Hepatocellular carcinoma — locoregional therapy¶
TL;DR — Locoregional therapy spans embolization, radioembolization, external-beam radiation, and combinations used with curative, bridging, downstaging, or disease-control intent. TACE remains the conventional BCLC B treatment for discrete arterialized tumors with preserved portal flow and liver reserve, but intermediate-stage HCC is too heterogeneous for automatic repetition (Reig 2022, PMID 34801630). TARE offers longer time to progression than TACE in pooled comparative data but no clear overall-survival advantage, and benefit depends heavily on patient selection and dosimetry (Brown 2023, PMID 35943116; Garin 2021, PMID 33166497). Phase 3 trials now show progression-free-survival gains when TACE is combined with durvalumab-bevacizumab or lenvatinib-pembrolizumab, but overall survival, toxicity, sequencing, and transplant implications determine whether these become durable standards (Sangro 2025, PMID 39798579; Kudo 2025, PMID 39798578). The key clinical skill is recognizing when local control remains useful and when repeated liver-directed treatment is causing harm or delaying systemic therapy.
Intent must be named¶
| Intent | Goal | Success measure |
|---|---|---|
| Definitive local treatment | Eradicate a localized tumor | Complete local response and durable control |
| Bridging | Prevent wait-list progression | Remain transplantable without decompensation |
| Downstaging | Reduce viable burden into transplant criteria | Sustained response and biological stability |
| Consolidation | Treat residual site after systemic response | Local control with preserved liver function |
| Disease control | Delay progression in liver-dominant disease | Time to progression, symptoms, liver preservation |
| Palliation | Reduce pain, bleeding, or mass effect | Symptom benefit with acceptable burden |
Without declared intent, response imaging and stopping decisions become incoherent.
TACE¶
TACE delivers arterial therapy to hypervascular tumor while restricting blood flow. Conventional TACE uses chemotherapy mixed with embolic material; drug-eluting-bead TACE uses calibrated beads carrying drug.
| Candidate feature | Favors TACE | Caution/contraindication |
|---|---|---|
| Tumor pattern | Discrete, arterialized, catheter-selectable | Diffuse infiltrative or massive replacement |
| Portal flow | Preserved | Main portal-vein compromise and poor collaterals |
| Liver function | Compensated | Decompensation, high bilirubin, refractory ascites |
| Performance | Preserved | Frailty or high symptom burden |
| Extrahepatic disease | Absent or clinically minor | Systemic-dominant disease |
| Vascular anatomy | Selective delivery feasible | Non-target embolization risk |
Selectivity matters: segmental treatment can maximize tumor dose while sparing non-tumor liver. Bilobar non-selective repetition accumulates ischemic injury.
TACE response and stopping¶
Viable enhancement, not only diameter, drives response assessment. Modified RECIST and LI-RADS treatment-response approaches focus on enhancing tumor.
Reassess after each treatment for:
- residual viable enhancement;
- new intrahepatic tumors;
- macrovascular invasion or extrahepatic spread;
- bilirubin, albumin, ascites, and performance status;
- whether another selective treatment is technically possible;
- whether treatment intent remains achievable.
“TACE failure/refractoriness” lacks one universal definition, but repeated progression, loss of selectivity, new vascular/spread pattern, or declining liver reserve should trigger migration rather than ritual repetition (Reig 2022, PMID 34801630).
TARE / selective internal radiation therapy¶
TARE delivers yttrium-90 microspheres arterially; tumor receives internal radiation while embolic effect varies by product and technique.
| Feature | TARE characteristic |
|---|---|
| Treatment | Often one or few administrations |
| Portal-vein thrombosis | Can be feasible in selected patients because macroembolic effect is limited |
| Radiation planning | Mapping angiography, lung-shunt assessment, target/non-target dosimetry |
| Response | Can evolve over months; early enhancement may persist |
| Toxicity | Radioembolization-induced liver disease, biliary injury, non-target irradiation |
An overall and individual-patient meta-analysis of 17 studies and 2,465 patients found no overall-survival difference between TARE and TACE; in three studies, mean time to progression was 17.5 versus 9.8 months, favoring TARE (mean difference 4.8 months, 95% CI 1.3–8.3) (Brown 2023, PMID 35943116).
