Skip to content

Hepatocellular carcinoma — surgical resection and ablation

TL;DR — Resection and local ablation can eradicate visible early HCC, but they leave the carcinogenic liver in place, so recurrence remains the dominant long-term problem. Resection generally offers better recurrence control for fit patients with technically removable tumors and adequate future liver remnant, while ablation causes less morbidity and can achieve comparable control for selected very small tumors (Feng 2015, PMID 24889505). Portal hypertension increases postoperative decompensation risk, but its effect depends on how it is measured and on modern selection and surgical technique (Aliseda 2024, PMID 38126757). Radiofrequency and microwave ablation had similar complete-ablation, survival, recurrence, and complication outcomes across five randomized trials (Yu 2021, PMID 33870454). No adjuvant systemic standard is established: IMbrave050's initial recurrence-free-survival benefit was not sustained at longer follow-up, and its updated benefit–risk profile did not support adjuvant atezolizumab-bevacizumab (Qin 2023, PMID 37871608; Yopp 2026, PMID 41580093).

Curative intent and its limits

Curative-intent local therapy means complete eradication of detectable tumor with no known extrahepatic disease. It does not mean removal of future HCC risk. Recurrence can arise from:

  • microscopic intrahepatic metastases from the treated tumor;
  • residual viable tumor at a margin;
  • vascular invasion not seen on imaging;
  • a new primary HCC from the cirrhotic field;
  • extrahepatic micrometastases.

Early recurrence is more often linked to original-tumor biology; late recurrence increasingly reflects de novo carcinogenesis.

Selecting resection

Domain Favorable feature Concern
Tumor Solitary, technically removable, no macrovascular invasion/spread Multifocal or infiltrative burden
Liver reserve Compensated, preserved synthetic function Decompensation or marginal reserve
Portal hypertension Absent or limited Varices, low platelets, high portal pressure
Anatomy Adequate future liver remnant; feasible margin Central location, major vascular/biliary involvement
Patient Fit for major surgery Frailty, severe cardiopulmonary disease
Strategy Salvage transplant remains possible Resection may complicate later transplant

AASLD guidance emphasizes preserved liver function, absence of clinically significant portal hypertension in typical candidates, and adequate future liver remnant, while allowing individualized expert-centre decisions (Singal 2023, PMID 37199193).

Portal hypertension

Portal hypertension predicts postoperative liver failure, ascites, and other morbidity, but studies use heterogeneous definitions. A 2024 meta-analysis limited to matched cohort and prospective studies found that the association with outcomes changes according to whether portal hypertension is measured invasively or inferred from indirect signs (Aliseda 2024, PMID 38126757).

Assessment Strength Limitation
Hepatic venous pressure gradient Physiologic measure Invasive; not universally available
Varices Clinically meaningful portal hypertension Sensitivity limited; influenced by prior treatment
Platelets + splenomegaly Accessible surrogate Non-specific and threshold-dependent
Elastography Continuous non-invasive risk Inflammation and platform affect values
Decompensation history Direct clinical consequence Late marker; may underestimate compensated pressure

The operative question is not simply “portal hypertension: yes/no,” but the probability of postoperative decompensation given extent of resection, remnant quality, and rescue options.

Resection approach

Anatomic resection removes the tumor-bearing portal territory; non-anatomic parenchymal-sparing resection preserves liver volume. The balance depends on location, cirrhosis severity, and recurrence strategy.

Laparoscopic resection may reduce blood loss, pain, and length of stay in selected lesions. Meta-analysis in cirrhotic HCC populations reports perioperative advantages, but most comparisons are observational and vulnerable to selection of smaller, peripheral tumors for laparoscopy (Kabir 2021, PMID 34757385).

