Methods of the clinical effectiveness review

A comprehensive search was undertaken to systematically identify clinical effectiveness literature relating to TheraSphere, SIR-Spheres and QuiremSpheres compared with each other, conventional transarterial therapy or established clinical management without selective internal radiation therapy, in patients with hepatocellular carcinoma. Randomised controlled trials were eligible for inclusion. Where randomised controlled trial evidence was insufficient to address the decision problem, non-randomised comparative studies and non-comparative studies were considered. In addition, a search for randomised controlled trials of comparator therapies was undertaken to strengthen the network of evidence.

Methods of network meta-analysis

A network meta-analysis was undertaken to estimate the relative effectiveness of the different treatments. Three network meta-analysis models were produced for the different populations of unresectable hepatocellular carcinoma patients: patients eligible for a transplant, patients ineligible for a transplant but eligible for conventional transarterial therapy and patients ineligible for conventional transarterial therapy.

The network meta-analysis in patients eligible for a transplant was not conducted. Clinical advice confirmed that there are short transplant waiting times in the UK, whereas these were much longer in the network trials. Therefore, the network may not be generalisable to UK practice. The network meta-analysis of patients eligible for conventional transarterial therapy was also not conducted because of the lack of good-quality evidence in this population.

Several network meta-analyses of patients who are ineligible for conventional transarterial therapy were conducted for both overall survival and progression-free survival outcomes in the per-protocol and intention-to-treat populations.

Methods of economic modelling

Owing to the limited clinical evidence in the early and intermediate patient groups, the focus of the Assessment Group’s economic analysis was on an advanced hepatocellular carcinoma population, in which high-quality randomised controlled trial evidence was available.

The Assessment Group built a fully probabilistic de novo model, which compared the three selective internal radiation therapy treatments with the systemic therapies lenvatinib (Kisplyx^®; Eisai Ltd, Tokyo, Japan) and sorafenib (Nexavar^®; Bayer plc, Leverkusen, Germany). The model structure comprised a decision tree representing the outcome of the work-up procedure transitioning into a three-state partitioned survival model. The main model structure is similar to that adopted in previous appraisals in advanced hepatocellular carcinoma, consisting of health states representing progression-free survival, post progression and death. The time horizon was 10 years. Costs and benefits were discounted at a rate of 3.5% per annum. Costs were valued at 2017/18 prices.

The model drew on data from the SorAfenib versus Radioembolization in Advanced Hepatocellular Carcinoma (SARAH) [Vilgrain V, Pereira H, Assenat E, Guiu B, Ilonca AD, Pageaux GP, et al. Efficacy and safety of selective internal radiotherapy with yttrium-90 resin microspheres compared with sorafenib in locally advanced and inoperable hepatocellular carcinoma (SARAH): an open-label randomised controlled Phase 3 trial. Lancet Oncol 2017;18:1624–36] and Selective Internal Radiation Therapy Versus Sorafenib in Locally Advanced Hepatocellular Carcinoma (SIRveNIB) (Chow PKH, Gandhi M, Tan SB, Khin MW, Khasbazar A, Ong J, et al. SIRveNIB: selective internal radiation therapy versus sorafenib in Asia-Pacific patients with hepatocellular carcinoma. J Clin Oncol 2018;36:1913–21) trials to estimate the relative effectiveness of selective internal radiation therapy and sorafenib; the base case assumed equivalence in efficacy for all selective internal radiation therapies. A hazard ratio derived from the network meta-analysis was applied to the sorafenib survival curve to estimate the efficacy of lenvatinib. Health state utilities were derived from the per-protocol subgroup of the SARAH trial for selective internal radiation therapy and systemic therapy patients. Resource use and cost inputs were derived primarily from the included trials, targeted literature searches, estimates presented in the companies’ evidence submissions, and previous National Institute for Health and Care Excellence technology appraisals.

Confidential Patient Access Schemes are available for a number of modelled technologies, including the comparator therapies lenvatinib and sorafenib and also for QuiremScout^® (Quirem Medical BV). All results in this report are based on list prices; separate analyses that include relevant Patient Access Scheme discounts are presented in Appendix 17.

