Preoperative prediction model for macrotrabecular-massive hepatocellular carcinoma based on contrast-enhanced CT and clinical characteristics: a retrospective study

Chutong He, Wanli Zhang, Yue Zhao, Jiamin Li, Ye Wang, Yao Wang, Nianhua Wang, Wenshuang Ding, Xinhua Wei, Ruimeng Yang, Xinqing JiangFrontiers in Oncology 2023 · DOI

Abstract

Combining contrast-enhanced CT (CECT) imaging features and clinical characteristics is an effective method for preoperatively identifying macrotrabecular-massive hepatocellular carcinoma (MTM-HCC), even in early-stage patients. The clinical-radiologic (CR) model developed in this study demonstrates high predictive performance and could potentially guide decision-making regarding aggressive therapies in MTM-HCC patients.

Introduction

Hepatocellular carcinoma (HCC) is a heterogeneous disease with various histological subtypes. Among these, the macrotrabecular-massive (MTM) subtype, recognized by the WHO in 2019, is particularly aggressive, associated with high recurrence rates and poor survival. Currently, definitive diagnosis requires postoperative pathological examination or invasive biopsy. There is a critical clinical need for non-invasive, preoperative methods to identify MTM-HCC using standard diagnostic tools like CT scans and blood markers.

Materials and Methods

This retrospective study analyzed 101 patients with pathology-proven HCC (35 MTM and 66 non-MTM) treated at The Second Affiliated Hospital, South China University of Technology, between January 2017 and November 2021.

Flowchart of the retrospective study showing patient selection and exclusion criteria.

Researchers evaluated three predictive models:

  1. Clinical (C) Model: Utilized blood biomarkers including alpha-fetoprotein (AFP) and hemoglobin (HGB).
  2. Radiologic (R) Model: Based on CT features within the LI-RADS (Liver Imaging Reporting and Data System) framework.
  3. Clinical-Radiologic (CR) Model: A hybrid model combining significant predictors from both clinical and radiologic data.

Results

The combined CR model demonstrated superior predictive performance compared to individual models, achieving an Area Under the Curve (AUC) of 0.793. Notably, in the subgroup of patients with early-stage disease (BCLC 0-A), the model's performance remained robust with an AUC of 0.801.

Multivariate analysis identified four independent predictors for MTM-HCC:

  • High Serum AFP (≥ 228 ng/mL; OR = 4.101)
  • High Hemoglobin (≥ 130.5 g/L; OR = 3.943)
  • Absence of an enhancing capsule (OR = 3.274)
  • Intratumor hypoenhancement (OR = 2.724)

Discussion

The association between high hemoglobin and MTM-HCC is a notable finding, potentially linked to tumor hypoxia-induced erythropoietin expression and angiogenesis. Radiologically, the absence of an enhancing capsule and the presence of intratumor hypoenhancement serve as indicators of tumor invasiveness and internal necrosis, respectively. Identifying these features preoperatively allows for the detection of aggressive subtypes even in early-stage patients, potentially shifting treatment strategies toward more aggressive surgical margins or intensive surveillance.

Conclusion

The CR model provides a non-invasive and effective tool for the preoperative prediction of MTM-HCC. While limited by its retrospective, single-center design and small sample size, the study highlights the potential of combining imaging and clinical data to improve patient stratification and treatment planning. Future research incorporating radiomics or deep learning may further enhance predictive accuracy.

Summary updated

Abstract

Combining CECT imaging features and clinical characteristics is an effective method for preoperatively identifying MTM-HCCs, even in early-stage patients. The CR model has high predictive performance and could potentially help guide decision-making regarding aggressive therapies in MTM-HCC patients.
BibTeX
@inproceedings{he2023preoperative,
  author = {Chutong He and Wanli Zhang and Yue Zhao and Jiamin Li and Ye Wang and Yao Wang and Nianhua Wang and Wenshuang Ding and Xinhua Wei and Ruimeng Yang and Xinqing Jiang},
  title = {{Preoperative prediction model for macrotrabecular-massive hepatocellular carcinoma based on contrast-enhanced CT and clinical characteristics: a retrospective study}},
  booktitle = {Frontiers in Oncology},
  year = {2023},
  doi = {10.3389/fonc.2023.1124069},
}
Figure