Approach Potential advantage Tradeoff
Open resection Broad exposure; complex vascular reconstruction Larger incision and recovery burden
Laparoscopic Less invasive; shorter recovery in selected cases Technical/location selection and learning curve
Robotic Articulation and minimally invasive access Cost and limited comparative evidence
Anatomic Removes portal territory and possible microscopic spread More parenchyma sacrificed
Non-anatomic Preserves liver Margin/territory tradeoff

Ablation modalities

Modality Mechanism Typical niche Main limitation
Radiofrequency ablation Alternating current produces thermal coagulation Very small early HCC Heat-sink near large vessels
Microwave ablation Electromagnetic heating Faster/larger thermal field Equipment and operator variation
Cryoablation Freeze-thaw injury Selected locations/centres Bleeding and cryoshock concerns
Irreversible electroporation Non-thermal electrical membrane disruption Selected tumors near critical structures Limited evidence and technical complexity
Ethanol injection Chemical necrosis Where thermal ablation unavailable More sessions and local recurrence
Stereotactic radiotherapy High-dose focused radiation Non-invasive alternative when thermal approach unsafe Radiation dose to diseased liver

Thermal ablation requires an adequate margin while protecting bile ducts, bowel, gallbladder, diaphragm, and major vessels. Hydrodissection or artificial ascites can create separation in selected cases.

Radiofrequency versus microwave ablation

A meta-analysis of five randomized trials included 413 RFA-treated and 431 MWA-treated patients, all with BCLC 0–A disease. Complete ablation was 96.7% versus 96.9%; no statistically significant differences were found in overall survival, recurrence-free survival, local recurrence, or complications (Yu 2021, PMID 33870454).

This supports modality choice based on tumor geometry, heat-sink risk, platform, and expertise rather than a presumption that one energy source is universally superior.

Resection versus ablation

The comparison is confounded because surgical candidates usually have better liver reserve and technically favorable disease. A meta-analysis of three randomized and 20 retrospective studies (15,482 patients) found higher overall and recurrence-free survival after resection, higher recurrence after RFA at some time points, and higher morbidity after resection, with no mortality difference (Feng 2015, PMID 24889505).

Clinical situation More likely favored Reason
Solitary resectable tumor, good reserve Resection Pathology, margin, recurrence control
Tumor ≤2 cm, favorable ablation location Ablation Low morbidity; high complete response
Major portal hypertension Ablation or transplant Avoid postoperative decompensation
Perivascular tumor MWA, resection, or radiation RFA heat-sink concern
Near central bile duct Resection or non-thermal/radiation strategy Thermal injury risk
Poor surgical fitness Percutaneous ablation or radiation Lower procedural burden

The decision should be reported as an individualized comparison, not as stage alone.

Pathology after resection

Resection supplies information unavailable after imaging-only ablation:

  • differentiation;
  • microvascular invasion;
  • satellite nodules;
  • capsule and margin;
  • combined HCC-cholangiocarcinoma;
  • background fibrosis and steatohepatitis;
  • tissue for research or molecular testing.

These findings refine recurrence risk but do not currently mandate a universally accepted adjuvant regimen.

Recurrence risk

Risk domain Adverse feature Biological interpretation
Tumor burden Larger size, multifocality More opportunity for dissemination
Invasion Micro/macrovascular invasion Metastatic competence
Marker High or rising AFP Tumor biology and burden
Pathology Poor differentiation, satellites Aggressive clone or intrahepatic spread
Treatment Positive/close or incomplete margin Residual local disease
Liver field Active viral disease, alcohol, metabolic injury New primary risk

Post-treatment surveillance must assess both the treated zone and the remaining liver. Etiologic therapy—HBV suppression, HCV cure, alcohol intervention, metabolic management—addresses the field but does not remove recurrence risk.

Adjuvant therapy

Historically, multiple adjuvant strategies failed to establish a durable standard. IMbrave050 randomized 668 patients at high recurrence risk after resection or ablation to atezolizumab plus bevacizumab or active surveillance. The first interim analysis reported improved recurrence-free survival (HR 0.72, adjusted 95% CI 0.53–0.98), with immature overall survival and expected anti-VEGF/immune toxicities (Qin 2023, PMID 37871608). At updated follow-up, the RFS HR moved to 0.90 (95% CI 0.72–1.12), OS remained immature at HR 1.26 (95% CI 0.85–1.87), and investigators concluded that the benefit–risk profile did not support adjuvant use (Yopp 2026, PMID 41580093).

