Nivolumab in Relapsed or Refractory Primary Central Nervous System Lymphoma: Multicenter, Retrospective Study
Article information
Abstract
Purpose
Given that 40%-50% of primary central nervous system lymphoma (PCNSL) tissues exhibit aberrancy on 9p24.1, immune checkpoint inhibitors (ICI) may work for the disease.
Materials and Methods
To define the role of ICIs in PCNSL, we carried out a nationwide retrospect analysis for 22 patients who had been treated with nivolumab monotherapy for relapsed or refractory PCNSL.
Results
The median age at diagnosis was 66, and male: female ratio was 1:1. Patients received nivolumab after a median of 3 lines (range, 2 to 6) of therapy and at the median age of 67 years (range, 37 to 82 years). Eleven patients (50%) were refractory to the last treatment prior to nivolumab. With a median follow-up duration of 22.3 months (95% confidence interval [CI], 13.1 to 31.5), nine patients (41%) had an objective response (6 complete responses, 3 partial responses), and the median duration of response was 20.9 months (95% CI, 1.7 to 40.0). The median progression-free survival and overall survival were 2.1 months (95% CI, 0.2 to 4.0) and 18.9 months (95% CI, 5.0 to 32.8), respectively. Nivolumab was generally well-tolerated as no patients required dose reduction and only two patients required delay of treatment.
Conclusion
Our study suggests that nivolumab can be a reasonable option with the durable response for RR PCNSL.
Introduction
Primary central nervous system lymphoma (PCNSL) is a diffuse large B-cell lymphoma (DLBCL) arising within the brain, spinal cord, leptomeninges, or eye which accounts for less than 1% of non-Hodgkin lymphomas and 2.4%-3% of all brain tumors [1]. Despite recent improvements in the frontline treatment [2,3], up to 60% of patients eventually relapse, and there has not been established standard treatment for them. Current treatment options for relapsed or refractory PCNSL (RR PCNSL) include conventional chemotherapy, novel agents including Bruton tyrosine kinase (BTK) inhibitors, or immunomodulatory drugs (IMiDs), whole brain radiotherapy (WBRT), or high-dose chemotherapy followed by autologous stem cell transplantation (ASCT). However, few of them result in high response rates or long-term remission, as the median overall survival (OS) is around 1 year [4]. Thus, there are huge unmet needs for these patients.
PCNSLs differ not only in the involved anatomical sites, but also in the genomic aberrancies from those of central nervous system (CNS) involvement of systemic DLBCL [5]. One of the notable features is that PCNSLs are enrichment in chromosome 9p24.1 copy number alterations (CNAs) and highly expressed programmed death-1 (PD-1) and its ligand programmed death-ligand 1 (PD-L1) [6]. These genomic aberrancies are also found in classical Hodgkin’s lymphoma (cHL) and primary mediastinal large B-cell lymphoma (PMBCL) [7], and several studies incorporating immune checkpoint inhibitors (ICIs) showed promising results for these diseases [8-10]. Accordingly, we may assume that ICIs may work for some patients with PCNSL as they share similar genomic aberrancies with cHL or PMBCL.
There have been limited data on treatment outcomes of ICIs for patients with RR PCNSL due to the rarity of the disease and the limited accessibility of the drugs [11,12]. To further elucidate the role of ICIs in PCNSL, we carried out a nationwide retrospective analysis of nivolumab-treated patients with RR PCNSL.
Materials and Methods
1. Patients
The inclusion criteria for this retrospective analysis were as follows: (1) histologically confirmed diagnosis of B-cell PCNSL; (2) received at least 1 cycle of nivolumab for RR PCNSL; (3) availability of medical reports, including age, sex, histologic subtype, prior treatment, risk factors, response, and survival outcomes. Patients with secondary CNS involvement of systemic lymphoma were excluded.
