Abstract
Background/Aim: Little is known about the prognostic role of the Glasgow prognostic score (GPS) in non-small cell lung cancer (NSCLC) patients treated with immunotherapy after platinum-based cytotoxic chemotherapy. Patients and Methods: This study used a lung cancer cohort of the Catholic Medical Center of Korea between January 2018 and September 2020. Results: A total of 78 patients with NSCLC treated with immunotherapy as second or further-line therapy were included. Higher GPS values were significant predictors of shorter immune-related progression-free survival (irPFS) and overall survival (OS). The hazard ratios for irPFS were 0.249 for programmed death-ligand 1 (PD-L1) expression ≥50% and 9.73 for a GPS of 2. Older age, lower PD-L1 expression and higher GPS values were independently associated with shorter OS. Conclusion: Higher GPS values were identified as a poor prognostic factor for OS and irPFS in NSCLC patients who received immunotherapy as second or further-line therapy.
The introduction of immune-checkpoint inhibitors (ICIs) into the therapy for non-small cell lung cancer (NSCLC) has transformed the therapeutic landscape of metastatic NSCLC (1). The expression of programmed death-ligand 1 (PD-L1) on tumor cells and tumor mutation burden (TMB) have been used to select patients suitable for ICIs but not all patients with these predictive factors benefit from ICIs (2). However, the proper predictive factors for using ICIs to treat NSCLC patients have not been developed.
Inflammation is an important factor in tumor progression and is associated with poor response to treatment. The response to cancer treatment depends not only on the tumor’s characteristics and tumor stage but also on patient-related factors including nutrition and inflammation status (3). Recently, systemic inflammatory response to predict progression and survival has been reported in patients with malignancies. Thus, cancer-related prognosis has been validated using inflammatory markers such as the neutrophil-lymphocyte ratio (NLR) or the lung immune prognostic index consisting of the NLR and lactate dehydrogenase (LDH) levels, and the systemic inflammation response index (2, 4).
The Glasgow prognostic score (GPS), reflecting the host systemic inflammatory response and immune status has been validated as a prognostic factor in many malignancies. The GPS is the combination of the C-reactive protein (CRP) concentration (>10 mg/l) and hypoalbuminemia (<35 g/l) (5, 6). The association of the GPS and ICIs in lung cancer patients has been explored previously; however, previous studies have had small sample sizes or evaluated the post-treatment GPS in lung cancer patients treated with ICIs.
Few studies have evaluated the prognostic value of the GPS in metastatic NSCLC patients treated with ICIs who received at least one regimen of cytotoxic chemotherapy before the administration of ICIs (7, 8). Recently Takamori et al. reported the clinical utility of pre-treatment GPS in advanced or recurrent NSCLC patients treated with ICIs but included patients treated with ICIs as first line treatment.(9) In this study, we retrospectively analyzed the survival and immune-related progression-free survival (irPFS) data of patients with metastatic NSCLC and explored the prognostic role of the GPS in these patients.
Patients and Methods
Patients. We used the lung cancer registry database of the Catholic Medical Center, Seoul, Republic of Korea. Since October 2014, seven hospitals of the Catholic University of Korea (Seoul St. Mary’s Hospital, Yeouido St. Mary’s Hospital, Eunpyeong St. Mary’s Hospital, Uijeongbu St. Mary’s Hospital, Bucheon St. Mary’s Hospital, Incheon St. Mary’s Hospital and St. Vincent’s Hospital) have consecutively enrolled lung cancer patients. Clinical information including stage, pathology, treatment modality and survival was systematically recorded by qualified managers to improve data accuracy. The researchers were permitted access by newly assigned serial numbers and anonymizing the dataset.
Patients were eligible for the study if they were diagnosed from January 2018 to March 2020 with histologically confirmed NSCLC and received ICIs as second-line or further-line therapy after treatment failure with platinum-based cytotoxic chemotherapy. Patients who received ICIs as first-line therapy, were diagnosed with small cell lung cancer, received post-chemoradiation consolidation treatment with durvalumab, were lost to follow up, or had no pre-treatment CRP and albumin levels were excluded. The study flow is summarized in Figure 1. The follow-up period ended on September 30, 2020.