This literature mixes one randomized trial with prospective and retrospective studies, stages BCLC A–C, and different dosimetry, so modality selection cannot rest on the pooled mean alone.
Dosimetry is treatment, not bookkeeping¶
DOSISPHERE-01 randomized patients with locally advanced HCC to personalized versus standard dosimetry for selective internal radiation therapy. Personalized tumor-dose planning improved response and survival, establishing that delivered absorbed dose materially affects outcome (Garin 2021, PMID 33166497).
Planning must consider:
- target tumor absorbed dose;
- non-tumor liver dose;
- lung shunt and lung dose;
- extrahepatic arterial branches;
- treated liver volume;
- baseline reserve and prior radiation.
A label of “TARE” without dosimetry can hide clinically different treatments.
TARE versus systemic therapy¶
SIRveNIB randomized Asia-Pacific patients with locally advanced HCC to selective internal radiation therapy or sorafenib and did not show superior overall survival for TARE; treatment exposure and toxicity profiles differed (Chow 2018, PMID 29498924). The negative comparison does not mean TARE lacks local activity; it means unselected substitution for systemic therapy did not improve the primary survival endpoint.
Dose-response analysis within SARAH linked higher tumor-absorbed dose with response and survival, helping explain why personalized dosimetry may identify a different therapeutic regime from fixed activity (Hermann 2020, PMID 32602828).
External-beam radiation¶
Modern conformal, stereotactic, and proton approaches can deliver ablative or consolidative doses while limiting uninvolved liver exposure.
| Use | Rationale | Main boundary |
|---|---|---|
| Small unresectable/unablatable HCC | Non-invasive local control | Nearby bowel and liver dose |
| Portal-vein tumor thrombus | Treat vascular disease | Liver reserve and target volume |
| Bridge/downstage | Control focal disease | Transplant-centre integration |
| Oligoprogression | Control resistant site during systemic therapy | Evidence mostly selected cohorts |
| Bone/adrenal metastasis palliation | Symptom control | Extrahepatic systemic burden |
RTOG 1112 compared stereotactic body radiotherapy followed by sorafenib with sorafenib alone in locally advanced HCC and reported improved outcomes for the radiation-containing strategy, supporting radiation as more than a last-resort modality in selected liver-dominant disease (Dawson 2025, PMID 39699905).
Radiation-induced liver disease risk depends on functional liver volume and baseline reserve. Child-Pugh worsening can erase local-control benefit.
TACE plus systemic therapy¶
Earlier combinations of TACE with systemic agents often failed to produce practice-changing survival benefit. The immunotherapy era has changed the rationale: TACE releases antigen and alters inflammation, while VEGF blockade and checkpoint inhibition may counteract pro-angiogenic and suppressive rebound.
| Trial | Regimen | Population | Reported primary message |
|---|---|---|---|
| EMERALD-1 | TACE + durvalumab ± bevacizumab versus TACE control | Embolization-eligible HCC | Durvalumab-bevacizumab arm improved PFS (Sangro 2025, PMID 39798579) |
| LEAP-012 | TACE + lenvatinib-pembrolizumab versus TACE + dual placebo | Unresectable, non-metastatic HCC | Combination improved PFS; OS follow-up remains central (Kudo 2025, PMID 39798578) |
These trials alter the BCLC B boundary, but unresolved questions remain:
- whether PFS gain becomes overall-survival benefit;
- whether control arms reflect best selective TACE practice;
- treatment duration and cost;
- hepatic versus immune/VEGF toxicity;
- selection of diffuse disease better treated systemically from the start;
- transplant timing after checkpoint exposure.
