Abstract
Aim: To investigate the association between positivity for programmed cell death-ligand 1 (PD-L1) in brain metastases (BM) and the prognosis or clinical factors in patients with non-small cell lung cancer (NSCLC). Materials and Methods: Thirty-two patients with surgically resected brain-metastatic NSCLC were enrolled. The PD-L1 expression in BM was analyzed using the antibody against human PD-L1 (clone SP142). The PD-L1 positivity was defined as PD-L1 expression on brain-metastatic tumor cells of ≥5%. Results: Seven (21.9%) out of 32 patients showed PD-L1 positivity in BM. The PD-L1-positive BM group had a significantly shorter brain-specific disease-free survival than the PD-L1-negative BM group (p<0.05). PD-L1 positivity in BM was significantly associated with a heavy smoking history and the administration of radiotherapy for BM before surgery (p<0.05 and p<0.05, respectively). Conclusion: The PD-L1 expression in BM from NSCLC may be associated with local recurrence following surgery, and the smoking- or radiotherapy-derived effects.
Immunotherapy targeting programmed cell death-1 (PD-1) or programmed cell death-ligand 1 (PD-L1) has been widely adopted in patients with non-small cell lung cancer (NSCLC) (1). The KEYNOTE-024 study showed that select patients with NSCLC with PD-L1 positivity of ≥50% had a significantly longer median progression-free survival in the pembrolizumab-treated group than in the chemotherapy-treated group (1). Our previous study showed that PD-L1 positivity in primary lung adenocarcinoma was significantly associated with both shorter disease-free (DFS) and overall survival (OS) (2). Although assessing the expression of PD-L1 in primary NSCLC is essential for predicting the prognosis, the clinical significance of PD-L1 positivity in metastatic lesions, especially in brain metastases (BMs), is still unknown. Thus, we aimed to elucidate the relationship between PD-L1 positivity in BMs and brain-specific DFS (BS-DFS) and OS in this study.
Recently, the relationship of the PD-L1 expression in primary NSCLC and clinical factors was also investigated (2). Our previous studies showed that PD-L1 positivity in primary NSCLC was significantly associated with many clinical factors such as male sex, smoking, advanced stage, the presence of pleural and vessel invasion, and wild-type epidermal growth factor receptor (EGFR) gene expression (2). However, the relationship between the PD-L1 expression in BM and clinical factors has not been clarified. Thus, we also investigated the association between PD-L1 positivity in BM and the clinical factors in patients with NSCLC.
Materials and Methods
Patients. From January 2005 to December 2016, 36 patients with NSCLC with BM underwent surgical resection of their brain tumors at the Department of Surgery and Science, Graduate School of Medical Sciences and Department of Neurosurgery, Kyushu University. Of these 36 patients, four who were diagnosed with small-cell lung cancer or did not have a definitive diagnosis were excluded. Of the remaining 32 patients, 15 had undergone surgical resection or a biopsy of the corresponding primary NSCLC at the Department of Surgery and Science or the Research Institute for Disease of the Chest, Graduate School of Medical Sciences, Kyushu University, respectively. The remaining 17 patients did not undergo surgery or biopsy for primary NSCLC. Therefore, 32 patients, including 15 cases in which both the primary and metastatic tumors were available, were enrolled in the present study.
We conducted a retrospective analysis of the age, sex, smoking status (pack-year index) and performance status (PS). Clinical or pathological T and N staging of the primary NSCLC was determined using the seventh edition of the TNM Classification of Malignant Tumors (3). The histopathological subtype, EGFR mutation status, number, size, administration of radiotherapy for BM before surgery and PD-L1 expression of the BM were also analyzed. The Institutional Review Board of our institution approved the current study (IRB No. 27-435).
Immunohistochemistry. The surgical specimens were submitted to immunohistochemical analysis in 32 cases with surgically resected BMs and in 15 cases with corresponding primary NSCLC. Formalin-fixed tissue sections were embedded in paraffin. Paraffin-embedded tissue sections were dewaxed with xylene and rehydrated through a graded series of ethanol. After cutting the tissue into 4-μm-thick slices, endogenous peroxidase was inhibited for 30 min with 3% H2O2 in methanol. The sections were pretreated in a decloaking chamber with Target Retrieval Solution (Dako, Glostrup, Denmark) for 15 min at 110°C, and incubated with a monoclonal antibody at 4°C overnight. The immune complex was detected by a DAKO EnVision Detection System (Dako). The sections were reacted in 3,3’-diaminobenzidine, and counterstained with hematoxylin. All of the slides were dehydrated in a graded series of alcohol and xylene. They were then mounted and cover slips were applied.