Study flow diagram. ICI, Immune-checkpoint inhibitors; SCLC, small cell lung cancer; CRP, C-reactive protein; NSCLC, non-small cell lung cancer; GPS, Glasgow prognostic score.
This study was approved by the Clinical Research Ethics Committee of the Catholic Medical Center (approval number: XC20RIDI0192). All methods were performed in accordance with the Declaration of Helsinki, participants were informed about the study, the handling of personal data and how confidentiality would be maintained in the management of material and in publishing and presenting results. Researchers were permitted to conduct this study by accessing dataset newly assigned with a serial number whose personal information was removed after ethical approval. We included only patients over 19 years old and written informed consent was obtained from all patients prior to registry enrollment.
Data. We extracted the following data from the patient medical records: patient demographics, smoking history, stage of lung cancer, Eastern Cooperative Oncology Group performance status, laboratory data, history of chemotherapy and/or radiation, survival status, and the dates of disease progression and death. Blood samples drawn within one week prior to ICI treatment were used to compile a pretreatment GPS for each patient using the laboratory values. The patients were classified into three groups based on GPS values as follows: (I) GPS of 2, elevated CRP level (>1.0 mg/dl) and hypoalbuminemia (<3.5 g/dl); (II) GPS of 1, elevated CRP level or hypoalbuminemia; and (III) GPS of 0, neither elevated CRP level nor hypoalbuminemia (5, 10). The serum CRP levels were measured using immunoturbidimetric assays (CRPL3, Roche Diagnostics, Indianapolis, IN, USA).
Statistical analysis. The patient baseline demographics and clinical outcomes were compared according to the GPS. We used Pearson’s chi-squared test to compare the discrete variables and the Student’s t-test or analysis of variance to compare the continuous variables. Kaplan-Meier survival analysis was used to compare OS and irPFS according to GPS. Differences in survival and irPFS were determined by the log-rank test. The Mann-Whitney test was used to compare the median values. Hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs) were calculated for the predictors that were significant in multivariate Cox regression analysis. A two-sided p-value of <0.05 was considered statistically significant. All statistical analyses were performed using SPSS for Windows software (ver. 20.0; IBM Corp., Armonk, NY, USA) (11).
Results
Patient characteristics. Overall, 122 NSCLC patients treated with ICIs were enrolled, of whom 44 met the exclusion criteria (treatment with ICIs as first-line therapy, small cell lung cancer treated with ICIs, treatment with durvalumab, and patients lost to follow-up, or those who had no pre-treatment CRP/albumin levels). Thus, 78 patients were finally included in our analysis (Figure 1). The mean age of the included subjects was 67.1±9.17 years (range=38.0-84.0 years). There were 64 (82.1%) males. Of all the included patients, 19 (24.4%), 35 (44.9%), and 24 (30.8%) were classified into the GPS 0, GPS 1, and GPS 2 groups, respectively. We compared these three groups and explored the clinical factors predicting treatment outcomes including OS and irPFS. The total and median follow-up times were 12.7, 14.2, 15.0, and 6.4 person-months, respectively.
The baseline characteristics of these three groups are summarized in Table I. The mean age and smoking history were not different between the three groups. The proportion of poor performance status, histologic type, treatment line of ICIs, driving mutations, and PD-L1 expression were similar between the three groups. However, the GPS 0 group tended to receive atezolizumab and the GPS 2 group tended to receive pembrolizumab as ICIs (p=0.042).
Baseline characteristics of NSCLC patients treated with immunotherapy according to GPS values.
Prognostic analysis. The GPS 2 group displayed a shorter median irPFS than the GPS 0 or GPS 1 group [23.0 (95% CI=12.43-33.57) days vs. 89.0 (95% CI=57.6-120.4) days vs. 107.0 (95% CI=41.92-172.1) days respectively, p<0.001]. Also, the GPS 2 group showed significantly shorter median OS than the GPS 0 or GPS 1 group [412.9 (95% CI=278.5-547.27) days vs. 593.7 (95% CI=504.3-683.2) days in GPS 0 and 768.5 (95% CI=655.7-881.2) days in GPS 1, p<0.001] (Figure 2).