Evidence synthesis across locoregional modalities¶
A 2024 systematic review and meta-analysis of randomized trials compared surgery, ablation, radiation, hepatic arterial infusion, TAE, TACE, and TARE for nonmetastatic HCC. Heterogeneity in stage, technique, and comparator limited a simple rank order, reinforcing that “locoregional therapy” is not one intervention (Patel 2024, PMID 39602117).
| Comparison problem | Why synthesis is difficult |
|---|---|
| Stage mix | Early, intermediate, and locally advanced disease have different aims |
| Liver reserve | Competing death modifies treatment effect |
| Technical quality | Selectivity, margins, and dose are not standardized |
| Response criteria | Size-based and viable-enhancement metrics differ |
| Crossover | Subsequent therapies dilute overall-survival contrasts |
| Expertise | Centre volume and planning influence outcomes |
Sequencing principles¶
- Preserve liver function as a co-primary endpoint.
- Use the most selective effective local treatment.
- Define response and the next decision before treatment.
- Avoid repeated ineffective embolization.
- Re-stage for vascular/extrahepatic progression.
- Consider systemic therapy early for diffuse or biologically aggressive disease.
- Coordinate checkpoint therapy with any transplant pathway.
Safety checkpoints¶
| Therapy | Pre-treatment concern | Post-treatment danger signal |
|---|---|---|
| TACE | Portal flow, bilirubin, renal function, arterial anatomy | Fever with infection, decompensation, ischemic biliary injury |
| TARE | Lung shunt, non-target vessels, dosimetry | Delayed jaundice/ascites, ulceration, pneumonitis |
| External radiation | Functional liver dose, bowel proximity | Child-Pugh worsening, gastrointestinal injury |
| VEGF combination | Varices/bleeding, hypertension, proteinuria | Bleeding, thrombosis, wound complications |
| Checkpoint combination | Autoimmune and transplant history | Immune hepatitis, rejection, multisystem toxicity |
Comparative evidence and the importance of the delivered dose¶
| Comparison | Quantified result | Interpretation |
|---|---|---|
| DEB-TACE versus conventional TACE | Seven studies/693 patients: complete-response OR 1.18 (95% CI 0.81–1.71); objective-response OR 1.40 (0.97–2.00) | No established efficacy superiority; systemic exposure and workflow may differ (Gao 2013, PMID 23282741) |
| TARE versus TACE, randomized meta-analysis | One-year OS OR 1.31 (95% CI 0.56–3.04); one-year PFS OR 0.23 (0.02–2.45) | Wide intervals and small trials preclude equivalence claims (Casadei Gardini 2018, PMID 30498358) |
| TRACE | 72 randomized participants; glass-microsphere TARE prolonged time to progression versus DEB-TACE | Single-centre phase 2 evidence; technique and dosimetry constrain transportability (Dhondt 2022, PMID 35258371) |
| Radiation-segmentectomy dose | Mean tumor dose 568 Gy and minimum dose to 100% of volume 213 Gy predicted response at 80% specificity | Delivered tumor dose, not administered activity alone, is biologically decisive (Cheng 2021, PMID 33795611) |
| SBRT versus RFA | Retrospective meta-analysis: one- and three-year local-control ORs favored SBRT, 0.42 and 0.54, while OS favored RFA | Selection of less fit/difficult tumors for SBRT confounds survival (Pan 2020, PMID 33194569) |
| SBRT versus TACE | OS HR 0.83 (95% CI 0.52–1.34); local-control HR 0.25 (0.09–0.67) | Better local control has not yet proven superior survival (Komiyama 2025, PMID 39515381) |
Personalized radioembolization dosimetry explains why technically similar trials can disagree. Review-level thresholds associated with response ranged 100–210 Gy for resin and 205–257 Gy for glass microspheres, while device, partition model, lung shunt, normal-liver constraint, and post-treatment verification all change the biological dose (Garin 2020, PMID 32545572). The positive DOSISPHERE result and dose-response analyses should therefore not be generalized to nondosimetric “Y-90” as a homogeneous intervention (Garin 2021, PMID 33166497; Cheng 2021, PMID 33795611).
