Introduction

The advent of immune checkpoint inhibitors (ICIs) has dramatically changed the treatment strategy of metastatic renal cell carcinoma (mRCC). Since a randomized clinical trial demonstrated a survival advantage of nivolumab, an anti-programmed death-1 (PD-1) monoclonal antibody, in the treatment of patients with mRCC [1], ICIs have been a mainstay treatment for advanced RCC [2]. In addition, therapeutic options have further expanded to include the combination of ICIs, such as nivolumab and ipilimumab, an anti-cytotoxic T-lymphocyte antigen-4 (CTLA-4) monoclonal antibody [3]. Most recently, several clinical trials have shown a survival benefit with the combined use of an ICI plus a multitargeted receptor tyrosine kinase inhibitor (TKI) [4, 5]. Thus, there will be more opportunities to use ICIs in the near future.

As the ideal treatment strategy for mRCC has become complicated in the ICI era, biomarkers predicting the clinical outcome of ICI treatment are more crucial than ever for the selection of appropriate treatment agents. Anti-programmed death-ligand 1 (PD-L1) has been expected to be a candidate predictive biomarker of nivolumab. However, while the previous reports have shown that the PD-L1 expression of tumor tissue may be associated with the clinical outcome of anti-PD-1 antibody treatment in several cancers [6,7,8], the survival benefit observed in a randomized clinical trial of mRCC patients treated with nivolumab was not associated with the PD-L1 status [1]. This contradiction may be attributed to several causes, such as variety in the PD-L1 expression pattern among cancer types, intratumoral heterogeneity and differences in the definition of PD-L1 positivity [8,9,10].

Recently, cancer-related systemic inflammation has been shown to be a major determinant of the disease progression and survival in most cancers [11, 12]. The neutrophil-to-lymphocyte ratio (NLR) has received focus as a biomarker for predicting the efficacy of nivolumab in several cancers, including mRCC. Several previous reports showed that a high NLR was significantly associated with worse clinical outcomes [13,14,15,16]. In addition, the level C-reactive protein (CRP), a common inflammatory marker, has also been studied as a prognostic biomarker of nivolumab in several cancers, such as lung cancer and melanoma [17,18,19], but its utility in mRCC patients treated with nivolumab has never been demonstrated.

The objectives of our study were to validate the utility of the NLR for predicting the clinical outcomes of our mRCC patients treated with nivolumab and to evaluate the impact of the CRP level on those patients.

Materials and methods

Patients

A total of 65 patients with metastatic or unresectable RCC who had been treated with nivolumab after 1 or more TKI regimens at Kobe University Hospital in Japan between December 2016 and February 2019 were included in this study. Informed consent to participate in the present study was obtained from all patients, and the study design was approved by the Research Ethics Committee of our institution (No. B190059), which was conducted in accordance with the Declaration of Helsinki.

Treatments and procedures

Nivolumab (3 mg/kg or 240 mg/body) was administered every 2 or 3 weeks until the occurrence of disease progression, unacceptable adverse events, withdrawal, or death. We collected the following data from the medical records of patients: patient demographics, histology, Karnofsky performance status (KPS), blood test results, and adverse events (AEs). Patients were classified into three risk categories (favorable-, intermediate-, and poor-risk groups) according to the International metastatic renal cell carcinoma Database Consortium (IMDC) classification [20]. The elevation of lactate dehydrogenase (LDH) was defined as a value of > 222 U/L, which is considered to be the upper limit of normal in our hospital. The treatment response to nivolumab was evaluated by computed tomography (CT) or bone scintigraphy at least once every 12 weeks and classified according to the response evaluation criteria in solid tumours (RECIST) 1.1. The objective response rate (ORR) was defined as the percentage of patients with confirmed complete or partial responses among all treated patients.

The NLR was derived from the absolute neutrophil and absolute lymphocyte counts of a full blood count. In present study, we used an NLR of 5 as the threshold value, as its clinical utility in predicting patient outcomes in a variety of cancers was examined in a previous systematic review [21]. In addition, the changes in the NLR and CRP level were compared between baseline and four weeks after the induction of nivolumab.

Statistical analyses

We assessed the objective response, progression-free survival (PFS) and overall survival (OS) of nivolumab treatment. Fisher’s exact test was performed to evaluate the association of the NLR and CRP values with an objective response. The PFS and OS were estimated using the Kaplan–Meier method, and we assessed several potential factors for predicting the PFS and OS with nivolumab using the Cox proportional hazards model.

In addition, we evaluated the Pearson product-moment correlation coefficient to analyze the correlation between the NLR and CRP level. The optimal threshold of the CRP level for predicting an NLR ≥ 5 was determined as the value maximizing the sum of the sensitivity and specificity in the receiver operating characteristic (ROC) analysis.

For all statistical analyses, we employed EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [22], which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). It is a modified version of R commander designed to add statistical functions frequently used in biostatistics. Each test was 2-sided, and a value of p < 0.05 was considered significant.