Clinical outcomes according to Glasgow prognostic score (GPS) in non-small cell lung cancer (NSCLC) patients treated with immunotherapy after platinum-based cytotoxic chemotherapy.
In the multivariate analysis of irPFS, the HR was 0.249 (95% CI=0.084-0.739, p=0.012) for PD-L1 expression ≥50% and 9.73 (95% CI=2.931-32.298, p<0.001) for GPS 2 (Table II). In multivariate analysis for OS, the HR was 0.189 for PD-L1 expression between 1 and 10%, 0.297 for PD-L1 expression ≥50%, 4.247 in the GPS 1 group, and 9.710 in the GPS 2 group (Table III).
Univariate and multivariate analysis for predicting progression-free survival in NSCLC patients treated with immunotherapy after platinum-based cytotoxic chemotherapy.
Univariate and multivariate analysis for predicting overall survival in NSCLC patients treated with immunotherapy after platinum-based cytotoxic chemotherapy.
Discussion
In the present study, we compared the clinical outcomes of NSCLC patients treated with ICIs as second-line or further-line therapy after treatment failure with platinum-based cytotoxic chemotherapy according to the GPS values and explored the prognostic role of GPS values for treatment outcomes. We found that the baseline characteristics were not different according to GPS. Patients with higher GPS values displayed shorter median irPFS and OS than those in the GPS 0 or GPS 1 groups. In multivariate analysis, lower PD-L1 expression and higher GPS was associated with shorter irPFS and OS.
Chronic inflammation is known to be associated with tumor development through the induction of oncogenic mutations, genomic instability, early tumor promotion, and enhanced angiogenesis (12). During inflammation, reactive oxygen and nitrogen species are created to combat pathogens, but also damage DNA, which in turn promote mutations that initiate cancer (13, 14). Inflammatory responses may act together with oncogenes by driving cells to undergo high levels of transcriptional activity that leads to genome instability (15). Tumorigenesis can be enhanced through increased genomic instability (16). Various cancers induce an inflammatory microenvironment (17). Nutrition is a critical component of immune responses (18). In lung cancer, systemic inflammation, malnutrition, and tumor immune microenvironments are associated with each other, and these are key determinants of tumor progression and treatment response (3). Elevated levels of circulating CRP could be a marker of the increased predisposition to malignancy due to chronic inflammation, a marker of occult cancer leading to inflammation, or both (17).
In the era of immunotherapy in cancer treatment, proper predictive factors of NSCLC patients for immunotherapy have not been developed. The expression of PD-L1 on tumor cells and the TMB have been used in qualifying patients to receive immunotherapy, but not all patients with these predictive factors benefit from immunotherapy (2). Patients with high TMB who received nivolumab and ipilimumab did not show significant survival benefit compared to those who received chemotherapy (19). The presence of immune cells in the anti-tumor immune response such as cluster of differentiation (CD) 8-positive cytotoxic T lymphocytes as well as CD4-positive memory and regulatory T lymphocytes has been postulated to be a prognostic marker of the disease course and predictors of activity or modulation of immune system function (20). For simpler prognostic markers, systemic inflammatory marker such as lung immune prognostic index are suggested as biomarker for ICIs in lung cancer. Poor combination of lung immune prognostic index and derived NLR or LDH value are associated with poorer outcomes in patients treated with ICIs (21).
Several studies have demonstrated that systemic inflammatory biomarkers in peripheral blood were predictive markers for treatment outcomes in different solid tumors including prostate, colorectal, and esophageal cancer, melanoma, and NSCLC (22-26). Although the exact biological basis for these findings has not been thoroughly elucidated, inflammatory cells such as neutrophils play a significant role in tumor development and progression via effects on tumor cells or other components of the tumor microenvironment, by secreting chemokines and cytokines such as transforming growth factor-β, interleukin-6 (IL-6) and matrix metalloproteinases (27, 28). CRP is a surrogate marker of IL-6, which is involved in the activation of immune cells, tumor migration and invasion and epithelial-to-mesenchymal transition (29, 30).
