Accessible and valuable imaging features for predicting early recurrence after resection in hepatocellular carcinoma at any stage in Vietnam: a cohort study using extracellular contrast agents

Introduction

Current guidelines for hepatocellular carcinoma (HCC) allow for a definitive diagnosis and treatment decisions based primarily on imaging features, eliminating the need for pathological confirmation (1-3). However, these noninvasive diagnostic strategies have limitations in preoperative prognostication due to a lack of information on biological aggressiveness (4). Biopsy has not been recommended for this purpose because of the risk of tumor seeding and the inability to represent the entire tumor due to HCC’s pathological heterogeneity (2). The choice of imaging modality—whether dynamic computed tomography (CT) or magnetic resonance imaging (MRI)—depends on the availability of imaging resources and the specific needs of each case (2).

A wide range of treatment options is available for HCC, making the selection of an appropriate treatment approach complex (4). Treatment decisions typically consider tumor stage and are further refined to address underlying liver disease, transplant eligibility, and the patient’s functional status (4). To optimize outcomes, clinical decision-making should also incorporate considerations of tumor biology. Liver resection offers a curative treatment strategy for resectable HCCs (2). Despite advancements in surgical techniques, prognosis for HCC patients remains poor, with early recurrence (ER) within the first two years still high at rates of 30% to 50%, which is associated with a worse prognosis than late recurrence (5). This has led to an increase in research focused on improving patient outcomes and addressing these prognostic challenges.

Growing evidence suggested that key pathological and molecular characteristics of HCC can be inferred from imaging features (6-8), providing accessible preoperative clues. Numerous studies have focused on MRI, especially with hepatobiliary contrast agents (8,9), although these are not always available in routine practice. Therefore, this study aimed to investigate the association between preoperative imaging features and ER using both dynamic MRI and CT, to develop a more accessible prognostic model. We present this article in accordance with the STROBE reporting checklist (available at https://amj.amegroups.com/article/view/10.21037/amj-24-155/rc).

Methods

Study design and participants

All patients with pathologically confirmed HCC who underwent curative-intent liver resection from January 2018 to December 2020 were eligible. Inclusion criteria were: (I) adult patients with histopathologically confirmed HCC; (II) patients undergoing anatomical resection with curative intent; and (III) patients with preoperative dynamic MRI or CT performed within one month before surgery. Exclusion criteria included: (I) history of other malignancies; (II) previous anti-tumor treatments, such as transarterial chemoembolization (TACE), targeted therapy, or radiotherapy; (III) two-stage hepatic resections; (IV) extrahepatic recurrence; (V) follow-up period less than two years post-surgery; and (VI) multiple HCCs with distinct imaging characteristics.

CT, MR imaging acquisition protocols and imaging feature analysis

The CT examinations were conducted on either 64-slice or 128-slice systems (SOMATOM^® Definition AS/AS+, Siemens Healthineers, Erlangen, Germany) using a dynamic contrast protocol (10). MRI was performed on either 1.5T (Magnetom Avanto, Siemens Healthcare Limited, Erlangen, Germany) or 3.0T (Magnetom Verio, Siemens Healthcare Limited, Erlangen, Germany) systems, utilizing a phased-array surface coil with 6 channels and dynamic imaging protocols. Extracellular or hepatobiliary contrast agents were administered as appropriate (10). For patients with multiple imaging studies, the most recent examination was used for analysis.

Liver CT or MRI examinations were performed with multiphase images (arterial, venous, and delayed phase) after the administration of intravenous contrast material. The arterial phase was 30–35 seconds after the intravenous contrast injection and the venous phase was 60–70 seconds; the delayed phase was 180 seconds after the intravenous contrast injection.

Two blinded radiologists with over 10 years of experience in liver imaging independently assessed preoperative imaging features on dynamic MRI or CT. Any discrepancies were resolved by consensus. The reviewers were aware that all patients had HCC but were blinded to other clinical, surgical, pathologic, and follow-up information.

