the training set, the optimal modified algorithms for EOB-MRI and ECA-MRI combined all identified significant MR characteristics: T2WI mild-to-moderate hyperintensity, nonperipheral washout (on PVP/TP for EOB-MRI; on PVP/DP for ECA-MRI), and restricted diffusion. Both modified algorithms maintained a comparable high specificity compared to the typical vascular pattern. However, the EOB-MRI-based algorithm demonstrated significantly higher sensitivity than the typical vascular pattern (training set, 0.812 vs. 0.400, p < 0.001; time-independent test set, 0.792 vs. 0.375, p = 0.006), an improvement not observed with the ECA-MRI-based algorithm. Furthermore, EOB-MRI-based modified algorithm demonstrated higher sensitivity than its ECA-MRI counterpart (training set, 0.812 vs. 0.425, p < 0.001; time-independent test set, 0.792 vs. 0.417, p = 0.004), with comparable specificity.

Conclusion

EOB-MRI with a modified algorithm exhibited superior diagnostic performance for subcentimeter recurrent HCC when compared with ECA-MRI.

Critical relevance statement

This study provides evidence comparing gadoxetic and extracellular MRI in diagnosing subcentimeter recurrent HCC, highlighting the superior diagnostic performance of EOB-MRI-based modified algorithm and its potential to optimize surveillance strategies and clinical decision-making.

Key Points

Early detection of subcentimeter recurrent HCC optimizes treatment windows.

EOB-MRI-based modified algorithm improves sensitivity while maintaining high specificity.

EOB-MRI-based modified algorithm expedites diagnosing subcentimeter recurrences, ultimately facilitating salvage treatment.

Graphical Abstract