Abstract
Background/Aim: Due to some interconnectedness at the molecular level, this study assessed the possible influence of laboratory parameters associated with systemic inflammatory environment on programmed death-ligand 1 (PD-L1) expression in non-small cell lung carcinoma (NSCLC). Patients and Methods: We assessed effects of c-reactive protein (CRP), albumin, haemoglobin, neutrophil, and lymphocyte levels on PD-L1 expression in NSCLC. Patient data were obtained retrospectively from LUCAS, the Czech registry of patients with lung carcinomas. Correlations of two continuous parameters (PD-L1 expression and laboratory parameters) were analysed by correlation coefficient. Differences in continuous parameters between two or more groups were tested by Mann–Whitney or Kruskal–Wallis tests. Independence of two categorical parameters was tested by chi-square test. Results: We demonstrated no influence of the investigated laboratory parameters on PD-L1 expression in NSCLC, either in continuous or categorical division of variables. Conclusion: Inflammatory laboratory parameters at time of NSCLC diagnosis are unlikely to affect the determination of PD-L1 expression.
Programmed death-ligand 1 (PD-L1) expression, despite its limitations, plays a crucial role in the use of immunotherapy in first-line treatment of metastatic non-small cell lung cancer (NSCLC), as well as in the consolidation with durvalumab therapy after chemoradiotherapy in stage III NSCLC (1). Further work has shown, however, that prediction as to the effectiveness of immunotherapy in NSCLC may be affected by other parameters (2). Predictive markers may include laboratory parameters associated with inflammation, as evidenced by our previous publication on this topic (3). A question thus arises as to whether the level of PD-L1 expression may be related in some way to the values of these inflammatory parameters. According to the available literature, certain common genetic and molecular pathways can be traced to support this idea (4-8). NF-ĸB is associated with tumour inflammation, and its effect on PD-L1 expression also has been shown (4, 5). A similar relationship has been demonstrated between some other inflammatory cytokines (e.g., IL-10, IL-17, or TNF-α) and PD-L1 expression (5, 6). At the genetic level, some sharing of transcription pathways using the JAK/STAT, RAS/MAPK, and PTEN-PI3K/AKT pathways may then be involved (5, 8).
Commonly determined parameters reflecting systemic inflammation include the so-called acute phase reactants. During inflammation, there is an increase in c-reactive protein (CRP) and neutrophils (Neu), a decrease in lymphocytes (Lym) and albumin (Alb), and gradually a decrease in haemoglobin (Hb) (9, 10). Tests for these parameters are readily available in the laboratory, and therefore we selected them for our analysis as a reflection of systemic inflammation. Changes between neutrophils and lymphocytes during systemic inflammation are reflected in the neutrophils-to-lymphocytes ratio (NLR), which usually rises during systemic inflammation (10).
The aim of this work was therefore to determine whether the level of PD-L1 expression may be related to the values of laboratory parameters associated with the systemic inflammatory response of the organism.
Patients and Methods
Group of patients. In this retrospective analysis, clinical data were analysed of patients with cytologically or histologically confirmed locally advanced or advanced (stage III or IV) NSCLC that had been assigned to LUCAS, a national register and non-interventional post-registration database of epidemiological and clinical data of patients with lung cancer in the Czech Republic (from high volume centres that allowed participation in this project). All patients had given their informed consent to be included into this database and for use of these data for scientific purposes.
The patients’ data were obtained from seven pneumo-oncology departments in the Czech Republic between 2018 and 2021. Laboratory parameters investigated in the present study included Hb, Neu, Lym, Alb, and CRP. NLR was also easily calculated as Neu:Lym. These parameters were found in medical databases and entered into the LUCAS register. The levels of these parameters were searched to correspond as closely as possible to the date of biopsy collection, from which PD-L1 expression was determined with a maximum deviation of +/–1 month from this date. PD-L1 expression levels and baseline patient characteristics were obtained from the LUCAS registry.
PD-L1 expression estimation. Immunohistochemical staining was performed on automated immunostainers (different types in different laboratories) on formalin-fixed (10% neutral buffered formalin; fixation time 24 h), paraffin-embedded tissue sections (3 nm thick). All laboratories used PD-L1 clone 22C3 pharmDx (Dako, Santa Clara, CA, USA). The specimens were visualized using the ultraView Universal DAB Detection Kit (Ventana Medical Systems, Oro Valley, AZ, USA).