The imaging evaluation included 13 features on a per-patient basis: (I) tumor number; (II) tumor growth patterns; (III) tumor size; (IV) non-rim arterial phase hyperenhancement (APHE); (V) non-peripheral washout; (VI) capsule appearance; (VII) Liver Imaging Reporting and Data System (LI-RADS) categories; (VIII) intratumoral arteries; (IX) tumoral necrosis; (X) intratumoral fat; (XI) corona enhancement; (XII) portal vein macro-invasion; and (XIII) biliary invasion (5-7,9,11). Detailed definitions of these imaging features are provided in Table 1.

Patient follow‑up

All patients underwent routine postoperative follow-up at one month, and every two months thereafter, including assessments with serum alpha-fetoprotein (AFP), liver function tests, and abdominal ultrasound. If recurrence was suspected based on serum AFP levels or ultrasound findings, dynamic MRI or CT was performed to confirm the diagnosis. Patients were followed up until either recurrence or September 1, 2023.

Recurrence was defined as the appearance of a new tumor compared to preoperative imaging, meeting one of the following criteria: (I) histologic confirmation of HCC; (II) characteristic imaging features of HCC based on the American Association for the Study of Liver Diseases (AASLD) criteria; (III) characteristic imaging features of tumor-in-vein as per AASLD guidelines; (IV) lipiodol deposition within the tumor following TACE; or (V) imaging appearance similar to the resected tumor. Recurrence was classified as ER if it occurred within two years post-resection (11). In contrast, non-ER was documented if recurrence occurred after two years or if no recurrence was observed during the follow-up period.

Statistical analysis

Data are presented as mean values ± standard deviation for normally distributed variables or as medians and interquartile ranges (IQRs) for non-normally distributed data. Qualitative variables are reported as frequencies and percentages. Statistical analyses were performed using Stata Statistical Software, release 17 (2021, StataCorp LLC, College Station, TX, USA).

The prognostic significance of all imaging features was evaluated using univariate Cox regression analyses. Independent variables with a P value of less than 0.05 in the univariate analysis were subsequently included in a multivariate Cox regression model. Survival outcomes were analyzed using the Kaplan-Meier method.

Ethical considerations

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This retrospective study was conducted at the Ethical Committee of University of Medicine and Pharmacy at Ho Chi Minh City (No. 696/HĐĐĐ-ĐHYD), with approval from the institutional review board. Informed consent is not applicable as this is a retrospective study; the requirement for informed consent was waived.

Results

Patient characteristics

Figure 1 describes the patient selection process.

Figure 1 Flowchart describing the patient selection process. CT, computed tomography; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging.

Detailed demographics and baseline clinical data of enrolled patients are summarized in Table 2.

A total of 214 patients [median age: 58.5 years; IQR, 49–66 years; 170 (79%) male] were included in this study. Of these, 83% (178/214) underwent contrast-enhanced CT, while 17% (36/214) underwent MRI. The median tumor size was 4.4 cm (IQR, 2.8–7.0 cm).

The median follow-up time was 30.5 months (IQR, 9.9–42.6 months). During this period, 43% (92/214) of patients experienced recurrence, with 39% (84/214) classified as ER.

Table 3 describes detailed characteristics of tumor.

Imaging predictors for ER

Thirteen imaging findings associated with ER were analyzed. Of these, ten imaging features were significantly associated with ER in univariable Cox regression analyses (Table 4). These ten features included multiple tumors [HR: 3.4; 95% confidence interval (CI): 2.2–5.3; P<0.001], non-smooth tumor margins (HR: 4.7; 95% CI: 2.4–9.1; P<0.001), tumor size greater than 5 cm (HR: 3.1; 95% CI: 2.0–4.9; P<0.001), non-peripheral washout (HR: 8.5; 95% CI: 1.2–61; P=0.03), LR-M category (HR: 4.3; 95% CI: 2.8–6.8; P<0.001), intratumoral arteries (HR: 3.2; 95% CI: 2.0–5.0; P<0.001), tumoral necrosis (HR: 1.7; 95% CI: 1.1–2.7; P=0.01), corona enhancement (HR: 2.2; 95% CI: 1.5–3.5; P<0.001), portal vein macro-invasion (HR: 3.9; 95% CI: 2.4–6.2; P<0.001), and biliary invasion (HR: 2.3; 95% CI: 1.4–3.9; P=0.002).

