OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2025.14931 OL-29-4-14931 Articles Baseline (modified) Glasgow prognostic score as a predictor of therapeutic response to immune checkpoint inhibitors in solid tumors: A systematic review and meta‑analysis YeHong 1 LiMeifang 2 Department of Respiratory and Critical Care Medicine, Wuhan Third Hospital, Tongren Hospital of Wuhan University, Wuhan, Hubei 430060, P.R. China Department of Brain Disease, Wuhan Hospital of Traditional Chinese Medicine, Wuhan, Hubei 430014, P.R. China Correspondence to: Professor Meifang Li, Department of Brain Disease, Wuhan Hospital of Traditional Chinese Medicine, 4 Huayuan Mountain, Wuchang, Wuhan, Hubei 430014, P.R. China, E-mail: limeifang1983lmf@163.com zhangdongdong@whu.edu.cn 04 2025 13 02 2025 29 4 184 15102024 15012025 Copyright: © 2025 Ye and Li. 2025 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

A systemic analysis was performed to evaluate the prognostic utility of the Glasgow prognostic score (GPS) and the modified (m)GPS in cancer patients treated with immune checkpoint inhibitors (ICI). The PubMed, Cochrane Library, EMBASE and Google Scholar databases were searched for entries added until May 1st, 2023, to obtain relevant articles for this study. The analysis examined several clinical outcomes, including overall survival (OS), progression-free survival (PFS), objective response rate and disease control rate (DCR). In this analysis, a total of 38 articles with 3,772 patients were included. The pooled results indicated that patients with high GPS levels had shorter OS [GPS 2 vs. 0, hazard ratio (HR): 4.35, P<0.001; GPS 1 vs. 0, HR: 2.00, P<0.001; GPS 2 vs. 1/0, HR: 2.62, P<0.001; GPS 2/1 vs. 0, HR: 2.60, P<0.001) and PFS (GPS 2 vs. 0, HR: 2.11, P=0.001; GPS 1 vs. 0, HR: 1.33, P=0.001; GPS 2 vs. 1/0, HR: 2.11, P<0.001; GPS 2/1 vs. 0, HR: 1.62, P<0.001], as well as a lower DCR [GPS 2 vs. 1/0, odds ratio (OR): 0.53, P<0.001, GPS 2/1 vs. 0, OR: 0.51, P<0.001]. It was also found that patients with high mGPS levels had poorer OS (mGPS 2 vs. 0, HR: 3.15, P<0.001; mGPS 1 vs. 0, HR: 1.70, P<0.001; mGPS 2 vs. 1/0, HR: 1.95, P=0.049; mGPS 2/1 vs. 0, HR: 3.14, P=0.041; continuous variables, HR: 1.52, P<0.001) and PFS (mGPS 2 vs. 0, HR: 2.70, P<0.001; mGPS 1 vs. 0, HR: 1.74, P=0.016; mGPS 2 vs. 1/0, HR: 1.91, P=0.044; continuous variables, HR: 1.29, P<0.001), and lower DCR (mGPS 2 vs. 1/0, HR: 0.46, P<0.001). In conclusion, the GPS and mGPS were reliable predictors of outcomes in cancer patients treated with ICIs.

 Glasgow prognostic score modified Glasgow prognostic score immune checkpoint inhibitors cancers outcomes Funding: No funding was received. Introduction

The immune evasion phenomenon is a crucial factor in the initiation and progression of cancer and is recognized as one of its primary characteristics (1). Immune checkpoints, which comprise co-inhibitory and stimulatory signals, regulate the immune system as well as shield tumor cells from immune attack (1–3). Immune checkpoint inhibitors (ICIs) have demonstrated remarkable and durable antitumor effects by inhibiting negative immunomodulatory signals (4). This approach has become the standard therapy for cancer patients (5). The Food and Drug Administration has approved multiple ICIs, including anti-programmed cell death 1 (PD-1), anti-PD-1 ligand 1 (PD-L1) and anti-cytotoxic T-lymphocyte associated protein 4 (CTLA-4) antibodies, for a wide range of cancer indications. In addition, several novel immune checkpoint molecules with therapeutic potential, such as T-cell immunoglobulin and ITIM domain, T-cell immunoglobulin and mucin-domain-containing-3, as well as lymphocyte activation gene-3, have been identified. Recently developed monoclonal antibodies against these novel targets have demonstrated encouraging therapeutic efficacy (4,6).

However, the response rate to ICI treatment varies greatly depending on the cancer type, typically ranging from 10 to 40%, and the majority of patients eventually progress despite an initial response (4,7). Besides, immune-related adverse effects from ICI therapy can be fatal or very severe (8). Early identification of individuals who do not respond to ICI treatment has recently become a hot topic in the treatment of malignant tumors in order to avoid ineffective treatment, reduce the risk of adverse reactions and the economic burden of patients (9,10). The higher costs of ICIs compared to traditional therapies, such as chemotherapy or radiotherapy, arise from expensive drug prices, prolonged treatment regimens, frequent monitoring, potential severe side effects and limited insurance coverage in certain cases. While ICIs offer substantial benefits in certain cancers, their financial impact remains a significant concern, particularly for patients and healthcare systems with limited resources (11).

Several predictive biomarkers have been explored for their association with ICI response, including the nutritional risk index, intratumoral PD-L1 expression, tumor mutational burden and T-cell infiltration metrics (12,13). However, each of these markers has notable drawbacks. The nutritional risk index is a continuous variable and there is no universally agreed-upon threshold for defining high or low risk, which complicates its clinical application. Intratumoral PD-L1 expression, whilst being the only biomarker with regulatory companion diagnostic approval by the FDA, exhibits significant intertumoral heterogeneity, and the variability in detection platforms leads to inconsistent predictive power (14,15). The tumor mutational burden requires complex and expensive sequencing technologies, limiting its widespread use in routine clinical practice. T-cell infiltration metrics are difficult to standardize across different studies and require invasive biopsy procedures, which are not always feasible, particularly in patients who are not candidates for repeated sampling. In addition, obtaining tumor samples prior to treatment initiation is often challenging due to technical and ethical considerations. Given these limitations, there is a clear need for alternative biomarkers that are more accessible and reliable.

Forrest et al (16) first developed the Glasgow Prognostic Score (GPS) as a prognostic tool for metastatic non-small cell lung cancer (NSCLC). A score of 1 is assigned for C-reactive protein (CRP) of >10 mg/l and/or albumin of <3.5 g/dl, and this culminates in patients being classified as low (0 points), intermediate (1 point) or high (2 points) risk (16). The modified (m)GPS has since improved upon the score's prognostic capabilities. A point is only awarded for low albumin if the CRP is elevated, thus more heavily weighting the inflammatory component of the score (17).

The GPS and mGPS have been extensively studied to predict postoperative outcomes in various cancers (18–22). Recently, Shimoyama et al (23) compared the prognostic effects of the prognostic nutritional index, GPS, mGPS, C-reactive protein-to-albumin ratio, neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, derived neutrophil-to-lymphocyte ratio, neutrophil-lymphocyte score, neutrophil-platelet score, platelet-lymphocyte score, lymphocyte-monocyte score, prognostic index, systemic immune-inflammation index, systemic inflammation response index, lung immune prognostic index and C-reactive protein albumin lymphocyte index for patients with gastric cancer (GC) treated with ICIs. The mGPS was found to have the strongest prognostic effect. However, this was only a single-center study, and the studies by Diker and Olgun (24), Yamamoto et al (25) and Freitas et al (26) confirm that the mGPS does not predict the efficacy of ICI treatment. Hence, the present study aimed to systematically evaluate the predictive value of the GPS and mGPS in ICI-treated patients with cancer. The findings of the current study can aid in developing effective treatment strategies that facilitate the administration of precise, cost-effective treatments with minimal adverse effects.