Statistical methods. Continuous parameters were described by valid N, mean, and standard deviation (SD), and median with 5th and 95th percentiles. Categorical data were described by absolute and relative frequencies. PD-L1 expression was analysed in two ways: continuously and categorically (0; 1-49; ≥50) according to a clinically valuable cut-off value. Laboratory parameters were analysed in two ways: continuously and categorically. Cut-offs of laboratory parameters were chosen according to the cut-offs in certificated laboratory values as follows: Hb – under or above the lower limit of normal values (with different limits for men and woman), Neu – under or above the upper limit of normal values, Leu – under or above the lower limit of normal values, Alb – under or above lower the limit of normal values, CRP – under the limit of normal value or over the limit. NLR cut-offs were determined in two ways: first as a median and second as the value 3, as in accordance with other literature and to enable the possibility for comparison (11).
Correlations of two continuous parameters (PD-L1 expression and laboratory parameters) were analysed by Spearman’s correlation coefficient. Differences of continuous parameters between two or more groups were tested by Mann–Whitney or Kruskal–Wallis test. Independence of two categorical parameters was tested by Pearson’s chi-square test.
Statistically significant results (relationships) were determined where p<0.05. Analyses were performed using IBM SPSS Statistics 24 (IBM, Armonk, NY, USA).
Results
Patient characteristics. In total, 1,148 patients were evaluated. The analysed dataset consisted of 61% men. Mean age was 67.7 years (median 68.8 years). At time of diagnosis, 31% of the patients were younger than 65 years of age. Eighty-nine patients (7.8%) had a tumour type other than adenocarcinoma or squamous: 59 NSCLC not otherwise specified (NOS); 22 adenosquamous, and 8 NSCLC neuroendocrine differentiation. Baseline patient characteristics are described in Table I.
Baseline patient characteristics.
Mean PD-L1 expression was 28.7 and median was 10.0. Among all patients, 29% had PD-L1 expression <1% (PD-L1 negative), 41% had PD-L1 expression between 1% and 49%, and 30% of patients had PD-L1 expression >49%.
The basic characteristics of laboratory parameter values are described in Table II. Because some data were missing, information as to the valid number of patients for each parameter is included in the Table.
Characteristics of laboratory parameters.
Relationship between basic patient characteristics and PD-L1 expression. PD-L1 expression was tested as a categorical parameter. There were no statistically significant relationships between basic patient characteristics and PD-L1 expression. The results are summarized in Table III.
Relationship between basic patient characteristics and PD-L1 expression.
Relationships between laboratory parameters and PD-L1 expression. This analysis was performed in four variants: First, measuring correlations (Spearman’s correlation coefficient) between laboratory parameters (as continuous parameters) and PD-L1 expression (as a continuous parameter). Second, measuring relationships between categories of laboratory parameters and PD-L1 expression as a continuous parameter (by Mann–Whitney test). Third, measuring relationships between laboratory parameters (as continuous parameters) and PD-L1 expression as a categorical parameter (by Kruskal–Wallis test). Fourth, measuring relationships between categories of laboratory parameters and PD-L1 expression as a categorical parameter (by Pearson’s chi-square test).
No statistically significant relationships were determined between the basic parameters and PD-L1 expression in these four analyses. The results are summarized in Table IV, Table V, Table VI, and Table VII.
Relationship between laboratory parameters (continuous parameters) and PD-L1 expression (continuous parameter).
Relationship between laboratory parameters (categorical parameters) and PD-L1 expression (continuous parameter).
Relationship between laboratory parameters (continuous parameters) and PD-L1 expression (categorical parameter).
Relationship between laboratory parameters (categorical parameters) and PD-L1 expression (categorical parameter).
Discussion
Although inflammatory parameters may be predictors of immunotherapy’s effectiveness, in our study we did not demonstrate a relationship between the levels of laboratory parameters associated with inflammation and levels of PD-L1 expression.
Some preclinical and clinical studies in other types of tumours have indicated a possible link between PD-L1 expression and the inflammatory environment, albeit with some contradictory results (8, 11, 12). Yassin et al. reported upregulation of PD-L1 in a mouse model of colorectal cancer after induction of colitis (12). Tang et al. then demonstrated in their study of 83 patients with oesophageal cancer treated with chemoradiotherapy a correlation between NLR and PD-L1 expression (8). Patients with PD-L1 positivity showed lower NLR than did patients that were PD-L1 negative. In contrast, Sanghamanon et al. demonstrated in a smaller study of 46 patients with cholangiocarcinoma the opposite relationship, as PD-L1 positivity was correlated with higher NLR (11).