In the multivariable Cox regression analysis, four features were independently associated with ER: multiple tumors (HR: 2.1; 95% CI: 1.3–3.5; P=0.004), non-smooth tumor margins (HR: 2.2; 95% CI: 1.0–4.6; P=0.047), LR-M category (HR: 3.3; 95% CI: 2.0–5.4; P<0.001), and intratumoral arteries (HR: 2.9; 95% CI: 1.6–5.0; P<0.001) (Table 4). The four prognostic imaging features associated with ER are illustrated in Figure 2.

Figure 2 Kaplan-Meier curves of the binary prognostic imaging features for early recurrence (≤2 years). The curves based on the number of tumors, non-smooth or ill-defined tumor margins, LR-M category and intratumoral arteries. The red line (multiple lesions/non-smooth margins/LR-M/presence of intratumoral arteries) declines more rapidly than the blue line (single lesion/smooth margins/LR-4 or LR-5/absence of intratumoral arteries), indicating that patients with multiple lesions/non-smooth margins/LR-M/presence of intratumoral arteries HCCs have a significantly shorter recurrence-free survival time compared to those with a single lesion/smooth margins/LR-4 or LR-5/absence of intratumoral arteries HCCs. HCC, hepatocellular carcinoma; LI-RADS (LR), Liver Imaging Reporting and Data System.

Discussion

HCC is a heterogeneous cancer with diverse outcomes (8,12), often showing a high recurrence rate after liver resection (4), especially in the early postoperative period. Imaging has increasingly been recognized as a tool for predicting the prognosis of HCC (5-7,9,11,13). This study aimed to identify easily assessable preoperative imaging features of HCC that could categorize patients as high-risk for ER after resection. Identifying these patients may contribute to more personalized decision-making and potentially improve survival rates.

Although Vietnam is an endemic region for HCC (14,15), research on this topic remains limited. Our study’s data may support future meta-analyses, as we carefully selected and defined variables according to the LI-RADS lexicon. Four imaging features were identified as independent risk factors for ER in HCC patients: multiple tumors, non-smooth or ill-defined margins, LR-M category, and intratumoral arteries.

Tumor multiplicity is a major factor in staging HCC. Multiple foci of HCCs have been reported to correlate positively with microvascular invasion (MVI) (7), which is associated with poor prognosis and ER. Although imaging cannot reliably distinguish between multicentric HCC and intrahepatic metastases in multifocal cases, the presence of multiple foci is still considered a negative prognostic factor (4,13). Contrast to multicentric HCC, intrahepatic metastases or satellite nodules typically present as multifocal lesions with a dominant mass being more common than multiple lesions of equal size. In our study, the presence of multiple non-dominant lesions contributed similarly to both ER and non-ER groups. Thus, we grouped all multiple focal lesions together, regardless of size.

Growth patterns such as non-smooth margin, ill-defined margin, confluent multinodular, or infiltrative type have been associated with the presence of MVI in HCC (7,13). Tumor margins in imaging may also reflect pathological growth behaviors, such as breaching through the tumor capsule or infiltration of tumor cells into the adjacent liver parenchyma (8). This may help explain why HCCs with non-smooth margins tend to exhibit greater aggressiveness and a higher risk of recurrence compared to other forms. Diffuse or infiltrative HCCs were classified under this growth pattern.

