Patient selection
The protocol was approved by the ethics committees of the Seventh Affiliated Hospital, Sun Yat-sen University and was in accordance with the Declaration of Helsinki. The requirement for written informed consent was waived due to its retrospective nature. The electronic medical records of consecutive patients with newly diagnosed HCC and MPV invasion between May 2019 and December 2021 were screened. The diagnosis of HCC was made according to the Chinese HCC guidelines [19]. HCC and MPV invasion was confirmed with the demonstration of a low-attenuation intraluminal mass expanding the main portal vein or filling defects in the MPV on contrast-enhanced computed tomography (CT) or magnetic resonance (MR) [27, 28]. The inclusion criteria were as follows: (a) age from 18 to 80 years; (b) Eastern Cooperative Oncology Group (ECOG) score less than 2; and (c) Child‒Pugh score less than 7. Patients were excluded if they (a) had massive ascites or hepatic encephalopathy or gastrointestinal bleeding or obstructive jaundice; (b) had an additional malignancy other than HCC; (c) had received previous treatments, including liver surgery, local interventional therapy or systemic therapy; or (d) had missing data.
The patients who received the treatment protocol were allocated to two groups: one group received systemic medical therapy alone (hereinafter. S group), while the other received systemic medical therapy plus additional TACE treatment (hereinafter. S-T group).
Systemic medical therapy
To minimize the influence of confounding treatments, all included patients in the present study received systemic medical therapy that included molecular targeted therapy plus immunotherapy. The molecular targeted agents include sorafenib, lenvatinib and apatinib. Immunotherapy consisted of PD-(L)1 inhibitors, such as pembrolizumab, camrelizumab and sintilimab. In the S group, molecular targeted agents and PD-(L)1 inhibitors were initially administered based on their standard dose and frequency within 1–3 days after the diagnosis of advanced HCC [29]. For patients in the S-T group, systemic treatment began 1–3 days after TACE. The protocol of administration of molecular targeted agents was followed as previously reported [10, 11]. Dose modification was based on grade 3 or 4 adverse events according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. All systemic medical treatments were continued until tumor progression or unacceptable adverse events (AEs).
Chemoembolization procedure
All patients received either conventional lipiodol-TACE (c-TACE) or drug-eluting bead TACE (d-TACE) treatments. All chemoembolization procedures were performed by two radiologists with more than 10 years of experience in interventional radiology (Y.Q.-Z., Z.-S.). Briefly, a 5-F catheter was introduced into the celiac artery and superior mesenteric artery, and angiography was performed to assess the tumor feeding artery and the patency of the portal vein. All chemoembolization procedures included superselective catheterization to the tumor-feeding arteries by using a 2.8-F microcatheter (Terumo, Japan).
For patients treated with c-TACE, 10 to 20 mL iodized oil (Lipiodol, Guerbet, France) and 20–60 mg epirubicin (Pfizer, New York, USA) were mixed. Depending on the tumor diameter and number and liver function, 2–20 mL of the emulsion was slowly infused into the liver tumor through the microcatheter. This was followed by performing embolization using gelfeom (350–560 µm, Alicon, Hangzhou, China).
The d-TACE protocol consisted of superselective injection of 2 mL of DC Bead (100–300 µm, Boston Science, USA) loaded with epirubicin (40–60 mg). The beads were mixed with a volume of contrast medium at a 1:2 ratio before delivery. If the tumor-feeding arterials were not fully embolized, embolization with gelfeom (350–560 µm) was performed supplementarily.
The TACE schedule was followed by an on-demand strategy as previously described [19, 30]. When viable tumors or new intrahepatic tumors were evident on contrast-enhanced CT or MR and the patients had preserved liver function, for instance, Child‒Pugh class A or B, TACE treatment was repeated.
Assessments and follow-up
The imaging follow-up was assessed by contrast-enhanced CT or MR at baseline and every 2 months after the initial treatment. The tumor response was calculated as the sum of complete response (CR) and partial response (PR). Additionally, to improve the objectivity of the tumor response, the AFP response was also assessed for patients with baseline AFP ≥ 400 ng/mL in the present study. AFP response was determined according to a previous report [30]: (a) CR (normalization), (b) PR (decrease by > 50% of the baseline value), (c) SD (between − 50% and + 50% of baseline value) or (d) PD (more than + 50% of baseline value). The interval between radiological assessment and AFP measurement was no more than 3 days. In the present study, the initial response was applied to assess tumor response because it is a robust predictor associated with OS. The initial response was defined as the first radiological evaluation at month 2 after the initial treatment [30, 31]. The radiological response was assessed according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) [32], and all assessments were performed by an independent radiologist (D.-M. L. with more than 10 years of experience in abdominal radiology) who was blinded to the clinical data. If the response classification was not definite, a senior radiologist (Q.-Y. L., with more than 20 years of experience in abdominal radiology) made the final decision.
The treatment-related adverse events were recorded according to CTCAE 5.0. The primary outcome in the present study was overall survival (OS), which was defined as the time from the start of treatment until death or the last follow-up. The second outcome was progression-free survival (PFS), which was defined as the time from the baseline radiological data to disease progression or death. The tumor response and treatment-related toxicities were also recorded.
Statistical analysis
All statistical analyses were performed using SPSS software (SPSS version 16.0, SPSS, Chicago, IL). Categorical variables are expressed as frequencies and percentages, and compared using χ2 test. Continuous variables are described as medians ± standard deviation, and compared using student t test. The Kaplan‒Meier method was used to calculate the PFS and OS between groups, and compared with the log-rank test. Univariate analyses were performed with the log-rank test. Variables with a P < 0.1 in the univariate analysis were entered into a multivariate analysis. A multivariate Cox analysis was used to identify risk factors that affected OS. All statistical tests were two-sided, and P < 0.05 was considered statistically significant.