Pembrolizumab Deemed Active Treatment for Brain Metastasis in NSCLC

Brain metastasis responses were elicited from treatment with pembrolizumab in patients with non–small cell lung cancer, in a PD-L1–positive cohort of a phase II clinical trial conducted by researchers at Yale Cancer Center. According to a YCC press release, pembrolizumab prolonged overall survival with limited toxicity in these patients.

Sarah B. Goldberg, MD, MPH

Brain metastasis responses were elicited from treatment with pembrolizumab (Keytruda) in patients with non—small cell lung cancer (NSCLC), in a PD-L1–positive cohort of a phase II clinical trial conducted by researchers at Yale Cancer Center (YCC). According to a YCC press release, pembrolizumab prolonged overall survival (OS) with limited toxicity in these patients.1

The long-term results and biomarker analysis were published in theLancet Oncology.

“Survival in this cohort of patients exceeds the historically documented survival for patients with brain metastasis from non—small cell lung cancer or NSCLC, which is a 2-year survival of about 14%,” said the study’s lead investigator Sarah B. Goldberg, MD, MPH, associate professor of medicine, Medical Oncology, YCC, in a statement.

Patients in the study were followed for a median of 8.3 months (range, 4.5-26.2). Overall, 42 patients with NSCLC were treated with pembrolizumab.2

For PD-L1—positive tumors (≥1%; cohort 1), confirmed brain metastasis response was observed in 11 patients (29.7%; 95% CI, 15.9%-47.0%). There were 7 partial responses and 4 complete responses, meeting the prespecified trial criteria for success. Overall responses were seen in 7 of the 37 patients in cohort 1 (18.9%; 95% CI, 8.0%-35.1%) for both brain and systemic responses.

Twenty-seven patients in the study were evaluable for both central nervous system (CNS) and systemic responses, and discordant outcomes were observed in 6 of them. In 3 patients (11%), progression occurred in the brain, and for the other 3, progression was systemic. The systemic responses lasted for a median duration of 6.9 months (range, 3.7-22.4). Thirty-one patients in cohort 1 either progressed or died. The median progression-free survival (PFS) for these patients was 1.9 months (95% CI, 1.8-3.7).

A post-hoc analysis was conducted to evaluate the subgroup of patients with brain metastasis and assess response according to baseline characteristics. The median duration of response in the brain for the 11 patients with CNS response was 5.7 months (range, 4.0-17.7). Just one patient in this subgroup had progression in the brain by the time of the last MRI on the study. The post-hoc median PFS was 2.3 months (95% CI, 1.9-not reached [NR]). Of the PD-L1—positive patients, 33% (95% CI, 19%-56%) were progression-free at 1 year.

In the overall study population, 26 of the 37 patients died by the time of data cutoff. The median overall survival was of 9.9 months (95% CI, 7.5-29.8). The estimated OS rate at 1 year was 40% (95% CI, 30%-64%) and at 2 years was 34% (95% CI, 21%-54%).

No new safety signals were observed with pembrolizumab in this study and AEs were consistent with that of other PD-L1 inhibitors. The neurological AEs were predominantly grades 1 and 2 in severity. Three patients each did have grade 3 cases of cognitive dysfunction, seizure, and stroke, however. Six serious treatment-related AEs occurred including pneumonitis, acute kidney injury, colitis, hypokalemia, and adrenal insufficiency, which was experienced by one patient each. Treatment discontinuation occurred in 11% of patients (n = 4) as a result of treatment-related AEs. Twenty-six patients died overall due to unknown causes.

Cohort 2 of the study included patients with PD-L1 expression less than 1% or unevaluable. In the primary analysis, no responses were seen with these patients. The cohort was also assessed in the exploratory analysis in comparison with cohort 1.

Thirty-four patients had available tumor tissue. For 32 of these patients, tumor biopsies were extracted before the start of the study and 2 were taken with intervening therapy. Four of these tumor tissue samples were from brain metastases. The biomarker analysis showed that better OS was achieved in patients with tumors showing PD-L1 expression of at least 1% in stromal or immune cells at 11.0 months (95% CI, 7.8-NR) compared with 2.7 months (95% CI, 1-NR;P= .031) in patients with no expression. A comparison between patients with tumor-cell PD-L1 expression of at least 1% and from those cohort 2 with tumor-cell PD-L1 expression less than 1% showed no significant difference in OS. The difference was 11.4 months (95% CI, 8.9-NR) versus 4.8 months (95% CI, 2.8-NR), respectively (P= .074). There was also no significant difference observed in PFS.

In an exploratory analysis, Goldberg et al assessed the effect of baseline tumor-infiltrating lymphocytes (TILs) on clinical outcomes. TILs had no impact on OS or PFS outcomes.

For use in the biomarker analysis, mRNA immune profiling was performed on 23 tumor biopsies, which showed higher amounts of several pro-inflammatory genes among patients who responded to treatment with pembrolizumab versus non-responders. These genes included key effector molecules and chemokines such as granzyme-B, CXC motif chemokines 9 and 10, and granulysin. No correlation between mRNA targets and PD-L1 expression was observed.

The 2-arm single-institution study evaluated for the primary end point of brain metastasis response, which consisted of partial and complete responses in CNS out of all evaluable patients by modified RECIST criteria. The secondary end points included OS, PFS, and safety/toxicity. The exploratory end points were PD-L1 expression and other biomarkers that were predictive of clinical activity.

To be eligible for cohort 1 the study, patients were required to have biopsy-proven metastatic NSCLC, an ECOG performance status of 2 or lower, PD-L1 expression in the tumor, a life expectancy of at least 3 months and adequate organ and marrow function. Patients could not have received prior treatment with PD-1, PD-L1, or PD-L2 inhibitors, nor could they have received other systemic therapies within 2 weeks of starting treatment in the study.

At baseline, several patients were excluded based on the study protocol. Eighty-six percent of patients (n = 26) had adenocarcinoma histology and 20% had tumors with oncogenic variants previously associated with a low response to checkpoint inhibitor therapy (n = 8). These oncogenic variants may have includedEGFR,ALK, orHER2mutations.

Pembrolizumab 10 mg/kg was administered intravenously every 2 weeks. Patients in the study received the treatment for up to 24 months or until disease progression or unacceptable toxicity. The median number of cycles patients received was 5.5 (range, 4-16.3).

“Pembrolizumab has activity in brain metastases from NSCLC that is similar to its systemic activity and can result in prolonged survival in a subset of patients. The potential benefit of immunotherapy alone is that all lesions are treated simultaneously and that such a treatment approach might reduce the incidence of toxicity from radiotherapy, including radiation necrosis. Studies of combination therapy such as with other immunotherapies, radiotherapy, or chemotherapy, are warranted to increase the frequency of responses in NSCLC patients with brain metastases,” wrote Goldberg et al.


  1. Yale Cancer Center reports first test using immunotherapy drug to treat advanced lung cancer shows benefit — and future promise [news release]. New Haven, Connecticut: Yale Cancer Center; April 13, 2020. Accessed April 15, 2020.
  2. Goldberg SR, Schalper KA, Gettinger SN, et al. Pembrolizumab for management of patients with NSCLC and brain metastases: long-term results and biomarker analysis from a non-randomised, open-label, phase 2 trial [Published Online April 3, 2020].Lancet Oncol. doi: 10.1016/ S1470-2045(20)30111-X