A rabbit monoclonal antibody to human PD-L1 (clone SP142, dilution 1:100; Spring Bioscience, Ventana, Tucson, AZ, USA) was used as the primary antibody (2). The evaluation of PD-L1 expression was independently performed by a surgeon (S.T.) and a pathologist (K.T.). The percentage of PD-L1-positive tumor cells was determined as four categories: <1%; 1%-4%; 5%-49%; and 50%-100%. Lesions with PD-L1 expression in ≥5% of total tumor cells were considered positive (4). Human placental sections were used as negative and positive controls.
Statistical analyses. The association between PD-L1 expression and clinical factors was analyzed using two-sided Fisher's exact test, Student's t-test and Mann–Whitney U-test where appropriate. The BS-DFS and OS were defined as the time (months) from the day of surgery until brain recurrence and death, respectively. The survival probabilities were estimated using the Kaplan–Meier method and log-rank test. The differences were considered to be statically significant when the p-value was less than 0.05. All of the statistical analyses were performed using the JMP® 13 software program (SAS Institute Inc., Cary, NC, USA).
Results
Characteristics of the patients enrolled in the current study. Table I shows the characteristics of the patients included in this study. The majority of patients (71.9%) were current or former smokers, and many (62.5%) had a smoking history of ≥40 pack-years. Multiple BMs were observed in 14 (43.8%) patients. The majority of patients (93.7%) were pathologically diagnosed with adenocarcinoma and squamous cell carcinoma, respectively. A sensitive EGFR mutation was detected in eight (25.0%) patients. No patients received molecular-targeted agents before surgical resection of BM, while six (18.8%) underwent chemotherapy prior to surgery. Four (12.5%) underwent radiotherapy for their BM followed by surgical resection. Four (26.7%) and seven (21.9%) patients presented PD-L1-positive primary and metastatic sites, respectively (4).
The characteristics of 32 patients with non-small cell lung cancer (NSCLC) with brain metastasis (BM).
Kaplan–Meier curves showing brain-specific disease-free survival (A) and the overall survival (B) of patients with non-small cell lung cancer according to the expression of programmed cell death-ligand 1 in brain metastases (cut-off: 5%).
The characteristics of patients with non-small cell lung cancer (NSCLC) with programmed cell death-ligand 1 (PD-L1) positivity in brain metastases (BM) (cut-off: 5%).
Impact of PD-L1 positivity in BM on the BS-DFS and OS. The PD-L1-positive BM group had a significantly shorter BS-DFS than the PD-L1-negative BM group (64% vs. 77% at 1 year, and 21% vs. 71% at 2 years; p<0.05). No significant difference in the OS was observed between the PD-L1-positive and -negative BM groups (p=0.33). The Kaplan–Meier curves are shown in Figure 1. PD-L1 positivity in primary NSCLC was not significantly associated with BS-DFS (p=0.100), although the sample size was small (n=15, data not shown).
Relationship between PD-L1 positivity in BM and clinical factors. Table II shows the relationship between PD-L1 positivity in BM and clinical factors. A heavy smoking history and the administration of radiotherapy for BM before surgery showed a significant trend toward association with PD-L1 positivity in BM (p<0.05 and p<0.05, respectively). In contrast, age, sex, PS, EGFR mutation status, systemic chemotherapy before surgery and PD-L1 positivity in primary NSCLC was not significantly associated with PD-L1 expression in BMs (p=0.749, p=0.374, p=1.000, p=1.000, p=0.101, and p=0.517, respectively).
