Skip to main content
Insights into Imaging
Download PDF
  • Original Article
  • Open access
  • Published:

Predicting histologic differentiation of solitary hepatocellular carcinoma up to 5 cm on gadoxetate disodium-enhanced MRI

Insights into Imaging volume 14, Article number: 3 (2023) Cite this article

Abstract

Background

To establish a preoperative score based on gadoxetate disodium-enhanced magnetic resonance imaging (EOB-MRI) and clinical indicators for predicting histologic differentiation of solitary HCC up to 5 cm.

Methods

From July 2015 to January 2022, consecutive patients with surgically proven solitary HCC measuring ≤ 5 cm at preoperative EOB-MRI were retrospectively enrolled. All MR images were independently evaluated by two radiologists who were blinded to all clinical and pathologic information. Univariate and multivariate logistic regression analyses were performed to identify significant clinicoradiological features associated with poorly differentiated (PD) HCC, which were then incorporated into the predictive score. The predictive score was validated in an independent validation set by area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy.

Results

A total of 182 patients were included, 42 (23%) with PD HCC. According to the multivariate analysis, marked hepatobiliary phase hypointensity (odds ratio [OR], 9.98), LR-M category (OR, 5.60), and serum alpha-fetoprotein (AFP) level > 400 ng/mL (OR, 3.58) were incorporated into the predictive model; the predictive score achieved an AUC of 0.802 and 0.830 on the training and validation sets, respectively. The sensitivity, specificity, and accuracy of the predictive score were 66.7%, 85.7%, and 81.3%, respectively, on the training set and 66.7%, 81.0%, and 77.8%, respectively, on the validation set.

Conclusion

The proposed score integrating two EOB-MRI features and AFP level can accurately predict PD HCC in the preoperative setting.

Key points

  • EOB-MRI features help capture the characteristics of tumor biology and heterogeneity.

  • EOB-MRI-based HCC differentiation score allowed accurate assessment of poor tumor differentiation preoperatively.

  • This scoring system might be useful for prompting tailored treatment selection.

Background

Primary liver cancer is the sixth most common malignancy and the fourth leading cause of cancer-related death globally, and hepatocellular carcinoma (HCC) accounts for around 75–85% of these cases [1]. Liver resection constitutes the backbone for curative treatment of HCC in patients with preserved liver function. However, its application is dampened by high incidence of postoperative recurrence that approaches 60–70% at 5 years [2].

Histologic differentiation is an established prognostic indicator in HCC. In specific, poorly differentiated HCC (PD HCC) has been associated with worse overall survival and/or disease-free survival after liver resection [3], liver transplantation [4], and radiofrequency ablation [5]. Therefore, accurate assessment of tumor differentiation, particularly in the preoperative stage, can help tailor individualized treatment decision-making and improve postoperative survival. Nevertheless, the degree of differentiation is accessible only through histologic examination following invasive biopsy or surgery, which is prone to sampling errors or can only be evaluated postoperatively.

Fortunately, imaging techniques can provide important information regarding tumor biology and heterogeneity in a noninvasive manner. For degree of differentiation, studies have shown that tumor signal intensity in the hepatobiliary phase (HBP) of gadoxetate disodium-enhanced magnetic resonance imaging (EOB-MRI) was associated with histologic grade of HCC [6,7,8,9]. Other imaging features, including the “washout” appearance [10], lower tumor-to-liver enhancement ratio in early arterial phase, absence of “capsule,” arterial phase peritumoral hyperenhancement [11], and peritumoral hypointensity in HBP, were also reported to correlate with moderately differentiated (MD) or PD HCC [9].

Although results so far have been promising, the evidence qualities of previous studies were limited by potentially biased radiology-pathology correlation, as most of these studies did not pose strict restrictions on tumor number and size [7, 9, 11]. Nevertheless, considering the markedly heterogenous nature of HCC (especially in multifocal and/or large tumors) [12, 13], sampling errors may have greatly impacted the results. For example, multifocal HCCs may have distinct degrees of differentiation among tumors, but most prior work only evaluated the dominant tumors. As for first-line curative-intent treatment options, solitary HCC ≤ 5 cm was within the Milan criteria, by considering possible influence of tumor differentiation on prognosis, it may help prompting tailored treatment selection among transplant, locoregional treatment, and resection. Therefore, the imaging indicators of tumor differentiation in solitary HCC ≤ 5 cm remain to be defined.

