Hepatocellular carcinoma (primary liver cancer) is classified in many ways. The Barcelona Clinic Liver Cancer (BCLC) group staging classifies the cancer based on patient's life expectancy. People with very early- or early-stage hepatocellular carcinoma have single tumour or three tumours of maximum diameter of 3 cm or less, Child-Pugh status A to B, and performance status 0 (fully functional). Management of hepatocellular carcinoma is uncertain.
To assess the comparative benefits and harms of different interventions used in the treatment of early or very early hepatocellular carcinoma through a network meta-analysis and to generate rankings of the available interventions according to their safety and efficacy. However, it was not possible to assess whether the potential effect modifiers were similar across different comparisons. Therefore, we did not perform the network meta-analysis and instead assessed the benefits and harms of different interventions versus each other or versus sham or no intervention using standard Cochrane methodology.
We searched CENTRAL, MEDLINE, Embase, Science Citation Index Expanded, and trials registers to September 2016 to identify randomised clinical trials (RCTs) on hepatocellular carcinoma.
We included only RCTs, irrespective of language, blinding, or publication status, in participants with very early- or early-stage hepatocellular carcinoma, irrespective of the presence of cirrhosis, portal hypertension, aetiology of hepatocellular carcinoma, size and number of the tumours, and future remnant liver volume. We excluded trials including participants who were previously liver transplanted. We considered interventions compared with each other, sham, or no intervention.
We calculated the odds ratio, mean difference, rate ratio, or hazard ratio with 95% confidence intervals using both fixed-effect and random-effects models based on available-participant analysis with Review Manager 5. We assessed the risk of bias according to Cochrane, controlled risk of random errors with Trial Sequential Analysis using Stata, and the quality of the evidence using GRADE.
Eighteen trials met the inclusion criteria for this review. Four trials (593 participants; 574 participants included for one or more analyses) compared surgery versus radiofrequency ablation in people with early hepatocellular carcinoma, eligible to undergo surgery. Fourteen trials (2533 participants; 2494 participants included for various analyses) compared different non-surgical interventions in people with early hepatocellular carcinoma, not eligible to undergo surgery. Overall, the quality of evidence was low or very low for all outcomes for both comparisons. Surgery versus radiofrequency ablationThe majority of participants had cirrhotic livers, and the hepatocellular carcinoma was of viral aetiology. The trials did not report the participants' portal hypertension status or whether they received adjuvant antiviral treatment or adjuvant immunotherapy. The average follow-up ranged from 29 months to 42 months (3 trials).There was no evidence of a difference in all-cause mortality at maximal follow-up for surgery versus radiofrequency ablation (hazard ratio 0.80, 95% confidence interval (CI) 0.60 to 1.08; 574 participants; 4 trials; I(2) = 68). Cancer-related mortality was lower in the surgery group (20/115 (17.4%)) than in the radiofrequency ablation group (43/115 (37.4%)) (odds ratio 0.35, 95% CI 0.19 to 0.65; 230 participants; 1 trial). Serious adverse events (number of participants) was higher in the surgery group (14/60 (23.3%)) than in the radiofrequency ablation group (1/60 (1.7%)) (odds ratio 17.96, 95% CI 2.28 to 141.60; 120 participants; 1 trial). The number of serious adverse events was higher in the surgery group (adjusted rate 11.3 events per 100 participants) than in the radiofrequency ablation group (3/186 (1.6 events per 100 participants)) (rate ratio 7.02, 95% CI 2.29 to 21.46; 391 participants; 2 trials; I(2) = 0%). None of the trials reported health-related quality of life. One trial was funded by a party with vested interests; three trials were funded by parties without any vested. Non-surgical interventionsThe majority of participants had cirrhotic livers, and the hepatocellular carcinoma was of viral aetiology. Most trials did not report the portal hypertension status of the participants, and none of the trials reported whether the participants received adjuvant antiviral treatment or adjuvant immunotherapy. The average follow-up ranged from 6 months to 37 months (11 trials). Trial participants, who were not eligible for surgery, were treated with radiofrequency ablation, laser ablation, microwave ablation, percutaneous acetic acid injection, percutaneous alcohol injection, a combination of radiofrequency ablation with systemic chemotherapy, a combination of radiofrequency ablation with percutaneous alcohol injection, a combination of transarterial chemoembolisation with percutaneous alcohol injection, or a combination of transarterial chemoembolisation with radiofrequency ablation.The mortality at maximal follow-up was higher in the percutaneous acetic acid injection (hazard ratio 1.77, 95% CI 1.12 to 2.79; 125 participants; 1 trial) and percutaneous alcohol injection (hazard ratio 1.49, 95% CI 1.18 to 1.88; 882 participants; 5 trials; I(2) = 57%) groups compared with the radiofrequency ablation group. There was no evidence of a difference in all-cause mortality at maximal follow-up for any of the other comparisons. The proportion of people with cancer-related mortality at maximal follow-up was higher in the percutaneous alcohol injection group (adjusted proportion 16.8%) compared with the radiofrequency ablation group (20/232 (8.6%)) (odds ratio 2.18, 95% CI 1.22 to 3.89; 458 participants; 3 trials; I(2) = 0%). There was no evidence of a difference in any of the comparisons that reported serious adverse events (number of participants or number of events). None of the trials reported health-related quality of life. Five trials were funded by parties without any vested interest; the source of funding was not available in the remaining trials.
The evidence was of low or very low quality. There was no evidence of a difference in all-cause mortality at maximal follow-up between surgery and radiofrequency ablation in people eligible for surgery. All-cause mortality at maximal follow-up was higher with percutaneous acetic acid injection and percutaneous alcohol injection than with radiofrequency ablation in people not eligible for surgery. There was no evidence of a difference in all-cause mortality at maximal follow-up for the other comparisons. High-quality RCTs designed to assess clinically important differences in all-cause mortality and health-related quality of life, and having an adequate follow-up period (approximately five years) are needed.