 Materials and methods <sec> <title>Literature search strategies

On May 1st, 2023, the article search was performed using the PubMed (https://pubmed.ncbi.nlm.nih.gov/), EMBASE (https://www.embase.com/) and Cochrane Library (https://www.cochranelibrary.com/) databases. Various search terms, including MeSH terms and keywords, were used to retrieve relevant studies (limited to English literature), such as ‘immune checkpoint inhibitors [MeSH]’, ‘PD-1 inhibitors’, ‘PD-L1 inhibitors’, ‘CTLA-4 inhibitors’, ‘Pembrolizumab’, ‘Nivolumab’, ‘Atezolizumab’, ‘Ipilimumab’, ‘Avelumab’, ‘Tremelimumab’, ‘Durvalumab’, ‘Cemiplimab’, ‘Glasgow prognostic score’ and ‘modified Glasgow prognostic score’. A detailed description of the search strategy is provided in Table SI. In addition, grey literature was explored using Google Scholar (https://scholar.google.cz/) and the reference lists of eligible studies were screened manually.

 Inclusion and exclusion criteria

The present study followed strict inclusion criteria, including articles that evaluated the prognostic value of the GPS or mGPS in patients with cancer who underwent treatment with ICIs. Only articles that discussed at least one of the following results were taken into account: Overall survival (OS), progression-free survival (PFS), objective response rate (ORR) and disease control rate (DCR). Conference abstracts, comments and case reports were excluded from the analysis.

To ensure the integrity and originality of the present findings, it was confirmed that no previous meta-analyses were included in the current meta-analysis. Instead, only original research studies were included, which provided primary data on the prognostic utility of the GPS and mGPS in ICI therapy. The rationale for this decision was to avoid the duplication of data, as meta-analyses often aggregate results from the same individual studies already included in the analysis. Including previous meta-analyses could lead to the repetition of data and potentially introduce bias or inflate effect sizes. By focusing solely on primary studies, the present meta-analysis ensures an independent and comprehensive evaluation of the available evidence.

 Data extraction and quality assessment

In this study, diverse information was gathered from the selected articles, including the names of the authors, study design, duration and location of the study, drugs used for treatment, cancer type, sample size, patient age and gender, follow-up time and outcomes. More weight was given to data from multivariate analyses of hazard ratios (HR) than univariate analyses. The Newcastle-Ottawa Scale (NOS) was utilized to assess the quality of observational research and literature with an NOS score of 6 or above was regarded as high-quality (27). All of the above steps were performed and cross-checked independently by two authors (HY and MFL).

 Statistical methods

Stata 15.0 software (StataCorp LP) was utilized for statistical analyses. To assess heterogeneity, the I2 was applied. The DerSimonian and Laird method was used for the random-effects model. Evaluated outcomes comprised hazard ratios (HRs) for OS and PFS, along with the odds ratio (OR) for the ORR and DCR. Pooled estimates and their corresponding 95% confidence intervals (CIs) were computed. HR >1 for OS and PFS indicated that the event of death is more likely to occur in the high group compared to the low group, implying a worse survival outcome. By contrast, an HR <1 would indicate a lower risk of death in the high group, suggesting a better survival outcome. OS was defined as the time from initiation of ICI treatment to death from any cause or the last date of contact. PFS was defined as the time from ICI treatment initiation to objective disease progression. The ORR was defined as the proportion of patients achieving a complete response (CR) or partial response (PR). The DCR was defined as the proportion of patients achieving a CR, PR or stable disease.

In the present study, the following comparisons were performed: (m)GPS 2 vs. 0, (m)GPS 1 vs. 0, (m)GPS 1/2 vs. 0 and (m)GPS 2 vs. 1/0. These comparisons were selected to evaluate the prognostic value of the GPS across varying levels of inflammatory and nutritional risk. Specifically, (m)GPS 2 vs. 0 and (m)GPS 1 vs. 0 are standard comparisons that assess the independent effects of moderate and severe inflammation on clinical outcomes. In addition, the comparison of (m)GPS 1/2 vs. 0 was performed to explore the combined impact of moderate and severe inflammation on clinical outcomes compared to patients with no significant inflammatory risk. This approach is particularly useful in studies with limited sample sizes or in instances where combining data provides greater statistical power. Finally, (m)GPS 2 vs. 1/0 was included to focus on the distinct impact of severe inflammation compared to patients with a lower inflammatory risk, providing a more nuanced understanding of the prognostic utility of the (m)GPS. These comparisons collectively ensure a comprehensive evaluation of the stratification capabilities of the (m)GPS, contributing to a deeper understanding of its prognostic relevance in cancer patients treated with ICIs.

Both Egger's and Begg's tests were conducted to estimate publication bias (27). If bias was found, the ‘trim and fill’ method was applied to assess its impact on the pooled results (28,29). In addition, a sensitivity analysis was performed by excluding each study independently to evaluate the robustness of the findings (28,29).

 Results <sec> <title>Characteristics of studies

Initially, 274 documents were retrieved. After eliminating duplicates (37 articles) and evaluating titles and abstracts (190 articles), 47 articles remained and were assessed in full-text. Among these, 38 articles were deemed eligible and a total of 3,772 patients were included in the final analysis (24–26,30–64). The study selection process is illustrated in Fig. 1 using a flowchart. The detailed characteristics of the eligible studies are presented in Table I. The risk of bias in all included studies was assessed using the NOS, with scores ranging from 6 to 8, indicating a low risk of bias (Tables I and SII). Of the 38 studies, 35 were retrospective and three were prospective. The included studies consisted of 13 on NSCLC, five on GC, five on head and neck squamous cell carcinoma and four on renal cell carcinoma. Furthermore, 22 studies assessed the predictive value of the GPS, 14 studies estimated the predictive utility of the mGPS and two studies explored the predictive potential of both metrics.

 Baseline GPS levels and OS/PFS

Through the analysis of data from 22 studies with 2,617 patients, the present study aimed to explore the association between GPS levels and OS in ICI-treated patients with solid tumors. The results revealed that patients with GPS 2 (I2=65.9%, P=0.003; HR: 4.35, 95% CI: 2.84–6.66, P<0.001; Fig. 2A) or GPS 1 (I2=44.0%, P=0.075; HR: 2.00, 95% CI: 1.52–2.64, P<0.001; Fig. 2B) had a shorter OS compared to patients with GPS 0, and patients with GPS 2 had a higher risk of death than patients with GPS 1.

Funnel, Begg's, and Egger's tests confirmed the publication bias of the above results (GPS 2 vs. 0: Begg's P=0.466, Egger's P=0.001; GPS 1 vs. 0: Begg's P=0.016, Egger's P=0.003; Fig. S1A and B). To account for potentially missing studies in the above analysis, the trim and fill method was applied. However, the results indicated that the pooled HR did not change significantly even after accounting for any potential missing studies (Fig. S1C and D). A sensitivity analysis was conducted to evaluate the robustness of the present findings by iteratively excluding each study and examining its impact on the pooled results. It was revealed that the exclusion of any individual study did not significantly affect the pooled HR (Fig. S2A and B).

In addition, it was also found that patients with GPS 2 had a worse prognosis compared to those with GPS 1/0 (I2=0.0%, P=0.724; HR: 2.62, 95% CI: 2.01–3.43, P<0.001; Fig. S3A). Patients with GPS 1/2 also had a higher mortality rate compared to patients with GPS 0 (I2=62.4%, P=0.021; HR: 2.60, 95% CI: 1.57–4.29, P<0.001; Fig. S3B). Begg's and Egger's tests confirmed the above results without significant publication bias (GPS 2 vs. 1/0: Begg's P=0.368, Egger's P=0.175; GPS 2/1 vs. 0: Begg's P=0.452, Egger's P=0.083). However, the funnel plot revealed publication bias in the GPS 2/1 vs. 0 group (Fig. S3C and D). A trim and fill method was used, and the results showed that the combined HR did not change significantly, even after accounting for any potential missing studies.