Thus, this study aimed to establish an easy-to-use score based on preoperative EOB-MRI and clinical indicators for predicting histologic differentiation of solitary HCC up to 5 cm.

Materials and methods

This single-center retrospective study was approved by the institutional review boards at West China Hospital, Sichuan University, and the requirements to obtain written informed consent were waived.

Patients

From July 2015 to January 2022, consecutive patients were enrolled according to the following inclusion criteria: (1) age ≥ 18 years; (2) with history of chronic hepatitis B viral infection and/or liver cirrhosis not attributable to congenital or vascular abnormalities; (3) undergoing EOB-MRI within 1 month prior to liver resection; (4) with surgically proven single HCC measuring ≤ 5 cm at preoperative EOB-MRI; and (5) with complete baseline clinical information (detailed below) within 2 weeks before surgery.

The exclusion criteria were as follows: (1) any previous treatment for HCC; (2) the postoperative pathology report was inadequate for determining tumor differentiation; and (3) the MR images were of insufficient quality (e.g., severe motion artifact) for image analysis.

Baseline preoperative clinical information, including patient demographics, underlying liver diseases, presence or absence of cirrhosis, Child–Pugh classes, and key laboratory data (e.g., alpha-fetoprotein [AFP], carbohydrate antigen 199 [CA199], alanine aminotransferase [ALT], aspartate aminotransferase [AST], total bilirubin [TBIL], albumin [ALB]), was recorded.

Image acquisition

The EOB-MR images were acquired on various 1.5 T or 3.0 T MR systems. The MRI sequences included: T2-weighted imaging (T2WI); diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) maps; in- and opposed-phase T1-weighted imaging; and dynamic T1-weighted imaging before and after injection of gadoxetate disodium (Xianai®, Zhengdatianqing Pharmaceutical Group; Primovist®, Bayer Schering Pharma AG) in the late arterial phase (AP), portal venous phase (PVP), transitional phase (TP), and HBP. Detailed imaging acquisition parameters are shown in Additional file 1: A1 and Table S1.

Image analysis

All de-identified MR images were independently reviewed by two fellowship-trained radiologists (with 5 and 7 years of experiences in liver MRI, respectively). The reviewers were aware that all patients had HCC but were blinded to all clinical, laboratory, and pathologic results. All disagreements between the two reviewers were resolved by a senior fellowship-trained radiologist who had over 20 years of experience in liver MRI.

The reviewers independently evaluated all the lesions regarding the presence/absence/degree of imaging features which had been reported to correlate with the degree of tumor differentiation, tumor burden, and HCC aggressiveness, including: (1) imaging features correlated with tumor differentiation (e.g., HBP signal intensity, degree of diffusion restriction); (2) all Liver Imaging Reporting and Data System (LI-RADS) version 2018 major and ancillary features (except for those related to growth and ultrasound visibility), LR-M features, TIV features, and LI-RADS category [14]; (3) imaging features profiling peritumoral changes (e.g., peritumoral HBP hypointensity [15], peritumoral T2WI hyperintensity, peritumoral biliary ductal dilation); and (4) other reported prognostic imaging features (e.g., intratumoral artery, non-smooth tumor margin, complete “capsule,” tumor growth pattern). The definitions and illustrations of the imaging features are detailed in Additional file 1: Table S2.

Histopathology

Information on tumor differentiation was extracted from routine pathological reports as the reference standard for tumor differentiation, which was recorded by another radiologist without knowing patient’s imaging and clinical data. Based on the standard operation procedure of our institution, a 7-site sampling procedure was performed to ensure adequate and reliable assessment of tumor differentiation. In specific, one piece of tissue was each sampled at the transition area between tumor and surrounding liver tissues at a ratio of 1:1 at 12, 3, 6 and 9 o’clock from the less bleeding and necrotic areas. One piece was sampled within the tumor area free from bleeding and necrosis, and one piece each was sampled, respectively, from proximal (≤ 1 cm to the tumor) and distant liver parenchyma (> 1 cm to the tumor). Tumor differentiation was determined according to the World Health Organization (WHO) criteria [16]. The higher pathological grade was recorded when the evaluated HCC had mixed tumor grades.