Discussion
To the best of our knowledge, this is the first report to show the clinical significance of the PD-L1 positivity in metastatic lesions. In the current study, we found PD-L1-positive BM in 21.9% of the 32 cases using the clone SP142 antibody. Mansfield et al. reported PD-L1-positive BM in approximately 32.9% of 73 cases using clone E1L3N antibody with the cut-off value at 5% (4). Our results were almost consistent with their report, taking into account the lower sensitivity of the SP142 clone reported by the Blueprint study (5). As shown in Figure 1, the PD-L1-positive BM group had a significantly shorter BS-DFS than the PD-L1-negative BM group. Given that the KEYNOTE-024 study revealed that select patients with NSCLC with at least 50% PD-L1 positivity in their primary lung cancer had a significantly longer progression-free survival in the pembrolizumab-treated group than in the chemotherapy-treated group (1), the analysis of PD-L1 positivity in BM may be useful for deciding postoperative therapeutic strategies or the follow-up interval period.
Because of the anatomical location of the samples, it is often difficult to investigate the PD-L1 expression in BM before treatment. With regard to the concordance rates of PD-L1 positivity between primary NSCLC and the corresponding BM, the current study and that of Mansfield et al. (4) showed that PD-L1 positivity in primary lesions was consistent with that of metastatic tumors in 75% and 86.2% of the total cases, respectively. Thus, the PD-L1 positivity in primary NSCLC may be sometimes be misleading for predicting the PD-L1 expression in BM (6). Furthermore, a recent report by Suda et al. using autopsy specimens showed that the heterogeneity of the expression of PD-L1 was dependent on the metastatic site (7). In the present study, a heavy smoking status and radiotherapy for the BM, but not the PD-L1 expression in primary NSCLC, were significantly associated with the PD-L1 positivity in BM (Table I). Although the sample size analyzed in this study was very small, substantial evidence exists to support the association between PD-L1 positivity and smoking status or radiotherapy. With regard to cigarette smoking, lung cancer in smokers was reported to be associated with a high burden of neoantigens, increased immunogenicity and up-regulation of PD-L1 (8-11). Regarding the relationship between PD-L1 expression and radiotherapy, pre-clinical studies have shown that PD-L1 expression was up-regulated both in vivo and in vitro after radiotherapy (12-14). In addition, some experimental studies have suggested that radiotherapy mediated the recruitment of CD8+ T-cells, which was reported to be associated with the intra-tumoral PD-L1 expression (15, 16). The results of those reports partly support our present findings.
The present study is associated with some limitations. One limitation is that this study investigated a small cohort of patients. Although the samples of BMs from NSCLC are rarely available, the small number of the cases makes it difficult to draw definite conclusions. Another limitation is that the PD-L1 analysis was performed using only one antibody (SP142). According to a previous report, the positivity rate using SP142 was found to be lower than that of other antibodies such as 28-8, 22C3 and SP263 (5). However, the rate of SP142-positivity was reported to be higher than that of other antibodies such as 28-8 and E1L3N in patients with small-cell lung cancer (17). Therefore, the PD-L1-positivity should be evaluated using other antibodies or cut-off values in future studies.
In conclusion, the expression of PD-L1 in BM was significantly associated with a shorter BS-DFS. A heavy smoking history and radiotherapy for BM before surgery were significantly associated with PD-L1 positivity in BM. These findings should be validated in future prospective studies with a larger patient population.
Acknowledgements
None.
Footnotes
Conflicts of Interest
Yoichi Nakanishi has received honoraria and research funding from Chugai Pharmaceutical Co. Ltd. Isamu Okamoto has received honoraria from Chugai Pharmaceutical Co. Ltd. All remaining Authors declare no conflicts of interest in regard to this study.
- Received October 24, 2017.
- Revision received November 9, 2017.
- Accepted November 13, 2017.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
In this issue
Jump to section
Related Articles
Cited By...
- PD-1/PD-L1 Blockers in NSCLC Brain Metastases: Challenging Paradigms and Clinical Practice
- Evaluation of Five Survival Scores in a Cohort of Elderly Patients With Cerebral Metastasis from Non-small Cell Lung Cancer
- Comparison of Diagnosis-specific Survival Scores for Patients With Cerebral Metastases from Malignant Melanoma Including the New WBRT-30-MM
- Potential Impact of the Interval Between Imaging and Whole-brain Radiotherapy in Patients With Relatively Favorable Survival Prognoses
- PD-L2 Expression as a Potential Predictive Biomarker for the Response to Anti-PD-1 Drugs in Patients with Non-small Cell Lung Cancer
- Molecular Sequence of Events and Signaling Pathways in Cerebral Metastases