Comparing Six Artery-Based Treatments for Liver Cancer: What a Major Review of 56 Randomized Trials Found

Table of Contents

• Key Points
• Why This Research Matters
• What Is Hepatocellular Carcinoma and How Is It Staged?
• The Six Transarterial Treatments Compared
• How the Study Was Conducted
• Study Quality, Bias, and Certainty of Evidence
• Key Findings: Tumor Response Rates
• Key Findings: Complete Response, Stable Disease, and Progression
• Key Findings: Side Effects and Safety
• Additional Analyses and Robustness Checks
• What This Means for Patients
• Limitations: What This Study Could Not Prove
• Practical Recommendations
• Frequently Asked Questions
• Source Information

Key Points

• In 56 randomized trials of 5,083 patients, DEB-TACE gave the most consistent benefit over conventional TACE: better tumor response, less progression, and fewer post-procedure symptoms.
• B-TACE showed the largest tumor-response advantage over conventional TACE, but this came from only three small trials with 214 participants and is described as hypothesis-generating, not proven.
• Evidence for DSM-TACE, plain embolization, and radioembolization remained too sparse or imprecise to declare any of them clearly superior across the board.
• Certainty of evidence for tumor response was low for every network comparison, and this review did not establish overall survival differences between the techniques.
• Treatment should be individualized through a multidisciplinary team, considering tumor burden, liver function, prior treatments, and personal goals rather than stage alone.

Why This Research Matters

Liver cancer is one of the deadliest cancers worldwide. According to GLOBOCAN 2022 estimates, about 866,000 people are diagnosed with liver cancer each year, and about 758,000 die from it. That makes liver cancer the sixth most common cancer and the third leading cause of cancer death globally.

Hepatocellular carcinoma (HCC) is the most common form of primary liver cancer, accounting for roughly 90% of all cases. Because it is so aggressive, doctors are constantly searching for better ways to control it and to tailor treatment to each patient.

For patients whose tumors cannot be removed by surgery, transarterial therapy (treatment delivered directly into the arteries feeding the liver tumor) is a mainstay. Yet a long-standing problem remains: doctors have several transarterial techniques to choose from, and until now it was unclear which one works best and which one causes the fewest side effects.

This new study set out to answer that question directly by comparing six techniques head to head, using the strongest type of evidence available.

What Is Hepatocellular Carcinoma and How Is It Staged?

Treatment for HCC is usually guided by the Barcelona Clinic Liver Cancer (BCLC) staging system. This system combines tumor size and number, liver function, the patient's overall fitness (called performance status), and the expected benefit of treatment.

Doctors divide HCC into broad stages. For very early and early-stage disease, surgery to remove the tumor (resection), liver transplantation, and local tumor destruction (ablation) remain the main potentially curative options. For advanced disease, systemic therapy (drugs that travel through the bloodstream, including targeted agents and immunotherapy) has become central.

The hardest group to manage is intermediate-stage HCC, particularly BCLC stage B. These patients vary enormously. Some have one large tumor, some have a few small nodules, and some have multifocal disease in both sides of the liver. Their liver function also ranges widely.

These differences matter because they affect how well a patient tolerates blocking blood flow to the tumor, how much chemotherapy reaches the tumor, and how much injury the surrounding liver suffers. As a result, lumping all these treatments together under a single broad label such as "TACE" or "locoregional therapy" is increasingly seen as inadequate for making real treatment decisions.

The Six Transarterial Treatments Compared

The researchers compared six strategies, each with a different mechanism of action.

• Conventional TACE (cTACE): The classic approach. Doctors inject a mixture of lipiodol (an oily contrast agent) and chemotherapy directly into the tumor's artery, then block the vessel with particles, often gelatin sponge.
• Drug-eluting bead TACE (DEB-TACE, also called DEM-TACE): Tiny beads loaded with chemotherapy are released into the tumor's artery. The beads deliver the drug slowly and steadily over time, which lowers chemotherapy exposure elsewhere in the body.
• Degradable starch microsphere TACE (DSM-TACE): Uses temporary, dissolvable particles that block blood flow for a short period. This may suit patients who need repeated treatments or who have limited liver reserve.
• Balloon-occluded TACE (B-TACE): A small balloon temporarily blocks the artery so that chemotherapy and lipiodol stay concentrated inside the tumor.
• Transarterial embolization (TAE): Cuts off the tumor's arterial blood supply without any chemotherapy, causing tumor cell death from starvation of blood and oxygen.
• Transarterial radioembolization (TARE): Delivers internal radiation using Yttrium-90, rather than relying mainly on blocking blood flow. It may help selected patients, including those with portal vein tumor thrombosis (cancer growing into the liver's main vein).

