The present meta-analysis compared the effectiveness of combining ICIs with RT vs. RT alone, and ICI + RT vs. ICI therapy alone for metastatic NSCLC. Randomized controlled trials (RCTs) and retrospective studies (RSs) were included and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed for methodological consistency (17). Analysis was performed according to the Cochrane Handbook for Systematic Reviews of Interventions, version 5.1 risk of bias tool (www.handbook.cochrane.org). Patient consent and ethical approval were not obtained, as the analysis solely relied on previously published original studies.

The studies included in the analysis adhered to specific criteria, which were as follows: i) Studies with patients diagnosed with metastatic NSCLC who underwent combination treatment of STR with ICIs targeting PD-1, PD-L1 and CTLA-4; and ii) studies that were RSs or RCTs, which ensured a variety of study designs for a comprehensive analysis. Furthermore, the following exclusion criteria were applied: i) Single-arm studies; ii) systematic reviews; iii) case reports; iv) non-English publications; v) studies lacking sufficient data for outcome estimation; and vi) animal studies.

Two researchers (SF and WS) independently screened each article based on the title and abstract, and the data collection process relied on preferred literature sources. Table I provides a comprehensive overview of the included studies and provides the following details: First author, year of publication, country of origin, study design, total patient count, type of therapy (ICI + RT vs. RT and ICI + RT vs. ICI), trial registration number, phase and outcomes. Furthermore, information on progression-free survival (PFS), grade 1 and 2 adverse events (AEs; including hypophysitis, uveitis, xerostomia, hypothyroidism, pneumonitis, hepatitis, enterocolitis, pancreatitis, autoimmune diabetes, adrenal insufficiency, arthralgia and dermatitis), complete remission (CR), progressive disease (PD), stable disease (SD) and partial remission (PR) outcomes was collected.

Statistical analysis was performed using the ReviewManager (RevMan) software, version 5.4 (The Cochrane Collaboration). Dichotomous variables were pooled using odds ratios (ORs) with 95% confidence intervals (CIs). The Mantel-Haenszel method was used to assess random effects models for OR. The inconsistency statistic (I²) was used to evaluate heterogeneity between studies, with pooled results considered significant and heterogeneous if I²>50%; in this case, a random effects model was applied. Additionally, potential publication bias was assessed using funnel plots. P<0.05 was considered to indicate a statistically significant difference.

A total of 1,250 records were screened, 85 of which were assessed in full. Moreover, 7 articles involving 682 patients diagnosed with metastatic NSCLC were included in the meta-analysis (Fig. 1). Among these patients, 384 patients were given ICI + RT or RT alone, whilst 298 patients received ICI + RT or ICI alone. The characteristics of the included studies are summarized in Table I. A total of three of the records were phase II trials. The studies included only patients with NSCLC. The ICIs used included anti-PD-1 (pembrolizumab and nivolumab), anti-PD-L1 (atezolizumab and durvalumab) and anti-CTLA-4 (ipilimumab and tremelimumab) agents, whilst SRT was the type of RT administered. Treatment effects were evaluated in individual studies according to Response Evaluation Criteria in Solid Tumours 1.1 (18). Efficacy was classified as CR, PR, SD or PD. PFS was calculated from treatment initiation to the time of disease progression or death. A total of three studies compared the efficacy between the ICI + RT arm and RT-alone arm, among which two studies (19,20) with 260 patients reported PFS, and no significant difference was observed between the two arms (heterogeneity: χ=2.35; df=1; P=0.13; I²=57% and test for overall effect: Z=0.07; P=0.94; Fig. 2A). In another study a subgroup analysis of patients diagnosed with NSCLC revealed a significant improvement in PFS [hazard ratio (HR)=0.49; 95% CI, 0.37–0.64; P<0.00001] in patients <65 years who received ICI + RT, but no significant difference in PFS was reported in patients >65 years (HR=0.86; 95% CI, 0.65–1.16; P=0.43) (21). However, neither grade 1 or 2 AEs recorded in two studies (20,22) that included 333 patients were significantly different after the application of ICI + RT compared with RT alone. However, a notable improvement in grade 1 AEs was recorded for the RT arm. Moreover, no heterogeneity was observed for grade 1 AEs (25 vs. 33%; heterogeneity: Τ²=0.00; χ²=0.51; df=1; P=0.47; I²=0%; and test for overall effect: Z=1.53; P=0.13) or grade 2 AEs (heterogeneity: χ²=0.48; df=1; P=0.49; I²=0% and test for overall effect: Z=0.34; P=0.74; Fig. 2B and C). The results indicate that the use of RT alone led to an increase in grade 1 AEs.

