Abstract
Background/Aim: Increasing availability of effective treatment options for metastatic renal cell carcinoma (mRCC) has highlighted the importance of identifying predictors of treatment response. Although PD-L1 expression in renal cancer has been reported as a predictor of treatment response and prognosis, its assessment by immunohistochemistry is invasive and difficult to perform repeatedly. Soluble PD-L1 (sPD-L1) has recently been proposed as a predictive biomarker for several tumour types. Therefore, we evaluated sPD-L1 levels in patients with mRCC treated with nivolumab and investigated its association with treatment response. Patients and Methods: We performed a prospective single-arm study in patients with mRCC treated with nivolumab as second line or later therapy. We measured serum sPD-L1 before and during treatment, classified patients based on baseline values (sPDL1 ≥0.23 ng/ml vs. <0.23 ng/ml) and compared outcomes between the two groups. Results: A total of 43 patients with mRCC were included in this study, with 17 (39.5%) classified as low sPD-L1 and 26 (60.5%) as high sPD-L1. The International Metastatic RCC Database Consortium risk score was significantly poorer in the high sPD-L1 group. The objective response rate was significantly higher (41.2% vs. 7.7%) and overall survival significantly longer (p=0.0323) in the low group compared to the high group. There were no significant differences in progression-free survival between the two groups. Conclusion: Our study findings indicate that sPD-L1 might be a predictor of treatment response to nivolumab in patients with mRCC.
Systemic therapy for metastatic renal cell carcinoma (RCC) has developed significantly in the past decade (1). Immune checkpoint inhibitors (ICI), such as programmed death-ligand 1 (PD-L1) inhibitors, programmed cell death protein 1 (PD-1) inhibitors, and cytotoxic T-lymphocyte associated antigen-4 (CTLA-4) inhibitors have become primary treatment options for patients with RCC (2). Phase III trials have shown that nivolumab, a human monoclonal antibody against PD-1, is effective in treating advanced RCC (2). However, some patients show excellent responses to ICIs, including nivolumab, while others show limited or no response (1). Therefore, the value of a biomarker to predict the therapeutic effect of ICI is increasingly apparent, although none have been identified to date.
PD-L1 is a transmembrane protein expressed in tumour cells, where it is implicated in tumour immune evasion (3), and in T cells, where it is associated with immune responses and inflammation (4). PD-L1 expression observed in tumour cells and tumour-infiltrating lymphocytes (TILs) by immunohistochemistry (IHC) has been reported to correlate with prognosis and treatment response (5-7). However, the evaluation of PD-L1 expression by IHC is limited by the invasive nature of tissue sampling, the difference in expression between primary and metastatic lesions, difficulty in quantification, and the change in expression with treatment (8).
Recently, it was reported that transmembrane PD-L1 has a soluble form which is produced by tumour cells or activated mature dendritic cells and associated with the diversity of the composition and function of the PD-1/PD-L1 signalling pathway (3, 9). Furthermore, the relationship between soluble PD-L1 (sPD-L1) in blood and treatment outcome has been reported for several carcinoma types (3, 10). Although the measurement of soluble PD-L1 is readily reproducible, the relationship between sPD-L1 and therapeutic efficacy in RCC remains unclear. Therefore, we assessed the relationship between sPD-L1 and treatment response and prognosis by evaluating sPD-L1 in patients with RCC treated with nivolumab.
Patients and Methods
Study design and setting. This was a prospective single-centre, single-arm, open-label invasive intervention study performed at the Kobe University Hospital. Patients with advanced or unresectable RCC treated with nivolumab as second-line or later therapy were included in this study and recruited from November 2017 to April 2019. Patients younger than 20 years and patients with active double cancer at the time of informed consent were excluded. Details of the study design, participants, schedule, and data management have been reported previously (11).
Ethical approval. Study approval was obtained from the Kobe University Clinical Research Ethical Committee (approval number C180067) before initiation. The clinical trial registration number was UMIN000027873. Written informed consent was obtained from all patients prior to participation.