Statistical analysis

Continuous variables were presented as mean ± standard deviation or median (interquartile range [IQR]) and analyzed with either the Student’s t test or the Mann–Whitney U test, whereas categorical variables were presented as the numbers of cases (percentages) and analyzed with either the Chi-square test or the Fisher 's exact test, where applicable.

Interobserver agreement between the two radiologists was assessed by calculating the intraclass correlation coefficient (ICC) for continuous imaging features, the Cohen’s κ value for binary imaging features, and the weighted κ value for categorical/ordinal imaging features, respectively.

Development of the predictive score for tumor differentiation

All eligible patients were randomly assigned into a training set and an independent validation set at a ratio of 7:3, while guaranteeing that the distributions of poorly and non-poorly differentiated HCCs between two sets were comparable.

In the training dataset, significant clinical and imaging predictors of PD HCC were selected by univariate logistic regression analysis. Afterward, controlling for patient age, gender, and underlying liver cirrhosis, all predictors with p < 0.1 at univariate logistic regression analysis were fit into a multivariate logistic regression model with backward stepwise method based on fivefold cross-validation, and the best-fit feature combination was obtained using the Akaike Information Criterion. To improve clinical utility and model simplicity, continuous variables were converted to categorical or binary ones according to normal range or clinical relevance. Inter-variable correlations were estimated by the pairwise Spearman’s correlation analysis; when significant collinearity was observed, variables with the largest odds ratio (OR) at univariate logistic regression analysis were kept for further analysis. An HCC differentiation score was developed based on the significant predictors at multivariate logistic regression analysis weighted by their β regression coefficients, with the largest β coefficient scaled as 10 points. The Youden’s index was used to analyze the receiver operating characteristic curve to determine the optimal threshold of the HCC differentiation score for predicting PD HCC.

Validation of the predictive score for tumor differentiation

The discriminative accuracy of the HCC differentiation score was evaluated with area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. Model calibration was assessed by the calibration curve with the Hosmer–Lemeshow test. Decision curve analysis was performed to investigate the clinical usefulness of the HCC differentiation score by quantifying the net benefits at different threshold probabilities.

The R software (version 3.5.1) and SPSS (version 26) were used to perform the statistical analysis. Two- tailed p ≤ 0.05 was considered statistically significant.

Results

Patient characteristics

A total of 182 patients (mean age ± standard deviation, 52.9 ± 11.2; range, 28–75; 145 men) were included in this study, among whom 128 and 54 patients were divided into the training and validation sets, respectively (Fig. 1).

Fig. 1
figure 1

Flowchart of patient selection. EOB-MRI, gadoxetate disodium-enhanced magnetic resonance imaging; HCC, hepatocellular carcinoma

Full size image

Baseline clinical characteristics are presented in Table 1; frequencies of EOB-MRI features are summarized in Table 2; comparisons in baseline clinical characteristics of patients between the training and validation sets are shown in the Additional file 1: Table S3, and interobserver agreement for all imaging features and LI-RADS categories are detailed in Additional file 1: Table S4.

Table 1 Baseline clinical characteristics of patients
Full size table
Table 2 Frequencies of EOB-MRI features
Full size table

On the training set, PD HCC was pathologically confirmed in 30 patients (23.4%; 30/128). No significant differences in baseline clinical parameters were observed between patients with PD HCC and those with well differentiated (WD) or MD HCC (p = 0.103–1.000), except for serum AFP level (p = 0.010). Moreover, rim arterial phase hyperenhancement, marked diffusion restriction, LR-M category and marked HBP hypointensity were significantly more frequent in PD HCC than in WD/MD HCC. Other EOB-MRI features were similar in both groups (p = 0.071–1.000).

On the validation set, PD HCC was pathologically confirmed in 12 patients (22.2%; 12/54). There were no significant differences in baseline clinical variables between patients with PD HCC and those with WD/MD HCC (p = 0.058–1.000), except for PT (p = 0.012). Additionally, the frequency of marked HBP hypointensity was significantly higher in PD HCC compared with WD/MD HCC. Other EOB-MRI features were not significantly different in both groups (p = 0.222–1.000).