A total of four studies including 298 patients evaluated the use of ICI + RT and ICI alone, with two studies (23,24) with 228 patients reporting the CR rate. A pooled analysis of CRs was performed using a random effects model, and no significant difference was observed between the two arms (heterogeneity: χ²=0.18; df=1; P=0.67; I²=0% and test for overall effect: Z=1.71; P=0.10; Fig. 3A). The PR rate of patients who were treated with a combination of ICI + RT compared with ICI alone was retrieved from four studies (23–26). Compared with patients who received ICI + RT, patients who received ICI alone achieved a significant PR (heterogeneity: χ²=2.13; df=3; P=0.54; I²=0% and test for overall effect: Z=2.66; P=0.01; Fig. 3B). A total of four studies (23–26) reported the SD rate, and no significant heterogeneity was observed among the studies (χ²=2.89; df=3; P=0.41; I²=0% and test for overall effect: Z=1.32; P=0.19; Fig. 3C). The pooled incidence of PD was reported in four studies (23–26) and the estimate favored ICI + RT, which was associated with a significantly decreased PD rate compared with ICI therapy alone (heterogeneity: Τ²=0.00; χ²=0.91; df=3; P=0.82; I²=8% and test for overall effect: Z=2.52; P=0.01; Fig. 3D).

The optimal sequencing of RT and ICIs remains a topic of ongoing investigation. Emerging research suggests that ICIs may exhibit maximal efficacy when administered either concurrently with or close to the time RT is given, as this approach helps to mitigate delayed antigen presentation in an immunotolerant environment. Moreover, the synergistic mechanism appears to vary depending on the specific type of ICI used. Anti-PD-1/PD-L1 agents are believed to enhance the activation of new T cells, whilst anti-CTLA4 antibodies function within the lymphatic system to suppress regulatory T cells and restore costimulatory signals, thereby enabling T cells to respond effectively to antigen presentation (27). Consequently, these processes have the potential to synergize with the T-cell activation and proliferation induced by RT, which culminates in a more robust immune response. Dovedi et al (12) compared RT administered concurrently with anti-PD-L1 therapy vs. RT followed by a 7-day interval before initiation of anti-PD-L1 treatment. The latter approach resulted in the ineffective activation of effector T cells and poor survival outcomes. Similarly, Young et al (28) evaluated the administration of RT (20 Gy ×1) alongside anti-CTLA4 at several time points relative to RT. Their findings revealed markedly improved survival rates when anti-CTLA4 was administered 7 days prior to RT. Furthermore, the study explored the use of anti-OX40, a secondary costimulatory checkpoint inhibitor, in conjunction with RT and reported optimal survival outcomes when ICI was given 1 day after RT. These findings underscore the importance of tailoring the sequencing of RT and ICIs based on the specific agent used. Currently, ongoing clinical trials, such as the RT sequence phase I trial (trial registration no. NCTO3307759), are assessing the optimal sequence of RT and ICIs, particularly in the context of metastatic NSCLC. In clinical practice, ICIs may be administered before, during or after SRT, with concurrent use defined as the administration of ICIs within 2-4 weeks before or after SRT. Favorable outcomes have been observed with this approach in terms of local control and overall survival (28,29). Based on current medical evidence, an interval of ≤4 weeks between SRT and ICI administration is recommended. Notably, the efficacy of ICIs may vary depending on the order and timing of their delivery in conjunction with SRT (28,29). Therefore, continued research in this area is essential to optimize treatment strategies and improve patient outcomes.

