LR is considered a viable treatment option for patients with HCC and CSPH [6], but its impact on survival remains controversial. Previous studies have suggested that LR can achieve acceptable outcomes in carefully selected patients with CSPH [21, 22]; however, most did not stratify patients based on tumor burden and liver function in detail. To address this limitation, we categorized patients using the JIS-score into three groups: very-early stage (JIS 0), early stage (JIS 1), and intermediate stage (JIS 2–4). This stratification revealed that CSPH was significantly associated with poorer survival in the very-early and intermediate-stage groups, whereas early stage HCC patients with CSPH had comparable outcomes to those without CSPH. These findings suggest that patients who have early stage HCC with CSPH may be suitable candidates for LR, while those with very-early or intermediate stages require more cautious consideration.

Several studies comparing LR outcomes in CSPH and non-CSPH patients have generally reported worse outcomes in the CSPH group [7, 8]. Bruix et al. [2] concluded that CSPH is a risk factor for postoperative liver dysfunction and worse 5-year survival even with preserved liver function. However, recent studies have shown that CSPH does not affect the prognosis of patients with early stage HCC [9, 10]. For instance, a prospective study found that laparoscopic LR in patients who had BCLC stage 0 or A with CSPH was safe and provided outcomes comparable to those of patients without CSPH [23]. These conflicting conclusions may arise from differences in patient demographics, small sample sizes, and insufficient adjustment for confounding factors. Two meta-analyses found that CSPH significantly affected OS after LR but also reported high heterogeneity among studies [3, 24]. Given this controversy, our study aimed to clarify the stage-specific impact of CSPH on surgical outcomes. Using one of the largest sample sizes to date, we stratified patients with HCC by JIS-score while adjusting for potential confounders. These methodological strengths enhance the robustness of our findings and provide valuable insights for future research.

Interestingly, CSPH had a less pronounced impact on OS in early stage HCC, which may be due to several factors. First, oncological factors appeared to be more favorable in the CSPH group. The JIS-score is determined by tumor size and Child–Pugh grade [12], and although Child–Pugh B was more common in patients with CSPH (24% vs. 2.2%, p < 0.001), tumors were smaller in diameter (2.3 vs. 3.2 cm, p < 0.001). Because of this oncological factor, non-anatomical resection and MIS were more frequently performed in the CSPH group, resulting in shorter operative durations and less blood loss. In patients with CSPH who are more likely to be eligible for MIS, a less-invasive procedure may have influenced short-term outcomes. Previous studies have reported that MIS is associated with better short-term outcomes, potentially contributing to comparable 90-day mortality rates and postoperative liver failure incidence between the CSPH and non-CSPH groups, despite patients with CSPH having worse liver function (Child–Pugh grade, MELD score, and ALBI score) [23, 25]. In this study, a subgroup analysis of open versus MIS in patients with CSPH also showed significantly better short-term outcomes with MIS (Supplementary Table S2). Second, unlike the very-early stage, where the T factor is fixed at 1, patients with early and intermediate-stage HCC include a broader range of tumor sizes and liver function statuses. This heterogeneity may explain why the negative prognostic impact of CSPH appeared diluted in the early stage group, in contrast to the very-early and intermediate stages, where liver function tends to be either consistently preserved or consistently impaired. Notably, 24% of patients with early stage CSPH were classified as Child–Pugh B and 52% were ALBI grade 2, compared with 65% and 80% in the intermediate-stage CSPH group, respectively. Although ALBI grades were significantly worse in the CSPH group than in the non-CSPH group for both early and intermediate stages, we accounted for these differences in our multivariate analysis. These findings suggest that within the same prognostic group, LR for early stage HCC may attenuate the adverse impact of CSPH—an effect not observed in intermediate-stage HCC. Additionally, despite similar post-recurrence treatments across all stages, deaths due to postoperative liver decompensation were not significantly increased in patients with early stage CSPH. Careful patient selection and surgical approach in early stage HCC may allow LR to balance surgical and hepatic risks, reducing the prognostic impact of CSPH. Conversely, patients with very-early stage HCC generally have preserved liver function (Child–Pugh and ALBI grades) and small tumors, leading to similar surgical approaches in both CSPH and non-CSPH groups. However, postoperative liver failure was more frequent in patients with CSPH, and mortality related to liver decompensation was significantly higher. In very-early stage HCC, non-anatomical and minimally invasive resections—surgical approaches associated with a lower physiological burden—were performed equally in both the CSPH and non-CSPH groups. However, these approaches may not have been sufficient to offset the adverse effects of portal hypertension on liver function in the CSPH group. Therefore, although short-term outcomes did not differ significantly, liver failure—primarily Grade A—appeared to be more frequent in the CSPH group. Additionally, patients with CSPH tended to have a higher prevalence of cirrhosis, although this difference was not statistically significant. We believe that the increased mortality in the CSPH group likely reflects the broader negative impact of CSPH on the natural progression of cirrhosis. In intermediate-stage HCC, as in early stage HCC, non-anatomical and minimally invasive approaches were more commonly used in the CSPH group. In the early and intermediate stages, wound infections and pneumonia—less severe complications that can prolong hospital stays—were more common in the non-CSPH group. Interestingly, despite higher rates of major complications and liver failure in patients with CSPH, those in the intermediate stage had shorter hospital stays. This paradox may be explained by the more frequent use of minimally invasive surgery in patients with CSPH in both the early and intermediate stages, as minimally invasive surgery is known to shorten hospital stays following LR. Nevertheless, short-term outcomes, including postoperative liver failure and 90-day mortality, were worse in the CSPH group, and more patients died of liver decompensation in the long term. These findings suggest that in intermediate-stage HCC, the impact of LR on residual liver function is greater in patients with CSPH, making it challenging to balance oncological benefits with hepatic risks. Future research is needed to develop new criteria for surgery in patients who have intermediate-stage HCC with CSPH, ensuring that good outcomes can be expected.