Studies relating to NSCLC contain similarly contradictory results. Ghanim et al. did not demonstrate an effect of CRP on PD-L1 expression in 61 malignant pleural effusions, most of whom were patients with NSCLC (13). Nardone et al. also demonstrated absence of any effect of inflammatory parameters (procalcitonin associated with erythrocyte sedimentation) on PD-L1 expression in their group of metastatic NSCLC patients (14). By contrast, Akamine et al. described higher PD-L1 positivity in patients with higher CRP among 508 patients with resected NSCLC (15). Unlike our study, these were patients with lower-stage tumours (a category not studied in our cohort). However, the study by Akamine et al. did not demonstrate an effect of disease stage on PD-L1 expression in a multivariate analysis (15). Also, Giatromanolaki et al. in their study on operable carcinomas, similarly to our study on stage III/IV tumours, confirmed absence of a stage effect on PD-L1 expression (16). Differences in findings between our study and that of Akamine et al. could relate to their using a different cut-off for the categorical evaluation of CRP and a different antibody in the PD-L1 determination [clone SP142 in the Akamine et al. study, which is not completely comparable with other antibodies (17) we had used]. An influence of geographical differences also cannot be ruled out. In agreement with our work, an absence of the effect of both CRP and albumin on PD-L1 expression is described by Bilginet et al. in their cohort of 217 patients also with locally advanced/metastatic NSCLC and in which case the same antibody (clone 22C3) was used to assess PD-L1 expression (18). In contrast to our results, Bilginet et al. described a significant relationship between PD-L1 expression (≥50%) and lower NLR (18). In contrast to our retrospective work, patients with known active infections and patients receiving medication that could affect leukocyte counts were excluded in the study by Bilniget et al. Given the nature of the LUCAS register, this was unfortunately not possible in our study. This, too, could have affected our data in some way. A different cut-off for NLR could also have played a role. Consistent with our results is a study by Tashima et al., who described no significant effect of NLR on PD-L1 expression in 83 patients with resected stage I NSCLC (19).
Overall, our study is the most robust work known for NSCLC on this topic, thus we aimed to address the possible general principle of our results. Interesting in this regard is a study by Mair et al. that examined the effect of inflammation on PD-L1 expression in brain tumours (20). The authors demonstrated an effect of inflammation on soluble PD-L1 in blood plasma, but this change did not affect the PD-L1 expression of tumour cells. Kasajima et al., meanwhile, considering the inflammatory microenvironment of neuroendocrine tumours, described a change in PD-L1 expression only on the immune cells within the tumour but not on the tumour cells themselves (21). Rangamuwa et al. described an increase in inflammatory parameters and an increase in PD-L1 expression in some patients after bronchoscopic thermal vapour ablation of NSCLC (22), but the increase in inflammatory parameters in the blood did not correlate with the increase in PD-L1 expression on tumour cells. Overall, it is possible that inflammatory parameters have an effect only on the tumour microenvironment rather than on the regular change in PD-L1 expression on the tumour cells themselves.
One of the limitations of our study is the retrospective data collection from the LUCAS registry, which is also limited to large clinical centres. This may have led to some bias when enrolling patients. On the other hand, we believe that, due to the robustness of our group, the risk of non-representativeness of the data within the Czech Republic is minimal. There also was no external validation of the measured laboratory parameters and PD-L1 expression, but all these values were determined by accredited methods within accredited workplaces with routine internal and external quality control. Use of different antibodies to detect PD-L1 also should play no role in our case, because all laboratories in our project used the same clone of PD-L1 antibodies. It is also necessary to mention a possible influence on the determination of PD-L1 expression by the date of laboratory markers, which, due to the retrospective nature of the study, did not reflect exactly the day of biopsy collection but only came as near to it as possible (within specified time limits). Lastly, in our group, due to the retrospective data, it was not possible to distinguish between an increase in inflammatory parameters due to active infection and to chronic tumour inflammation. However, the proportion of active inflammation and tumour inflammation is often poorly distinguished in clinical practice, as well, and thus our work reflects clinically real data of everyday practice.
A strength of our study lies in its large number of evaluated patients, which endows the statistics with the required robustness. We also evaluated inflammatory parameters and PD-L1 expression in addition to the usual categorically determined groups using cut-offs as continuous parameters. This substantially reduces the risk of improperly selected cut-offs and allows for easier comparison with follow-up studies.
In conclusion, it can be stated that, as indicated by the results of our work, a possible effort to influence the laboratory parameters associated with inflammation (e.g., using antibiotic treatment or nutritional support) is unlikely to lead to a change in PD-L1 expression. The reasons why inflammatory parameters may affect the effectiveness of immunotherapy will therefore need to be sought in other contexts.
Acknowledgements
This study was supported by a grant of the Ministry of Health of the Czech Republic - Conceptual Development of Research Organization (Faculty Hospital in Pilsen - FNPl, 00669806, Pilsen, Czech Republic).
Footnotes
Authors’ Contributions
Conceptualization: MS; Methodology: MS, JB; Investigation: MS, MD, OF, MM, MH, OV, PZ, MS, JB, DK. Statistical analyses: MS, MS; Writing – Original Draft Preparation: MS, JB; Writing – Review & Editing: MS, JB.
Conflicts of Interest
The Authors declare that they have no conflicts of interest in relation to this study.
- Received January 20, 2022.
- Revision received February 6, 2022.
- Accepted February 7, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.