Next, 15 studies with 2,026 patients were analyzed to investigate the relationship between GPS levels and PFS in cancer patients treated with ICIs. The findings indicated that patients with GPS 2 (I2=68.1%, P=0.014; HR: 2.11, 95% CI: 1.38–3.24, P=0.001; Fig. 3A) or GPS 1 (I2=0.0%, P=0.836; HR: 1.33, 95% CI: 1.13–1.56, P=0.001; Fig. 3B) had a significantly poorer PFS compared to those with GPS 0, and patients with GPS 2 had a higher risk of progression than patients with GPS 1.

Funnel plot, Begg's and Egger's tests demonstrated the publication bias of the above analysis (GPS 2 vs. 0: Begg's P=0.086, Egger's P=0.023; GPS 1 vs. 0: Begg's P=0.086, Egger's P=0.030; Fig. S4A and B). To address the potential issue of missing studies, the trim and fill method was employed. However, the pooled HR remained unchanged even after accounting for any missing studies (Fig. S4C and D). The results of the sensitivity analysis showed that excluding any of the studies did not have a significant impact on the pooled HR (Fig. S5A and B).

Furthermore, the present analysis revealed that patients with a GPS score of 2 had a significantly worse prognosis than those with a GPS score of 1 and 0 (I2=0.0%, P=0.509; HR: 2.11, 95% CI: 1.61–2.78, P<0.001; Fig. S6A). Furthermore, patients with a GPS score of 1 and 2 exhibited a higher progression rate compared to patients with a GPS score of 0 (I2=0.0%, P=0.414; HR: 1.62, 95% CI: 1.26–2.09, P<0.001; Fig. S6B). Funnel plot, Begg's and Egger's tests indicated no significant publication bias in the above results (GPS 2 vs. 1/0: Begg's P=0.086, Egger's P=0.171; GPS 2/1 vs. 0: Begg's P=0.308, Egger's P=0.059; Fig. S6C and D).

 Baseline GPS levels and DCR/ORR

The analysis proceeded to examine the link between GPS levels and response to ICI treatment in cancer patients through 10 studies with 1,433 patients. The results demonstrated that patients with GPS 2 had a lower DCR compared to those with GPS 1/0 (OR: 0.53, 95% CI: 0.40–0.69, P<0.001; Fig. 4A). Patients with GPS 1/2 also had a lower DCR compared to patients with GPS 0 (OR: 0.51, 95% CI: 0.39–0.67, P<0.001; Fig. 4B). The above analysis did not find a significant publication bias (GPS 2 vs. 1/0: Begg's P=0.711, Egger's P=0.111; GPS 2/1 vs. 0: Begg's P=0.462, Egger's P=0.167; Fig. 4C and D). The results of the sensitivity analysis showed that excluding any of the studies did not have a significant impact on the pooled HR (Fig. 4E and F). In addition, high GPS scores were also significantly associated with a lower ORR in cancer patients (GPS 2 vs. 1/0, I2=52.5%, P=0.077, OR: 0.53, 95% CI: 0.28–1.00, P=0.049; GPS 2/1 vs. 0, I2=15.3%, P=0.316, OR: 0.31, 95% CI: 0.16–0.62, P<0.001; Fig. S7).

 Baseline mGPS levels and OS/PFS

A total of 16 studies with 1,474 patients were analyzed to explore the association between mGPS levels and OS in patients with solid tumors treated with ICIs. The results showed that patients with mGPS 2 (I2=42.0%, P=0.125; HR: 3.15, 95% CI: 1.95–5.03, P<0.001; Fig. 5A) or mGPS 1 (I2=0.0%, P=0.905; HR: 1.70, 95% CI: 1.19–2.43, P=0.004; Fig. 5B) had poorer survival compared to patients with mGPS 0 and patients with mGPS 2 had a higher risk of death than patients with mGPS 1.

Funnel plot, Begg's and Egger's tests revealed no publication bias in the above findings (mGPS 2 vs. 0: Begg's P=1.000, Egger's P=0.805; mGPS 1 vs. 0: Begg's P=0.707, Egger's P=0.969; Fig. 5C and D). Furthermore, a sensitivity analysis was employed to test the robustness of the results by excluding each study iteratively and assessing its impact on the pooled outcomes. It was found that the exclusion of any single study did not substantially affect the pooled HR (Fig. 5E and F).

The present analysis further revealed that patients with mGPS 2 had a poorer prognosis than those with mGPS 1/0, with an HR of 1.95 (95% CI: 1.00–3.82, P=0.049; Fig. 6A), and patients with mGPS=1/2 also had a higher mortality rate compared to patients with mGPS=0, with an HR of 3.14 (95% CI: 1.05–9.39, P=0.041; Fig. 6B). Begg's and Egger's tests confirmed no significant publication bias was observed (mGPS 2 vs. 1/0: Begg's P=1.000, Egger's P=0.378; mGPS 2/1 vs. 0: Begg's P=0.308, Egger's P=0.388; Fig. 6C and D). Finally, it was found that when mGPS is a continuous variable, a larger mGPS value is associated with shorter patient OS (HR: 1.52, 95% CI: 1.30–1.78, P<0.001; Fig. 6E).

The association between mGPS and PFS in cancer patients treated with ICIs was then explored using data from 13 studies with 1,216 patients. It was demonstrated that patients with mGPS 2 (I2=27.1%, P=0.241; HR: 2.70, 95% CI: 1.75–4.16, P<0.001; Fig. 7A) or mGPS=1 (I2=22.8%, P=0.269; HR: 1.74, 95% CI: 1.11–2.72, P=0.016; Fig. 7B) had poorer PFS and patients with mGPS 2 had a higher risk of progression than patients with mGPS 1.

The results of funnel plot, Begg's and Egger's tests indicated that there was no publication bias in the aforementioned findings (mGPS 2 vs. 0: Begg's P=0.462, Egger's P=0.129; mGPS 1 vs. 0: Begg's P=1.000, Egger's P=0.289; Fig. 7C and D). Furthermore, a sensitivity analysis was performed to evaluate the robustness of the results by iteratively excluding each study and examining its influence on the pooled outcomes. It was discovered that the exclusion of any individual study did not significantly alter the pooled HR (Fig. 7E and F).

The present analysis further showed that patients with mGPS 2 had a worse PFS compared to those with mGPS=1/0 (HR: 1.91, 95% CI: 1.02–3.57, P=0.044; Fig. 8A). The funnel plot, Begg's and Egger's tests showed no significant publication bias (mGPS 2 vs. 1/0: Begg's P=1.000, Egger's P=0.954, Fig. 8B). It was also found that when mGPS is a continuous variable, a larger mGPS value was associated with worse PFS of patients (HR: 1.29, 95% CI: 1.14–1.46, P<0.001; Fig. 8C).

 Baseline mGPS levels and DCR

An analysis was performed to investigate the relationship between mGPS levels and response to ICI therapy in cancer patients using data from three studies with 465 patients. The findings demonstrated that cancer patients with mGPS 2 exhibited a lower DCR (I2=0.0%, P=0.644; OR: 0.46, 95% CI: 0.31–0.70, P<0.001; Fig. S8A) of ICI therapy than those with mGPS 1/0. Sensitivity analysis confirmed that the results were stable and reliable (Fig. S8B).

 Discussion

The present investigation aimed to explore the predictive value of the GPS and mGPS in cancer patients receiving ICIs. Through this analysis of relevant studies, a strong relationship between elevated GPS and mGPS and worse OS and PFS, as well as a lower DCR, was found. As a cost-effective, easy-to-use tool, the GPS or mGPS status can be used as a predictor to identify patients who are likely to experience favorable clinical outcomes.

