OBJECTIVES: To evaluate the benefits and harms of gene therapy in people with hepatocellular carcinoma, irrespective of sex, administered dose, and type of formulation.

SEARCH METHODS: We identified randomised clinical trials through electronic searches in The Cochrane Hepato-Biliary Group Controlled Trials Register, CENTRAL, MEDLINE, Embase, LILACS, Science Citation Index Expanded, and Conference Proceedings Citation Index-Science. We searched five online clinical trial registries to identify unpublished or ongoing trials. We checked reference lists of the retrieved studies for further trials. The date of last search was 20 January 2023.

SELECTION CRITERIA: We aimed to include randomised clinical trials assessing any type of gene therapy in people diagnosed with hepatocellular carcinoma, irrespective of year, language of publication, format, or outcomes reported.

DATA COLLECTION AND ANALYSIS: We followed Cochrane methodology and used Review Manager to prepare the review. The primary outcomes were all-cause mortality/overall survival (whatever data were provided), serious adverse events during treatment, and health-related quality of life. The secondary outcomes were proportion of people with disease progression, adverse events considered non-serious, and proportion of people without improvement in liver function tests. We assessed risk of bias of the included trials using RoB 2 and the certainty of evidence using GRADE. We presented the results of time-to-event outcomes as hazard ratios (HR), dichotomous outcomes as risk ratios (RR), and continuous outcomes as mean difference (MD) with their 95% confidence intervals (CI). Our primary analyses were based on intention-to-treat and outcome data at the longest follow-up.

MAIN RESULTS: We included six randomised clinical trials with 364 participants. The participants had unresectable (i.e. advanced inoperable) hepatocellular carcinoma. We found no trials assessing the effects of gene therapy in people with operable hepatocellular carcinoma. Four trials were conducted in China, one in several countries (from North America, Asia, and Europe), and one in Egypt. The number of participants in the six trials ranged from 10 to 129 (median 47), median age was 55.2 years, and the mean proportion of males was 72.7%. The follow-up duration ranged from six months to five years. As the trials compared different types of gene therapy and had different controls, we could not perform meta-analyses. Five of the six trials administered co-interventions equally to the experimental and control groups. All trials assessed one or more outcomes of interest in this review. The certainty of evidence was very low in five of the six comparisons and low in the double-dose gene therapy comparison. Below, we reported the results of the primary outcomes only. Pexastimogene devacirepvec (Pexa-Vec) plus best supportive care versus best supportive care alone There is uncertainty about whether there may be little to no difference between the effect of Pexa-Vec plus best supportive care compared with best supportive care alone on overall survival (HR 1.19, 95% CI 0.78 to 1.82; 1 trial (censored observation at 20-month follow-up), 129 participants; very low-certainty evidence) and on serious adverse events (RR 1.42, 95% CI 0.60 to 3.33; 1 trial at 20 months after treatment, 129 participants; very low-certainty evidence). The trial reported quality of life narratively as "assessment of quality of life and time to symptomatic progression was confounded by the high patient dropout rate." Adenovirus-thymidine kinase with ganciclovir (ADV-TK/GCV) plus liver transplantation versus liver transplantation alone There is uncertainty about whether ADV-TK/GCV plus liver transplantation may benefit all-cause mortality at the two-year follow-up (RR 0.39, 95% CI 0.20 to 0.76; 1 trial, 45 participants; very low-certainty evidence). The trial did not report serious adverse events other than mortality or quality of life. Double-dose ADV-TK/GCV plus liver transplantation versus liver transplantation alone There is uncertainty about whether double-dose ADV-TK/GCV plus liver transplantation versus liver transplantation may benefit all-cause mortality at five-year follow-up (RR 0.40, 95% CI 0.22 to 0.73; 1 trial, 86 participants; low-certainty evidence). The trial did not report serious adverse events other than mortality or quality of life. Recombinant human adenovirus-p53 with hydroxycamptothecin (rAd-p53/HCT) versus hydroxycamptothecin alone There is uncertainty about whether there may be little to no difference between the effect of rAd-p53/HCT versus hydroxycamptothecin alone on the overall survival at 12-month follow-up (RR 3.06, 95% CI 0.16 to 60.47; 1 trial, 48 participants; very low-certainty evidence). The trial did not report serious adverse events or quality of life. rAd-p53/5-Fu (5-fluorouracil) plus transarterial chemoembolisation versus transarterial chemoembolisation alone The trial included 46 participants. We had insufficient data to assess overall survival. The trial did not report serious adverse events or quality of life. E1B-deleted (dl1520) adenovirus versus percutaneous ethanol injection The trial included 10 participants. It did not report data on overall survival, serious adverse events, or health-related quality of life. One trial did not provide any information on sponsorship; one trial received a national research grant, one trial by the Pedersen foundation, and three were industry-funded trials. We found five ongoing randomised clinical trials.