Elevated serum carcinoembryonic antigen (> 5 ng/mL) was more frequently observed for patients in the training dataset (7.8% vs. 0%, p = 0.035) than those in the validation dataset. No difference in other baseline clinical characteristics was detected between the training and testing dataset (p = 0.134–0.784).

Development of the HCC differentiation score on the training set

Significant clinical and imaging predictors of PD HCC, including patient demographics, etiology, Child–Pugh stage, tumor markers (i.e., AFP, CA199, CEA), laboratory indexes (e.g., ALT, AST, TBIL), LI-RADS v2018 feature and other imaging features which had been reported to correlate with the degree of tumor differentiation, tumor burden, and HCC aggressiveness (e.g., marked HBP hypointensity, peritumoral hypointensity in HBP, and complete capsule), were analyzed in the univariate analysis. And it identified 6 variables predictive of PD HCC on the training set, including LR-M category (OR, 7.92; p = 0.005), marked HBP hypointensity (OR, 12.38; P < 0.001), serum AFP level > 400 ng/mL (OR, 3.18; p = 0.012), rim arterial phase hyperenhancement (OR, 6.33; p = 0.016), and marked diffusion restriction (OR, 3.26; p = 0.043). According to the multivariate analysis (Table 3), LR-M category (odds ratio [OR], 5.60; p = 0.054), marked hepatobiliary phase hypointensity (OR, 9.98; P < 0.001), and serum AFP level > 400 ng/mL (OR, 3.58; p = 0.021) were incorporated into the HCC differentiation score and are illustrated in Fig. 2. The total score was calculated by adding the individual points of each variable, ranging from 0 to 23 points. According to Youden’s index, the optimal threshold for predicting PD HCC was 6.5 points. Typical patients with PD HCC or MD HCC are shown in Fig. 3.

Table 3 Univariable and multivariable logistic regression analyses of predictors for histological differentiation of hepatocellular carcinoma on training set
Full size table
Fig. 2
figure 2

a Illustration of the HCC differentiation score; b definitions and representative images of the MRI feature associated with the HCC differentiation score; c illustration of LR-M criteria. HCC, hepatocellular carcinoma; AFP, serum alpha-fetoprotein; HBP, hepatobiliary phase; LR, Liver Imaging Reporting and Data System

Full size image
Fig. 3
figure 3

Typical cases of poorly and moderately differentiated hepatocellular carcinoma (HCC). (a) A 61-year-old man with poorly differentiated HCC. (A) Rim hyperenhancement is visible on the arterial phase (arrow). (B) The portal venous phase shows an incomplete enhancing tumor capsule (arrow). (C) The signal intensity of the lesion on the hepatobiliary phase is less than vessels, showing marked hypointensity, and peritumoral hypointensity is also seen on the hepatobiliary phase (arrow). (D) Axial T2-weighted imaging shows a hyperintense signal lesion (arrow). (E) Diffusion weighted imaging shows a marked-restricted-diffusion lesion (arrow). (F) Photomicrograph shows the poorly differentiated HCC (hematoxylin–eosin stain; original magnification, × 100). (b) A 31-year-old man with moderately differentiated HCC. (A) Non-rim hyperenhancement is visible on the arterial phase (arrow). (B) The tumor present washout on the portal venous phase (arrow). (C)The lesion shows mild hypointensity on the hepatobiliary phase (arrow). (D) Axial T2-weighted imaging shows a hyperintense signal lesion (arrow). (E) Diffusion-weighted imaging shows a marked-restricted-diffusion lesion (arrow). (F) Photomicrograph shows the moderately differentiated HCC (hematoxylin–eosin stain; original magnification, 100)

Full size image

Score assessment and validation

On the training set, the HCC differentiation score exhibited an AUC of 0.802 (0.703–0.900) with sensitivity, specificity, PPV, NPV, and accuracy of 66.7%, 85.7%, 58.8%, 89.4%, 81.3%, respectively. Calibration plot demonstrated a good consistency between the score-predicted probabilities and the actual PD HCC estimates on the training set (p = 0.883; Additional file 1: Fig. S1a). Decision curves revealed that the predictive score provided a larger net benefit than that of assuming all patients had PD HCC when the threshold probability was greater than 0.1 on the training set (Additional file 1: Fig. S2a).