Naming in this field has historically been inconsistent. A treatment using lipiodol plus an embolic particle but no chemotherapy has sometimes been labeled TAE, while lipiodol plus chemotherapy plus an embolic particle is lipiodol-based TACE. This distinction is clinically important, because the chemotherapy drugs, the emulsion preparation, the embolic material, how selectively the catheter is placed, and how often treatment is repeated can all change both effectiveness and toxicity.

How the Study Was Conducted

This was a systematic review and network meta-analysis of randomized controlled trials. A network meta-analysis is a statistical technique that allows researchers to compare treatments that have never been tested directly against each other, by linking them through shared comparison groups.

The protocol was registered in advance in PROSPERO (registration number CRD420261392214). The researchers followed the PRISMA 2020 reporting standards and the PRISMA extension for network meta-analyses.

The team searched PubMed, Embase, Web of Science, the Cochrane Central Register of Controlled Trials (CENTRAL), and the Chinese National Knowledge Infrastructure database (CNKI) from their inception through May 1, 2026. They also screened Google Scholar, reference lists of eligible studies, and relevant reviews and guidelines.

To be included, a study had to meet four criteria:

1. Patients had HCC diagnosed by imaging, pathology, or accepted clinical guideline criteria.
2. The study was a randomized controlled trial.
3. At least two transarterial strategies were compared.
4. At least one extractable efficacy or safety outcome was reported.

Studies were excluded if they were observational, non-randomized, single-arm, duplicate publications, or lacked usable data. Studies mixing tumor types were excluded unless the HCC data could be separated out. Two reviewers independently screened records and extracted data, with a third reviewer resolving disagreements.

The primary outcome was objective response rate (ORR), which is the percentage of patients whose tumors shrank measurably. Secondary outcomes included disease control rate (DCR), complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Safety outcomes focused on post-embolization syndrome — the cluster of symptoms that commonly follows these procedures — specifically nausea or vomiting, abdominal pain, and fever.

Statistical analysis used a random-effects model. Response outcomes were reported as risk ratios (RRs), and progression and safety outcomes as odds ratios (ORs), each with 95% confidence intervals (CIs). Treatments were ranked using SUCRA (surface under the cumulative ranking curve), a 0–100% score where a higher value means a treatment is more likely to rank near the top. All analyses were run in Stata version 19.

Study Quality, Bias, and Certainty of Evidence

The researchers assessed risk of bias using the Cochrane RoB 2 tool, which examines five areas: the randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selective reporting. Certainty of evidence was graded with the CINeMA framework, which covers within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and incoherence.

Of the 56 randomized trials, 10 were judged at overall low risk of bias (about 1 in 6) and 46 had "some concerns" (about 4 in 5). No trial was rated at overall high risk of bias. The most common weaknesses were incomplete reporting of how randomization and blinding were handled, plus uncertainty in how some outcomes were measured or reported.

Two important reassuring findings emerged. First, statistical tests found no significant global inconsistency for the primary outcome (p = 0.9513 for ORR), and node-splitting analyses found no significant local inconsistency, with all major comparisons showing p values above 0.05. Second, comparison-adjusted funnel plots showed no clear evidence of small-study effects or publication bias for any outcome assessed.

Overall network heterogeneity was low for tumor response measures. The tau-squared values (a measure of between-study variability, where lower is more consistent) were 0.0153 for ORR, 0.0028 for DCR, 0.0355 for CR, less than 0.001 for PR, and 0.0429 for PD. Heterogeneity was notably higher for abdominal pain (tau-squared = 0.2819).

CINeMA assessment found low certainty of evidence for every network comparison for objective response rate. A small number of comparisons for disease control, progression, nausea or vomiting, and fever reached moderate or high certainty, but the overall evidence base remained limited by sparse nodes, imprecision, and risk-of-bias concerns.

