A total of 1,187 hepatocellular carcinoma (HCC) patients who underwent open hepatectomy with intraoperative portal pressure measurement at the First Affiliated Hospital of Guangxi Medical University between January 2013 and January 2020 were initially enrolled. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (Approval No. 2023-E488-01). The inclusion and exclusion criteria were as follows:

Inclusion criteria: (1) patients who underwent open hepatectomy with intraoperative portal pressure measurement at our hospital between January 2013 and January 2020; (2) age ≥ 18 and < 70 years; (3) postoperative histopathological confirmation of hepatocellular carcinoma.

Exclusion criteria: (1) patients with postoperative recurrence, ruptured HCC, previous upper abdominal surgery, any preoperative interventional therapy (such as TACE or HAIC), portal vein ligation, splenectomy, or splenic artery ligation; (2) presence of portal vein or splenic vein tumor thrombus, macrovascular invasion, or arteriovenous fistula; (3) lack of complete preoperative contrast-enhanced CT imaging; (4) missing preoperative or postoperative laboratory tests (including complete blood count, liver function, coagulation profile, etc.); (5) coexistence of other malignancies; (6) loss to follow-up after surgery.

After applying the above criteria, a total of 730 HCC patients were ultimately included in the study. (Figure 1)

Clinical and imaging parameters were extracted from the institutional electronic medical records system of the First Affiliated Hospital of Guangxi Medical University. The collected data included: patient demographics; etiology of liver disease; serological testing for liver fluke infection performed using an enzyme-linked immunosorbent assay to detect serum IgG antibodies against Clonorchis sinensis; laboratory parameters (e.g., platelet count, liver function tests, INR), from which APRI and FIB-4 scores and Child–Pugh grade were derived; preoperative imaging metrics measured from CT scans, including portal vein diameter (PVD), splenic vein diameter (SVD), liver volume (LV), and spleen volume (SV); surgical details including anatomical vs. non-anatomical and extensive vs. non-extensive hepatectomy, intraoperative blood loss, operation time, and concurrent procedures (splenectomy or splenic artery ligation); the core study variable of intraoperative portal venous pressure (PVP); tumor characteristics (e.g., size, number, greatest tumor diameter [GTD], microvascular invasion [MVI], AFP levels, abnormal prothrombin levels (also known as des-gamma-carboxy prothrombin, DCP), BCLC and CNLC stage); and outcome measures covering short-term endpoints (e.g., postoperative bilirubin, hospital stay, Clavien–Dindo complication grade) and long-term survival (overall survival [OS] and recurrence-free survival [RFS]).

Patients were regularly followed up through outpatient clinic visits, imaging reviews, and telephone interviews until the cutoff date of December 31, 2023. In this study, recurrence-free survival (RFS) was defined as the time from the date of surgery to the date of recorded tumor recurrence or the last follow-up. Overall survival (OS) was defined as the time from the date of surgery to the date of patient death or the last follow-up. The median follow up duration for the entire cohort was 52.3 months (IQR 29.5–76.1). For the PHT group, the median follow up was 49.8 months (IQR 28.3–74.2), and for the non PHT group, it was 52.8 months (IQR 29.8–76.4).

Preoperative contrast-enhanced CT scans were analyzed using 3D-Slicer (www.slicer.org) with the Total Segmentation plugin to calculate: portal vein diameter (PVD) at the confluence of the left and right branches; splenic vein diameter (SVD) at its proximal origin (two independent measurements averaged for final values); total liver volume (LV); and spleen volume (SV). Two board-certified radiologists independently performed manual ROI segmentation using ITK-SNAP (http://www.itksnap.org), and tumor-to-liver volume ratios were calculated for functional liver remnant assessment.

Intraoperative portal venous pressure (PVP) was measured under general anesthesia via right gastroepiploic vein cannulation using a saline-filled manometer (zero reference: mid-axillary line). PHT was defined as PVP ≥ 25 cm H₂O (20).