On the validation set, the predictive score yielded an AUC of 0.830 (0.693–0.966) with sensitivity, specificity, PPV, NPV and accuracy of 66.7%, 81.0%, 50.0%, 89.5%, 77.8%, respectively (Table 4). Calibration plot showed that the score-predicted probabilities were closely consistent with the actual PD HCC estimates on the validation set (p = 0.963; Additional file 1: Fig. S1b). In terms of the clinical utility, the predictive score exhibited a larger net benefit than that of assuming all patients had PD HCC when the threshold probability was greater than 0.08 on the validation set (Additional file 1: Fig. S2b).

Table 4 Diagnostic performance of models for predicting histological differentiation of hepatocellular carcinoma
Full size table

Discussion

In patients with surgically-confirmed solitary HCC ≤ 5 cm, we found that two EOB-MR imaging features (marked hepatobiliary phase hypointensity and the LR-M category) along with serum AFP > 400 ng/mL were significantly associated with poor tumor differentiation. Based on these indicators, we developed and validated an easy-to-use scoring system which allowed accurate preoperative assessment of tumor differentiation.

Marked HBP hypointensity, defined as tumor signal intensity in the HBP lower than that of liver and similar to or lower than that of intrahepatic vessels, was present in 43 (24%) patients and significantly associated with poor tumor differentiation. This finding was in line with previous studies, in which lower tumor signal intensity in the HBP was associated with worse differentiation [7, 8]. One possible explanation for this outcome is that the expressions of organic anion transporters on the cell membrane, which are responsible for the uptake of gadoxetic disodium and signal intensity in the HBP, decrease gradually as the tumor de-differentiates [17]. However, the assessments of HBP signal intensity were mostly performed via quantitative analyses based on manual regions of interest placement in prior works [7, 8]. Despite allowing accurate quantifications of signal intensities, this approach may suffer from suboptimal generalizability, as signal intensity in the HBP is largely affected by vendor and acquisition parameter variations. In contrast, using intrahepatic vessels as a reference, we explored a reproducible and effective semi-quantitative way of measuring relative tumor signal intensity on HBP images, which could aid in preoperative evaluation of HCC differentiation.

The LR-M category is defined as lesions that are probably or definitely malignant but not HCC specific. Any targetoid appearances or nontargetoid mass with one or more features (including infiltrative appearance, marked diffusion restriction, necrosis or severe ischemia, and other feature that suggests non-HCC malignant), were considered to be sufficient for LR-M categorization [14]. And it was another imaging indicator of PD HCC. Our results were in consistent with the study of Shin et al., who found that the LR-M HCCs had poorer histologic differentiation than those LR-4/5 tumors [18]. In addition to worse differentiation, the LR-M category has also been reported to correlate with other aggressive pathomolecular characteristics of HCC, including increased stemness features (i.e., expression of CK19) [19], more pronounced macrotrabecular pattern, more frequent microvascular invasion and sinusoid-like microvascular pattern, and a more hypoxic and fibrotic tumor microenvironment [20]. As a result, LR-M category was widely-proved as an independent risk factor for worse prognosis in HCC [18, 21].

Serum AFP level > 400 ng/mL was the only laboratory indicator significantly associated with poor tumor differentiation in our study, which was in line with existing evidence. As the most established biomarker in HCC, higher serum AFP levels have been associated with increased tumor aggressiveness (e.g., worse tumor differentiation [22], increased incidence of microvascular invasion [23], macrotrabecular-massive subtype [20, 24], proliferative HCC [25]), and worse posttreatment prognosis [26,27,28]. Our work further confirmed its utility, alone and in combination with other imaging features, in identifying PD HCC.

Our findings had important clinical implications. Methodologically, considering marked intra- and intertumoral heterogeneity in HCC, satisfactory radiology–pathology correlation was achieved via the rigorous multi-site sampling method as we only enrolled patients with solitary tumors measuring ≤ 5 cm. Of note, identification of poorly-differentiated elements in small tumors might have been more challenging compared with tumors > 5 cm since these components are less frequently found in smaller HCCs [29]. However, our model still demonstrated AUCs over 0.80 on both the training and testing datasets. Moreover, all three indicators included in the final scoring system were identified as significantly associated with PD HCC in three out of the five folds of cross-validation during modeling. These outcomes further confirmed the robustness and effectiveness of our findings. In terms of clinical utility, all patients included in the current study were within the Milan criteria, of whom ablation, resection, and transplant are matched first-line curative-intent treatment options. In this context, the HCC differentiation score could serve as an effective therapeutic decision-making tool. Specifically, during the preoperative work-ups, our findings may help informing neoadjuvant treatment [30], as well as prompting resection over radiofrequency ablation even for small tumors less than 2 cm [31] in patients with PD HCCs according to the scoring system.

