Background: Immune checkpoint inhibitors produce prolonged responses in select non-small cell lung cancer (NSCLC) patients, however the identification of patients most likely to benefit is difficult. Pretreatment derived neutrophil-to-lymphocyte ratio (dNLR) is an easily calculated marker available in routine clinical care that has shown prognostic value in many cancer treatment settings, but its association with survival in NSCLC patients treated with immune-checkpoint inhibitors is less understood. Patients and Methods: We retrospectively reviewed 72 NSCLC patients receiving either nivolumab or pembrolizumab between 3/1/15 and 3/1/17 with a median follow - up time of 5.1 months. Patients were compared using Cox proportional hazards models to detect an association between pretreatment dNLR < 3 vs ≥3 on overall survival (OS), progression-free survival (PFS) and overall response rate. Results: Median age was 65 (range: 41 - 86), 65% were male, 40% received ≥ 2 prior systemic therapies and 14% had an Eastern Cooperative Oncology Group Performance Status (ECOG PS) ≥ 2. Pretreatment dNLR ≥ 3 was independently associated with shortened OS (median 3.6 vs 8.5 months; HR: 5.4; 95% CI: 2.0 - 14.6; p = 0.001) and PFS (median 2.1 vs 3.4; HR: 2.3; 95% CI: 1.1 - 4.8; p = 0.027). Conclusion: Pretreatment dNLR ≥ 3 was independently associated with inferior survival in NSCLC treated with immune checkpoint inhibitors in routine practice. Prospective verification of this marker is warranted as it could serve as an inexpensive and widely-available marker for identifying NSCLC patients most likely to benefit from PD-1 inhibitors.
Immune checkpoint inhibitor therapies have been approved for multiple malignancies and have heralded a new era for cancer therapies. New data is rapidly materializing and clinical practices and protocols are changing quickly. Immune checkpoint inhibitors can prevent cancer cells from inhibiting the immune system’s natural antineoplastic response by blocking inhibitory signaling pathways involving interactions between programmed death 1 (PD-1) receptors and its ligands (PD-L1 and PD-L2), thus enhancing the activity of T cells within the tumor microenvironment [
Two programmed death 1 (PD-1) inhibitors, nivolumab and pembrolizumab, have particularly revolutionized the treatment of previously treated NSCLC. Nivolumab extended overall survival (OS) in phase III randomized controlled trials compared to second line docetaxel in both squamous-cell and non-squamous cell NSCLS [
Most efforts to identify such a marker have focused on PD-L1 expression. This biomarker offers some prognostic information for patients beginning treatment with nivolumab or pembrolizumab, but cannot predict non-benefiting patients [
As the predictive role of PD-L1 expression is nebulous, additional markers can help create a more robust system for identifying patients more likely to have a favorable clinical outcome. The pretreatment neutrophil-to-lymphocyte ratio (NLR) has been correlated with overall survival (OS) and progression free survival (PFS) in a wide range of malignancies, including melanoma patients treated with the immune-checkpoint inhibitor ipilimumab [
We conducted a retrospective analysis of all NSCLC patients receiving either nivolumab or pembrolizumab within our health system between 3/1/15 and 3/1/17. Patients were excluded if they received less than two cycles of therapy. We included both PD-1 inhibitors to maximize sample size, although analysis on the nivolumab-only cohort was also done. Data collected from the electronic medical record include Eastern Cooperative Oncology Group Performance Status (ECOG PS), tumor histology, driver mutation status including EGFR, ALK and ROS-1, PD-L1 expression, sites of metastases at initiation of PD-1 inhibitor, previous treatments and baseline complete blood count. Data on immune-related adverse events (irAEs), such as pneumonitis, dermatitis, hepatitis, arthritis, colitis and others, were graded using the National Cancer Institute’s Common Terminology Criteria for Adverse Events version 4.0 (CTCAE v4.0) by a single investigator (J.K.) (
Pretreatment dNLR was calculated via the formula: dNLR = ANC/(WBC − ANC) [
Treatment response was evaluated using either the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 or the PET Response in Solid Tumors v 1.0 [
The Ochsner Medical Center institutional review board approved the project and informed consent was waived in this retrospective study.