Intratumoral arteries have been correlated with increased tumor angiogenesis, a higher risk of MVI, poor tumor differentiation, and the macrotrabecular-massive (MTM) subtype of HCC (7,16). These factors are indicative of biological aggressiveness, unfavorable molecular alterations, and poor prognostic value (5,6).

HCCs classified as LR-M exhibit shorter disease-free survival and overall survival following surgical resection compared to those classified as LR-4 or LR-5, as demonstrated in Eastern cohorts (6,7). Scirrhous HCC and MTM-HCC show a relatively poor prognosis and often present with LR-M features, such as rim-like enhancement and necrosis (7,12). Our findings are consistent with previous studies, highlighting the poorer prognosis associated with LR-M tumors (5). Because LR-M features encompassed both targetoid and non-targetoid appearances, leading to the classification of infiltrative HCCs within this group. Infiltrative HCCs frequently presented with tumor in veins and satellite nodules.

Other imaging features were evaluated in our study but did not demonstrate enough significance for inclusion in the multivariate model. Portal vein macro-invasion and tumor size greater than 5 cm were considered two of the most important risk factors associated with ER. In univariate analysis, our study produced results consistent with those of other studies (17). However, in multivariate analysis, our findings indicated that tumor size and macrovascular invasion were no longer independent prognostic factors for ER. This could be because these factors are associated with other variables. Our study supports the hypothesis that larger tumors and macrovascular invasion are linked to other factors that increase the risk of recurrence, rather than the hypothesis that they have independent prognostic value.

All four predictive features identified in our study were consistent with those reported by Jiang et al. (5), although the HRs in multivariable analyses varied. Additionally, the LR-M category in our study exhibited a higher HR compared to Cannella et al.’s findings (9), potentially due to differences in the definition of recurrence. While both Jiang et al. and our study focused on ER, Cannella et al. included both early and late recurrence.

The four imaging features identified as predictors of ER are readily observable in preoperative imaging, making them feasible to implement widely in clinical practice. By considering these features, clinicians can more accurately predict ER in individual patients, integrate findings into the treatment decision-making process, and potentially adjust therapeutic strategies or intensify postoperative follow-up. High-risk patients for ER may also be candidates for clinical trials of adjuvant therapy, although there is currently no standard adjuvant therapy for surgically treated HCC patients (2,11). Early detection of recurrent tumors is crucial for subsequent therapeutic options, such as TACE or systemic therapy, and occasionally non-anatomical minor resection. This is particularly important since remnant liver volume may not support a second hepatectomy.

Limitations of our study include its single-center design, which may not fully represent the Vietnamese population. Additionally, the proposed model has not been validated internally or externally. Our study does not explore further information on the progression of ER cases or subsequent treatments. Finally, imaging features were collected from both CT and MRI modalities, potentially influencing the frequency of certain features due to modality sensitivity, such as intratumoral fat and necrosis. However, this approach aligns with current guidelines and is more practical for clinical use.

Conclusions

In summary, our study of 214 patients undergoing curative-intent liver resection for HCC identified several preoperative imaging features as significant prognostic factors for predicting ER (≤2 years). These findings offer a valuable opportunity to tailor therapies more precisely to individual patients, advancing beyond the current one-size-fits-all approach in HCC management.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://amj.amegroups.com/article/view/10.21037/amj-24-155/rc

Data Sharing Statement: Available at https://amj.amegroups.com/article/view/10.21037/amj-24-155/dss

Peer Review File: Available at https://amj.amegroups.com/article/view/10.21037/amj-24-155/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://amj.amegroups.com/article/view/10.21037/amj-24-155/coif). The 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. This retrospective study was conducted at the Ethical Committee of University of Medicine and Pharmacy at Ho Chi Minh City (No. 696/HĐĐĐ-ĐHYD), with approval from the institutional review board. Informed consent is not applicable as this is a retrospective study; the requirement for informed consent was waived.