Several limitations of the current study should be noted. First, as a single-institutional retrospective study with relatively small sample size, no external validation was available to assess the performance and generalizability of our HCC differentiation score. Second, most of our enrolled patients had chronic hepatitis B virus infection. However, HCCs developed in non-HBV background (more prevalent in Western countries) harbor distinct pathomolecular characteristics from those in HBV patients; thus, our findings mandate further validations in HCC patients free of HBV infections. Third, aiming to propose an easy-to-use and interpretable scoring system for routine clinical adoptions, we only analyzed qualitative/semi-quantitative imaging indicators without consideration of complex and high-dimensional quantitative analyses. Another limitation is the relative subjectivity when evaluating imaging features, and the interreader agreement of some imaging features is poor in our work. Besides, our results are only applicable to solitary HCC less than 5 cm. Finally, we did not perform survival analysis as the prognostic role of tumor differentiation in HCC has been well established [29, 32]. However, further large-scale multi-center studies enrolling patients with more variegated etiologies of underlying liver diseases and adequate patient follow-up are warranted to validate and refine our findings.

In conclusion, in patients with solitary HCC ≤ 5 cm, we constructed and validated an easy-to-use HCC differentiation score based on two EOB-MR imaging features (marked HBP hypointensity and the LR-M category) and serum AFP level. This scoring system allowed accurate assessment of poor tumor differentiation in the preoperative setting and thus might be useful for prompting neo-adjuvant therapy and tailored treatment selection among resection, transplant, and ablation.

Availability of data and materials

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

ADC:

Apparent diffusion coefficient

AFP:

Alpha-fetoprotein

ALB:

Albumin

ALT:

Alanine aminotransferase

AP:

Arterial phase

AST:

Aspartate aminotransferase

AUC:

Area under the receiver operating characteristic curve

CA199:

Carbohydrate antigen 199

DWI:

Diffusion-weighted imaging

EOB-MRI:

Gadoxetate disodium-enhanced magnetic resonance imaging

HBP:

Hepatobilliary phase

HCC:

Hepatocellular carcinoma

ICC:

Intraclass correlation coefficient

IQR:

Interquartile range

LI-RADS/LR:

Liver imaging reporting and data system

MD:

Moderately differentiated

NPV:

Negative predictive value

OR:

Odds ratio

PD:

Poorly differentiated

PPV:

Positive predictive value

PVP:

Portal venous phase

T2WI:

T2-weighted imaging

TBIL:

Total bilirubin

TP:

Transitional phase

WD:

Well differentiated

References

  1. Sung H, Ferlay J, Siegel RL et al (2021) Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 71(3):209–24

    Article  Google Scholar 

  2. Kulik L, El-Serag HB (2019) Epidemiology and management of hepatocellular carcinoma. Gastroenterology 156(2):477–491

    Article  Google Scholar 

  3. Ruff SM, Rothermel LD, Diggs LP et al (2020) Tumor grade may be used to select patients with multifocal hepatocellular carcinoma for resection. HPB (Oxford) 22(7):1004–1010

    Article  Google Scholar 

  4. Decaens T, Roudot-Thoraval F, Badran H et al (2011) Impact of tumour differentiation to select patients before liver transplantation for hepatocellular carcinoma. Liver Int 31(6):792–801

    Article  Google Scholar 

  5. Lu Z, Sun Z, Liu C et al (2021) Prognostic nomogram for hepatocellular carcinoma with radiofrequency ablation: a retrospective cohort study. BMC Cancer 21(1):751

    Article  CAS  Google Scholar 

  6. Choi JW, Lee JM, Kim SJ et al (2013) Hepatocellular carcinoma: imaging patterns on gadoxetic acid-enhanced MR Images and their value as an imaging biomarker. Radiology 267(3):776–786