To provide a comprehensive comparison of the present findings with previous meta-analyses, including Zhang et al (65), Wu et al (66) and Wang et al (67), survival outcomes and response metrics were analyzed across a broader range of studies and cancer types. Compared to Zhang et al (65), who reported HRs for OS of 3.89 (GPS 2 vs. 0) and 1.48 (GPS 1 vs. 0), the present analysis yielded higher pooled HRs for OS (4.354 for GPS 2 vs. 0, and 2.004 for GPS 1 vs. 0), indicating a potentially stronger prognostic value of the GPS. Similarly, for PFS, the HR determined in the present study for GPS 2 vs. 0 was 2.113, compared to Zhang et al (65)'s 1.48, highlighting a more pronounced association in the larger dataset of the present study. In terms of the DCR, a unique aspect of the current study, it was found that higher GPS and mGPS were significantly associated with lower DCR and ORR, metrics not previously explored by Zhang et al (65). This additional analysis provides new insights into the predictive value of the GPS/mGPS beyond survival outcomes. Furthermore, unlike prior meta-analyses that combined pre- and post-treatment GPS/mGPS data or failed to stratify by baseline characteristics, the current study maintains methodological rigor by stratifying analyses, reducing heterogeneity and enhancing clinical interpretability (65–67).

GPS or mGPS values are determined based on CRP and albumin, which reflect the level of inflammation and nutritional status of the body. Consistent evidence suggests that hypoalbuminemia is a significant predictor of poor prognosis in cancer patients (68–70). Serum albumin has been shown to have an immunomodulatory role, as it can bind to prostaglandin E2 (PGE2), which is associated with the downregulation of tumor necrosis factor-α derived from macrophages and contributes to immunosuppression (71–74). Of note, administering a human albumin solution has been found to counteract the effects of PGE2, leading to improved immune function (72). According to earlier research on the immunological role of serum CRP, elevated CRP levels are linked to lymphocytopenia, an elevation of proinflammatory cytokines, and decreased T lymphocyte and macrophage responses (75–77). The maintenance of tumor growth and tumor invasiveness are both facilitated by elevated proinflammatory cytokines (78,79). Inflammation mediators have been demonstrated to impede the synthesis of albumin, whereas oxidative stress may induce the denaturation of albumin, thereby fostering a precipitous decline in serum albumin concentrations among individuals with an inflammatory milieu (80–82). An additional pivotal aspect in tumorigenesis pertains to the tumor microenvironment. By orchestrating the recruitment of T lymphocytes, tumor-associated macrophages and circulating cytokines, inflammatory mediators can profoundly modulate various facets of tumor biology, encompassing cell proliferation and angiogenesis, as well as tumor invasion or metastasis (2,83,84).

Because of this, GPS and mGPS levels can significantly predict the efficacy of ICI therapy in cancer patients. In addition, GPS and mGPS serve as objective, highly reproducible ways to more accurately classify patients based on a three-index-scoring indicator. As a result, due to its well-established impact on the host's inflammation and nutritional status as well as cancer, the GPS and mGPS could serve as useful tools in predicting the therapeutic outcomes of ICIs in cancer patients. Individualized and timely nutritional and immunological interventions may improve the prognosis of patients with high baseline GPS and mGPS.

Notably, most studies included in the present analysis were retrospective, which may have limited the statistical robustness. In addition, the majority of data were derived from studies conducted in China and Japan, which may restrict the generalizability of the findings. The limited number of studies also precluded subgroup analyses for each cancer type and for different classes of ICIs. Therefore, further high-quality research with larger sample sizes, particularly multicenter prospective studies, is essential to validate and enhance the accuracy of the present results.

In conclusion, the current meta-analysis indicated that the GPS and mGPS may be reliable predictors of outcomes in cancer patients treated with ICIs.

 Supplementary Material Supporting Data Supporting Data Acknowledgements

Not applicable.

 Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

 Authors' contributions

HY and ML conceived and designed the study. HY and ML were responsible for the collection and assembly of data, data analysis and interpretation. HY was involved in writing the manuscript and ML revised the manuscript. HY and ML checked and confirm the authenticity of the raw data (pertaining to the pooled dataset). All authors have read and approved the final manuscript.