Key Findings: Tumor Response Rates

The objective response network included 48 studies, 96 treatment arms, and 4,282 patients, with all six treatment types represented. The 56 included trials spanned publication years from 1992 to 2026, with individual trial sizes ranging from 24 to 289 patients and a median follow-up of about 12 months (range roughly 1 to 51 months). Most studies came from China, with others from Europe, North America, Japan, and other regions.

B-TACE ranked first for tumor response, with a SUCRA of 99.9% and a 99.8% probability of being the best treatment. Compared with cTACE, B-TACE produced a risk ratio of 3.50 (95% CI 1.92–6.36, p < 0.001). In plain terms, this suggests a much higher chance of tumor shrinkage. However, this estimate rested on only three small trials — two of which enrolled patients with hypovascular (poor blood supply) or mixed liver tumors — so the authors label it hypothesis-generating rather than confirmatory.

DEB-TACE ranked second (SUCRA 76.5%) and showed a consistent, statistically significant advantage over cTACE: risk ratio 1.36 (95% CI 1.27–1.47, p < 0.001). This means about a 36% relative increase in the chance of tumor response, a result unlikely to be due to chance.

TARE ranked third (SUCRA 61.9%). Its point estimate also favored better response over cTACE (RR 1.23), but the confidence interval crossed the line of no effect (95% CI 0.98–1.53, p = 0.069). Because the p value was above 0.05, this result is compatible with no real difference.

The authors stress that evidence for the DSM-TACE and B-TACE nodes was sparse. DSM-TACE was represented by just two randomized trials published in 2020 and 2021, with 121 randomized participants, of whom 63 received DSM-TACE. Neither trial existed when an earlier major review was conducted in 2017. B-TACE was represented by three randomized trials with 214 participants, of whom 104 received B-TACE. Together, three post-2017 randomized trials involving 181 participants newly informed these two nodes.

Key Findings: Complete Response, Stable Disease, and Progression

Beyond overall response, the review examined several other tumor outcomes. Each tells a slightly different part of the story.

• Disease control rate (DCR): B-TACE and DEB-TACE ranked highest, with SUCRA values of 98.9% and 74.9%. Compared with DEB-TACE, cTACE had a lower disease control rate (RR 0.86, 95% CI 0.83–0.90, p < 0.001). Compared with B-TACE, cTACE was also lower (RR 0.68, 95% CI 0.54–0.86, p = 0.001).
• Complete response (CR): DSM-TACE and B-TACE had the highest SUCRA values, at 78.9% and 77.8%. Compared with DEB-TACE, cTACE had a lower complete response rate (RR 0.68, 95% CI 0.59–0.79, p < 0.001). The authors caution that evidence for DSM-TACE here was sparse.
• Partial response (PR): B-TACE and DSM-TACE ranked highly, with SUCRA values of 89.2% and 84.7%. Compared with DEB-TACE, cTACE had a lower partial response rate (RR 0.77, 95% CI 0.71–0.84, p < 0.001).
• Stable disease (SD): cTACE ranked highest, with a SUCRA of 91.6%. Importantly, the authors warn that stable disease most likely reflects an intermediate state — tumors that have neither shrunk nor grown — and should not be read on its own as a treatment advantage.
• Progressive disease (PD): B-TACE, DSM-TACE, and DEB-TACE ranked favorably. The odds of disease progression with cTACE versus DEB-TACE were 2.65 times higher (OR 2.65, 95% CI 2.10–3.33, p < 0.001). For cTACE versus B-TACE, the odds were 4.52 times higher (OR 4.52, 95% CI 2.03–10.09, p < 0.001).

Pairwise comparisons for disease control, complete response, partial response, stable disease, and progression pointed in the same directions. Still, the B-TACE and DSM-TACE comparisons rest on limited randomized evidence and should be treated as exploratory.

Tumor sizes in the included trials were not confined to small tumors. Reported central tumor sizes ranged from 3.85 to 13.45 cm in cTACE arms, 4.10 to 13.16 cm in DEB-TACE arms, and 8.60 to 9.10 cm in B-TACE arms. Many arms did not report tumor size at all. This means the evidence cannot be limited to tumors under 3–4 cm, but it also prevents reliable conclusions for specific tumor sizes.

Key Findings: Side Effects and Safety

This is where DEB-TACE stood out most clearly. The safety analysis focused on post-embolization syndrome, the collection of symptoms that commonly follows these procedures.

