Dynamics of blood neutrophil-related indices during nivolumab treatment may be associated with response to salvage chemotherapy for non-small cell lung cancer: A hypothesis-generating study.

Several recent studies have shown that salvage chemotherapy following PD-1 blockade produces high antitumor activity in some patients with non-small lung cancer (NSCLC). However, the underlying synergistic mechanisms remain uncertain. The blood neutrophil-to-lymphocyte ratio (NLR) and absolute neutrophil count (ANC) can reflect the number of circulating myeloid-derived suppressor cells and tumor-associated neutrophils. The immunosuppressive status of the tumor microenvironment could be monitored by the time-series patterns of NLR and ANC. The dynamics of NLR and ANC during nivolumab treatment were retrospectively explored in 15 patients: 8 patients receiving subsequent salvage chemotherapy (2 groups: 3 non-responders and 5 responders), and 7 responders to nivolumab alone (2 groups: 4 partial response and 3 complete response). The dynamics of NLR and ANC during nivolumab differed among these four groups (NLR P = 0.045, ANC P = 0.067). NLR and ANC during nivolumab treatment increased over time in non-responders to salvage chemotherapy, with an inverse relationship between drug response and NLR or ANC at four to six weeks among the four groups. We hypothesize that the early dynamics of NLR and ANC during nivolumab may be associated with the late efficacy of subsequent salvage chemotherapy. Further studies involving a large cohort are needed to confirm these findings, which could provide insight into the role of myeloid immunosuppressor cells in combination PD-1 blockade and chemotherapy.

Introduction

The combination of immunotherapy and chemotherapy, including salvage chemotherapy following PD‐1 inhibitor therapy, is a promising treatment for patients with non‐small cell lung cancer (NSCLC).1, 2, 3 We recently reported that salvage chemotherapy following nivolumab increased the response in some patients with NSCLC, regardless of tumor PD‐L1 status.4 In addition, the presence of minimal PD‐1 expression on tumor‐infiltrating lymphocytes (TILs) before treatment may be associated with improved response to salvage chemotherapy.4 However, we did not clarify the mechanisms that created these synergistic effects during nivolumab treatment. Several studies have shown that the neutrophil‐to‐lymphocyte ratio (NLR) and the absolute neutrophil count (ANC) in the blood can predict the efficacy of PD‐1 inhibitors in patients with NSCLC.5, 6, 7, 8, 9, 10 NLR and ANC have also been reported to reflect the number of circulating myeloid‐derived suppressor cells (MDSCs) and tumor‐associated neutrophils,11, 12, 13, 14, 15 which inhibit the function of antitumor T cells.16, 17 The immunosuppressive status in the tumor microenvironment could be monitored using the dynamics of NLR and ANC. Thus, the time‐series patterns of NLR and ANC during nivolumab treatment were retrospectively explored in lung cancer patients who received subsequent salvage chemotherapy to find an immunological clue to design future studies.

Methods

This single‐institute, exploratory study was performed retrospectively, following our previously published study.4 In brief, patients with recurrent or advanced NSCLC underwent salvage chemotherapy following nivolumab treatment between April 2016 and December 2017. Patients were excluded from the analysis if their lung cancer harbored EGFR mutations or ALK rearrangements, if the patients had documented infection, or if they had any other neoplasm requiring treatment. As controls, patients who achieved a complete response (CR) or a partial response (PR) to nivolumab alone were also selected. Patients received 3 mg/kg of nivolumab every two weeks. All patients were followed‐up to October 2018. The best tumor responses were assessed according to Response Evaluation Criteria in Solid Tumors version 1.1. In the present study, responders to salvage chemotherapy were defined as those with CR or PR, while non‐responders were those with stable disease (SD) or progressive disease. An automated hematology analyzer, the Sysmex XE‐5000 (Sysmex Corporation, Kobe, Japan), was used to evaluate the circulating leukocyte count. NLR was calculated as the ANC divided by the absolute lymphocyte count.

In the present study, four groups were analyzed according to drug response: non‐responders and responders to salvage chemotherapy following nivolumab therapy, and patients with PR and CR to nivolumab alone. The differences in the time‐series patterns of NLR and ANC among the groups were evaluated by repeated‐measures analysis of variance. The ANC values were log‐transformed for normalization. The Jonckheere–Terpstra test was used to evaluate the relationship between drug response and the neutrophil‐related index. A two‐tailed P value < 0.05 was considered significant. Statistical analyses were performed using R, version 3.5.1 (The R Foundation, Vienna, Austria). The institutional review board of Sasebo City General Hospital approved the study protocol.