 Ethics approval and consent to participate

Not applicable.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References HanahanDWeinbergRAHallmarks of cancer: The next generationCell144646674201110.1016/j.cell.2011.02.01321376230 YuanQDengDPanCWeiTWuZZhangBLiSYinPShangDIntegration of transcriptomics, proteomics, and metabolomics data to reveal HER2-associated metabolic heterogeneity in gastric cancer with response to immunotherapy and neoadjuvant chemotherapyFront Immunol13951137202210.3389/fimmu.2022.95113735990657 PardollDMThe blockade of immune checkpoints in cancer immunotherapyNat Rev Cancer12252264201210.1038/nrc323922437870 RibasAWolchokJDCancer immunotherapy using checkpoint blockadeScience35913501355201810.1126/science.aar406029567705 YeFDewanjeeSLiYJhaNKChenZSKumarAVishakha BehlTJhaSKTangHAdvancements in clinical aspects of targeted therapy and immunotherapy in breast cancerMol Cancer22105202310.1186/s12943-023-01805-y37415164 MoradGHelminkBASharmaPWargoJAHallmarks of response, resistance, and toxicity to immune checkpoint blockadeCell18453095337202110.1016/j.cell.2021.09.02034624224 O'DonnellJSTengMWLSmythMJCancer immunoediting and resistance to T cell-based immunotherapyNat Rev Clin Oncol16151167201910.1038/s41571-018-0142-830523282 DolladilleCEderhySSassierMCautelaJThunyFCohenAAFedrizziSChrétienBDa-SilvaAPlaneAFImmune checkpoint inhibitor rechallenge after immune-related adverse events in patients with cancerJAMA Oncol6865871202010.1001/jamaoncol.2020.396032297899 ZhangLFengJKuangTChaiDQiuZDengWDongKZhaoKWangWBlood biomarkers predict outcomes in patients with hepatocellular carcinoma treated with immune checkpoint Inhibitors: A pooled analysis of 44 retrospective sudiesInt Immunopharmacol118110019202310.1016/j.intimp.2023.11001936933492 ZhaoJLiDXieSDengXWenXLiJWuZYangXLiMTangYNomogram for predicting prognosis of patients with metastatic melanoma after immunotherapy: A Chinese population-based analysisFront Immunol131083840202210.3389/fimmu.2022.108384036618343 VergnenegreAChouaidCEconomic analyses of immune-checkpoint inhibitors to treat lung cancerExpert Rev Pharmacoecon Outcomes Res21365371202110.1080/14737167.2021.186379033306411 SchalperKACarletonMZhouMChenTFengYHuangSPWalshAMBaxiVPandyaDBaradetTElevated serum interleukin-8 is associated with enhanced intratumor neutrophils and reduced clinical benefit of immune-checkpoint inhibitorsNat Med26688692202010.1038/s41591-020-0856-x32405062 ZhangLWangKKuangTDengWHuPWangWLow geriatric nutritional risk index as a poor prognostic biomarker for immune checkpoint inhibitor treatment in solid cancerFront Nutr101286583202310.3389/fnut.2023.128658338024341 RenJWangALiuJYuanQIdentification and validation of a novel redox-related lncRNA prognostic signature in lung adenocarcinomaBioengineered1243314348202110.1080/21655979.2021.195152234338158 HavelJJChowellDChanTAThe evolving landscape of biomarkers for checkpoint inhibitor immunotherapyNat Rev Cancer19133150201910.1038/s41568-019-0116-x30755690 ForrestLMMcMillanDCMcArdleCSAngersonWJDunlopDJEvaluation of cumulative prognostic scores based on the systemic inflammatory response in patients with inoperable non-small-cell lung cancerBr J Cancer8910281030200310.1038/sj.bjc.660124212966420 ProctorMJMorrisonDSTalwarDBalmerSMO'ReillyDSFoulisAKHorganPGMcMillanDCAn inflammation-based prognostic score (mGPS) predicts cancer survival independent of tumour site: A glasgow inflammation outcome studyBr J Cancer104726734201110.1038/sj.bjc.660608721266974 NagashimaYFunahashiKKagamiSUshigomeMKanekoTMiuraYYoshidaKKodaTKuriharaAWhich preoperative immunonutritional index best predicts postoperative mortality after palliative surgery for malignant bowel obstruction in patients with late-stage cancer? A single-center study in Japan comparing the modified Glasgow prognostic score (mGPS), the prognostic nutritional index (PNI), and the controlling nutritional status (CONUT)Surg Today532230202310.1007/s00595-022-02534-335781553 SchuettfortVMGustKD'AndreaDQuhalFMostafaeiHLaukhtinaEMoriKRinkMAbufarajMKarakiewiczPIImpact of the preoperative modified glasgow prognostic score on disease outcome after radical cystectomy for urothelial carcinoma of the bladderMinerva Urol Nephrol74302312202210.23736/S2724-6051.21.04216-634114784 HsuehSWLiuKHHungCYKuoYCTsaiCYHsuJTHungYSTsangNMChouWCSignificance of the glasgow prognostic score in predicting the postoperative outcome of patients with stage III gastric cancerJ Clin Med81448201910.3390/jcm809144831547247 VashistYKLoosJDedowJTachezyMUzunogluGKutupAYekebasEFIzbickiJRGlasgow prognostic score is a predictor of perioperative and long-term outcome in patients with only surgically treated esophageal cancerAnn Surg Oncol1811301138201110.1245/s10434-010-1383-720981494 YamadaSFujiiTYabusakiNMurotaniKIwataNKandaMTanakaCNakayamaGSugimotoHKoikeMClinical implication of inflammation-based prognostic score in pancreatic cancer: Glasgow prognostic score is the most reliable parameterMedicine (Baltimore)95e3582201610.1097/MD.000000000000358227149487 ShimoyamaRImamuraYUryuKMaseTOhtakiMOhtaniKShiragamiMFujimuraYHayashiMShinozakiNMinamiHInflammation-based prognostic markers in patients with advanced or recurrent gastric cancer treated with nivolumab: Tokushukai REAl-world Data project 02 (TREAD 02)Mol Clin Oncol2190202410.3892/mco.2024.278839421231 DikerOOlgunPAssociation of the immune-inflammation-nutritional parameters with immune checkpoint inhibitor outcomes in patients with advanced non-small cell lung cancerJ Oncol Sci84353202210.37047/jos.2021-87477 YamamotoYMatsuyamaHMatsumotoHSakanoSFujiNObaKYamamotoMKamiryoYHiraginoTNagaoKPrognostic value of risk stratification using blood parameters for nivolumab in Japanese patients with metastatic renal-cell carcinomaJpn J Clin Oncol50214220202010.1093/jjco/hyz16831755525 FreitasCJacobMTavaresNCruz-MartinsNSouto-MouraCAraújoDNovais-BastosHSantosVFernandesGMagalhãesAModified glasgow prognostic score predicts survival among advanced non-small cell lung carcinoma patients treated with anti-PD1 agentsAnticancer Drugs32567574202110.1097/CAD.000000000000106033661189 ZhangLMaWQiuZKuangTWangKHuBWangWPrognostic nutritional index as a prognostic biomarker for gastrointestinal cancer patients treated with immune checkpoint inhibitorsFront Immunol141219929202310.3389/fimmu.2023.121992937545502 ZhangLLiXWangKWuMLiuWWangWPrognostic impact of body composition in hepatocellular carcinoma patients with immunotherapyAnn Med562395062202410.1080/07853890.2024.239506239189472 ZhangLXiaZLiZZhangJWangKWangWInfluence of body fat tissue on outcomes in patients undergoing hepatectomy or liver transplantationInt J Surg111116711812024(Epub ahead of print)10.1097/JS9.000000000000186438920322 AliWASHuiPMaYWuYZhangYChenYHongSYangYHuangYZhaoYDeterminants of survival in advanced non-small cell lung cancer patients treated with anti-PD-1/PD-L1 therapyAnn Transl Med91639202110.21037/atm-21-170234988148 ArakiTTateishiKSoneharaKHirotaSKomatsuMYamamotoMKandaSKuraishiHHanaokaMKoizumiTClinical utility of the C-reactive protein:albumin ratio in non-small cell lung cancer patients treated with nivolumabThorac Cancer12603612202110.1111/1759-7714.1378833434414 BrownJTLiuYShabtoJMMartiniDRavindranathanDHitronEERusslerGACaulfieldSYantorniLJoshiSSModified glasgow prognostic score associated with survival in metastatic renal cell carcinoma treated with immune checkpoint inhibitorsJ Immunother Cancer9e002851202110.1136/jitc-2021-00285134326170 BrownJTLiuYShabtoJMMartiniDJRavindranathanDHitronEERusslerGACaulfieldSYantorniLBJoshiSSBaseline modified glasgow prognostic score associated with survival in metastatic urothelial carcinoma treated with immune checkpoint inhibitorsOncologist26397405202110.1002/onco.1372733634507 ChikuieNHamamotoTUedaTTaruyaTKonoTFuruieHIshinoTTakenoSBaseline neutrophil-to-lymphocyte ratio and glasgow prognostic score are associated with clinical outcome in patients with recurrent or metastatic head and neck squamous cell carcinoma treated with nivolumabActa Med Okayama75335343202134176937 FujiwaraRTakemuraKFujiwaraMYuasaTYasuokaSKomaiYNumaoNYamamotoSYoneseJModified glasgow prognostic score as a predictor of prognosis in metastatic renal cell carcinoma treated with nivolumabClin Genitourin Cancer19e78e83202110.1016/j.clgc.2020.10.00733279413 ImaiHKishikawaTMinemuraHYamadaYIbeTYamaguchiOMouriAHamamotoYKanazawaKKasaiTPretreatment glasgow prognostic score predicts survival among patients with high PD-L1 expression administered first-line pembrolizumab monotherapy for non-small cell lung cancerCancer Med1069716984202110.1002/cam4.422034414673 KangHSShinAYYeoCDKimSKParkCKKimJSKimSJLeeSHKimJWSignificance of glasgow prognostic scores in NSCLC patients treated with immunotherapy after platinum-based cytotoxic chemotherapyIn Vivo3534233430202110.21873/invivo.1264234697178 KasaharaNSunagaNTsukagoshiYMiuraYSakuraiRKitaharaSYokoboriTKairaKMogiAMaenoTPost-treatment glasgow prognostic score predicts efficacy in advanced non-small-cell lung cancer treated with anti-PD1Anticancer Res3914551461201910.21873/anticanres.1326230842182 KasajimaMIgawaSManakaHYamadaKAkazawaYManabeHYagamiYYamamotoHItoHKaizukaNThe glasgow prognostic score predicts outcomes of pembrolizumab or atezolizumab monotherapy in patients with pretreated non-small cell lung cancerOncology1016976202310.1159/00052696436103811 KawakamiHSunakawaYInoueEMatobaRNodaKSatoTSuminakaCYamakiMSakamotoYKawabataRSoluble programmed cell death ligand 1 predicts prognosis for gastric cancer patients treated with nivolumab: Blood-based biomarker analysis for the DELIVER trialEur J Cancer1841020202310.1016/j.ejca.2023.02.00336889037 KurosakiTKawakamiHMitaniSKawabataRTakahamaTNonagaseYFumitaSOzakiTChibaYTamuraTNakagawaKGlasgow prognostic score (GPS) and tumor response as biomarkers of nivolumab monotherapy in third- or later-line setting for advanced gastric cancerIn Vivo3419211929202010.21873/invivo.1198932606164 MadedduCBusquetsSDonisiCLaiEPrettaALópez-SorianoFJArgilésJMScartozziMMacciòAEffect of cancer-related cachexia and associated