• Nausea or vomiting: Patients receiving cTACE had 58% higher odds of nausea or vomiting than those receiving DEB-TACE (OR 1.58, 95% CI 1.19–2.11, p = 0.002).
• Abdominal pain: cTACE was linked to 55% higher odds of abdominal pain compared with DEB-TACE (OR 1.55, 95% CI 1.08–2.22, p = 0.017).
• Fever: cTACE was linked to 47% higher odds of fever compared with DEB-TACE (OR 1.47, 95% CI 1.12–1.93, p = 0.006).

The nausea or vomiting network was made up mainly of cTACE, DEB-TACE, DSM-TACE, and TAE, and the authors note there was insufficient evidence to draw firm conclusions about DSM-TACE and TAE for this outcome.

Additional Analyses and Robustness Checks

Good reviews test whether their conclusions hold up under different assumptions. This one ran many such checks.

Sensitivity analyses repeated the response analysis while restricting it to trials published from 2010 onward, to English-language reports only, and to trials at overall low risk of bias. The researchers also ran a post hoc analysis removing the two clinically indirect B-TACE studies that enrolled hypovascular or mixed-tumor populations, excluded treatment nodes supported by two or fewer studies, and excluded studies with "some concerns" in key risk-of-bias domains.

They assessed whether the network was "transitive" — meaning whether the studies were similar enough to compare fairly. They compared potential effect modifiers across studies, including publication period, BCLC stage, Child-Pugh class (a measure of liver function), tumor size, follow-up length, age, sex, viral cause of liver disease, ECOG performance status, alpha-fetoprotein level, and tumor number.

Exploratory analyses included univariable meta-regression using study-level characteristics and subgroup analysis based on the proportion of patients with Child-Pugh class A liver function. The researchers also examined direct DEB-TACE versus cTACE comparisons, plus chemotherapy class and microsphere class as possible modifiers of treatment effect. All subgroup and meta-regression results were treated as exploratory only, because they rely on study-level data rather than individual patient data and are vulnerable to ecological bias.

What This Means for Patients

The headline message is straightforward: DEB-TACE showed the most consistent overall benefit of any strategy tested against conventional TACE. It improved tumor response, reduced the risk of disease progression, and caused fewer of the common post-procedure symptoms — nausea, vomiting, abdominal pain, and fever.

This combination of better response and better tolerability is meaningful for patients. Fewer symptoms after treatment can mean shorter hospital stays, faster recovery, and the ability to continue with further treatment cycles if needed.

The B-TACE results are exciting but preliminary. A risk ratio of 3.50 for tumor response is large, but it comes from three small trials, two of which included patients whose tumors were not typical of the usual intermediate-stage population. For now, this finding should be viewed as a signal for future research, not a reason to change practice.

The authors also conclude that comparisons of DSM-TACE, B-TACE, TAE, and TARE against cTACE remain insufficient to establish a definitive overall advantage. In other words, none of these can yet be declared clearly superior across the board.

Limitations: What This Study Could Not Prove

Several important limitations temper these findings.

• Low certainty of evidence for tumor response. CINeMA grading rated all network comparisons for objective response rate as low certainty. Sparse data, imprecise estimates, and risk-of-bias concerns all contributed.
• Sparse treatment nodes. DSM-TACE rested on two trials (121 participants), and B-TACE on three (214 participants). Small numbers make estimates unstable.
• Clinically indirect evidence. Two B-TACE trials enrolled patients with hypovascular or mixed liver tumors that could not be fully separated from the main analysis.
• Heterogeneous populations and procedures. BCLC stage, Child-Pugh class, tumor size, publication period, and follow-up varied widely across trials. Tumor size went unreported in many arms.
• Mostly "some concerns" on bias. Only 10 of 56 trials were at overall low risk of bias; the rest had concerns, mainly about randomization and blinding reporting.
• Exploratory subgroup findings. All subgroup and meta-regression results used study-level, not patient-level, data and can be misleading.
• Randomized-only restriction. By limiting the network to randomized trials, the authors prioritized internal validity but accepted a sparser network and less information on uncommon or delayed harms.

Critically, this review compared tumor response and side effects. It did not establish overall survival differences between these techniques, which is what matters most to many patients.

Practical Recommendations

Based on this evidence, here is what patients and families can reasonably take away.