Combination era: progression benefit is not automatically cure¶
TACE can release antigen and VEGF, providing a rationale for checkpoint/antiangiogenic combinations. PETAL enrolled only 15 patients but showed pembrolizumab 30 days after TACE was feasible without dose-limiting synergistic toxicity; 93% experienced treatment-related adverse events (Pinato 2024, PMID 38578610). TACTICS-L provided prospective single-arm evidence for lenvatinib-TACE in selected Child-Pugh A disease without vascular invasion or spread, but lacks a randomized control (Kudo 2024, PMID 38344448).
The phase-3 era changed the evidentiary level. EMERALD-1 and LEAP-012 improved progression-free survival with TACE-based systemic combinations, while the magnitude, overall-survival maturity, toxicity, duration of systemic therapy, and applicability to selective low-burden TACE remain debated (Sangro 2025, PMID 39798579; Kudo 2025, PMID 39798578). A meta-analysis pooling three phase-3 programs supports a PFS class signal, but pooling different immune/VEGF regimens cannot prove identical benefit-risk profiles (Wang 2025, PMID 41268044).
Stopping rules protect the liver¶
Repetition can convert Child-Pugh A into B/C before systemic therapy begins. Among 65 TACE-refractory patients in one cohort, 27 deteriorated to Child-Pugh B/C within a year; baseline bilirubin predicted deterioration (Park 2020, PMID 32209803). Japanese criteria define refractoriness after at least two consecutive insufficient responses despite drug/feed-artery reassessment, while European practice emphasizes “untreatable progression,” liver function, vascular invasion, spread, and performance status (Kudo 2014, PMID 25427730; Raoul 2014, PMID 24945002). Neither definition is universally validated; both are safeguards against procedural momentum.
Controversies¶
- TACE versus TARE. Average randomized outcomes are similar, but TARE offers longer time to progression and outpatient convenience in some settings; dosimetry and patient selection may matter more than the modality label (Dhondt 2022, PMID 35258371; Brown 2023, PMID 35943116).
- SBRT as substitute. Local control is strong for lesions unsafe for ablation or embolization, but retrospective OS comparisons reflect selection and do not establish a universal hierarchy (Li 2023, PMID 37541936; Komiyama 2025, PMID 39515381).
- Combination therapy. PFS improvement may delay stage migration, yet added systemic toxicity and immature OS mean that low-burden, selectively embolizable disease may not require triple therapy (Wang 2025, PMID 41268044).
- When to stop TACE. Fixed procedure counts are crude; response, ALBI/Child-Pugh change, bilirubin, distribution, and new invasion/spread should trigger migration before irreversible decompensation (Piscaglia 2018, PMID 29662837; Park 2020, PMID 32209803).
Open questions¶
- Which BCLC B phenotype should receive systemic combination rather than TACE first (Reig 2022, PMID 34801630)?
- Will EMERALD-1 and LEAP-012 PFS gains produce overall-survival benefit (Sangro 2025, PMID 39798579; Kudo 2025, PMID 39798578)?
- Can personalized dosimetry be standardized across TARE platforms and centres (Garin 2021, PMID 33166497)?
- What is the optimal stopping rule after incomplete TACE response?
- Which radiation-systemic sequence maximizes local control without decompensation (Dawson 2025, PMID 39699905)?
Related pages¶
- Staging and treatment allocation — defines intent and migration.
- Liver transplantation — uses locoregional therapy for bridge/downstage.
- Systemic therapy — supplies combination and migration options.
- Diagnosis and imaging — standardizes response.
- Red flags and safety concerns — expands treatment toxicity.