Demographic characteristics, safety and efficacy were described with descriptive statistics as either relative frequencies (percentages) or medians with ranges for quantitative variables. dNLR cut off values at intervals of 0.5 were compared
using Receiver Operator Curves (ROC) with analysis of the Area Under the Curve (AUC) using OS as an end-point. Kaplan-Meier analysis was used to estimate OS and PFS with differences in curves analyzed with log-rank test. Cox proportional hazard regression models were used to estimate hazard ratios (HR) and P-values for the relationship between dNLR and OS and PFS. The final multivariate model included age, sex, race, ECOG PS, smoking history, squamous histology, number of prior medical therapies and site of metastatic disease. Exploratory analysis for the contribution of ANC and ALC was conducted both by dividing the cohort with previously established cutoffs (750 for ANC; 1000 for ANC) and by dividing the cohort into quartiles based on either their ANC or ALC [
For all analysis, p < 0.05 was considered significant. The statistical analysis was performed in Stata version 14.2 (Stata Corp., College Station, Texas).
Between March 2015 and April 2017, 72 NSCLC patients were treated with at least 2 cycles of a PD-1 inhibitor at our institution. Among them, 54 (75%) received nivolumab, and 18 (25%) received pembrolizumab. Median follow up from the initiation of therapy was 5.1 months, 34 (47.2%) of patients died as of April 2017. Patient characteristics are listed in
Among the 57 (79.1%) patients with non-squamous histology, 47 (82%) had molecular testing for EGFR mutation, 42 (72%) for ALK translocation and 32 (56%) for ROS-1 translocation. Of the 15 (21%) patients with squamous cell histology, 3 (20%) had EGFR mutation testing and 2 (13%) had ALK translocation testing.
The ROC curve established the optimal dNLR cutoff by comparing the sensitivity and specificity of different thresholds. The optimal cutoff was determined to be 3; AUC was 0.722 (p < 0.006). Median survival in the entire cohort was 7.9 months. Patients with dNLR < 3 had superior OS in the unadjusted analysis (median 8.5 vs 3.6 months; HR 2.35, 95% CI 1.15 - 4.82; p = 0.002). In multivariate analysis, elevated ECOG PS (≥2), the presence of liver metastases, squamous cell histology and dNLR ≥ 3 were significantly associated with shortened OS, baseline dNLR < 3 was associated with prolonged OS (HR: 5.4; 95% CI: 2.0 - 14.6; p = 0.001) (
Median PFS was 2.7 months, 54 (75%) subjects had progressed at the time of analysis. dNLR < 3 was associated with superior PFS on crude analysis (3.4 vs 2.0 months; HR 1.75, 95% CI 0.96 - 3.17; p = 0.005). Baseline dNLR and treatment were the only variables associated with PFS in multivariate analysis. In the final multivariate model for PFS, elevated dNLR remained independently associated with prolonged PFS (HR: 2.3; 95% CI: 1.1 - 4.8; p = 0.027). In a multivariate analysis of the nivolumab-only cohort, dNLR ≥ 3 was also independently associated with PFS (HR: 2.8; 95% CI: 1.2 - 46.4; p = 0.018).