Results were presented in terms of incremental net monetary benefit versus the least costly option in each scenario. Fully incremental cost-effectiveness ratios were also produced. Uncertainty was accounted for using probabilistic and deterministic sensitivity analyses. The base case was based on 20,000 model iterations using Monte Carlo sampling methods.

Results of the clinical effectiveness review

Seven randomised controlled trials, seven prospective comparative studies, five retrospective comparative studies and one non-comparative case series were included in the review of clinical effectiveness.

Network meta-analysis results

The base-case network meta-analysis was in adults with unresectable hepatocellular carcinoma who were categorised as Child–Pugh class A and ineligible for conventional transarterial therapy in the per-protocol population. Three studies were included: two randomised controlled trials comparing SIR-Spheres with sorafenib (SARAH and SIRveNIB) and one randomised controlled trial comparing lenvatinib with sorafenib (REFLECT: Kudo M, Finn RS, Qin S, Han KH, Ikeda K, Piscaglia F, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial. Lancet 2018;391:1163–73). The results provided no evidence that the random-effects model should be preferred. Therefore, the results of the fixed-effects model were used for the base-case and scenario analyses.

There were no meaningful differences in overall survival between any of the three treatments in the per-protocol or intention-to-treat populations. In the per-protocol population, SIR-Spheres showed a non-significant marginal improvement in overall survival when compared with sorafenib (hazard ratio 0.94, 96% credible interval 0.77 to 1.14), although the credible interval indicates that this result is uncertain. SIR-Spheres was ranked as the most efficacious therapy, with a probability of being the best of 0.61. Sorafenib was ranked as the worst treatment, with a probability of being the best of 0.16. Lenvatinib was ranked as the second best, with a probability of being the best of 0.22.

To produce an efficacy estimate for TheraSphere, a sensitivity analysis included the only study that directly compared TheraSphere with SIR-Spheres for Child–Pugh class A patients ineligible for conventional transarterial therapy (Biederman DM, Titano JJ, Tabori NE, Pierobon ES, Alshebeeb K, Schwartz M, et al. Outcomes of radioembolization in the treatment of hepatocellular carcinoma with portal vein invasion: resin versus glass microspheres. J Vasc Interv Radiol 2016;27:812–21.e2). Adding this study had a substantial effect on the network meta-analysis results. In the per-protocol population, TheraSphere showed a significant improvement in overall survival when compared with SIR-Spheres (hazard ratio 0.44, 95% credible interval 0.20 to 0.84), sorafenib (hazard ratio 0.41, 95% credible interval 0.20 to 0.77) and lenvatinib (hazard ratio 0.40, 95% credible interval 0.18 to 0.78). However, these results may be biased and unreliable as the Biederman et al. study is a low-quality retrospective study reporting a very strong treatment effect on overall survival for TheraSphere compared with SIR-Spheres (hazard ratio 0.40, 95% credible interval 0.20 to 0.78). A sensitivity analysis excluding the Asia-Pacific SIRveNIB study from the network meta-analysis had very little impact on the results for overall survival in the per-protocol and intention-to-treat populations compared with the base case; there were no significant differences in treatment effects for any comparisons.

Results of economic modelling

The Sirtex and BTG Ltd (hereafter BTG) company submissions each present the methods and results of two separate economic evaluations that split the population potentially eligible for selective internal radiation therapy into two groups: patients eligible for conventional transarterial therapy and patients ineligible for conventional transarterial therapy. In the corrected version of the BTG conventional transarterial therapy-eligible population, the probabilistic incremental cost-effectiveness ratio for selective internal radiation therapy compared with drug-eluting bead transarterial chemoembolisation was £24,647. In the corrected version of the BTG conventional transarterial therapy-ineligible population, the probabilistic incremental cost-effectiveness ratio for TheraSphere compared with regorafenib (Stivarga^®, Bayer plc, Leverkusen, Germany) was £69,070. The economic assessment in the conventional transarterial therapy-eligible population submitted by Sirtex was a cost-minimisation analysis, and found that the costs of selective internal radiation therapy overlapped significantly with those of conventional transarterial therapy. The base-case economic analysis submitted for the conventional transarterial therapy-ineligible population by Sirtex was in a subgroup of patients with low tumour burden and preserved liver function. The results of the presented probabilistic analysis predicted that SIR-Spheres dominated sorafenib (lower costs and higher quality-adjusted life-years).