Interpretation requires:

  • updated recurrence-free survival after longer follow-up;
  • overall survival;
  • treatment after recurrence;
  • balance between delaying recurrence and exposing cured patients to toxicity;
  • subgroup consistency by resection versus ablation and etiology;
  • biomarker identification.

The reversal is a direct warning that an early positive interim endpoint is not equivalent to proven cure improvement.

Salvage transplantation

Resection can be used first in patients with preserved liver function, with transplantation reserved for recurrence or decompensation. Advantages include avoiding or delaying organ use; disadvantages include interval progression beyond transplant criteria, operative complexity, and dropout.

Salvage strategy is most coherent when transplant eligibility is assessed before resection and recurrence surveillance is reliably delivered.

Follow-up after local cure

Time Assessment purpose
Early post-ablation/resection imaging Confirm complete treatment and establish baseline
Repeated cross-sectional liver imaging Detect local, intrahepatic, vascular, and extrahepatic recurrence
AFP when informative Track marker-producing disease
Liver-function review Detect decompensation and reassess transplant candidacy
Etiology management Reduce ongoing field injury

Exact intervals vary by guideline, risk, and treatment. Surveillance after HCC treatment is more intensive than primary surveillance because recurrence probability is higher.

Comparative evidence: local cure is not one intervention

Comparison Quantified result Main threat to interpretation
Resection versus RFA, early randomized evidence Three RCTs within Milan criteria favored resection for OS, HR 1.41 (RFA relative to resection) Older ablation technology and heterogeneous selection (Qi 2014, PMID 24172183)
SURF-RCT 302 randomized patients with ≤3 tumors, each ≤3 cm; surgery and RFA had no material OS advantage at final analysis Predominantly solitary, small Japanese tumors; expert-center execution (Kawaguchi 2025, PMID 40554738)
RFA versus microwave Seven RCTs, 921 patients: complete response RR 1.01 (95% CI 0.99–1.02); no survival difference through five years Device, power, operator, and tumor-location heterogeneity (Facciorusso 2020, PMID 33339274)
SBRT versus RFA for recurrent small HCC Local-progression HR 0.45 (95% CI 0.24–0.87); two-year local PFS 92.7% versus 75.8% Single clinical context; local control does not establish OS superiority (Xi 2025, PMID 39693584)
Laparoscopic versus open major resection Nine comparative studies, 1,173 patients; laparoscopy added 74.1 minutes on average but reduced several short-term morbidity measures Non-randomized selection and center learning curves (Wang 2019, PMID 31694596)
Anatomic versus non-anatomic resection 38 studies, 9,122 patients; only one randomized study and substantial baseline imbalance Apparent recurrence benefit may reflect selection and parenchymal reserve (Jiao 2020, PMID 32413500)

The apparent conflict between older meta-analyses favoring resection and SURF is informative. Pooled non-randomized studies preferentially assign smaller, more accessible tumors and frailer patients to ablation, while surgery supplies pathology and wider margins but consumes more liver. In a randomized population dominated by solitary tumors ≤2 cm, this selection gradient narrows (Wang 2014, PMID 24404166; Kawaguchi 2025, PMID 40554738).