Treatment responses were available for 63 patients. The ORR was 20.6%, with one patient experiencing complete response (CR), 12 partial response (PR), 28 stable disease (SD) and 22 progressive disease (PD) as best response. dNLR was not associated with response to treatment in multivariate analysis (OR: 5.45; 95% CI: 0.34 - 88.65; p = 0.233), nor were the other variables included in the
Characteristic | N (%) |
---|---|
Age | |
Median | 65 |
Range | 41 - 86 |
<75 | 60 (83.3) |
Male | 47 (65.3) |
Race | |
White | 37 (51.4) |
Black | 30 (41.7) |
Other | 3 (4.2) |
ECOG PSa | |
0 | 26 (36.1) |
1 | 36 (50) |
2 | 10 (13.9) |
3 | 0 (0) |
Smoking historyb | |
Light/never | 18 (25) |
Heavy | 51 (70.8) |
Histology | |
Non-squamous | 57 (79.2) |
Squamous | 15 (20.8) |
Targetable driver mutationc | |
EGFR | 9 (12.5) |
ALK | 0 |
ROS-1 | 0 |
Site of metastases at initiation of therapy | |
Bone | 20 (27.8) |
Liver | 13 (18.1) |
Brain | 11 (15.3) |
Line of therapy | |
2 | 36 (50) |
3 | 20 (27.8) |
4+ | 9 (12.5) |
Therapy | |
Nivolumab | 54 (75) |
Pembrolizumab | 18 (25) |
dNLR | |
<3 | 55 (76.4) |
≥3 | 17 (23.6) |
aECOG PS: Eastern Cooperative Oncology Group Performance Status; bHeavy (≥10 pack-years), light/never (<10 pack-years); cEGFR: Epidermal Growth Factor Receptor; ALK: Anaplastic Lymphoma Kinase; ROS1: ROS1 oncogene.
Overall Survival | Progression-Free Survival | |||||
---|---|---|---|---|---|---|
Parameter | HR | P | 95% CI | HR | P | 95% CI |
dNLR ≥ 3 | 5.41 | 0.001 | 2.01 - 14.56 | 2.29 | 0.027 | 1.10 - 4.79 |
Age | 0.96 | 0.113 | 0.92 - 1.01 | 0.99 | 0.496 | 0.96 - 1.02 |
Sex | 0.94 | 0.879 | 0.40 - 2.18 | 0.92 | 0.809 | 0.46 - 1.84 |
Race | 0.56 | 0.222 | 0.22 - 1.42 | 0.89 | 0.729 | 0.47 - 1.68 |
Smoking | 0.38 | 0.002 | 0.20 - 0.70 | 0.73 | 0.195 | 0.46 - 1.17 |
Treatment line | 1.41 | 0.115 | 0.92 - 2.16 | 1.45 | 0.024 | 1.05 - 1.99 |
ECOG PS | 1.79 | 0.031 | 1.06 - 3.05 | 1.50 | 0.062 | 0.98 - 2.30 |
Liver metastases | 4.46 | 0.005 | 1.58 - 12.61 | 1.76 | 0.162 | 0.80 - 3.90 |
Brain metastases | 0.57 | 0.378 | 0.17 - 1.97 | 0.89 | 0.809 | 0.36 - 2.21 |
Bone metastases | 2.10 | 0.088 | 0.90 - 4.93 | 1.23 | 0.55 | 0.62 - 2.46 |
Squamous histology | 5.58 | 0.001 | 2.06 - 15.14 | 1.74 | 0.106 | 0.89 - 3.40 |
final logistic regression model (age, sex, race, smoking status, ECOG PS, treatment line, liver, brain, bone metastases or histology).