Results

Patients’ characteristics

The clinical characteristics of the 65 patients are shown in Table 1. The median age was 68 years (range 44–87 years), and most patients were male (n = 47; 72.3%), had a KPS ≥ 80% (n = 47; 72.3%), had undergone nephrectomy (n = 56; 86.2%), and had been diagnosed with clear cell carcinoma (n = 47; 72.3%). According to the IMDC classification, three patients had a favorable risk, 34 (52.3%) had an intermediate risk and 28 (43.1%) had a poor risk. Thirty-six (55.4%) patients had received only one previous TKI, and the remaining 29 (44.6%) patients had received ≥ 2 such therapies. The median values of the NLR, platelet count and LDH and CRP levels were 3.1 (range 1.1–10.0), 24.4 (×104/µL, range 7.9–47.6), 199 (IU/L, range 104–2104) and 0.48 (mg/dL, 0.0–20.0), respectively. Immune-related adverse events (irAEs) of any grade occurred in 40 (62.5%) patients.

Table 1 Patient and tumor characteristics
Full size table

Treatment response and survival outcomes

As of database lock, the response assessment to nivolumab was available in 64 of 65 patients. Of these, 3 (4.7%) patients achieved a complete response, and 12 (18.7%) achieved a partial response; the ORR was 23.4% (Table 2). The median PFS of nivolumab was 7.2 months (95% confidence interval [CI] 2.6–9.7), and the median OS was not reached during the median 9.5 months of observation in this study (Fig. 1a, b).

Table 2 Treatment response in mRCC patients treated with nivolumab
Full size table
Fig. 1
figure 1

Kaplan–Meier estimates of a the progression-free survival and b overall survival among mRCC patients treated with nivolumab

Full size image

Impact of the NLR and CRP on clinical outcomes

As shown in Fig. 2, both the PFS and OS of patients with NLR ≥ 5 was significantly shorter in comparison to those with NLR < 5. The median PFS of patients with NLR < 5 and NLR ≥ 5 was 7.9 months and 1.1 months (Fig. 2a, p < 0.001), and the median OS was not reached and 6.0 months (Fig. 2b, p < 0.001), respectively.

Fig. 2
figure 2

Kaplan–Meier estimates of a the progression-free survival and b overall survival among mRCC patients treated with nivolumab based on the NLR stratified by a cut-off value of 5

Full size image

Next, our analysis of the correlation between the NLR and CRP level revealed a moderate correlation between the two variables (Fig. 3a, correlation coefficient: 0.568), and the optimal cut-off value of the CRP level for predicting an NLR ≥ 5, as estimated by an ROC analysis, was 2.14 mg/dL (Fig. 3b, sensitivity: 0.714, specificity: 0.843).

Fig. 3
figure 3

a Results of a correlation analysis between the NLR and CRP level b receiver operating characteristic curve (ROC) of CRP for predicting an NLR ≥ 5 at the induction of nivolumab, and c results of a correlation analysis between the change in the NLR and CRP level

Full size image

The PFS and OS curves based on the CRP level stratified at 2.1 mg/dL are shown in Fig. 4a, b. The PFS of patients with a CRP level ≥ 2.1 mg/dL was significantly shorter than that of patients with a CRP level of < 2.1 mg/dL (Fig. 4a, median PFS: 2.1 months and 7.9 months, respectively; p = 0.004). Similarly, the OS in patients with a CRP level of ≥ 2.1 mg/dL was significantly shorter than that in patients with a CRP level of < 2.1 mg/dL (Fig. 4b, median OS: 10.3 months and not reached, respectively; p < 0.001).

Fig. 4
figure 4

Kaplan–Meier estimates of a the progression-free survival and b overall survival among mRCC patients treated with nivolumab based on the CRP level stratified by a cut-off value of 2.1 mg/dL

Full size image

We then evaluated the association between the clinical characteristics and PFS of patients treated with nivolumab. As shown Table 3, a poor risk on IMDC classification, NLR ≥ 5, and elevated LDH were identified as factors significantly associated with poor PFS, with hazard ratios (HRs) of 3.08 (95% CI 1.40–6.81; p = 0.005), 4.54 (95% CI 1.93–10.7; p < 0.001) and 3.02 (95% CI 1.43–6.40; p = 0.004), respectively. However, the CRP level was not associated with PFS.

Table 3 Results of univariate and multivariate analyses for the progression-free survival in mRCC patients treated with nivolumab
Full size table

In the analysis of the relationship between the clinical characteristics and OS of nivolumab-treated patients (Table 4), we observed that a history of nephrectomy, the NLR and the CRP level significantly contributed to the predicted OS. While NLR ≥ 5 (HR: 4.88, 95% CI 1.35–17.7; p < 0.001) and CRP ≥ 2.1 mg/dL (HR: 3.89, 95% CI 1.01–15.0; p = 0.049) were significantly associated with poor OS, a poor risk on IMDC classification was not.