2. Statistical analysis
The primary objective of the study was to evaluate the efficacy and the safety of nivolumab treatment in patients with RR PCNSL. The international workshop to standardize baseline evaluation and response criteria for primary CNS lymphoma were used to determine response [13], and the subgroup comparisons were performed using Pearson’s λ2 tests. Progression-free survival (PFS) was calculated from the date of the first administration of nivolumab to the date of progression or death. Duration of response (DoR) was calculated from the date of the first confirmed response to the date of progression. OS was calculated from the first administration of nivolumab to the date of death. All of these parameters were calculated using the Kaplan-Meier method. The safety profile was recorded when dose reduction or delay of treatment occurred according to the National Cancer Institute Common Terminology Criteria for Adverse Events ver. 5.0. For all statistical analyses, p < 0.05 was considered significant, and they were performed using IBM SPSS statistics for Windows ver. 21.0 (IBM Corp., Armonk, NY).
3. Compliance with ethical standards
All procedures performed in this study were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The exemptions for obtaining the informed consent were approved by the institutional review board of each institute.
Results
1. Patients’ characteristics
From November 2018 to September 2022, 22 consecutive patients from four tertiary institutes in Korea were included in the analysis. There were 11 males and 11 females each. Their median age at initial diagnosis was 66 years (range, 34 to 81 years). All patients had DLBCL, and in 15 patients whose immunohistochemical subtypes could be identified, three (20.0%) and 12 (80.0%) patients had germinal center B-cell (GCB) and non-GCB subtypes, respectively. In terms of the backbone of frontline treatment, 16 patients received rituximab plus high-dose methotrexate (HD-MTX), procarbazine, and vincristine (MPV), and five patients received MPV without rituximab, and one patient received HD-MTX monotherapy. Six and two patients had received ASCT and WBRT as consolidation therapy.
Before nivolumab treatment, patients had received a median of 3 lines (range, 2 to 6) of prior treatment. Salvage treatment included etoposide (n=13), platinum-containing regimens (n=12), re-administration of HD-MTX (n=9), cytarabine (n=9), lenalidomide (n=6), WBRT (n=8), ibrutinib (n=3), and ASCT (n=2). Rituximab was combined in eight patients. Eleven patients (50.0%) had refractory disease before nivolumab treatment.
Patients had started nivolumab treatment after a median duration of 15.7 months (95% confidence interval [CI], 10.3 to 21.1) from the initial diagnosis of PCNSL at their median age of 67 (range, 37 to 82 years). According to the International Extranodal Lymphoma Study Group (IELSG) risk stratification [14], four (18.2%), 12 (54.5%), and six (27.3%) patients were classified as low-, intermediate-, and high-risk groups, respectively. The details are described in Table 1.
Characteristics of baseline and at the time of nivolumab treatment
2. Outcomes of the nivolumab treatment
All patients started a nivolumab treatment nivolumab at a dose of 3 mg/kg every 2 weeks, and the median treatment cycle was 3 (range, 1 to 23). The overall response rate (ORR) was 41% (9/22), with six complete responses and three partial responses noted. The median time to first response was 3.0 months (95% CI, 0.7 to 5.3). All four patients in the IELSG low-risk group responded (100%), while only five out of 18 patients (28%) in the IELSG intermediate- and high-risk group responded (p=0.028). After a median follow-up duration of 22.3 months (95% CI, 13.1 to 31.5), the estimated median PFS was 2.1 months (95% CI, 0.2 to 4.0) and the 1-year PFS rate was 36.9% (±10.9%). For the nine responders, the estimated median DoR was 20.9 months (95% CI, 1.7 to 40.0). The estimated median OS was 18.9 months (95% CI, 5.0 to 32.8), and the 1-year OS rate was 53.1% (±13.9%) (Fig. 1). As expected, responders showed a strong tendency for prolonged OS (36.3 vs. 8.6 months, p=0.058). While earlier use (3L vs. 4L or later) of nivolumab was not associated with a higher response rate (40% vs. 42%), it showed better OS trend (36.3 vs. 8.6 months, p=0.164).
