Association between pretreatment neutrophil‐to‐lymphocyte ratio and immune‐related adverse events due to immune checkpoint inhibitors in patients with non‐small cell lung cancer

Abstract Background Immune checkpoint inhibitors (ICIs) have revolutionized the treatment of advanced or recurrent non‐small cell lung cancer (NSCLC). They cause immune‐related adverse events (irAEs), but the underlying mechanisms and predictors remain to be fully elucidated. In this retrospective study, we investigated the association between pretreatment neutrophil‐to‐lymphocyte ratio (NLR) and the occurrence of irAEs. Methods The study involved 115 patients with NSCLC who started ICI‐only treatment in our hospital between January 2016 and April 2020. Results Forty‐five patients (39.1%) had irAEs, and pretreatment NLR was significantly lower in the irAEs group than in the non‐irAEs group (2.8 vs. 4.1; p = 0.036). The cutoff value of the NLR was 2.86 (area under curve, 0.62; sensitivity, 0.56; specificity, 0.71), and the incidence rate of irAEs was significantly higher in the NLR < 2.86 group than in the NLR ≥2.86 group (p = 0.004; odds ratio [OR]: 3.12; 95% confidence interval [CI]: 1.43–6.84). The multivariate analysis showed that the NLR was significantly associated with the occurrence of irAEs (p = 0.016; OR: 2.69; 95% CI: 1.21–6.01). Conclusions Low pretreatment NLR may be a predictive factor for the occurrence of irAEs. By focusing on the potential risk of irAEs in patients with a low pretreatment NLR, irAEs can be appropriately managed from an early period.


INTRODUCTION
Immune checkpoint inhibitors (ICIs) have paved a new era for the treatment of advanced or recurrent non-small cell lung cancer (NSCLC). Nivolumab, pembrolizumab, atezolizumab, durvalumab, and ipilimumab have been approved for the treatment of NSCLC, and monotherapy and combination therapy have been established as a standard course of care. 1 ICIs exert antitumor effects by binding to inhibitory receptors expressed on T cells, or their ligands, and blocking the inhibitory signals of the immune system. 2 An increase in immune functions induced by ICIs may cause immunerelated adverse events (irAEs) due to the disruption of immune function homeostasis. 3 irAEs pose a risk of poor prognosis in severe cases, and ICI-related mortality rate is in the range of 0.36%-1.23%. 4 Appropriate treatment from an early period is required when patients experience irAEs; while there have been some reports on the predictive factors of irAEs, [5][6][7][8][9] there are no established factors.
The neutrophil-to-lymphocyte ratio (NLR) is an index that reflects systemic inflammation. It can be easily calculated from the results of inexpensive routine blood tests; the prognostic value of NLR has been proven in cardiovascular diseases and infections. 10,11 Reportedly, the NLR can serve as a prognostic factor for solid tumors [12][13][14] and a predictor of clinical benefits of ICIs. [15][16][17] Furthermore, its relationship with irAEs has been investigated in recent years 8,9 ; however, there is still no unified view. In this study, we retrospectively investigated the association between NLR and the occurrence of irAEs in patients with advanced/recurrent NSCLC treated with ICIs.

Patients
This retrospective study involved 115 patients with NSCLC who started treatment with ICIs between January 2016 and April 2020 in the National Hospital Organization Kyushu Medical Center. We excluded patients who had been treated with other ICIs before starting ICIs, patients who received ICIs in combination with chemotherapy, and patients whose peripheral blood neutrophils and lymphocytes were not measured within one week before the start of ICIs. We retrospectively surveyed the medical records of patients and investigated the irAEs. The NLR was calculated as follows: neutrophil count (/μl) / lymphocyte count (/μl). The Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 was used to assess irAEs. The observation period was from January 2016 to September 2020.

Ethical approval
This study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki and the ethical guidelines for Medical and Health Research Involving Human Subjects. This study was approved by the investigational review board of the ethics committee of Kyushu Medical Center (approval no. 20C200).