Results

A total of 37 patients received nivolumab, including salvage chemotherapy, during the study period. As of October 2018, 8 patients (3 CR, 4 PR, and 1 SD) were still being treated with nivolumab. The other 19 patients underwent best supportive care following nivolumab treatment. The remaining 10 patients received salvage chemotherapy following nivolumab treatment; two were excluded because of the presence of concurrent cancer or EGFR mutation. Thus, 8 patients treated with salvage chemotherapy (3 non‐responders and 5 responders) were analyzed. Coincidentally, the 8 patients who received salvage chemotherapy did not show any objective response to the preceding nivolumab treatment. As mentioned above, 7 patients who responded to nivolumab alone (3 CR and 4 PR) were also analyzed. The patients’ characteristics are summarized in Table 1.

Table 1

Patient characteristics (n = 15)

Variable	Non‐responders to salvage Cx (n = 3)	Responders to salvage Cx (n = 5)	Patients with PR to Nivo (n = 4)	Patients with CR to Nivo (n = 3)
Age, years
Range	64–73	62–75	58–66	70–76
Gender
Male	2	4	4	3
Female	1	1	0	0
ECOG PS
0	0	1	1	0
1	3	4	3	3
Smoking status
Former smoker	3	5	4	3
Never‐smoker	0	0	0	0
Histologic subtype
Adenocarcinoma	0	2	2	1
Squamous cell carcinoma	2	2	2	1
Others	1	1	0	1
Stage
III	0	1	1	0
IV	2	3	2	0
Recurrent	1	1	1	3
Tumor PD‐L1 expression
≥ 50%	1	1	0	1
1–49%	1	1	2	1
< 1%	1	2	0	0
Not available	0	1	2	1
Line of nivolumab
Second‐line	2	1	1	0
Third‐line	1	2	1	2
Fourth‐line	0	2	2	1
Best response to nivolumab
CR	0	0	0	3
PR	0	0	4	0
SD	0	1	0	0
PD	3	4	0	0
Regimen of salvage Cx
S‐1	3	2	Not done	Not done
Carboplatin + nab‐paclitaxel	0	2	Not done	Not done
Docetaxel + ramucirumab	0	1	Not done	Not done
Best response to salvage Cx
CR	0	0	Not done	Not done
PR	0	5	Not done	Not done
SD	2	0	Not done	Not done
PD	1	0	Not done	Not done

CR, complete response; Cx, chemotherapy; ECOG PS, Eastern Cooperative Oncology Group performance status; Nivo, nivolumab; PD, progressive disease; PR, partial response; SD, stable disease.

The dynamics of NLR during nivolumab differed significantly among the four groups (P = 0.045) (Fig 1a). NLR tended to increase over time in non‐responders, while it was sustained in responders, with decreasing NLR in patients with PR and CR. ANC showed a similar trend over time (P = 0.067) (Fig 1b). A significant inverse relationship was observed between drug response and six‐week NLR (P = 0.048), and between drug response and six‐week ANC (P = 0.044) (Fig 2a). The four‐week NLR (P = 0.021) and the four‐week ANC (P = 0.071) showed a similar relationship (Fig 2b).

Figure 1

(a) The dynamics of the neutrophil‐to‐lymphocyte ratio (NLR) among the four groups: non‐responders and responders to salvage chemotherapy following nivolumab therapy, and patients with a partial response (PR) and a complete response (CR) to nivolumab alone. The vertical bars indicate standard errors. The dynamics of NLR are significantly different among the four groups (P = 0.045) () Non‐responders, () Responders, () Patients with PR, and () Patients with CR. (b) The dynamics of the absolute neutrophil count (ANC). The values of ANC were log‐transformed for normalization. The dynamics of log ANC over time show a similar trend (P = 0.067) () Non‐responders, () Responders, () Patients with PR, and () Patients with CR.

Figure 2

The relationships of drug response with the neutrophil‐to‐lymphocyte ratio (NLR) and the absolute neutrophil account (ANC) among the four groups: non‐responders and responders to salvage chemotherapy following nivolumab therapy, and patients with a partial response (PR) and a complete response (CR) to nivolumab alone. The vertical bars indicate standard errors. The values of ANC were log‐transformed for normalization. The relationships between drug‐response and the (a) six‐week NLR (P = 0.048) and six‐week ANC (P = 0.044) and (b) the four‐week NLR (P = 0.021) and four‐week ANC (P = 0.071) show a similar relationship. () NLR, and () ANC.