    Article  Google Scholar 

  7. Haimerl M, Utpatel K, Götz A et al (2021) Quantification of contrast agent uptake in the hepatobiliary phase helps to differentiate hepatocellular carcinoma grade. Sci Rep 11(1):22991

    Article  CAS  Google Scholar 

  8. Peng Z, Jiang M, Cai H et al (2016) Gd-EOB-DTPA-enhanced magnetic resonance imaging combined with T1 mapping predicts the degree of differentiation in hepatocellular carcinoma. BMC Cancer 16:625

    Article  Google Scholar 

  9. Huang K, Dong Z, Cai H et al (2019) Imaging biomarkers for well and moderate hepatocellular carcinoma: preoperative magnetic resonance image and histopathological correlation. BMC Cancer 19(1):364

    Article  Google Scholar 

  10. Witjes CDM, Willemssen FEJA, Verheij J et al (2012) Histological differentiation grade and microvascular invasion of hepatocellular carcinoma predicted by dynamic contrast-enhanced MRI. J Magn Reson Imaging 36(3):641–647

    Article  Google Scholar 

  11. Rong D, Liu W, Kuang S et al (2021) Preoperative prediction of pathologic grade of HCC on gadobenate dimeglumine-enhanced dynamic MRI. Eur Radiol 31(10):7584–7593

    Article  CAS  Google Scholar 

  12. Dagogo-Jack I, Shaw AT (2018) Tumour heterogeneity and resistance to cancer therapies. Nat Rev Clin Oncol 15(2):81–94

    Article  CAS  Google Scholar 

  13. Friemel J, Rechsteiner M, Frick L et al (2015) Intratumor heterogeneity in hepatocellular carcinoma. Clin Cancer Res 21(8):1951–1961

    Article  CAS  Google Scholar 

  14. CT/MRI liver imaging reporting and data system version 2018. American College of Radiology Web site. https://www.acr.org/Clinical-Resources/Reporting-andData-Systems/LI-RADS/CTMRI-LI-RADS-v2018. Accessed 1 Dec 2018.

  15. Chou Y, Lao I, Hsieh P et al (2019) Gadoxetic acid-enhanced magnetic resonance imaging can predict the pathologic stage of solitary hepatocellular carcinoma. World J Gastroenterol 25(21):2636–2649

    Article  Google Scholar 

  16. WHO Classification of tumors: digestive system tumours. 5th edn. Lyon, France. Available https://whobluebooks.iarc.fr/publications/index.php. Accessed 19 Jan 2021

  17. Ringe KI, Husarik DB, Sirlin CB, Merkle EM (2010) Gadoxetate disodium-enhanced MRI of the liver: part 1, protocol optimization and lesion appearance in the noncirrhotic liver. AJR Am J Roentgenol 195(1):13–28

    Article  Google Scholar 

  18. Shin J, Lee S, Kim SS et al (2021) Characteristics and early recurrence of hepatocellular carcinomas categorized asLR-M: comparison with those categorized as LR-4 or 5. J Magn Reson Imaging 54(5):1446–1454

    Article  Google Scholar 

  19. Choi SY, Kim SH, Park CK et al (2018) Imaging features of gadoxetic acid-enhanced and diffusion-weighted MR imaging for identifying cytokeratin 19-positive hepatocellular carcinoma: a retrospective observational study. Radiology 286(3):897–908

    Article  Google Scholar 

  20. Feng Z, Li H, Zhao H et al (2021) Preoperative CT for characterization of aggressive macrotrabecular-massive subtype and vessels that encapsulate tumor clusters pattern in hepatocellular carcinoma. Radiology 300(1):219–229

    Article  Google Scholar 

  21. An C, Park S, Chung YE et al (2017) Curative resection of single primary hepatic malignancy: liver imaging reporting and data system category LR-M portends a worse prognosis. AJR Am J Roentgenol 209(3):576–583

    Article  Google Scholar 

  22. Montal R, Andreu-Oller C, Bassaganyas L et al (2019) Molecular portrait of high alpha-fetoprotein in hepatocellular carcinoma: implications for biomarker-driven clinical trials. Br J Cancer 121(4):340–343