Dosage of ICIs
Nivolumab was administered every two weeks at a dose of 3 mg/kg until August 31, 2018, and 240 mg/bodyweight after November 1, 2018. During the two months between September 2018 and October 2018, a dose of either 3 mg/kg or 240 mg/bodyweight was chosen for each patient. Pembrolizumab was administered at a dose of 200 mg/bodyweight every three weeks, and atezolizumab was administered at 1200 mg/bodyweight every three weeks.

Statistical analysis
In this study, patients were divided into two groups, namely, the irAEs and non-irAEs groups. We used Fisher's exact test and the chi-square test to compare the qualitative data between the groups, and Mann-Whitney U test to compare the quantitative data. Patients were dichotomized according to the cutoff values of NLR < 2.86 versus ≥2.86. The NLR cutoff value before the start of treatment was calculated using receiver operating characteristic (ROC) curve analysis. Multivariable logistic regression analysis was used to identify the predictive factors of irAEs.
Overall survival (OS) was defined as the period from the initiation of ICIs to patient death from any cause or to the last date of confirmation of survival based on the medical records. The OS was compared between the groups using the log-rank test, and Holm method of the post-hoc test was used to compare the OS among the four groups. Results with p < 0.05 were considered statistically significant. All statistical analyses were performed using EZR software version 1.53. 18

Patient characteristics
This study involved 115 patients with advanced NSCLC, except for one patient with stage I who was inoperable because of poor lung function and was treated with an ICI after the failure of stereotactic irradiation. Forty-five patients (39.1%) experienced irAEs (Table 1), with four patients experiencing multiple irAEs. There were no significant differences between the groups in terms of age, sex, Eastern Cooperative Oncology Group performance status (ECOG-PS), smoking status, histopathological diagnosis, driver gene alteration, tumor proportion score, tumor stage, type of ICI, previous chemotherapy, previous thoracic radiotherapy, taking systemic corticosteroids, white blood cell count, neutrophil count, and lymphocyte count. The median (range) pretreatment NLR was 2.8 (0.9-12.0) and 4.1 (0.8-10.7) in the irAEs and non-irAEs groups, respectively, and the difference was significant (p = 0.036). Table 2 shows the occurrence of irAEs and their severity. The most frequent irAEs was thyroid-related events, followed by skin-related events and interstitial lung diseases. Skin-related events were rash and pruritus; none of the patients had vitiligo. The severity of thyroid-related events was grade ≤ 2, and the treatment was interrupted in four patients, but treatment discontinuation was not required for any patient. The severity of interstitial lung disease was grade ≥ 3 in 5 patients, and treatment interruption and discontinuation were required for two and six patients, respectively. Other irAEs that required the discontinuation of treatment were colitis, encephalitis, hepatopathy, cardiacrelated events, thrombocytopenia, and hypoadrenocorticism. Colitis was considered to be immune-related enteritis in all patients; all patients were treated with corticosteroids, except one patient, who was treated with infliximab. With respect to encephalitis, two patients presented decreased consciousness level of grade 3; with regard to cardiac-related events, pericardial tamponade was observed in one patient and acute myocardial infarction in one patient. Two patients presented with immunogenic thrombocytopenia and one patient developed symptoms two months after ICI discontinuation.  Non-irAEs (n = 70) n (%) irAEs (n = 45) n (%) p-value  Table 3). There was no significant difference between the grade of irAEs and level of NLR (grade 1, 2 vs. > 3; p = 0.577, date not shown).