References¶
- Reig M, et al. BCLC strategy for prognosis prediction and treatment recommendation: the 2022 update. J Hepatol. 2022;76:681-693. PMID 34801630
- Brown AM, et al. TACE versus TARE for patients with hepatocellular carcinoma: overall and individual patient level meta-analysis. Cancer Med. 2023;12:2590-2599. PMID 35943116
- Garin E, et al. Personalised versus standard dosimetry approach of selective internal radiation therapy in locally advanced HCC (DOSISPHERE-01). Lancet Gastroenterol Hepatol. 2021;6:17-29. PMID 33166497
- Chow PKH, et al. SIRveNIB: selective internal radiation therapy versus sorafenib in Asia-Pacific patients with hepatocellular carcinoma. J Clin Oncol. 2018;36:1913-1921. PMID 29498924
- Hermann AL, et al. Relationship of tumor radiation-absorbed dose to survival and response in HCC treated with yttrium-90 in SARAH. Radiology. 2020;296:673-684. PMID 32602828
- Dawson LA, et al. Stereotactic body radiotherapy vs sorafenib alone in hepatocellular carcinoma: NRG/RTOG 1112. JAMA Oncol. 2025;11:136-144. PMID 39699905
- Sangro B, et al. Durvalumab with or without bevacizumab with TACE in hepatocellular carcinoma (EMERALD-1). Lancet. 2025;405:216-232. PMID 39798579
- Kudo M, et al. TACE combined with lenvatinib plus pembrolizumab versus dual placebo for unresectable non-metastatic HCC (LEAP-012). Lancet. 2025;405:203-215. PMID 39798578
- Patel KR, et al. Locoregional therapies for hepatocellular carcinoma: a systematic review and meta-analysis. JAMA Netw Open. 2024;7:e2447995. PMID 39602117
- Gao S, et al. Doxorubicin-eluting bead versus conventional TACE for unresectable HCC: a meta-analysis. Hepatogastroenterology. 2013;60:813-820. PMID 23282741
- Dhondt E, et al. Y-90 radioembolization versus DEB-TACE: TRACE randomized phase 2 trial. Radiology. 2022;303:699-710. PMID 35258371
- Casadei Gardini A, et al. Radioembolization versus chemoembolization for unresectable HCC: meta-analysis of randomized trials. Onco Targets Ther. 2018;11:7315-7321. PMID 30498358
- Cheng B, et al. Tumor dose-response threshold after Y-90 radiation segmentectomy. Nucl Med Commun. 2021;42:892-898. PMID 33795611
- Wehrenberg-Klee E, et al. Patient selection and outcomes of Y-90 in HCC. Tech Vasc Interv Radiol. 2019;22:70-73. PMID 31079713
- Garin E, et al. Personalised dosimetry in radioembolisation for HCC. Cancers. 2020;12. PMID 32545572
- Pan YX, et al. SBRT versus RFA in HCC: a meta-analysis. Front Oncol. 2020;10:1639. PMID 33194569
- Li LQ, et al. SBRT for small HCC: systematic review and network meta-analysis. Clin Oncol (R Coll Radiol). 2023;35:652-664. PMID 37541936
- Komiyama S, et al. SBRT versus TACE for HCC: systematic review and meta-analysis. Radiother Oncol. 2025;202:110614. PMID 39515381
- Kudo M, et al. TACE plus lenvatinib in intermediate HCC: TACTICS-L phase 2 trial. Liver Cancer. 2024;13:99-112. PMID 38344448
- Pinato DJ, et al. Pembrolizumab following TACE: PETAL phase 1b study. Clin Cancer Res. 2024;30:2433-2443. PMID 38578610
- Wang L, et al. TACE combined with immunotherapy and targeted therapy: meta-analysis of phase 3 trials. Ther Adv Gastroenterol. 2025;18:17562848251396424. PMID 41268044
- Park KH, et al. Risk factors for liver-function deterioration after TACE refractoriness. Korean J Gastroenterol. 2020;75:147-156. PMID 32209803
- Kudo M, et al. TACE failure/refractoriness: JSH-LCSGJ criteria 2014 update. Oncology. 2014;87 Suppl 1:22-31. PMID 25427730
- Piscaglia F, et al. Patient selection for TACE: importance of benefit-risk assessment. Liver Cancer. 2018;7:104-119. PMID 29662837
- Raoul JL, et al. How to define TACE failure or refractoriness: a European perspective. Liver Cancer. 2014;3:119-124. PMID 24945002