Research outline. After registration, general blood tests and serum sPD-L1 were measured in patients prior to treatment with nivolumab. In addition, the target lesion was evaluated by computed tomography (CT) and/or magnetic resonance imaging (MRI) before treatment. Nivolumab was started at 240 mg every 2 weeks in all patients. During the treatment period, serum sPD-L1, general blood test parameters, and target lesions images were evaluated at 4, 12, 24 and 48 weeks after the start of treatment. Treatment was discontinued according to the physicians’ decision when disease progression or intolerable adverse effects were observed. The status of each patient with respect to response rate, disease progression, and death was followed for at least 2 years.
Patient characteristics. Patient characteristics including age, sex, body mass index (BMI), and Karnofsky Performance Status (KPS) were recorded. Tumour characteristics including International Metastatic RCC Database Consortium (IMDC) risk group (12), tumour stage at diagnosis, number, and type of metastatic organs before nivolumab treatment, histology of the primary tumour, and treatment line of nivolumab were also recorded.
sPD-L1 measurements. For measurement of sPD-L1, peripheral blood samples were centrifuged, and the resulting serum was stored at −80°C until analysis by ELISA using the HI-1000 system (Sysmex, Hyogo, Japan) (13).
Outcome measurements. Target lesions were assessed in all patients by radiographic studies before and during treatment with nivolumab. Objective response was evaluated by complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) based on the RECIST criteria (14). The objective response rate (ORR) was defined as the sum of the percentages of patients who had the best overall response of CR or PR to nivolumab. Progression-free survival (PFS) was defined as the period from first administration of nivolumab until confirmation of exacerbation. Overall survival (OS) was defined as the period from first administration of nivolumab until confirmation of death.
Outcomes. Cases were classified based on concentration of sPD-L1 ≥0.23 ng/ml vs. <0.23 ng/ml. The cut-off value for sPD-L1 was set at 0.23 ng/ml, which was the median concentration of soluble PD-L1 measured in patients with RCC in a previous study (15). We compared the outcomes of PFS, OS, and ORR between the low and high sPD-L1 groups.
Statistical analysis. Continuous variables were reported as the mean and standard deviation (SD) if the distribution met normality criteria or as the median and interquartile range (IQR) if the distribution did not meet normality criteria. The Mann-Whitney U-test and Kruskal-Wallis test were used to analyse continuous variables. The chi-squared test was used to analyse categorical variables. Recurrence-free survival and overall survival were determined using Kaplan-Meier analysis and evaluated using the log-rank test and Cox regression analysis. Values of p<0.05 were considered statistically significant. All statistical analyses were performed using EZR (16), a graphical user interface for R, designed to add statistical functions frequently used in biostatistics.
Results
A total of 43 patients with mRCC were included in the present study, of which 17 (39.5%) were classified as low sPD-L1 patients and 26 (60.5%) as high sPD-L1 patients. The median follow-up was 26.2 months, with 44.3 months as the longest follow-up. Patient characteristics are detailed in Table I. The median age was 66 years, and 32 patients (74.4%) were male. Forty patients (93.0%) were post-nephrectomy. All patients had distant metastases prior to treatment with nivolumab, and 30 patients (69.8%) had lung metastases. Histological analysis of the primary tumour showed that 37 patients had clear cell tumour histology while five patients had non-clear status. Nivolumab was second-line treatment in 28 cases and third-line or later in the remaining 15 cases.
Patient characteristics.
Table II shows the comparison of sPD-L1 by patient characteristics. sPD-L1 was significantly higher in the group aged over 65 years and in males. In addition, sPD-L1 was significantly higher in the group with multiple metastatic organs. Furthermore, there was a significant difference in sPD-L1 according to IMDC risk. There was no significant difference in sPD-L1 for KPS, BMI, neutrophil-lymphocyte ratio, treatment line of nivolumab, or histology of the primary tumour.
Soluble PD-L1 levels according to patient characteristics.
Comparison of oncological outcomes between the two groups is shown in Table III. ORR was significantly higher in the low sPD-L1 group compared with the high sPD-L1 group (41.2% vs. 7.7%; p=0.0191). There were two cases of CR for the overall best response in the low group, but none in the high group. In the high group, the overall best response was PD in 15 cases (57.7%). The median PFS was 5.7 (2.3-38.6) months in the low group and 2.3 (1.8-17.8) months in the high group, although the difference between the groups was not significant (Figure 1). However, median OS was not reached (NR) (34.7-NR) in the low group and was 24.6 (10.9-NR) months in the high group and was significantly longer in the low group (p=0.0323) (Figure 2).
