Fig 1.
Trial selection process according to the PRISMA statement.
Fig 2.
Straight black lines denote direct head-to-head randomized comparisons. Numbers refer to the number of RCTs with direct comparisons available for each link and the size of circles is proportional to the pooled sample size (patients) available for each treatment node.
Table 1.
Included randomized controlled trials and baseline patient demographics and index tumour characteristics.
Fig 3.
Forest plots (random effects) of direct frequentist analyses (RevMan, Cochrane). Risk of bias assessment by the Cochrane Collaboration tool is presented as well.
Fig 4.
Patient survival network meta-analysis (Random effects forest plot).
Different treatments are reported in order of efficacy ranking according to the SUCRA statistic. Black circles denote the posterior median and the black lines denote the associated 95% CrI. Numbers represent hazard ratios (HR) and 95% CrIs. The combination of TACE and ablation was found to be the most effective treatment (SUCRA 95%).
Fig 5.
Projected survival curves for each treatment were fitted with an exponential model up to 5 years. Conventional TACE was the most common comparator in the overall network of evidence and was used as the anchor treatment because it had the largest sample size. Absolute survival estimates of TACE at different time points were calculated with a standard random effects proportional model weighted by patient sample for each trial (black circles). Median patient survival (half-life) for each treatment was then calculated by combining the fitted hazard rate (exponential decay constant) of the anchor treatment with the pairwise posterior median HR calculated by the Bayesian model for the respective treatment.
Fig 6.
Forest plots (random effects) of direct frequentist analyses of patient survival (RevMan, by Cochrane). Risk of bias assessment by the Cochrane Collaboration tool is presented as well.
Fig 7.
Objective response network meta-analysis (Random effects plot).
Different treatments are reported in order of efficacy ranking according to the SUCRA statistic. Black circles denote the posterior median and the black lines denote the associated 95% CrI. Numbers represent odds ratios (OR) and 95% CrIs. The combination of TACE and ablation was found to be the most effective treatment (SUCRA 99%).
Fig 8.
Patient survival and objective response.
Two-dimensional ranking of different treatments according to patient survival (y-axis) and objective response (x-axis) based on the cumulative rank probabilities (SUCRA; %). Note the linear correlation (linear regression fit R2 = 0.926) between the 2 outcome metrics.
Fig 9.
Forest plots (random effects) of direct frequentist analyses of patient survival (RevMan, Cochrane). Risk of bias assessment by the Cochrane Collaboration tool is presented as well.
Fig 10.
Serious adverse events network meta-analysis (Random effects forest plot).
Different treatments are reported in order of safety ranking according to the SUCRA statistic. Black circles denote the posterior median and the black lines denote the associated 95% CrI. Numbers represent odds ratios (OR) and 95% CrIs. TARE was found to be the safest treatment (SUCRA 90%).
Fig 11.
Strength and quality of evidence.
QoE was graded as recommended for network meta-analyses on the basis of clinical diversity (between-trial heterogeneity of patient characteristics and/or study design), indirectness (absence of direct randomized comparisons), and imprecision (we chose a threshold of information fraction <50%). Effective sample size n for each comparison is shown along with information fraction (IF; %) in parentheses (compared to n = 560 for a hypothetical well-powered randomized study to detect a survival benefit of HR = 0.70 at 2 years). Color-coded representation of QoE; very low (light gray), low (yellow), moderate (green). There were no cases of high QoE observed.