Indications and criteria for splenic modulation procedures: For HCC patients complicated by portal hypertension (PHT), the decision to perform a concomitant splenic modulation procedure (splenectomy or splenic artery ligation) during hepatectomy was made according to the following criteria.

Indications for Splenic Modulation: The procedure was indicated only when all the following conditions were met: (1) Adequate hepatic function (Child-Pugh class A or strictly selected class B with a score ≤7); (2) Significantly elevated portal venous pressure (intraoperative measurement ≥35 cm H₂O); (3) Presence of marked hypersplenism, particularly when the platelet count was < 80 × 10⁹/L.

Selection of Surgical Technique: The specific technique was chosen on an individualized basis. Splenectomy was performed for patients with intraoperative findings or preoperative evidence of moderate-to-severe esophagogastric varices or massive splenomegaly, provided that the peri-splenic anatomy was clear and the procedure was not anticipated to significantly increase surgical risk or duration. Splenic artery ligation served as an alternative, suitable for cases with milder varices, or when the planned hepatectomy itself was highly extensive, time-consuming, and the patient’s condition did not permit further prolongation or expansion of the surgery.

Contraindications: The presence of Child-Pugh-C cirrhosis, uncontrolled active infection, or severe coagulopathy was considered an absolute contraindication for splenic modulation.

Propensity score matching (1:1) was used to balance 18 baseline covariates between the PHT and non-PHT cohorts, including demographic (sex), biochemical (platelet count, total bilirubin, albumin, AST, ALT, AST/ALT ratio, APRI, INR, FIB-4), clinical (Child–Pugh score, presence of ascites), and imaging metrics (spleen volume, tumor/non-tumor liver volume ratio, portal and splenic vein diameters, maximal tumor diameter). To this end, the propensity score was calculated by enrolling these 18 variables into a logistic regression equation. Subsequently, patients in the non-PHT and PHT groups were matched 1:1 based on their raw propensity score without replacement, using the greedy matching method with a fixed caliper width of 0.5 standard deviation.

Continuous variables were presented as mean ± standard deviation for normally distributed data and compared using Student’s t test, or as median (interquartile range [IQR]) and compared using the Mann–Whitney U test for non-normally distributed data. Categorical variables were expressed as counts (percentages) and compared using Pearson’s chi-square test or Fisher’s exact test, as appropriate. Survival outcomes were analyzed using the Kaplan–Meier method and compared with the log-rank test. Missing data (< 5% of variables) were handled by mean imputation. Statistical significance was defined as p < 0.05.

The study cohort comprised 730 patients underwent open hepatectomy for resectable hepatocellular carcinoma (HCC), stratified by intraoperative portal venous pressure measurements into portal hypertension (PHT, n = 88, 12%) and non-PHT (n = 642, 88%) groups. Significant intergroup disparities (p < 0.05) were observed across 17 parameters: demographic (gender), biochemical (platelet count, total bilirubin, albumin, AST, AST/ALT ratio, APRI, INR, FIB-4), clinical (Child-Pugh score, ascites), and imaging metrics (spleen volume, tumor/non-tumor liver volume ratios, portal/splenic vein diameters, maximal tumor diameter, CNLC stage) (Table 1).

Notably, the PHT cohort demonstrated smaller tumor volume, reduced maximal diameters, and higher CNLC stage Ia prevalence compared to non-PHT counterparts. To mitigate selection bias, propensity score matching (1:1) generated balanced cohorts (59 pairs) with comparable baseline characteristics (Table 2).

Within the PHT group (n=88), 34 patients (38.6%) had platelet counts <80×10⁹/L, indicative of laboratory evidence of hypersplenism. Massive splenomegaly was defined as a spleen volume >1000 cm³ on preoperative CT volumetry (21, 22), and was present in 19 patients (21.6%). Moderate-to-severe esophagogastric varices were defined as grade 2 or higher on preoperative endoscopy or as the presence of varices with red color signs on imaging/operative notes (23), and were documented in 27 patients (30.7%).