Comparison of oncological outcomes between the low and high soluble PD-L1 groups.
The Kaplan-Meier curve estimates of progression-free survival of metastatic renal cell carcinoma patients treated with nivolumab stratified by the soluble PD-L1 level cut-off of 0.23 ng/ml.
The Kaplan-Meier curve estimates of overall survival of metastatic renal cell carcinoma patients treated with nivolumab stratified by the soluble PD-L1 level cut-off of 0.23 ng/ml.
In Table IV, the impact of various clinicopathological factors on OS was evaluated using Cox regression analysis. Univariate analysis identified the significant factors as IMDC risk, sPD-L1, and number of metastatic organs for OS. Furthermore, on multivariate analysis of these significant factors, IMDC risk was shown to have independent prognostic value. On multivariate analysis, sPD-L1 was not an independent predictor of OS, although the hazard ratio for sPD-L1 was 2.30.
Univariate and multivariate Cox regression analysis for overall survival.
sPD-L1 was evaluated after 4 weeks of nivolumab treatment in 40 patients. In patients with an objective response, there was no significant difference in sPD-L1 between baseline and week 4 of treatment. In addition, the change in sPD-L1 from baseline to week 4 of treatment was not significantly different between the groups with and without an objective response.
Discussion
In the present study, we evaluated serum sPD-L1 in patients with mRCC treated with nivolumab and assessed the correlation between sPD-L1 and treatment outcome. We observed a significantly longer OS in patients with low sPD-L1 before treatment initiation. In addition, the ORR was significantly higher in the low sPD-L1 group. Therefore, sPD-L1 before nivolumab treatment may represent a predictor of prognosis and treatment response in patients with mRCC.
PD-L1 expression by IHC has previously been reported as a predictive factor in RCC therapy. In renal tumours, PD-L1 expression by IHC has been shown to be associated with clinicopathological features, including high-stage disease, high-grade disease, increased tumour size, and tumour necrosis (5, 6, 17). Furthermore, it has been reported that outcomes were significantly poorer in patients with PD-L1-positive renal cancer (5). The relationship between PD-L1 expression and the efficacy of systemic therapy has also been reported in several studies (7, 18, 19). PD-L1 expression has been associated with poor treatment response and prognosis in patients treated with tyrosine kinase inhibitors (TKI) (7) and with nivolumab as second-line therapy (19).
PD-L1 expression evaluated by IHC is therefore a useful predictor of treatment response, but this method is invasive and difficult to repeat (8). However, the evaluation of sPD-L1 is minimally invasive and repeatable using blood samples. A number of recent studies have reported the relationship between sPD-L1 and renal cancer. Kushlinskii et al. (20) evaluated the association between sPD-L1 in serum and oncological features in 106 patients with RCC and observed that advanced disease stage, larger tumours, lymph node metastasis, and distant metastasis were significantly associated with high sPD-L1. Similarly, Frigola et al. (15) measured sPD-L1 in 172 patients with clear cell RCC and found that sPD-L1 was higher in patients with larger tumours, advanced stage and grade tumours, and tumours with necrosis. In the present study, sPD-L1 was significantly higher in the group with multiple metastatic organs, and there was a significant difference in sPD-L1 according to IMDC risk. Similar to previous reports, sPD-L1 was higher in patients with advanced disease. It is important to note that sPD-L1 in the present study was stratified by patient background in patients with mRCC after systemic therapy.
The level of sPD-L1 appears to be related to prognosis; previous studies have reported that sPD-L1 predicts prognosis in patients with clear cell RCC (21) and postoperative recurrence in patients with non-metastatic RCC (22). In our study, high sPD-L1 in serum was associated with poor OS in patients with mRCC treated with nivolumab as second-line or later therapy. In renal cancer patients, risk classifications, such as the Memorial Sloan Kettering Cancer Center (MSKCC) and IMDC models have been used widely as clinical prognostic factors (12, 23). In particular, the IMDC risk group has been shown to be a valuable predictor of prognosis in renal cancer patients treated with TKI or ICI (24). In the present study, the IMDC risk group was a strong predictor of OS in mRCC patients treated with nivolumab. However, sPD-L1 was not found to be an independent predictor of OS in the multivariate analysis, although the hazard ratio for sPD-L1 was 2.30. In studies with greater numbers of patients and matched IMDC risk, sPD-L1 may be found to be an independent predictor of prognosis.