changes in nutritional status, inflammatory status, and muscle mass on immunotherapy efficacy and survival in patients with advanced non-small cell lung cancerCancers (Basel)151076202310.3390/cancers1504107636831431 MatsubaraTTakamoriSHaratakeNToyozawaRMiuraNShimokawaMYamaguchiMSetoTTakenoyamaMThe impact of immune-inflammation-nutritional parameters on the prognosis of non-small cell lung cancer patients treated with atezolizumabJ Thorac Dis1215201528202010.21037/jtd.2020.02.2732395289 MatsukiTOkamotoIFushimiCSawabeMKawakitaDSatoHTsukaharaKKondoTOkadaTTadaYHematological predictive markers for recurrent or metastatic squamous cell carcinomas of the head and neck treated with nivolumab: A multicenter study of 88 patientsCancer Med950155024202010.1002/cam4.312432441463 MatsuoMYasumatsuRMasudaMTohSWakasakiTHashimotoKJiromaruRManakoTNakagawaTInflammation-based prognostic score as a prognostic biomarker in patients with recurrent and/or metastatic head and neck squamous cell carcinoma treated with nivolumab therapyIn Vivo36907917202210.21873/invivo.1278035241549 MinichsdorferCGleissAAretinMBSchmidingerMFuerederTSerum parameters as prognostic biomarkers in a real world cancer patient population treated with anti PD-1/PD-L1 therapyAnn Med5413391349202210.1080/07853890.2022.207066035535695 MinoharaKMatobaTKawakitaDTakanoGOguriKMurashimaANakaiKIwakiSHojoWMatsumuraANovel prognostic score for recurrent or metastatic head and neck cancer patients treated with nivolumabSci Rep1116992202110.1038/s41598-021-96538-734417539 MortensenREJHolmströmMOLisleTLHasselbyJPWillemoeGLMetÖMarie SvaneIJohansenJNielsenDLChenIMAndersenMHPre-existing TGF-β-specific T-cell immunity in patients with pancreatic cancer predicts survival after checkpoint inhibitors combined with radiotherapyJ Immunother Cancer11e006432202310.1136/jitc-2022-00643236948507 NamikawaTYokotaKTaniokaNFukudomeIIwabuJMunekageMUemuraSMaedaHKitagawaHKobayashiMHanazakiKSystemic inflammatory response and nutritional biomarkers as predictors of nivolumab efficacy for gastric cancerSurg Today5014861495202010.1007/s00595-020-02048-w32542414 NiwaKKawakitaDNagaoTTakahashiHSaotomeTOkazakiMYamazakiKOkamotoIHiraiHSaigusaNMulticentre, retrospective study of the efficacy and safety of nivolumab for recurrent and metastatic salivary gland carcinomaSci Rep1016988202010.1038/s41598-020-73965-633046752 NoguchiGNakaigawaNUmemotoSKobayashiKShibataYTsutsumiSYasuiMOhtakeSSuzukiTOsakaKC-reactive protein at 1 month after treatment of nivolumab as a predictive marker of efficacy in advanced renal cell carcinomaCancer Chemother Pharmacol867585202010.1007/s00280-020-04088-y32537714 OguraYKataokaNKunimatsuYTachibanaYSugimotoTTaniNSatoIHiroseKKatoDTakedaTPredictors of survival among Japanese patients receiving first-line chemoimmunotherapy for advanced non-small cell lung cancerThorac Cancer1297105202110.1111/1759-7714.1372033124197 SakaiDOmoriTFumitaSFujitaJKawabataRMatsuyamaJYasuiHHiraoMKawaseTKishiKReal-world effectiveness of third- or later-line treatment in Japanese patients with HER2-positive, unresectable, recurrent or metastatic gastric cancer: A retrospective observational studyInt J Clin Oncol2711541163202210.1007/s10147-022-02162-435489010 TadaTKumadaTHiraokaAKariyamaKTaniJHirookaMTakaguchiKAtsukawaMFukunishiSItobayashiENew prognostic system based on inflammation and liver function predicts prognosis in patients with advanced unresectable hepatocellular carcinoma treated with atezolizumab plus bevacizumab: A validation studyCancer Med1269806993202310.1002/cam4.549536484470 TakamoriSTakadaKShimokawaMMatsubaraTFujishitaTItoKToyozawaRYamaguchiMOkamotoTYoneshimaYClinical utility of pretreatment glasgow prognostic score in non-small-cell lung cancer patients treated with immune checkpoint inhibitorsLung Cancer1522733202110.1016/j.lungcan.2020.11.02633341085 TanakaTYoshidaTMasudaKTakeyasuYShinnoYMatsumotoYOkumaYGotoYHorinouchiHYamamotoNOheYPrognostic role of modified glasgow prognostic score in elderly non-small cell lung cancer patients treated with anti-PD-1 antibodiesRespir Investig617481202310.1016/j.resinv.2022.10.00336460585 TokuyamaNTakegawaNNishikawaMSakaiAMimuraTKushidaSTsumuraHYamamotoYMikiITsudaMPretreatment glasgow prognostic score as a predictor of outcomes in nivolumab-treated patients with advanced gastric cancerPLoS One16e0247645202110.1371/journal.pone.024764533635904 UekiYTakahashiTOtaHShodoRYamazakiKHoriiAPredicting the treatment outcome of nivolumab in recurrent or metastatic head and neck squamous cell carcinoma: Prognostic value of combined performance status and modified Glasgow prognostic scoreEur Arch Otorhinolaryngol27723412347202010.1007/s00405-020-05945-532239313 WasamotoSImaiHTsudaTNagaiYMinemuraHYamadaYUmedaYKishikawaTShionoAKozuYPretreatment glasgow prognostic score predicts survival among patients administered first-line atezolizumab plus carboplatin and etoposide for small cell lung cancerFront Oncol121080729202310.3389/fonc.2022.108072936741711 YamamotoYYatsudaJShimokawaMFujiNAokiASakanoSYamamotoMSugaATeiYYoshihiroSPrognostic value of pre-treatment risk stratification and post-treatment neutrophil/lymphocyte ratio change for pembrolizumab in patients with advanced urothelial carcinomaInt J Clin Oncol26169177202110.1007/s10147-020-01784-w32948925 YangZZhangDZengHFuYHuZPanYChenJWangJZhangYZhouZInflammation-based scores predict responses to PD-1 inhibitor treatment in intrahepatic cholangiocarcinomaJ Inflamm Res1557215731202210.2147/JIR.S38592136238770 ZaitsuJYamashitaYIshikawaASaitoAKagimotoAMimuraTHirakawaTMitoMFukuharaKSenooTSystemic inflammatory score predicts response and prognosis in patients with lung cancer treated with immunotherapyAnticancer Res4136733682202110.21873/anticanres.1515834230166 TakegawaNHirabayashiTTanakaSNishikawaMTokuyamaNMimuraTKushidaSTsumuraHYamamotoYMikiITsudaMThe impact of nutritional status in nivolumab-treated patients with advanced esophageal cancerPLoS One18e0285365202310.1371/journal.pone.028536537146004 TanakaKHirakawaHSuzukiMHigaTAgenaSHasegawaNKawakamiJToyamaMHigaTKinjyoHBiomarkers for predicting anti-programmed cell death-1 antibody treatment effects in head and neck cancerCurr Oncol3054095424202310.3390/curroncol3006041037366893 ZhangYChenSChenHLiWA comprehensive analysis of glasgow prognostic score (GPS)/the modified glasgow prognostic score (mGPS) on immune checkpoint inhibitor efficacy among patients with advanced cancerCancer Med123848202310.1002/cam4.494035702873 WuJWangHYangRLiuDLiWPrognostic significance of the modified Glasgow Prognostic Score in NSCLC patients undergoing immune checkpoint inhibitor therapy: A meta-analysisFront Oncol141449853202410.3389/fonc.2024.144985339464717 WangHYangRChengCWangSLiuDLiWPrognostic value of the glasgow prognostic score in non-small cell lung cancer patients receiving immunotherapy: A meta-analysisNutr Cancer76187195202410.1080/01635581.2023.229452238140926 YaoZHTianGYYangSXWanYYKangYMLiuQHYaoFLinDJSerum albumin as a significant prognostic factor in patients with malignant pleural mesotheliomaTumour Biol3568396845201410.1007/s13277-014-1938-525051913 GuptaDLisCGPretreatment serum albumin as a predictor of cancer survival: A systematic review of the epidemiological literatureNutr J969201010.1186/1475-2891-9-6921176210 MiuraKHamanakaKKoizumiTKitaguchiYTeradaYNakamuraDKumedaHAgatsumaHHyogotaniAKawakamiSClinical significance of preoperative serum albumin level for prognosis in surgically resected patients with non-small cell lung cancer: Comparative study of normal lung, emphysema, and pulmonary fibrosisLung Cancer1118895201710.1016/j.lungcan.2017.07.00328838406 YangJPetersenCEHaCEBhagavanNVStructural insights into human serum albumin-mediated prostaglandin catalysisProtein Sci11538545200210.1110/ps.2870211847277 ChinaLMainiASkeneSSShabirZSylvestreYColasRALyLBecares SallesNBellotiVDalliJAlbumin counteracts immune-suppressive effects of lipid mediators in patients with advanced liver diseaseClin Gastroenterol Hepatol16738747.e7201810.1016/j.cgh.2017.08.02728859868 ChoeWHBaikSKProstaglandin E2-mediated immunosuppression and the role of albumin as its modulatorHepatology6110801082201510.1002/hep.2764425482406 ArroyoVMoreauRTying up PGE2 with albumin to relieve immunosuppression in cirrhosisNat Med20467469201410.1038/nm.355324804750 Du ClosTWMoldCC-reactive protein: An activator of innate immunity and a modulator of adaptive immunityImmunol Res30261277200410.1385/IR:30:3:26115531769 NozoeTMatsumataTSugimachiKPreoperative elevation of serum C-reactive protein is related to impaired immunity in patients with colorectal cancerAm J Clin Oncol23263266200010.1097/00000421-200006000-0001110857890 CoussensLMWerbZInflammation and cancerNature420860867200210.1038/nature0132212490959 AbramovitchRMarikovskyMMeirGNeemanMStimulation of tumour growth by wound-derived growth factorsBr J Cancer7913921398199910.1038/sj.bjc.669022310188881 CannaKHilmyMMcMillanDCSmithGWMcKeeRFMcArdleCSMcNicolAMThe relationship between tumour proliferative activity, the systemic inflammatory response and survival in patients undergoing curative resection for colorectal cancerColorectal Dis10663667200810.1111/j.1463-1318.2007.01416.x18005189 CoffeltSBde VisserKECancer: Inflammation lights the way to metastasisNature5074849201410.1038/nature1306224572360 BitoRHinoSBabaATanakaMWatabeHKawabataHDegradation of oxidative stress-induced denatured albumin in rat liver endothelial cellsAm J Physiol Cell Physiol289C531C542200510.1152/ajpcell.00431.200415872008 WuJDingPWuHYangPGuoHTianYMengLZhaoQSarcopenia: Molecular regulatory network for loss of muscle mass and functionFront Nutr101037200202310.3389/fnut.2023.103720036819699 CondeelisJPollardJWMacrophages: Obligate partners for tumor cell migration, invasion, and metastasisCell124263266200610.1016/j.cell.2006.01.00716439202 DingPGuoHHeXSunCLoweSBentleyRZhouQYangPTianYLiuYEffect of skeletal muscle loss during neoadjuvant imatinib therapy on clinical outcomes in patients with locally advanced GISTBMC Gastroenterol22399202210.1186/s12876-022-02479-436028812

