Fig 1.
PRISMA flow chart of the study selection.
Fig 2.
Summary plot of reporting quality assessment based on ARRIVE Essential 10.
Green indicates fully reported items, Yellow indicates partially reported items, and Red indicates items not reported (generated using the robvis tool). Overall, the studies only partially reported the information required by the ARRIVE guidelines.
Fig 3.
Orchard plots of meta-analysis results for primary tumor regression.
a) Tumor volume, and b) Tumor weight. On each plot, the thick black line indicates 95% confidence intervals (CI), the extended black line indicates 95% prediction intervals (PI), and the midpoint is denoted by a diamond indicating the pooled normalized mean difference (NMD) effect estimate. Each circle represents a comparison and its diameter denotes the weight each comparison bears in the pooled effect size based on precision, meaning the larger the sample size the comparison has, the more precise it is and the higher its weight in the pooled effect. I2 reports the quantitative value for heterogeneity (98.05% and 98.8% respectively). The number of comparisons included in the meta-analysis is denoted by k, and with the number of unique studies denoted in brackets.
Fig 4.
Orchard plots of univariate meta-regression analysis for primary tumor volume.
a) Tumor model, b) Animal strain, c) Induction method, d) Cell line application route, and e) Drug administration route. The thick black line indicates 95% confidence intervals (CI). Each circle represents a comparison and its diameter denotes the weight each comparison bears in the pooled effect size based on precision. No difference was observed between subgroups and the p-value for test of moderators was not statistically significant (P = 0.45, P = 0.67, P = 0.10, P = 0.63, P = 0.12, respectively). The number of comparisons included in each level of the variable of interest is denoted by k, and with the number of unique studies denoted in brackets.
Fig 5.
Orchard plots of metastatic tumor regression.
a) Metastatic nodules number, and b) Metastatic tumor area based on bioluminescence measurement. On both plots, the thick black line indicates 95% confidence intervals (CI), the extended black line indicates 95% prediction intervals (PI), and the midpoint denoted by a diamond indicating the pooled standardized mean difference (SMD). Each circle represents a comparison and the size of it denotes the weight each comparison bears in the pooled effect size based on precision meaning the bigger sample size the comparison has, the more precise it is. I2 reports the quantitative value for heterogeneity (84.32% and 68.1% respectively). The number of comparisons included in the meta-analysis is denoted by k, and with the number of unique studies denoted in brackets.
Fig 6.
Orchard plots of univariate meta-regression analysis for metastatic nodules number.
a) Tumor model, b) Animal strain, c) Induction method, d) Cell line application route, and e) Drug administration route. The thick black line indicates 95% confidence intervals (CI). Each circle represents a comparison and its diameter denotes the weight each comparison bears in the pooled effect size based on precision. No difference was observed between subgroups and the p-value for test of moderators was not statistically significant (P = 0.42, P = 0.59, P = 0.55, P = 0.81, P = 0.39, respectively). The number of comparisons included in each level of the variable of interest is denoted by k, and with the number of unique studies denoted in brackets.
Fig 7.
Funnel plots and Trim and Fill analysis of immunotherapy outcomes.
a) Primary tumor volume, and b) Metastasis nodules number. The closed dots indicate the observed studies. The open dots indicate the missing studies imputed by the trim and fill method (n = 32 and n = 13 for a and b, respectively). The x-axis shows the observed outcome as normalized mean difference or the standardized mean difference with the y-axis representing the standard error or the inverse square root of the sample size, respectively. The two dotted lines on either side of plot (a) represent the 95% confidence intervals. The short-dash vertical line indicates the adjusted effect size including potentially missing studies under the random effects model: (a) NMD = 54.35%, and (b) SMD = 1.88 SD.