Some RCC patients are resistant to TKI or ICI therapy. Therefore, predictors of systemic treatment for RCC are essential. In previous studies, high sPD-L1 levels have been associated with poor PFS and OS in mRCC patients treated with TKIs, such as sunitinib and axitinib (22, 25). However, few studies to date have examined the relationship between treatment efficacy of ICI and sPD-L1 in RCC patients. Incorvaia et al. (26) reported the relationship between plasma sPD-L1 and treatment efficacy using PFS and objective response in 21 patients with mRCC treated with nivolumab as second-line therapy. Although the report was based on a small number of cases, significantly prolonged PFS was observed in patients with higher baseline sPD-L1. In the present study, however, there was no significant difference in PFS by serum sPD-L1, although ORR was significantly better in the group with lower sPD-L1. In addition, OS was significantly longer in the low sPD-L1 group, a finding which is not aligned with the previous study by Incorvaia et al. (26). However, they did observe, as in the present study, that sPD-L1 was higher in patients with more metastases and poorer IMDC risk. It is challenging to explain the prolonged PFS observed in these patients. Differences in matrix type for measuring sPD-L1 (plasma versus serum) and the inclusion of patients that were not in an IMDC poor risk group might have led to differences in the results between our studies. Nivolumab is a PD-1 immune checkpoint inhibitor antibody that selectively blocks the interaction between PD-1 and its ligands PD-L1 and PD-L2 (27); however, ORR and OS were significantly poorer in patients with higher sPD-L1 in the present study. This finding might be attributable to the poor prognosis of patients with high sPD-L1 (21) and the low ORR for nivolumab treatment when given as second-line or later therapy.
The evaluation of sPD-L1 is minimally invasive and repeatable, and is therefore suitable for comparing sPD-L1 before and after treatment. There are limited reports on the relationship between changes in sPD-L1 during treatment for RCC and treatment response. Incorvaia et al. (26) showed no significant difference in sPD-L1 between baseline and week 4 of treatment in RCC patients who responded to treatment with nivolumab. In our study, there was no significant difference in sPD-L1 between baseline and week 4 of treatment in patients who had OR to nivolumab. The present study found no association between treatment response and changes in sPD-L1. Although the role of sPD-L1 in cancer remains unclear, sPD-L1 level may affect not only disease status but also inflammatory and immune responses (9, 28).
Our study had some limitations, including its prospective design and lack of randomization. Furthermore, our findings are based on a small patient cohort, and patient backgrounds between the two groups were not matched. Despite these limitations, our findings provide further evidence and support for the relationship between sPD-L1 and prognosis in patients with mRCC treated with nivolumab with long-term follow-up.
Conclusion
Previously, it had been suggested that PD-L1 expression may represent a predictive factor for prognosis. In the present study, we found that sPD-L1 might be a predictor of treatment response to nivolumab in patients with mRCC. It is important to predict the efficacy of systemic therapy for mRCC on the selection of treatment options. The measurement of sPD-L1 is minimally invasive and uncomplicated, and its usefulness as a predictive biomarker warrants further investigation.
Acknowledgements
We thank Clare Cox, Ph.D., from Edanz (www.edanz.com/ac) for editing a draft of this manuscript.
Footnotes
Authors’ Contributions
Study concept and design: N.W. and N.H.; acquisition of data: N.W. and Y.B.; analysis and interpretation of data: N.W. and N.H.; drafting of the manuscript: N.W. and N.H.; critical revision of the manuscript for important intellectual content: T.H., T.T. and J.F.; statistical analysis: N.W. and N.H.; funding acquisition: N.H. and M.F.; supervision: Y.N. and M.F.
Conflicts of Interest
This study was funded by Ono Pharmaceutical Co., Ltd and Bristol-Myers Squibb, Co. M.F. and N.H. received research funding from Ono Pharmaceutical Co., Ltd and Bristol-Myers Squibb, Co. The other Authors have no conflicts of interest to declare in relation to this study. There are no Authors working for Ono Pharmaceutical Co., Ltd and Bristol-Myers Squibb, Co.
- Received December 10, 2022.
- Revision received December 28, 2022.
- Accepted December 30, 2022.
- Copyright © 2023 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