Discussion
In this study, pretreatment NLR was significantly lower in the irAEs group than in the non-irAEs group, with a cutoff value of 2.86. Additionally, the occurrence rate of irAEs was significantly higher in the NLR < 2.86 group than in the NLR ≥2.86 group. This result is in agreement with the results of Pavan et al. 8 and Eun et al. 9 who showed that a pretreatment NLR of <3 is a risk factor for the occurrence of irAEs. In these studies, the NLR cutoff value was set to 3. This value was not calculated as a risk factor for the development of irAEs, but was considered from previous studies reporting the association between NLR and prognosis of solid tumors and ICI treatment outcomes. [19][20][21][22][23] We focused on irAEs and calculated the cutoff value of NLR, and the value was close to 3. Pavan et al. 8 reported the findings of a study in Europe and Eun et al. 9 investigated multiple cancer  types. It has been reported that the types of irAEs and their occurrence rates differ depending on race and cancer type. 24,25 Our findings demonstrated the association between low pretreatment NLR and high occurrence rates of irAEs in Japanese patients with NSCLC. As far as we know, this is the first study to have demonstrated the significance of NLR in predicting irAEs in Japanese patients with NSCLC. On the basis of these results, we speculate that the NLR may be an important factor for predicting irAEs, regardless of race and cancer type. We showed that OS of patients with a pretreatment NLR of <2.86 was significantly longer than that of patients with a pretreatment NLR of ≥2.86. This indicates that patients with a pretreatment NRL of <2.86 are at a high risk of developing irAEs, but good clinical outcomes can be expected. In addition, the OS of patients with NLR < 2.86/irAEs was significantly prolonged compared with that of patients with NLR ≥2.86/non-irAEs. Although no significant difference was observed, the survival curve suggested that the OS of patients in the NLR < 2.86/irAEs group may be longer than that of patients in the NLR < 2.86/non-irAEs group. This difference may be because a lower NLR has been associated with a better outcome of ICI treatment, 16,17 and the development of irAEs has been found to be related to survival benefits in patients with NSCLC. [26][27][28] Lymphocytes, mainly T cells, are involved in the antitumor immune response and suppress the growth of tumor cells. 29 In contrast, neutrophils promote tumor development by acting on tumor cells and the tumor microenvironment, 30 and peripheral neutrophil counts have been reported to correlate with intertumoral neutrophil counts. 31 Furthermore, the increase in peripheral neutrophils adversely affects the cytotoxic activity of lymphocytes and may attenuate the antitumor response. 30 The main mechanism of the occurrence of irAEs is considered to be the damage to self-tissues due to the activation of lymphocytes that react with self-antigens following the administration of ICIs. 32 Low NLRs reflect low peripheral neutrophil counts and high lymphocyte counts, indicating the maintenance of antitumor response of the immune system as well as the increased risk of the presence of lymphocytes that inappropriately react to self-antigens. Therefore, we consider that a low pretreatment NLR leads to the improvement of positive outcomes of ICI treatment and the increased risk of irAEs.
Biomarkers for irAEs have been extensively investigated. To date, such biomarkers including HLA genotypes, soluble serum proteins, and gut microbiota have been reported to predict irAEs. [33][34][35] However, evaluation of these biomarkers is associated with clinical challenges, such as accessibility, complexity, and high cost. Compared with these biomarkers, NLR is easy to measure and inexpensive. The NLR could be a clinically cost-effective and useful predictive factor for irAEs.
There were some limitations to the study, including its retrospective nature, single-institutional design, and limited sample size. Large-scale, multi-institutional, retrospective, and prospective studies on the relationship between the NLR and the occurrence of irAEs are warranted. Additionally, we cannot deny that the association between a low NLR and higher incidence of irAEs may be caused by the longer OS in a low NLR group; longer OS usually means exposure to more doses of ICIs. The relationship between the duration of exposure or the cumulative dose of ICIs and irAEs has not been clarified, and this should be investigated in the future.
In conclusion, a low NLR may be a predictive factor for the onset of irAEs. The NLR can be measured in an inexpensive and easy manner through routine clinical practice. By focusing on the potential risk of irAEs in patients with a low pretreatment NLR, irAE management can be appropriately managed from an early period. We consider that this could maximize the benefits of ICI treatment.