The results of the Assessment Group’s base-case analysis (probabilistic) suggested that TheraSphere is cost-saving relative to both SIR-Spheres and QuiremSpheres. However, incremental costs between TheraSphere and SIR-Spheres were small, and pairwise net monetary benefit was close to zero (–£182). QuiremSpheres was associated with substantial incremental costs of £6615 relative to both TheraSphere and SIR-Spheres (exclusive of Patient Access Scheme). Pairwise net monetary benefit between QuiremSpheres and TheraSphere in the Assessment Group’s base case was, therefore, negative, at –£6599. In analyses presented in Appendix 17, which include available Patient Access Scheme discounts, QuiremSpheres remained more costly than both TheraSphere and SIR-Spheres; thus, the pairwise net monetary benefit remained negative.

In a fully incremental analysis at list price, none of the three selective internal radiation therapies was predicted to be cost-effective at any willingness-to-pay threshold, being more costly and less effective than lenvatinib. The predicted net monetary benefit for lenvatinib compared with TheraSphere (the lowest-costing selective internal radiation therapy) was –£2154. In a pairwise comparison of sorafenib with TheraSphere, the incremental cost-effectiveness ratio for sorafenib was £31,974 per quality-adjusted life-year gained, with an estimated net monetary benefit of –£150 (implying that TheraSphere is cost-effective compared with sorafenib at a willingness-to-pay threshold of £30,000).

In a fully incremental analysis conducted including confidential Patient Access Scheme discounts, lenvatinib remained the most cost-effective therapy and dominated all selective internal radiation therapies, generating greater health benefits at lower costs. In pairwise comparisons of sorafenib with each selective internal radiation therapy, sorafenib also dominated all selective internal radiation therapies.

A number of scenarios were produced to explore the effect of using data from more restrictive but clinically effective subpopulations, downstaging to potentially curative therapy, different resource use, cost assumptions and data sources. When the modelled population was limited to only those with a low tumour burden and preserved liver function, the incremental cost-effectiveness ratios for TheraSphere and SIR-Spheres were £17,165 and £18,783, respectively, per quality-adjusted life-year gained versus the most cost-effective systemic therapy at list price. The most optimistic incremental cost-effectiveness ratios were produced when downstaging to curative therapy was permitted in this more selective population; incremental cost-effectiveness ratios for TheraSphere and SIR-Spheres decreased to £1440 and £2339, respectively. However, there was no scenario in which selective internal radiation therapy was predicted to be cost-effective at a willingness-to-pay threshold of £30,000 when confidential Patient Access Scheme discounts were included.

Implications for service provision

The existing evidence cannot provide decision-makers with clear guidance on the comparative effectiveness of treatments in early- and intermediate-stage hepatocellular carcinoma.

In the advanced-stage hepatocellular carcinoma population, two large randomised trials have assessed the comparative effectiveness of SIR-Spheres with sorafenib, showing that selective internal radiation therapy has effectiveness similar to that of sorafenib.

None of the selective internal radiation therapies is cost-effective at any willingness-to-pay threshold, being more costly and less effective than lenvatinib; this is the case both at list price and using Patient Access Schemes.

Suggested research priorities

No strong conclusions can be drawn in the early- and intermediate-stage hepatocellular carcinoma populations owing to considerable uncertainty in estimates of effectiveness and high risk of bias. A priority for further research is, therefore, the conduct of studies in these populations.

The low tumour burden/albumin–bilirubin 1 subgroup potentially represents a group of patients for whom selective internal radiation therapy may be beneficial when compared with sorafenib. Future work considering this subgroup may, therefore, be useful.

There is currently very limited evidence on the comparative effectiveness of alternative selective internal radiation therapies. Future high-quality studies evaluating alternative selective internal radiation therapies would be beneficial.