ICIs have revolutionized the management of NSCLC by targeting CTLA-4 and PD-1, along with their corresponding ligands B7-1 (CD-80)/B7-2 (CD-86) and (PD-L1)/PD-L2 (30). In a comprehensive analysis by Chicas-Sett et al (31), which included 18 publications and encompassed 1,736 patients who underwent treatment with SRT and ICIs, this combination demonstrated notable success rates, with a 71% local control rate and a 41% success rate in eliciting the abscopal effect. Furthermore, combination treatment led to a significant improvement in overall survival (OS; (12.4 months; P=0.006). Clinical trials evaluating several combinations of SRT and ICIs in patients with metastasis are supported by theoretical and preclinical evidence that has shown a synergistic relationship between these modalities. However, controversies persist regarding the optimal sequence of administration, SRT fractionation patterns and dosages, selection of appropriate ICIs, therapeutic effects and side effects, among other factors. Although most studies have reported enhanced local control and OS with acceptable AE rates, it is essential to note that these findings primarily stem from retrospective cohort analyses with a limited population size; these studies have predominantly included patients with primary histology indicative of malignant melanoma and have used ipilimumab more than other ICIs. These inherent limitations pose challenges in extrapolating published results and understanding their true impact, especially for brain metastases originating from other primary tumors (32).

Immunotherapy (IT) and ICIs have revolutionized cancer treatment, but they also cause a range of AEs. Common AEs associated with IT and ICIs include fatigue, cutaneous eruptions, pruritus and hormonal imbalances, such as hypophysitis and thyroid dysfunction, particularly with anti-CTLA-4 agents, as well as gastrointestinal issues (primarily colitis). Respiratory complications that range from mild coughing or shortness of breath to pneumonia are also prevalent (33). Moreover, reports have indicated potential detrimental effects on the heart and nervous system and the likelihood of experiencing these AEs appears to increase with the combination of RT and IT (33). In one study, AEs exceeding grade 3 with RT + ICI combination therapy ranged from 10–17% and 29–38% in patients who received anti-PD-1/PD-L1 therapy and anti-CTLA-4 therapy, respectively (31). However, these findings did not significantly differ from those observed with IT alone. For instance, in the PACIFIC trial, the incidence of grade 3-4 AEs was greater in the durvalumab group than in the placebo group, as patients treated with durvalumab also had a greater incidence of lung disease (34). Early discontinuation of durvalumab was primarily attributed to grade 3-4 lung disease and radiation-induced lung disease. Additional randomized trials have further supported these findings and have shown no significant difference in grades ≥3 and ≥4 pulmonary AEs when RT/IT was compared with placebo/RT (31,35). Nevertheless, caution is warranted due to the potential AEs of the RT + IT combination, such as inflammation in the colon with abdominopelvic RT or nerve damage with RT of brain metastases. SRT may mitigate certain AEs compared with conformal RT (36).

No indication of publication bias was found as every study fell within the 95% CI range (Fig. 4). Subsequently, Egger's test demonstrated statistical evidence of funnel plot symmetry. However, the results did not show any evidence of publication bias for SD (heterogeneity: χ²=2.89; df=3; P=0.41; I²=0; Fig. 3C).

In the present study, the Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1 risk of bias tool (The Cochrane Collaboration) was used to objectively assess the quality of the trials. The potential for bias was evaluated in the following areas: i) Sequence generation; ii) allocation concealment; iii) blinding; iv) incomplete data; v) selective reporting; and vi) other factors (Fig. 5). ‘High-risk’ trials were those with risks of bias for ≥1 important domains. A trial was deemed to be ‘low risk’ if it had a low bias risk across all important domains. Alternatively, the trial was deemed ‘unclear’. Disagreements between the researchers were resolved by discussion with the corresponding author.