Flow diagram for identifying eligible studies.

Forest plots for the relationship between GPS and overall survival. (A) GPS 2 vs. 0; (B) GPS 1 vs. 0. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; GPS, Glasgow prognostic score.

Forest plots for the relationship between GPS and progression-free survival. (A) GPS 2 vs. 0; (B) GPS 1 vs. 0. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; GPS, Glasgow prognostic score.

(A and B) Forest plots for the relationship between GPS and disease control rate. (A) GPS 2 vs. 1/0; (B) GPS 2/1 vs. 0. (C and D) Funnel plots. (C) GPS 2 vs. 1/0; (D) GPS 2/1 vs. 0. (E and F) Sensitivity analysis for the association between GPS and disease control rate. (E) GPS 2 vs. 1/0; (F) GPS 2/1 vs. 0. OR, odds ratio; CI, confidence interval; DL, DerSimonian and Laird; GPS, Glasgow prognostic score; Lnhr, the natural logarithm of the hazard ratio; s.e., standard error.

(A and B) Forest plots for the relationship between mGPS and overall survival. (A) mGPS 2 vs. 0; (B) mGPS 1 vs. 0. (C and D) Funnel plots. (C) mGPS 2 vs. 0; (D) mGPS 1 vs. 0. (E and F) Sensitivity analysis for the association between mGPS and overall survival. (E) mGPS 2 vs. 0; (F) mGPS=1 vs. 0. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; mGPS, modified Glasgow prognostic score; Lnhr, the natural logarithm of the hazard ratio; s.e., standard error.

Forest plots for the relationship between mGPS and overall survival. (A) mGPS 2 vs. 1/0; (B) mGPS 2/1 vs. 0; (C and D) Funnel plots. (C) mGPS 2 vs. 1/0; (D) mGPS 2/1 vs. 0. (E) continuous variables. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; mGPS, modified Glasgow prognostic score; Lnhr, the natural logarithm of the hazard ratio; s.e., standard error.

(A and B) Forest plots for the relationship between mGPS and progression-free survival. (A) mGPS 2 vs. 0; (B) mGPS 1 vs. 0. (C and D) Funnel plots. (C) mGPS 2 vs. 0; (D) mGPS 1 vs. 0. (E and F) Sensitivity analysis for the association between mGPS and progression-free survival. (E) mGPS 2 vs. 0; (F) mGPS 1 vs. 0. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; mGPS, modified Glasgow prognostic score; Lnhr, the natural logarithm of the hazard ratio; s.e., standard error.

Forest plots for the relationship between mGPS and progression-free survival. (A) mGPS 2 vs. 1/0; (B) Funnel plots of mGPS 2 vs. 1/0; (C) continuous variables. HR, hazard ratio; CI, confidence interval; DL, DerSimonian and Laird; mGPS, modified Glasgow prognostic score; Lnhr, the natural logarithm of the hazard ratio; s.e., standard error.

Main characteristics of the studies included.

