Introduction
In recent years, more research evidence on immunotherapy has been emerging in the field of oncology. Immune checkpoint inhibitors (ICIs) have demonstrated notable efficacy in both perioperative and advanced cancer treatment (1,2). Nevertheless, the problems of hyper-progression and serious adverse reactions after receiving immunotherapy cannot be ignored (3,4). At present, the expression level of programmed death ligand-1 (PD-L1) is the main predictive biomarker for immunotherapy. However, a study has found that a small proportion of patients with negative PD-L1 expression remain sensitive to immunotherapy (5). In addition, there is still a lack of unified standards for the detection method of PD-L1. The prognostic value of tumor mutational burden (TMB) as a biomarker for immunotherapy has been validated in clinical studies and using real-world data (6,7). However, due to the lack of a unified detection standard, the inability to fully reflect the tumor immune microenvironment, the differences between different tumor types and the limitations of dynamic changes, the application of TMB is limited (6). Therefore, PD-L1 expression and TMB cannot predict the prognosis and efficacy of immunotherapy fully and effectively. It is therefore imperative to explore new predictive markers for identifying patients with a potentially poor prognosis at an earlier stage and to provide supportive treatment to improve treatment outcomes to further expand the population benefiting from immunotherapy.
The value of inflammatory markers and nutritional status in evaluating tumor prognosis has gradually become a research hotspot. Specifically, the prognostic scoring systems based on inflammatory markers, such as the Glasgow prognostic score (GPS), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), immunoinflammatory index (SII), lymphocyte-to-monocyte ratio (LMR) and prognostic nutritional index (PNI), have been shown to have predictive value in the prognosis of antitumor therapy (3,8–10). The GPS, originally proposed by Forrest et al (11), is a scoring system based on a combination of C-reactive protein (CRP) and albumin levels. Albumin is a biomarker of nutritional status (12), whereas CRP is a sensitive indicator of the inflammatory response and has also been shown to promote the formation of an immunosuppressive tumor microenvironment and the growth of tumor cells (13). CRP can be used as an independent prognostic factor for a variety of malignant tumors such as gastric and pancreatic cancer (14,15). Recently, CRP has also been found to predict the efficacy of programmed cell death protein 1 (PD-1) treatment (16). Hypoalbuminemia is a manifestation of cachexia in patients with advanced tumors and severe inflammatory damage can also lead to hypoalbuminemia. Decreased albumin can weaken the ability of cells to activate immunity and antioxidant (17,18). Thus, the GPS combines the two indicators, CRP and albumin, to more effectively evaluate the prognosis of patients with tumors.
The GPS criteria are as follows: CRP elevation (CRP >10 mg/l) combined with hypoalbuminemia (albumin <35 g/l) is assigned a score of 2, only one abnormality is assigned a score of 1 and a score of 0 is assigned when both indicators are normal (19). In 2007, McMillan et al (20) refined the scoring system to create the modified GPS (mGPS). The improved scoring system emphasizes that individuals with elevated CRP and normal albumin are scored as 1 point. Studies have shown that the mGPS system is more effective in reflecting systemic inflammatory responses than isolated inflammatory indicators (21,22). GPS and mGPS have gradually been proven to have value in evaluating the prognosis of patients receiving immunotherapy (23,24). However, with the continuous emergence of new research, the predictive value of GPS and mGPS for immunotherapy deserves further validation.
In the present study, the literature on the association between GPS or mGPS and the prognosis of patients with advanced cancer receiving immunotherapy was reviewed, and a meta-analysis was conducted to demonstrate the role of GPS or mGPS in predicting overall survival (OS) or progression-free survival (PFS).
Discussion
GPS and mGPS have been shown to have prognostic value in multiple cancer types and different treatment stages (46,47). Although other inflammatory indicators, such as NLR, PLR, PNI and SII, have some predictive value, they lack a comprehensive assessment of nutritional status and are less consistent across cancer types. Specifically, Yamanouchi et al (48) evaluated the effects of inflammatory and nutritional indicators such as GPS, NLR and PNI on the prognosis of patients with metastatic breast cancer, and only GPS was indicated to be an independent predictor of OS. GPS was also shown to be an independent indicator of prognosis in patients with hepatocellular carcinoma, and was superior to other inflammatory prognostic scores in terms of prognostic power (49). Shimoyama et al (50) evaluated the prognostic value of 17 inflammatory markers in patients with advanced or recurrent gastric cancer who received opdivo and found that mGPS had the strongest association with patient outcome. GPS or mGPS also have the characteristics of being cheap, convenient and easy to obtain, and have great application potential (46).
The present meta-analysis, which included data from 18 studies involving 1,355 patients with cancer, assessed the association between GPS or mGPS and survival in patients receiving ICIs. The results indicated that a higher GPS or mGPS contributed to poorer survival in patients with cancer, and subgroup analyses also demonstrated that patients with GPS or mGPS 1 and GPS or mGPS 2 exhibited poorer OS and PFS than those with GPS 0. However, due to the heterogeneity of some of the data, caution is required in interpreting these results.