    Article  CAS  Google Scholar 

  23. Jiang H, Wei J, Fu F et al (2022) Predicting microvascular invasion in hepatocellular carcinoma: a dual-institution study on gadoxetate disodium-enhanced MRI. Liver Int 42(5):1158–1172

    Article  CAS  Google Scholar 

  24. Mule S, Galletto PA, Tenenhaus A et al (2020) Multiphase liver MRI for identifying the macrotrabecular-massive subtype of hepatocellular carcinoma. Radiology 295(3):562–571

    Article  Google Scholar 

  25. Kang HJ, Kim H, Lee DH et al (2021) Gadoxetate-enhanced MRI features of proliferative hepatocellular carcinoma are prognostic after surgery. Radiology 300(3):572–582

    Article  Google Scholar 

  26. Ji GW, Zhu FP, Xu Q et al (2020) Radiomic features at contrast-enhanced CT Predict recurrence in early stage hepatocellular carcinoma: a multi-institutional study. Radiology 294(3):568–579

    Article  Google Scholar 

  27. Mehta N, Heimbach J, Harnois DM et al (2017) Validation of a risk estimation of tumor recurrence after transplant (RETREAT) score for hepatocellular carcinoma recurrence after liver transplant. JAMA Oncol 3(4):493–500

    Article  Google Scholar 

  28. Zhu AX, Kang YK, Yen CJ et al (2019) Ramucirumab after sorafenib in patients with advanced hepatocellular carcinoma and increased α-fetoprotein concentrations (REACH-2): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 20(2):282–296

    Article  CAS  Google Scholar 

  29. Shinkawa H, Tanaka S, Kabata D et al (2021) The prognostic impact of tumor differentiation on recurrence and survival after resection of hepatocellular carcinoma is dependent on tumor size. Liver Cancer 10(5):461–472

    Article  CAS  Google Scholar 

  30. Akateh C, Black SM, Conteh L et al (2019) Neoadjuvant and adjuvant treatment strategies for hepatocellular carcinoma. World J Gastroenterol 25(28):3704–3721

    Article  Google Scholar 

  31. Heimbach JK, Kulik LM, Finn RS et al (2018) AASLD guidelines for the treatment of hepatocellular carcinoma. Hepatology 67(1):358–380

    Article  Google Scholar 

  32. Liao S, Su T, Jeng Y et al (2019) Clinical manifestations and outcomes of patients with sarcomatoid hepatocellular carcinoma. Hepatology 69(1):209–221

    Article  Google Scholar 

Download references

Funding

This work was supported by the National Natural Science Foundation of China (Grant Nos. 82101997, 81971571) and the Science and Technology Department of Sichuan Province (Grant Nos. 2021YFS0021, 2021YFS0141).

Author information

Author notes

  1. Ting Yang and Hong Wei contributed equally to this work

Authors and Affiliations

  1. Department of Radiology, West China Hospital, Sichuan University, No. 37 Guoxue Alley, Chengdu, 610041, China

    Ting Yang, Hong Wei, Yun Qin, Jie Chen, Hanyu Jiang & Bin Song

  2. Big Data Research Center, University of Electronic Science and Technology of China, Chengdu, 610000, Sichuan, China

    Yuanan Wu

  3. Department of Radiology, Sanya People’s Hospital, Sanya, Hainan, China

    Bin Song

Authors
  1. Ting Yang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Hong Wei
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yuanan Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yun Qin
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jie Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Hanyu Jiang
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Bin Song
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

HW and HJ contributed to the study conception and design. YQ and JC performed the material preparation and TY collected data. YW interpreted data and performed statistical analysis. The first draft of the manuscript was written by TY and then substantively revised by HW under the supervision of BS and HJ, and all authors read and approved the final manuscript.

Corresponding authors

Correspondence to Hanyu Jiang or Bin Song.

Ethics declarations

Ethics approval and consent to participate

The study had ethics and institutional approvals in place.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1. 

Supplementary materials.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yang, T., Wei, H., Wu, Y. et al. Predicting histologic differentiation of solitary hepatocellular carcinoma up to 5 cm on gadoxetate disodium-enhanced MRI. Insights Imaging 14, 3 (2023). https://doi.org/10.1186/s13244-022-01354-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13244-022-01354-w

Keywords