Efficacy and safety comparison of oncolytic viruses and systemic therapies in cancer: a systematic review with network meta-analysis and trial sequential analysis

Oncolytic viruses (OVs) selectively replicate in and lyse tumor cells while stimulating antitumor immunity. However, their relative efficacy and safety compared with systemic therapies (ST), and the optimal combination strategies, remain uncertain. This network meta-analysis (NMA) and trial sequential analysis (TSA) aimed to compare different OV regimens with ST. We systematically searched PubMed, Embase, Ovid, Web of Science, and ClinicalTrials.gov for phase II/III randomized controlled trials (RCTs) evaluating OVs up to August 30, 2024. Primary outcomes were objective response rate (ORR) and overall survival (OS); the secondary outcome was grade ≥ 3 adverse events (≥ 3 AEs). Pairwise meta-analysis, Bayesian random-effects NMA, frequentist cross-validation (netmeta), and TSA were performed. Sensitivity analyses included 95% prediction intervals, prior sensitivity analysis, and subgroup analyses stratified by tumour immune microenvironment. The certainty of evidence was assessed using GRADE. Eighteen RCTs involving 2,922 participants across 11 cancer types were included. In the NMA, talimogene laherparepvec (T-VEC) ranked highest for ORR (network OR 5.48, 95% CI 2.36–12.73 versus systemic therapies [ST]), followed by H101+ST (OR 4.20, 95% CI 1.60–11.0) and T-VEC+ST (OR 2.92, 95% CI 1.17–7.27), both significantly higher than ST; however, most other comparisons had wide confidence intervals crossing 1, indicating substantial imprecision. T-VEC showed a non-significant trend toward improved OS (HR 0.75, 95% CI 0.49–1.15). In subgroup analyses, combinations of OVs with immune checkpoint inhibitors (ICIs) showed higher ORR (OR 2.92, 95% CI 1.48–5.73) and a trend toward better OS (HR 0.83, 95% CI 0.56–1.23) compared with OV monotherapy or chemotherapy/targeted therapy combinations. TSA indicated that the cumulative evidence for ORR remains insufficient (required information size approximately 14,782), whereas the boundary for increased risk of grade ≥ 3 adverse events with OVs versus ST was crossed, with T-VEC showing the highest AE risk (OR 2.08, 95% CI 0.36–12.0 in the network). The GRADE assessment rated the certainty of evidence for ORR and OS as very low. This NMA suggests that T-VEC showed the highest ORR (OR 5.48, 95% CI 2.36–12.73 vs ST), with a non-significant trend toward better OS (HR 0.75, 95% CI 0.49–1.15). However, most network comparisons yielded extremely wide credible intervals crossing 1, and the certainty of evidence for all outcomes was rated as very low by GRADE, primarily due to clinical heterogeneity, imprecision, and risk of bias. TSA indicated that the cumulative evidence for ORR remains insufficient (required information size ≈ 14,782), while the boundary for increased risk of grade ≥ 3 AEs with OVs versus STs was crossed. Oncolytic viruses were generally associated with a numerically higher incidence of grade ≥ 3 adverse events than STs. These findings should be interpreted with caution and require confirmation in additional large-scale RCTs. Although TSA confirms that definitive conclusions await larger confirmatory trials, these findings provide a rationale for prioritising cancer-specific confirmatory studies in immunologically responsive tumour types. Despite their promise, oncolytic viruses currently face challenges from increased grade ≥ 3 adverse events and the limited number of high-quality RCTs. Future studies should focus on larger trials and optimized combination strategies, particularly with immune checkpoint inhibitors, to better define their clinical role and improve the benefit–risk profile.