Recurrence prevention: positive subsets and negative totals

Strategy Evidence Current inference
HBV antiviral therapy High viral load predicted recurrence, RR 1.85 (95% CI 1.41–2.42), in 20 studies/8,204 patients; nucleos(t)ide therapy reduced recurrence Treating HBV is etiologic care, though most data are observational (Zhou 2014, PMID 24791945)
Adjuvant TACE, selected HBV/high-risk RCT of 280 patients: RFS HR 0.68 with adjuvant TACE Benefit may depend on MVI, multifocality, viral context, and protocol (Wang 2018, PMID 29420221)
Adjuvant TACE, broader stage I/II Phase 3, 332 patients: RFS HR 0.88 (95% CI 0.62–1.24), p=0.468 Contradicts blanket postoperative TACE and supports biological selection (Ma 2025, PMID 39808820)
Adjuvant sintilimab for MVI Randomized phase 2, 198 patients Positive recurrence signal requires phase-3 confirmation and OS maturity (Wang 2024, PMID 38242982)
Adjuvant sorafenib Meta-analysis found no robust OS or RFS superiority Advanced-disease activity does not imply eradication of micrometastatic disease (Li 2020, PMID 32022577)
Adjuvant checkpoint blockade overall Meta-analysis pooled 11 studies, but nine were retrospective; IMbrave050's updated RFS HR was 0.90 (95% CI 0.72–1.12) Heterogeneity and a nonsustained phase-3 signal prevent class-wide conclusions (Hu 2024, PMID 38854716; Yopp 2026, PMID 41580093)

Interferon illustrates the same problem: antiviral, antiproliferative, and immunologic plausibility did not translate into a simple tolerable standard, with the high-dose arm of a randomized trial stopped after all first six patients discontinued for toxicity (Lo 2007, PMID 17522506). Post-resection trials must therefore separate recurrence caused by occult metastatic clones from new primaries arising in the diseased liver field.

Controversies

  • Resection versus ablation. Randomization supports equivalence in a narrow small-tumor population, while broader meta-analysis may favor resection at five years; neither justifies ignoring location, portal hypertension, liver volume, or salvage transplant (Qi 2014, PMID 24172183; Kawaguchi 2025, PMID 40554738).
  • Anatomic resection. Portal-territory removal is mechanistically attractive for microscopic spread, but meta-analyses are dominated by non-randomized cohorts and conflict with the need for parenchymal preservation (Jing-Dong Li 2011, PMID 21940339; Ye 2012, PMID 22901120; Jiao 2020, PMID 32413500).
  • SBRT as ablation. SBRT can improve local control when thermal access is poor, but pooled OS comparisons are confounded by worse baseline fitness in radiation cohorts (Zhang 2021, PMID 34278818; Shin 2022, PMID 36170689).
  • Adjuvant therapy. Positive high-risk subsets coexist with negative broader trials and the nonsustained IMbrave050 signal; recurrence-free survival, treatment after recurrence, toxicity, and mature OS must all be considered (Wang 2018, PMID 29420221; Ma 2025, PMID 39808820; Yopp 2026, PMID 41580093).

Open questions

  • Which very small tumors have equivalent long-term control with ablation and resection in randomized contemporary practice (Feng 2015, PMID 24889505)?
  • Can physiologic portal-pressure assessment identify patients safely resectable despite surrogate portal hypertension (Aliseda 2024, PMID 38126757)?
  • Does adjuvant atezolizumab-bevacizumab improve overall survival after longer follow-up, and which patients benefit (Qin 2023, PMID 37871608)?
  • Can circulating tumor DNA distinguish metastatic early recurrence from future de novo tumors?
  • When should radiation replace thermal ablation for technically hazardous lesions?