Discussion

This explorative, hypothesis‐generating study provided the following important findings and suggestions. The dynamics of NLR and ANC during nivolumab treatment differed among the patients receiving nivolumab alone and salvage chemotherapy. An inverse relationship was observed between drug response and these neutrophil‐related indices at four to six weeks after the initiation of nivolumab.

The dynamics of NLR and ANC during nivolumab treatment may be associated with the efficacy of the subsequent salvage chemotherapy. Many studies have shown that low baseline NLR and ANC predict the efficacy of nivolumab therapy in various cancers.6, 18, 19, 20 More recent studies have shown that an early decrease in NLR during treatment with PD‐1 inhibitors predicts its efficacy.5, 6, 7, 8, 9, 10 NLR and ANC have been reported to correlate with the number of circulating MDSCs and tumor‐associated neutrophils,11, 12, 13, 14, 15 which suppress the function of antitumor T cells.16, 17 Circulating MDSC levels were inversely associated with response rate and survival after treatment with nivolumab in cancer patients.21, 22 In tumor‐bearing mouse models, the chemokine‐related substances that recruit neutrophils and MDSCs into the tumor suppressed the antitumor activity of PD‐1 blockade.23, 24

In contrast, our previous studies showed that the presence of a few PD‐1+ TILs was associated with the response to salvage chemotherapy, as well as nivolumab treatment.4, 24, 25 However, we did not clarify why preceding nivolumab therapy showed less antitumor activity in the responders to salvage chemotherapy. A recent study showed that the neutrophil lineage was the most abundant immune cell type within NSCLC tissues, where neutrophil content was inversely correlated with the content of CD8+ T cells.26 In addition, disease progression during first‐line chemotherapy for NSCLC was reported to correlate with high levels of circulating MDSCs.27 Combining these findings with the results of our previous studies,4, 25, 28 we hypothesize that both the exhaustion levels of cytotoxic T cells and interference with myeloid cells may affect the efficacy of nivolumab and subsequent salvage chemotherapy (Fig 3).

Figure 3

A hypothetical model of the possible underlying mechanisms of nivolumab and subsequent salvage chemotherapy, based on the results of our previous and present studies.4, 25, 28 PD‐1 expression levels on tumor‐infiltrating lymphocytes (TILs) probably indicate the exhaustion of lymphocytes. The dynamics of the neutrophil‐to‐lymphocyte ratio (NLR) and absolute neutrophil count (ANC) during nivolumab treatment possibly reflect the number of myeloid‐derived suppressor cells (MDSCs) and tumor‐associated neutrophils (TANs) in the tumor. These combinations may contribute to the efficacy of each therapy. Cx, chemotherapy.

The combination of PD‐1 blockade and chemotherapy has shown clinical benefits in patients with advanced NSCLC as first‐line treatment.1, 29 However, the complete response rates were not necessarily high, therefore further immunotherapeutic strategies are needed to cure lung cancer. Myeloid immunosuppressor cells are considered a major obstacle to cancer immunotherapy.30 Many clinical trials with agents targeting myeloid immunosuppressor cells are ongoing.31 For example, the nuclear hormone liver‐X receptor agonist that depletes MDSCs enhances the antitumor activity of tumor‐antigen specific T cells as monotherapy or in combination with PD‐1 blockade in mice and cancer patients.32 The dynamics of blood NLR and ANC might become a surrogate biomarker to identify patients who could benefit from immunotherapy targeting the myeloid immunosuppressor cells.

The present explorative study has several limitations, including a small population of heterogenous patients and the lack of immunohistochemical evaluation. The present findings are potentially subject to selection bias and should be interpreted cautiously.

In conclusion, the present findings create a hypothesis that the early dynamics of NLR and ANC during nivolumab treatment may be associated with the late efficacy of subsequent salvage chemotherapy following nivolumab treatment. Further studies involving a large cohort are required to verify our hypothesis, which could shed light on the role of myeloid immunosuppressor cells when the combination of PD‐1 blockade and chemotherapy is used.

Disclosure

No authors report any conflicts of interest.