First author, year Study design Study period Study region ICI treatment Cancer type Sample size Age, years Gender (male/female) Outcome Follow-up, months (median/mean) NOS score (Refs.)
Tanaka, 2023 R 12/2015-04/2019 Japan Nivolumab, Pembrolizumab NSCLC 51 79 (75–87)a 40/11 GPS (OS, PFS, ORR, DCR) 7b 7 (56)
Takegawa, 2023 R 03/2020-09/2021 Japan Nivolumab EC 37 67 (46–84)a 32/5 GPS (OS, PFS) 23.3b 7 (63)
Madeddu, 2023 P 03/2017-08/2021 Italy Nivolumab, Pembrolizumab NSCLC 74 69.3 ± 11.3c 54/20 mGPS (OS, PFS) 24b 8 (42)
Kasajima, 2023 R 03/2017-01/2022 Japan Pembrolizumab, Atezolizumab NSCLC 80 71 (44–86)a 56/24 GPS (OS, PFS, ORR, DCR) 11.1b 7 (39)
Wasamoto, 2023 R 08/2019- 05/2021 Japan Atezolizumab SCLC 84 71 (43–89)a 70/14 GPS (OS, PFS, ORR, DCR) 12.9b 8 (59)
Diker, 2022 R 03/2017-10/2021 Cyprus ICIs NSCLC 102 67 (35–88)a 89/13 mGPS (OS, PFS) 8.4b 8 (24)
Zaitsu, 2021 R 10/2016-04/2020 Japan Nivolumab, Pembrolizumab, Atezolizumab LC 73 70.9±9.4c 52/21 mGPS (OS, PFS, DCR) - 7 (62)
Takamori, 2021 R 01/2016-12/2019 Japan Nivolumab, Pembrolizumab, Atezolizumab NSCLC 304 66 (31–88)a 242/62 GPS (OS, PFS, DCR), mGPS (OS, PFS, DCR), 13.8b 8 (55)
Ogura, 2021 R 02/2019-07/2020 Japan Pembrolizumab, Atezolizumab NSCLC 34 72 (55–81)a 29/5 mGPS (OS, PFS) - 6 (52)
Kang, 2021 R 01/2018-03/2020 Korea Pembrolizumab, Nivolumab, Atezolizumab NSCLC 78 67.1±9.2c 64/14 GPS (OS, PFS) - 7 (37)
Imai, 2021 R 02/2017-06/2019. Japan Pembrolizumab NSCLC 142 70 (47–86)a 117/25 GPS (OS, PFS, ORR, DCR) 15.7b 8 (36)
Freitas, 2021 R - Portugal Nivolumab, Pembrolizumab NSCLC 77 65 (44–87)a 55/22 mGPS (OS, PFS) - 7 (26)
Araki, 2021 R 01/2015-12/2019 Japan Nivolumab NSCLC 113 69 (36–86)a 87/26 mGPS (OS) - 7 (31)
Ali, 2021 R 12/2015-08/2017 China Pembrolizumab, Nivolumab, Camrelizumab, Atezolizumab NSCLC 73 54 (28–73)a 51/22 mGPS (OS, PFS) 21.2b 8 (30)
Matsubara, 2020 R 01/2018-03/2019 Japan Atezolizumab NSCLC 24 64.5±9.7c 17/7 mGPS (OS) - 6 (43)
Kasahara, 2019 R 01/2016-06/2018 Japan Nivolumab, Pembrolizumab NSCLC 47 33/14d 37/17 GPS (OS, PFS, DCR) 14.7b 7 (38)
Kawakami, 2023 P - Japan Nivolumab GC 439 70 (26–90)a 321/118 GPS (OS, PFS, ORR, DCR) - 7 (40)
Sakai, 2022 R 09/2017-03/2020 Japan Nivolumab GC 100 71 (38–89)a 78/22 GPS (OS) 5.0b 7 (53)
Tokuyama, 2021 R 02/2015-06/2019 Japan Nivolumab GC 45 65 (40–81)a 31/14 GPS (OS, ORR) - 6 (57)
Namikawa, 2020 R 10/2017-12/2019 Japan Nivolumab (49–86)a GC 29 71 19/10 GPS (OS, PFS) - 6 (49)
Kurosaki, 2020 R 10/2017-03/2019 Japan Nivolumab GC 80 71 (43–87)a 67/13 GPS (OS, PFS, ORR, DCR) 5.1b 7 (41)
Matsuo, 2022 R 04/2017-032020 Japan Nivolumab HNSCC 164 65 (23–87)a 127/37 GPS (OS, ORR, DCR) 12.6b 8 (45)
Minohara, 2021 R 04/2017-10/2019 Japan Nivolumab HNC 126 68 (35–90)a 104/22 mGPS (OS, PFS) 7.5b 7 (47)
Chikuie, 2021 R 05/2017-10/2019 Japan Nivolumab HNSCC 56 66 (31–90)a 40/16 GPS (OS, PFS) - 7 (34)
Ueki, 2020 R 03/2017-03/2019 Japan Nivolumab HNSCC 42 68 (43–79)a 35/7 GPS (OS) 9.6b 7 (58)
Matsuki, 2020 R 05/2017-08/2018 Japan Nivolumab HNSCC 88 49/39d 71/17 mGPS (OS, PFS, DCR) 6.1b 7 (44)
Yamamoto, 2021 R 2015-2019 Japan Pembrolizumab UC 121 74 (50–86)a 87/34 GPS (OS), mGPS (OS) 7.9e 8 (60)
Brown, 2021 (UC) R 2015-2018 USA Atezolizumab, Pembrolizumab, Nivolumab UC 53 70 (32–86)a 45/8 mGPS (OS, PFS) 27.1b 8 (33)
Fujiwara, 2021 R 09/2013-08/2019 Japan Nivolumab RCC 45 62 (55–69)b - GPS (OS) 26.4b 7 (35)
Brown, 2021 (RCC) R 2015-2020 USA ICIs RCC 156 64 (23–90)a 108/48 mGPS (OS, PFS) 24.2b 8 (32)
Yamamoto, 2020 R 2016-2018 Japan Nivolumab RCC 65 69 (39–83)a 51/14 GPS (OS) 11.6b 7 (25)
Noguchi, 2020 R - Japan Nivolumab RCC 64 69 (30–86)a 51/13 GPS (PFS) 8.3b 6 (51)
Minichs-dorfer, 2022 R 01/2015-11/2016 Austria Pembrolizumab, Nivolumab Pan-cancer 114 60 (22–88)a 74/40 GPS (OS, PFS) - 7 (46)
Tada, 2023 R 09/2020-03/2022 Japan Atezolizumab HCC 421 74 (68–79)f 340/81 GPS (OS, PFS) 8.7b 7 (54)
Yang, 2022 R 02/2019-02/2021 China Nivolumab, Pembrolizumab, Toripalimab, Camrelizumab, Sintilimab ICC 73 57 (31–75)a 49/24 GPS (OS) 11.2e 8 (61)
Mortensen, 2023 P - Denmark Nivolumab ± Ipilimumab PC 32 68 (47–80)a 18/14 mGPS (OS, PFS) 7.3b 7 (48)
Tanaka, 2023 R 04/2017-12/2020 Japan Nivolumab HNSCC 42 61 (26–81)a 36/6 GPS (ORR, DCR) - 6 (64)
Niwa, 2020 R 05/2017-09/2019 Japan Nivolumab SGC 24 56 (29–82)a 19/5 mGPS (OS, PFS) 6.5b 6 (50)

Median (range);

median;

mean ± standard deviation;

≥65 vs. <65;

mean;

median (interquartile range). R, retrospective study; P, prospective study; ICIs, immune checkpoint inhibitors; NSCLC, non-small cell lung cancer; LC, lung cancer; GC, gastric cancer; HNSCC, head and neck squamous cell carcinoma; HNC, head and neck cancer; UC, urothelial carcinoma; RCC, renal cell carcinoma; ICC, intrahepatic cholangiocarcinoma; PC, pancreatic cancer; SGC, salivary gland carcinoma.