Cancer prognosis is associated with not only tumor staging but also patient-related factors such as nutritional and functional decline. CRP represents systemic inflammation, and albumin reflects both systemic inflammation and the amount of lean tissue (31, 32). GPS is a reliable independent prognostic factor in patients with various malignancies and also a marker for predicting prognosis, even in surgery, chemoradiation, and various subgroups incapable of surgery (5, 6). In 15 studies including >2,000 patients, GPS was associated with increased weight loss, poor performance status, increased comorbidity, increased proinflammatory and angiogenic cytokines, and complications from cancer treatments (5). In resectable lung cancer, the pre- and post-operative GS predicts adverse survival outcomes in patients with resected stage I NSCLC (33, 34).
Few previous studies have been conducted on the association between GPS and clinical outcomes in lung cancer patients treated with ICIs. Taichi et al. reported that modified GPS values checked before treatment were associated with shorter OS in NSCLC patients treated with atezolizumab (8). In another study, post-treatment GPS predicted anti-programmed cell death protein 1 (PD1) treatment (nivolumab or pembrolizumab) efficacy in NSCLC patients (7). Recently, Takamori et al. reported that pretreatment GPS were associated with shorter OS and PFS in NSCLC patients treated with ICIs as first line or higher (9). These studies included populations from single institutions and reported the results of groups treated with anti-PD1 or anti-PD-L1 antibody, or those treated with ICIs as first line therapy. Our study had a moderate sample size and analyzed the prognostic role of pre-treatment GPS values in all of the NSCLC patients treated with anti-PD1 or anti-PD-L1 antibodies after failure of platinum-based cytotoxic chemotherapy.
Limitations. This study had some limitations. Firstly, it was a retrospective study. Nonetheless, our study was based on a lung cancer cohort with a moderate sample size using medical records that were faithfully collected from the time of enrollment, and the data were rechecked by authorized data managers. Therefore, data including the baseline characteristics and clinical outcomes were of high quality and reliable. Also, we enrolled lung cancer patients from seven teaching hospitals in the Republic of Korea, so our data represent the Korean general population to some extent. Secondly, since our analyzed results were based on clinical parameters, we could not exactly elucidate the mechanisms of GPS on NSCLC patients treated with immunotherapy. However, hematologic biomarkers are promising predictors of the response to ICIs due to their convenience and accessibility in clinical practice.
Conclusion
Patients with higher GPS values before treatment showed shorter median irPFS and OS in NSCLC patients treated with ICIs as a second-line or further-line therapy. The pre-treatment serum GPS, along with PD-L1 expression, is a promising value to identify NSCLC patients who could benefit more from ICIs as a second-line or further-line therapy after treatment failure with platinum-based cytotoxic chemotherapy. Further large-scale studies are warranted to validate its clinical value.
Acknowledgements
The Authors would like to acknowledge and thank Bora Kwon, data manager for help with data collection and management of lung cancer registry.
Footnotes
This article is freely accessible online.
Author’s Contributions
HSK: conception, design, definition of intellectual content, data analysis, statistical analysis, manuscript preparation and manuscript review; CDY: conception, design, definition of intellectual content, data analysis and manuscript review; SHL: conception, design, definition of intellectual content, data analysis and manuscript review; JWK: conception, design, definition of intellectual content, data analysis, manuscript preparation and manuscript review; AYS: literature search, clinical studies, data acquisition and manuscript review; SKK: literature search, clinical studies, data acquisition, manuscript preparation and manuscript review; CKP: literature search, clinical studies, data acquisition and manuscript review; JSK: literature search, clinical studies, data acquisition and manuscript review; SJK: literature search, clinical studies, data acquisition, manuscript preparation and manuscript review.
Conflicts of Interest
All Authors have completed the ICMJE uniform disclosure form. The Authors have declared that no competing interests exist.
- Received July 29, 2021.
- Revision received September 23, 2021.
- Accepted September 27, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