To further validate the impact of CSPH, we analyzed postoperative OS stratified by the BCLC staging system (very-early: BCLC-0; early: BCLC-A; intermediate: BCLC-B) [6]. The results were consistent with those from the JIS-score analysis (Supplementary Tables S3–5), confirming that CSPH was an independent risk factor for OS in very-early and intermediate-stage patients but not in early stage patients, reinforcing the minimal impact of CSPH on OS in early stage HCC. The JIS-score has been reported to be superior to BCLC staging in predicting prognosis, because BCLC staging encompasses a broad spectrum of patients within each category, particularly in the early stage group (BCLC-A) [11]. Therefore, JIS scoring may provide a more precise assessment of CSPH’s impact on LR outcomes, especially in early stage HCC.

This study has several limitations. First, as a retrospective study, selection and confounding biases are inevitable, even with multivariate adjustments. Future multicenter prospective studies with standardized CSPH diagnostic criteria are needed to validate our findings. Second, while CSPH can be diagnosed using non-invasive criteria, the invasive measurement of HVPG (≥ 10 mmHg) is considered the gold standard for detecting portal hypertension [26]. Although HVPG measurement is ideal, its clinical application is challenging; therefore, non-invasive criteria were used in this study [17, 18]. Third, this study was conducted at a single institution, which may limit the generalizability of the findings to other centers. We selected only patients with HCC and CSPH and analyzed risk factors for outcomes within this group (Supplementary Table S6). Our findings indicate that variceal grade is not an independent risk factor for OS, whereas tumor diameter and ALBI grade are significant predictors. However, these factors alone may not be sufficient to establish new surgical selection criteria. Therefore, we believe that future multicenter studies are needed to further refine and validate these criteria. Finally, CSPH was assessed only preoperatively without evaluating postoperative changes, underscoring the need for future studies to assess postoperative CSPH progression.

In conclusion, patients with early stage HCC (JIS 1 or BCLC-A) and CSPH have long-term prognoses comparable to those without CSPH. However, CSPH is associated with poorer survival in patients with very-early and intermediate-stage HCC undergoing LR. Surgery should be actively considered for patients with early stage HCC regardless of CSPH, while a more cautious evaluation is warranted for patients with very-early or intermediate stages. For patients with very-early or intermediate-stage HCC and CSPH, alternative strategies, such as radiofrequency ablation, transarterial chemoembolization with chemotherapy, or liver transplantation, should be carefully considered alongside LR.