Letter to “Preoperative Prediction of Microvascular Invasion by EOB-MRI Imaging for Surgical Decision”

Dear Editor,

We read with great interest the article written by Kim et al. [1] on “Preoperative Prediction of Microvascular Invasion with Gadoxetic Acid-Enhanced Magnetic Resonance Imaging in Patients with Single Hepatocellular Carcinoma: The Implication of Surgical Decision on the Extent of Liver Resection.” They identified preoperative risk factors for microvascular invasion (MVI) in patients with small (≤5 cm) hepatocellular carcinoma (HCC) and analyzed the relationship between the extent of hepatectomy and the recurrence-free survival stratified by the preoperative MVI risk.

Several studies have tried to identify preoperative predictive factors for MVI [2], which represent the aggressiveness of the HCC and are associated with a higher risk of tumor recurrence after hepatectomy. To construct an MVI predictive model, the authors selected 11 items from the preoperative clinicopathologic characteristics and 7 items from gadoxetic acid-enhanced MRI (EOB-MRI) features. EOB-MRI has higher sensitivity [3] and staging capability [4] for diagnosing HCC because its contrast presents hepatic functional imaging at the hepatobiliary phase (HBP). We have a few concerns about the contributing risk predictors. The imaging features adapted in the model were tumor size, non-smooth tumor margin, and arterial peritumoral enhancement, with the latter being more evident on MRI with extracellular contrast than with hepatocyte-specific contrast (e.g., gadoxetic acid). The validation of these parameters for predicting preoperative MVI by MRI with extracellular contrast is an area that requires further investigation.

Additionally, several published risk factors that capable of classifying MVI with an AUROC up to 0.90 [5] were not included in the multivariable model. Previous studies have identified clinical and radiological risk factors from gadoxetic-enhanced MRI, highlighting peritumoral hypo-intensity on the HBP [5], the true diffusion coefficient (D) value, and satellite nodules [6] as independent predictors of MVI status. In this study, it remained unclear why LR-M features and peritumoral HBP low signal intensity, which showed significance in the univariate analysis, were not included in the multivariable analyses for evaluation. Furthermore, the authors might consider validating other published MVI prediction models and examining the relationship between surgical extent and prognosis.

The most novel aspect of this study is the generation of an MVI-risk predictor for stratifying different surgical methods in subgroup analysis. The authors found that the cumulative recurrence rate in the high MVI-risk population was significantly lower in patients receiving major hepatectomy than minor hepatectomy. Because this was a retrospective study, the decision regarding the extent of surgical resection was subject to selection bias; therefore, the baseline characteristics of these 2 patient groups (major and minor resection) should be provided and adjusted in the multivariable analyses. For example, antiviral usage in patients with hepatitis B or C would reduce the risk of recurrence, and this information should be provided. Although the authors found that major hepatectomy provided RFS benefits for the high MVI-risk group, it did not offer overall survival benefits. Consequently, we question whether major hepatectomy in this subgroup is truly better.

The ultimate goal of this study is to determine the optimal extent of surgery to improve overall survival. It is important to consider that MVI may not be the sole or most critical factor influencing prognosis. Prior to developing more accurate preoperative MVI prediction models, it would be beneficial for the authors to analyze survival rates associated with different surgical extents in patients confirmed to have MVI via histopathology. This approach would help identify cases of MVI missed by the current predictive model and assess the impact on survival based on the surgical method employed. Furthermore, the authors can obtain insights into whether recurrence following different surgical methods is attributable to factors other than MVI.

Finally, we would like to congratulate the authors for this publication. This is the first research to validate the predicted preoperative MVI risk for the surgical extension of hepatectomy, which is novel and clinically important. Further prospective validation of this clinical pathway is to be expected.