Table 4 Results of univariate and multivariate analyses for the overall survival in mRCC patients treated with nivolumab
Full size table

According to previous reports [23, 24], we focused on the change in the NLR and CRP but not their actual values for the analysis of factors associated with an objective response. There was also a moderate correlation between the change in the NLR and that in the CRP level (Fig. 3c, correlation coefficient: 0.599). As shown Fig. 5, patients with a ≥ 25% decrease in their NLR and CRP level showed a significantly better response to nivolumab than those without a ≥ 25% decrease in these values, with odds ratios of 9.54 (95% CI 2.09–49.8, p = 0.001) and 4.36 (95% CI 1.03–18.9, p = 0.032), respectively. The actual NLR and CRP values had no impact on the objective response (data not shown).

Fig. 5
figure 5

Treatment response to nivolumab in mRCC patients according to the decrease in the NLR and CRP. ORR objective response rate, CR complete response, PR partial response, SD stable disease, PD progressive disease

Full size image

Discussion

In this retrospective analysis, we showed that the NLR and CRP level were significantly associated with clinical outcomes in mRCC patients treated with nivolumab. To our knowledge, this was the first report of the utility of the CRP level as a prognostic biomarker of nivolumab in patients with mRCC.

Systemic inflammation is recognized as a key determinant of the outcome in patients with cancer [21]. Due to their widespread availability in clinical practice, subtypes of white blood cells, such as neutrophils and lymphocytes, have been used parameters of the cancer-related inflammatory response. Neutrophilia has been shown to be associated with a poor survival in several cancer types [25,26,27]. Jeyakumar et al. reported that tumor-associated neutrophils may enhance angiogenesis, tumor growth and progression to a metastatic phenotype [28]. An elevated neutrophil count is included as a risk factor in the IMDC classification (anaemia, thrombocytosis, neutrophilia, hypercalcaemia, KPS < 80%, and < 1 year from diagnosis to treatment) which is often used to predict the clinical outcomes in patients with mRCC [20].

On the other hand, lymphocytes have been shown to be effective suppressors of cancer progression and to reflect host immunity [29]. Cytotoxic T lymphocytes (CTL) are known to induce apoptosis of cancer cells by Fas signaling: an interaction between CD95L molecules on the CTL and CD95 molecules on target tumor cells [30]. The association between lymphocytes and cancer outcomes has also been investigated. Fogar et al. reported that lymphopenia was associated with negative outcomes in patients with pancreatic cancer [31]. In addition, previous reports have shown that tumor-infiltrating lymphocytes were associated with a reduced tumor recurrence and favorable prognosis [32,33,34].

However, Basem et al. suggested the superiority of the NLR to leukocyte counts (e.g., neutrophils and lymphocytes), because the NLR is stable compared to absolute counts, which can be influenced by various physiological, pathological and physical factors. In addition, the NLR can represent both inflammatory and immune pathways [35]. We, therefore, used the NLR for our assessments in the present study. Our study showed that while a high NLR was associated with a poor OS, a poor risk according to the IMDC classification, which was established in the TKI era, was not associated with a poor OS. The NLR may be suitable for inclusion in a new classification system in the ICI era if such a system was to be established.

The CRP level is another commonly used inflammatory maker that has been studied as a predictive biomarker for mRCC patients treated with cytokine agents or TKIs [36, 37]. Interleukin-6, which mainly regulates the CRP synthesis in the liver, is thought to play important roles in stimulating angiogenesis [38] and inhibiting apoptosis of cancer [39], indicating that the CRP level reflects the tumor microenvironment and aggressiveness of tumors. In the present study, we showed for the first time that the CRP level was positively correlated with the NLR and that an elevated CRP level was significantly associated with a poor prognosis of nivolumab in mRCC patients, suggesting that the CRP level may also be a prognostic biomarker in mRCC patients treated with nivolumab, similar to the NLR.

In addition, the present study showed that the change in the NLR and CRP level but not their actual values were biomarkers for predicting the treatment response. Consistent with our results, a previous report on mRCC patients showed that a decreased CRP level was a parameter for predicting the anti-tumor effect of TKIs [23, 24]. In ICI treatment, Lalani et al. demonstrated an association between a decrease in the NLR and an objective response, although not significantly so [40]. However, the efficiency of using changes in inflammatory markers to predict the treatment response of ICIs in mRCC remains unclear, and further validation will be needed.

The present study had several limitations. First, the study was small in size and retrospective, with a relatively short follow-up duration. To validate our results, further prospective, large-scale studies will be needed. Second, the cut-off of the NLR and CRP level should be validated further for their future clinical application. In addition, the CRP level is not routinely evaluated in some hospitals [41]; as such, the CRP level will be most beneficial for predicting the treatment response in settings, where it is routinely evaluated, such as Japan. Furthermore, we showed a moderate correlation between the NLR and CRP. Thus, the results of the multivariate analysis of factors associated with for PFS and OS should be interpreted with care, because there may have been a multicollinearity problem [42].

In conclusion, we found that high values of the NLR and CRP level were significantly associated with a poor OS among mRCC patients treated with nivolumab. Regarding the treatment response, decreases in the NLR and CRP level but not their actual values were identified as significant factors for predicting an objective response. Our data suggest that the CRP level may be a useful marker for predicting the clinical outcome of ICI treatment, along with the NLR. These findings should be further investigated in a larger prospective cohort.