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. Fowler KJ, Chernyak V, Ronot M, et al. Hepatocellular Carcinoma: It Is Time to Focus on Prognosis. Radiology 2023;307:e220884. [Crossref] [PubMed]
2. Singal AG, Llovet JM, Yarchoan M, et al. AASLD Practice Guidance on prevention, diagnosis, and treatment of hepatocellular carcinoma. Hepatology 2023;78:1922-65. [Crossref] [PubMed]
3. Nadarevic T, Giljaca V, Colli A, et al. Computed tomography for the diagnosis of hepatocellular carcinoma in adults with chronic liver disease. Cochrane Database Syst Rev 2021;10:CD013362. [PubMed]
4. Reig M, Forner A, Rimola J, et al. BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. J Hepatol 2022;76:681-93. [Crossref] [PubMed]
5. Jiang H, Qin Y, Wei H, et al. Prognostic MRI features to predict postresection survivals for very early to intermediate stage hepatocellular carcinoma. Eur Radiol 2024;34:3163-82. [Crossref] [PubMed]
6. Wei H, Yang T, Chen J, et al. Prognostic implications of CT/MRI LI-RADS in hepatocellular carcinoma: State of the art and future directions. Liver Int 2022;42:2131-44. [Crossref] [PubMed]
7. Ronot M, Chernyak V, Burgoyne A, et al. Imaging to Predict Prognosis in Hepatocellular Carcinoma: Current and Future Perspectives. Radiology 2023;307:e221429. [Crossref] [PubMed]
8. Fowler KJ, Burgoyne A, Fraum TJ, et al. Pathologic, Molecular, and Prognostic Radiologic Features of Hepatocellular Carcinoma. Radiographics 2021;41:1611-31. [Crossref] [PubMed]
9. Cannella R, Matteini F, Dioguardi Burgio M, et al. Association of LI-RADS and Histopathologic Features with Survival in Patients with Solitary Resected Hepatocellular Carcinoma. Radiology 2024;310:e231160. [Crossref] [PubMed]
10. Chernyak V, Fowler KJ, Kamaya A, et al. Liver Imaging Reporting and Data System (LI-RADS) Version 2018: Imaging of Hepatocellular Carcinoma in At-Risk Patients. Radiology 2018;289:816-30. [Crossref] [PubMed]
11. Chan AWH, Zhong J, Berhane S, et al. Development of pre and post-operative models to predict early recurrence of hepatocellular carcinoma after surgical resection. J Hepatol 2018;69:1284-93. [Crossref] [PubMed]
12. Loy LM, Low HM, Choi JY, et al. Variant Hepatocellular Carcinoma Subtypes According to the 2019 WHO Classification: An Imaging-Focused Review. AJR Am J Roentgenol 2022;219:212-23. [Crossref] [PubMed]
13. Yoneda N, Matsui O, Kobayashi S, et al. Current status of imaging biomarkers predicting the biological nature of hepatocellular carcinoma. Jpn J Radiol 2019;37:191-208. [Crossref] [PubMed]
14. Nguyen-Dinh SH, Do A, Pham TND, et al. High burden of hepatocellular carcinoma and viral hepatitis in Southern and Central Vietnam: Experience of a large tertiary referral center, 2010 to 2016. World J Hepatol 2018;10:116-23. [Crossref] [PubMed]
15. Zhang CH, Cheng Y, Zhang S, et al. Changing epidemiology of hepatocellular carcinoma in Asia. Liver Int 2022;42:2029-41. [Crossref] [PubMed]
16. Rhee H, Cho ES, Nahm JH, et al. Gadoxetic acid-enhanced MRI of macrotrabecular-massive hepatocellular carcinoma and its prognostic implications. J Hepatol 2021;74:109-21. [Crossref] [PubMed]
17. Yan WT, Li C, Yao LQ, et al. Predictors and long-term prognosis of early and late recurrence for patients undergoing hepatic resection of hepatocellular carcinoma: a large-scale multicenter study. Hepatobiliary Surg Nutr 2023;12:155-68. [Crossref] [PubMed]