GPS, based on CRP and albumin levels, has been proven to have independent prognostic value in various cancer types in previous years (46,51). CRP and albumin are sensitive and reliable indicators that reflect the inflammatory state and immune nutritional status of a cancer population, and have been widely confirmed as biomarkers of poor prognosis in cancer (46). CRP is an acute phase response protein that can serve as a biomarker for infection and tissue damage (52). CRP can promote an inflammatory response (53), thereby inhibiting immune function and promoting tumor proliferation and metastasis (54). At the same time, CRP can also suppress the immune response, accelerate tumor migration and tumor microenvironment formation (54), which may weaken the role of ICIs. Previous studies have found that CRP can directly inhibit T cells and dendritic cells (55,56), thereby affecting the action of ICIs by impacting innate and acquired immunity in patients with cancer. The CRP level before treatment can not only predict the therapeutic effect of ICIs (57), but can also predict the survival of patients with advanced urothelial cancer or non-small cell lung cancer receiving immunotherapy (58,59). The serum albumin level is a notable indicator of nutritional status, and low albumin levels are a manifestation of malnutrition and poor general status (60). Hypoproteinemia reflects poor nutritional status and chronic inflammation, and can serve as a biomarker for cancer prognosis (61). Studies have identified serum albumin as a potential biomarker for evaluating the efficacy of ICIs, whether alone or in combination with chemotherapy (61,62). The results of the present study showed that a high GPS was associated with unfavorable OS and PFS, suggesting poor survival outcomes for patients treated with ICIs, regardless of hypoproteinemia or elevated CRP. However, the OS results were heterogeneous. Sensitivity analysis showed that the study by Kasajima et al (34) contributed significantly to heterogeneity. The study was divided into GPS 1–2 and GPS 0 groups, with a tumor type of non-small cell lung cancer, treatment regimen of pembrolizumab or altezomumab monotherapy and a treatment line of second-line or subsequent-line therapy (34). There were no significant differences in clinical features between the study by Kasajima et al (34) and other included studies. The heterogeneity in the OS results was reduced after subgroup analysis based on the scores. This may be due to the subgroups being divided reasonably, which makes the studies within each subgroup more homogeneous in certain key features. The grouping of different GPSs is likely to be the main cause of heterogeneity.
Extensive evidence suggests that weight loss and poor physical fitness are related to systemic inflammatory responses in advanced cancer (3,63,64). Significant inflammatory response can lead to hypoalbuminemia (65), and cancer-related inflammation can impair albumin synthesis by altering cytokine production and thereby increasing microvascular permeability (66). Therefore, hypoalbuminemia is more likely to be secondary to elevated CRP levels. There is growing evidence that inflammation has a role in cancer development, and that inflammation may also be accelerated by the cancer itself due to increased catabolism and malnutrition (67,68). McMillan et al (20) revised the GPS to mGPS to more accurately predict the prognosis of various cancer types. mGPS weakens the effect of albumin, emphasizing the importance of CRP. GPS 0 is determined by the CRP and albumin levels, but mGPS 0 is determined by CRP alone, regardless of albumin levels (20). For mGPS 1, elevated CRP is rated as 1 even if albumin levels are normal (20), but GPS 1 includes patients with hypoalbuminemia who do not have elevated CRP levels. The rating criteria for GPS 2 and mGPS 2 are the same and both combine high CRP and hypoproteinemia, which may indicate strong systemic inflammation and poor nutritional status. The results of the present study demonstrated that mGPS, like GPS, can serve as a predictor of OS and PFS in cancer populations receiving immunotherapy. However, there was high heterogeneity in the PFS data. Sensitivity analysis did not identify any study that had a notable impact on heterogeneity. After subgroup analysis by country and score, significant heterogeneity still existed, suggesting that the present analysis failed to fully explain the source of heterogeneity and further exploration and verification are needed.
The evaluation of GPS or mGPS only requires routine blood tests, which is low cost, highly standardized and suitable for clinical promotion. The evaluation can be adapted to different cancer types or population characteristics, while other static indicators lack such optimization space. For example, mGPS further improves the predictive sensitivity and specificity of GPS by adjusting the critical value of CRP and albumin. The cut-off values may vary across different diseases and even within different cohorts of the same disease. Therefore, cut-off values can be determined clinically according to actual disease conditions, such as tumor type and ethnicity. Compared with PD-L1 and TMB, which focus on the local characteristics of tumors, the core advantages of GPS and mGPS lie in their comprehensiveness, universality and economy, and are more suitable for widespread clinical promotion and the construction of a comprehensive prognostic evaluation system. GPS or mGPS can be used in combination with immunotherapy-related markers to provide multidimensional information. For example, a high inflammatory state may impair efficacy in patients with high PD-L1 expression, and the combination of GPS or mGPS and PD-L1 detection may optimize patient stratification. Systemic inflammation may indirectly regulate immunotherapy response by affecting intestinal flora, and GPS or mGPS can be used as a proxy indicator of systemic inflammation in comprehensive analysis.
The present study has some limitations. First, the included studies were retrospective and showed notable methodological diversity. Second, although subgroup and sensitivity analyses were performed, heterogeneity was not completely eliminated. Heterogeneity may be caused by multiple factors, such as tumor type, ICI regimens, treatment lines, the diversity of the study population and an inconsistent follow-up time. The small number of studies within subgroups also resulted in insufficient statistical power to accurately assess differences between subgroups. Third, the absence of patient-level data may limit the assessment of certain baseline characteristics and affect the interpretation of results. Therefore, there is still a need to design clinical studies to further validate the relevance of GPS and mGPS in predicting the survival of each patient with cancer. In particular, further research is needed on specific factors, such as differences in the effectiveness of interventions in different populations.
In conclusion, high GPS or mGPS, that is, either high CRP or low albumin, may have adverse effects on the OS and PFS of patients with cancer receiving ICIs. In clinical practice, patients can be risk stratified according to GPS or mGPS prior to immunotherapy, thereby enhancing patient management and improving treatment outcomes. Future prospective cohort studies are needed to validate the results of the present study, and further studies in combination with other prognostic markers are needed to improve the prediction accuracy.