References

  1. Feng Q, et al. Efficacy and safety of percutaneous radiofrequency ablation versus surgical resection for small hepatocellular carcinoma: a meta-analysis of 23 studies. J Cancer Res Clin Oncol. 2015;141:1-9. PMID 24889505
  2. Aliseda D, et al. The impact of portal hypertension assessment method on the outcomes of hepatocellular carcinoma resection: a meta-analysis of matched cohort and prospective studies. Ann Surg. 2024;280:46-55. PMID 38126757
  3. Yu Q, et al. Percutaneous microwave ablation versus radiofrequency ablation of hepatocellular carcinoma: a meta-analysis of randomized controlled trials. Abdom Radiol (NY). 2021;46:4467-4475. PMID 33870454
  4. Kabir T, et al. Laparoscopic versus open resection of hepatocellular carcinoma in patients with cirrhosis: meta-analysis. Br J Surg. 2022. PMID 34757385
  5. Qin S, et al. Atezolizumab plus bevacizumab versus active surveillance in patients with resected or ablated high-risk hepatocellular carcinoma (IMbrave050). Lancet. 2023;402:1835-1847. PMID 37871608
  6. Singal AG, et al. AASLD Practice Guidance on prevention, diagnosis, and treatment of hepatocellular carcinoma. Hepatology. 2023;78:1922-1965. PMID 37199193
  7. Qi X, et al. Radiofrequency ablation versus hepatic resection for small HCC: a meta-analysis of randomized trials. J Clin Gastroenterol. 2014;48:450-457. PMID 24172183
  8. Kawaguchi Y, et al. Surgery versus ablation for HCC: SURF randomized and prospective cohort trials. J Clin Oncol. 2025;43:2628-2638. PMID 40554738
  9. Ye JZ, et al. Recurrence after anatomic versus nonanatomic resection for HCC: a meta-analysis. Asian Pac J Cancer Prev. 2012;13:1771-1777. PMID 22901120
  10. Wang K, et al. Adjuvant sintilimab in resected high-risk HCC: a randomized phase 2 trial. Nat Med. 2024;30:708-715. PMID 38242982
  11. Ma T, et al. Adjuvant TACE after curative HCC resection: a randomized phase 3 trial. Hepatology. 2025;82:1112-1121. PMID 39808820
  12. Wang Z, et al. Adjuvant TACE for HBV-related HCC after resection: a randomized trial. Clin Cancer Res. 2018;24:2074-2081. PMID 29420221
  13. Lo CM, et al. Randomized trial of postoperative adjuvant interferon after HCC resection. Ann Surg. 2007;245:831-842. PMID 17522506
  14. Zhou Y, et al. Antiviral therapy decreases recurrence of HBV-related HCC after resection: a meta-analysis. World J Surg. 2014;38:2395-2402. PMID 24791945
  15. Xi M, et al. RFA versus SBRT for recurrent small HCC: a randomized trial. J Clin Oncol. 2025;43:1073-1082. PMID 39693584
  16. Zhang H, et al. RFA versus SBRT for HCC: a meta-analysis. Future Oncol. 2021;17:4027-4040. PMID 34278818
  17. Wang Y, et al. RFA versus hepatic resection for small HCC: meta-analysis of randomized and nonrandomized trials. PLoS One. 2014;9:e84484. PMID 24404166
  18. Facciorusso A, et al. Microwave versus radiofrequency ablation for HCC: meta-analysis of randomized trials. Cancers. 2020;12. PMID 33339274
  19. Wang ZY, et al. Laparoscopic versus open major liver resection for HCC: a meta-analysis. BMC Cancer. 2019;19:1047. PMID 31694596
  20. Jiao S, et al. Anatomic versus non-anatomic resection for HCC: a meta-analysis. Int J Surg. 2020;80:243-255. PMID 32413500
  21. Li Z, et al. Sorafenib for prevention of HCC recurrence after resection: a meta-analysis. Rev Esp Enferm Dig. 2020;112:201-210. PMID 32022577
  22. Hu L, et al. Adjuvant immune-checkpoint inhibitors after high-risk HCC resection: a systematic review. Front Oncol. 2024;14:1374262. PMID 38854716
  23. Shin HS, et al. SBRT versus RFA as initial treatment of small HCC. Eur J Gastroenterol Hepatol. 2022;34:1187-1194. PMID 36170689
  24. Yopp A, et al. Updated data from IMbrave050: adjuvant atezolizumab plus bevacizumab for high-risk hepatocellular carcinoma. J Hepatol. 2026;84:1102-1111. PMID 41580093
  25. Jing-Dong L, et al. Prognosis in patients with small hepatocellular carcinoma: a meta-analysis. Hepatogastroenterology. 2011;58:1708-1712. PMID 21940339