To evaluate the contribution of ANC and ALC to the survival benefits of low dNLR, we evaluated the effect of each variable on survival by dividing the cohort into quartiles. Of note, only 41 patients had ALC data available. ANC quartiles were 1.50 - 3.30 (n = 18), 3.30 - 4.72 (n = 18), 4.76 - 6.60 (n = 18) and 6.65 - 16.42 (n = 18); ALC quartiles were 0.45 - 0.89 (n = 11), .91 - 1.27 (n = 10), 1.31 - 1.67 (n = 10) and 1.67 - 2.93 (n = 10). As shown in
The frequency and grade of the irAEs observed are summarized in
Immune checkpoint inhibitors have radically altered the therapeutic compendia for metastatic NSCLC. Indeed, nivolumab and pembrolizumab have been shown to prolong OS with favorable toxicity profiles [
Adverse event | N (%) | |
---|---|---|
Any Grade | Grade 3/4 | |
Total events | 19 | 4 |
Number of patients experiencing event | 16 (22.2) | 4 (5.6) |
Dermatitis | 5 (6.9) | |
Hypothyroidism | 4 (5.5) | |
Pneumonitis | 4 (5.5) | 2 (2.8) |
Mucositis | 2.(2.8) | |
Transaminitis | 1 (1.4) | |
Hypophysitis | 1 (1.4) | 1 (1.4) |
Myalgia | 1 (1.4) | |
Colitis | 1 (1.4) | 1 (1.4) |
therapies. In this 72 patient retrospective analysis of NSCLC patients treated with immunotherapy, pretreatment dNLR < 3 was associated with prolonged survival and PFS. Due to the widespread availability of dNLR in routine clinical practice, we suggest that dNLR should continue to be validated as a potential marker for identifying patients most likely to respond to immune checkpoint inhibitors.
Neutrophilia with relative lymphocytopenia has been shown to portend a worse prognosis in a multitude of solid tumors [
Recently, Bagley et al found NLR > 5 to be associated with inferior OS and PFS in nivolumab-treated NSCLC patients [
Our findings did differ, however, in the relative contributions of ANC and ALC. We found no difference in survival in groups separated by ANC quartiles, while there was marked separation in survival curves when patients were separated by both ALC quartiles or divided by a previously established ALC cutoff [
Our study population had inferior OS and PFS to that of the NSCLC registration trials [
This study has multiple limitations. We included two drugs, albeit with identical mechanisms, as it was felt that maximizing sample size yielded more powerful results, particularly in the exploratory analysis. To address this, analysis on the nivolumab-only cohort was conducted in parallel for our central conclusions. Additionally, we conducted formal RECIST measurements for our treatment response analysis, which would not account for pseudoprogression. Due to the predominant use of traditional RECIST criteria over immune-related response criteria in NSCLC PD-1 inhibitor research, and low incidence of pseudoprogression described in real-world populations, it was decided that RECIST was an appropriate, formalized means to track progression and response in this study [
Pretreatment dNLR < 3 was independently associated with prolonged OS and PFS in patients receiving a PD-L1 inhibitor for NSCLC. Given the possible ability of NLR to identify patients most likely to benefit from these therapies, we encourage prospective, adequately powered studies to validate our findings.
・ Not all NSCLC patients derive benefit from immune checkpoint inhibitors, and so a pretreatment clinical marker could be useful for stratifying patients most likely to benefit.
・ dNLR is an inexpensive and widely-available marker that is easily calculated from a complete blood count.
・ To date, the association of NLR and survival of patients receiving an immune checkpoint inhibitor has largely only been evaluated in melanoma patients.
・ Our study consisted of 72 NSCLC patients, 75% received nivolumab and 25% received pembrolizumab. Patients were dichotomized according to pretreatment dNLR < 3 vs ≥3.
・ dNLR ≥ 3 was independently associated with shortened OS (median 3.6 vs 8.5 months) and PFS (median 2.1 vs 3.4) in cox proportional hazards models.
・ dNLR should be prospectively verified as a potential marker to identify NSCLC patients most likely to benefit from PD-1 inhibitors.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Kucharczyk, J., Sullivan, C., Lu, J., Kolomensky, A., Peters, E. and Matrana, M.R. (2018) Prognostic and Predictive Value of Pretreatment Derived Neutrophil-to-Lymphocyte Ratio in Non-Small-Cell Lung Cancer Patients Treated with an Immune Checkpoint Inhibitor. Journal of Cancer Therapy, 9, 23-34. https://doi.org/10.4236/jct.2018.91004