Pre-matching analyses revealed significant disparities in surgical parameters between cohorts. The portal hypertension (PHT) group demonstrated shorter operative durations (median 200 vs 244 minutes, p = 0.026) but greater intraoperative hemorrhagic burden (500 vs 350 mL, p < 0.001). Postoperatively, PHT patients exhibited prolonged hyperbilirubinemia (days 3–5 peak: 24.6 vs 17.7 μmol/L, p = 0.004) and extended hospitalization (12 vs 9 days, p < 0.001). Although Clavien-Dindo grade≥II complications occurred more frequently in the PHT cohort (15.9% vs 10.3%), this difference lacked statistical significance (Table 3 and Supplementary Table S1).

Propensity score-matched cohorts (n = 59 per group) demonstrated persistent surgical disparities, with the portal hypertension (PHT) group maintaining shorter operative durations (median 185 vs 240 min, p = 0.020) and greater intraoperative blood loss (500 vs 300 mL, p = 0.025) compared to non-PHT counterparts. Notably, postoperative hyperbilirubinemia (days 3–5: 24.0 vs 18.6 μmol/L) lost statistical significance post-matching (p = 0.554). However, prolonged hospitalization remained evident in PHT patients (11 vs 10 days, p = 0.004). No significant intergroup differences were observed in transfusion requirements (RBC: 3.00 vs 3.25 U, p = 0.638; FFP: 600 vs 600 mL, p = 0.782), microvascular invasion prevalence (27.1% vs 37.3%, p = 0.325), or Clavien-Dindo grade ≥III complications (10.2% vs 6.78%, p = 0.741) (Table 3).

In the propensity score-matched cohort undergoing Anatomical hepatectomy (PHT: n = 25 vs non-PHT: n = 33), the portal hypertension group demonstrated significantly higher intraoperative blood loss (median 500 vs 250 mL, p = 0.015) and a clinically notable increase in Clavien-Dindo grade ≥III complications (20.0% vs 3.03%, p = 0.075), although overall complication rates remained comparable (p = 0.882). No intergroup differences were observed in operative duration (243 vs 267 min, p = 0.377), transfusion requirements (RBC: 2.75 vs 4.00 U, p = 0.960; FFP: 550 vs 500 mL, p = 0.650), microvascular invasion prevalence (28.0% vs 33.3%, p = 0.882), postoperative hyperbilirubinemia (25.2 vs 18.6 μmol/L, p = 0.718), or hospitalization length (10 vs 10 days, p = 0.554).

Patients undergoing non-Anatomical hepatectomy with portal hypertension (PHT, n = 34) demonstrated significantly prolonged postoperative hospitalization compared to non-PHT counterparts (median 12 [IQR 10.2-15] vs 10 [7-11] days, p < 0.001), despite comparable intraoperative metrics including blood loss (400 vs 350 mL, p = 0.508) and operative duration (180 vs 225 min, p = 0.070). No significant intergroup differences were observed in transfusion requirements (RBC: 3.00 vs 2.25 U; FFP: 600 vs 600 mL), microvascular invasion rates (26.5% vs 42.3%), or Clavien-Dindo grade ≥III complications (2.94% vs 11.5%).

In patients undergoing extensive hepatectomy, the portal hypertension (PHT) cohort exhibited significantly greater intraoperative blood loss compared to non-PHT counterparts (median 900 vs 200 mL, p = 0.011), despite comparable surgical durations (330 vs 253 min, p = 0.094) and transfusion requirements (RBC: 2.75 vs 4.00 U, p = 0.207; FFP: 600 vs 500 mL, p = 0.801). No intergroup differences were observed in postoperative outcomes, including hyperbilirubinemia (26.5 vs 27.1 μmol/L, p = 0.602), hospitalization length (9.5 vs 10 days, p = 0.403), or severe complications (Clavien-Dindo ≥III: 25.0% vs 0%, p = 0.164).