1. Ask whether DEB-TACE is an option for you. Among the strategies studied, it had the most reliable balance of tumor response and tolerability.
2. Discuss B-TACE as promising but unproven. If your team offers it, ask whether your tumor type matches the small trials that produced the striking results.
3. Request multidisciplinary review. Both international and Chinese guidelines emphasize that treatment should not be decided by stage alone but individualized through a multidisciplinary team, with adjustments as tumor behavior, liver function, and response change.
4. Ask about your liver function. Child-Pugh class and liver reserve strongly influence which transarterial option is safe for you.
5. Prepare for post-embolization symptoms. Nausea, vomiting, abdominal pain, and fever are common. Ask what will be done to prevent and manage them.
6. Ask about survival data, not just response. Tumor shrinkage does not always translate directly into longer life.
7. Consider a clinical trial. Several questions here — especially about DSM-TACE, B-TACE, TAE, and TARE — remain open and need larger randomized studies.

Finally, remember that these are population-level findings. Your individual treatment plan should reflect your tumor burden, liver function, prior treatments, and personal goals.

Frequently Asked Questions

Is B-TACE a better option than conventional TACE?

Balloon-occluded TACE showed the largest tumor-response advantage over conventional TACE, but that came from only three small trials with 214 participants, two of which included patients with atypical tumors. The authors call this finding hypothesis-generating rather than proven, so it is a signal for future research, not a reason to change treatment now.

Which treatment caused fewer side effects after the procedure?

Compared with DEB-TACE, conventional TACE was linked to higher odds of nausea or vomiting, abdominal pain, and fever. These are symptoms of post-embolization syndrome, the cluster that commonly follows these procedures. The nausea network mainly involved conventional TACE, DEB-TACE, DSM-TACE, and plain embolization, with too little evidence for firm conclusions about the latter two.

Does tumor shrinkage mean I will live longer?

Not necessarily. This review compared tumor response and side effects, not overall survival, so it did not establish survival differences between these techniques. Tumor shrinkage does not always translate directly into longer life. Ask your team about survival data, not just response rates, when discussing which transarterial option may suit you.

What does a risk ratio of 1.36 mean for tumor response?

A risk ratio compares how often a response occurred in one group versus another. A value of 1.36 means patients receiving DEB-TACE were about 36 percent more likely to have their tumor shrink measurably than those receiving conventional TACE. The confidence interval was 1.27 to 1.47, so this difference was statistically significant.

How certain is the evidence behind these comparisons?

Certainty was graded low for every network comparison of objective response rate, limited by sparse data, imprecise estimates, and risk-of-bias concerns. Of 56 trials, 10 were at overall low risk of bias and 46 had some concerns, mostly about how randomization and blinding were reported. No trial was rated at overall high risk of bias.

What should I ask my treatment team before choosing a transarterial therapy?

Ask whether DEB-TACE is an option for you, and whether your tumor type matches the small B-TACE trials. Request multidisciplinary review, since guidelines emphasize individualizing treatment rather than deciding by stage alone. Ask how your liver function affects safety, how post-procedure symptoms will be managed, and whether a clinical trial is suitable.

When should a patient with intermediate-stage hepatocellular carcinoma considering TACE seek a second opinion?

A second opinion is reasonable when the recommended transarterial treatment is not DEB-TACE, which showed the most consistent benefit against conventional TACE, or when B-TACE is offered, since its large tumor-response advantage rests on only three small trials and is hypothesis-generating. It also helps when the plan is based on stage alone rather than your liver function, tumor burden, and prior treatments, or when survival data have not been discussed. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Comparative efficacy and safety of transarterial treatment strategies for hepatocellular carcinoma: a systematic review and network meta-analysis of randomized controlled trials.

Authors: Chen J, Huang G, Mo F, Liao Z, Zhao C.

Publication details: Frontiers in Medicine, volume 13, article 1881070. Published August 24, 2026. DOI: 10.3389/fmed.2026.1881070. Received May 14, 2026; revised August 1, 2026; accepted August 10, 2026. Open access under a Creative Commons Attribution (CC BY) license.

Systematic review registration: PROSPERO CRD420261392214.

Note: This patient-friendly article is based on peer-reviewed research. It summarizes a systematic review and network meta-analysis of 56 randomized controlled trials involving 5,083 patients. It is intended for educational purposes and does not replace personalized medical advice from your own treatment team.