Patients undergoing non-extensive hepatectomy with portal hypertension (PHT, n = 51) demonstrated reduced operative durations (median 180 vs 240 min, p = 0.004) but prolonged postoperative hospitalization (11 vs 10 days, p = 0.001) compared to non-PHT counterparts (n = 48). No significant intergroup disparities were observed in intraoperative blood loss (400 vs 300 mL, p = 0.240), transfusion volumes (RBC: 3.00 vs 2.00 U; FFP: 600 vs 600 mL), microvascular invasion rates (27.5% vs 39.6%), or Clavien-Dindo ≥III complications (7.84% vs 8.33%). Postoperative hyperbilirubinemia showed comparable trajectories (21.6 vs 18.1 μmol/L, p = 0.435) (Table 4).

Propensity score-matched analyses demonstrated comparable long-term survival outcomes between PHT and non-PHT cohorts across all evaluated timepoints. Pre-matching 5-year overall survival (OS) rates were 37.8% (PHT) versus 40% (non-PHT) (p = 0.750), with recurrence-free survival (RFS) at 46.7% versus 41.2% (p = 0.370). Post-matching analyses revealed persistent non-significant disparities: 5-year OS of 42.4% (PHT) versus 36.3% (non-PHT) (p = 0.520), and RFS of 24.4% versus 44.6% (p = 0.120). Notably, the 3-year OS paradoxically favored PHT patients post-PSM (58.1% vs 46.4%), though this difference remained statistically insignificant. These findings collectively indicate that portal hypertension status does not independently predict long-term survival outcomes following hepatectomy (Figure 2; Supplementary Table S2).

Subgroup analyses revealed distinct survival patterns based on hepatectomy types. For anatomical hepatectomies, comparable 5-year overall survival (OS) and recurrence-free survival (RFS) were observed between portal hypertension (PHT) and non-PHT cohorts both pre- (OS: 48.9% vs 39%, p = 0.550; RFS: 28.2% vs 40.7%, p = 0.430) and post-propensity score matching (PSM) (OS: 40.2% vs 45.5%, p = 0.720). Conversely, non-anatomical hepatectomy demonstrated significant RFS disparities post-PSM (26.2% vs 67%, p = 0.035). Extensive hepatectomies showed no survival differences across all stages (5-year OS post-PSM: 42% vs 71.4%, p = 0.830). Notably, non-extensive hepatectomy exhibited significantly inferior RFS in PHT patients post-PSM (18.3% vs 49%, p = 0.030, suggesting residual portal hypertension pathophysiology adversely impacts recurrence in parenchymal-sparing procedures (Figure 3; Supplementary Table S3).

In patients with portal hypertension (PHT) undergoing open hepatectomy, three spleen management strategies were analyzed: hepatectomy alone (n = 68), hepatectomy with splenectomy (n = 11), and hepatectomy with splenic artery ligation (n = 9). Given limited sample sizes precluding propensity score matching, direct comparisons revealed no survival differences between splenectomy and splenic artery ligation cohorts (5-year OS: 67.0% vs 87.5%, p = 0.800; RFS: 77.8% vs 71.4%, p = 0.440). However, both intervention groups demonstrated superior outcomes compared to hepatectomy-alone controls, with splenectomy achieving 67.0% vs 27.3% 5-year OS (p = 0.0028) and 77.8% vs 35.2% RFS (p = 0.035), while splenic artery ligation showed 87.5% vs 27.3% OS (p = 0.0028) and 71.4% vs 35.2% RFS (p = 0.035). These findings suggest adjunctive splenic modulation may mitigate PHT-related survival deficits, though larger cohorts are needed for validation (Figure 4; Supplementary Tables S4, S5).