Kaplan-Meier curves for patients. (A) Progression-free survival (PFS) and overall survival (OS). (B) Duration of response (DoR). CI, confidence interval.
Nivolumab treatment was generally well-tolerated as no patients required dose reduction and only two patients experienced delay of treatment (1 infection and 1 generalized seizure). The generalized seizure was eventually revealed to have been caused by progressive disease, thus the patient stopped treatment.
After nivolumab treatment, 11 patients received subsequent treatments, including four cases of rituximab plus lenalidomide, three cases of WBRT, two cases of ibrutinib, one case of intrathecal chemotherapy and high-dose chemotherapy each. At last follow-up, 12 patients were alive, and 10 patients died (9 due to lymphoma progression and one due to coronavirus disease 2019 pneumonia 2 years after the end of nivolumab treatment).
Discussion
Despite recent therapeutic progress including incorporating rituximab and various consolidation treatment modalities, up to 60% of patients with PCNSL will experience refractory or relapsed disease following first-line treatment. Several treatments, including re-treatment with HD-MTX, rituximab, or alkylating agents such as temozolomide or thiotepa, and novel agents including BTK inhibitors, or IMiDs have been introduced, yet none of them has been established as a standard-of-care, which prompted genomic approaches in therapeutic area.
Targetable genomic aberrancies of PCNSL include structural variations at chromosome 9p24.1 with consequent PD-L1 overexpression [15]. Chapuy et al. [6] reported that 52% (33/63) of PCNSL patients had CNAs in chromosome 9p24.1. Ou et al. [16] also reported that 32 out of 48 (66.7%) patients expressed intermediate to high PD-L1 expression and 38 out of 42 (90.5%) exhibited intermediate to high tumor mutational burden. In another series of 71 PCNSL tumor specimens, PD-1 and PD-L1 expressions were found in 16 (23%) and 42 (59%), respectively [17]. In a Korean study, PD-L1 expression was found in 67% (35/68) of PCNSL samples and it was positively associated with serum levels of soluble PD-L1, which resulted in poor OS and PFS [18]. Finally, Takashima et al. [19] conducted RNA-sequencing on PCNSL samples from 31 patients and found that high expression of inhibitory genes including PD-1 and LAG-3 was associated with poor prognosis. As there are several conflicting reports [20,21], further studies including proper modalities for measurement of PD-L1 and prognostication are required.
Certain subtype of lymphomas which exhibited CNAs in chromosome 9p24.1 and PD-1/PD-L1 expression showed promising efficacies with ICIs. For instance, cHL, in which chromosome 9p24.1 CNAs are found over 90% of cases [22], ICIs significantly improved outcomes both in newly-diagnosed and relapsed/refractory patients [8,9]. Another subtype of lymphoma that shares similar clinicogenetics features with cHL is PMBCL [7], and ICIs showed positive results in relapsed/refractory settings [10,23].
These findings encouraged the introduction of ICIs in PCNSL treatment. In the very first report by Nayak et al. [11], four patients with recurrent PCNSL and one patient with CNS recurrence of primary testicular lymphoma (PTL) had been treated with nivolumab at 3 mg/kg every 2 weeks. All five patients responded (4 complete response [CR], 1 partial response [PR]) with improvement of neurologic function, and they lived without progression at least 13 months. In a Russian case series, eight patients with PCNSL and one patient with CNS involvement of PTL were treated with nivolumab 100 mg every 2 weeks [12]. Four (44%) and three (33%) patients achieved partial and complete response, respectively. The 2-year PFS and OS rates were 26% and 44%, respectively. In one case report, nivolumab was administered intrathecally to overcome blood-brain barrier [24], which achieved 12 months of complete response. There are several studies dealing with ICI-based combinations in RR PCNSL patients. Ambady and colleagues reported a case series in which six patients with RR PCNSL had been treated with rituximab plus ICIs (5 pembrolizumab, 1 nivolumab) [25]. Three patients (50%) achieved CR while the other three patients immediately progressed. Our study shows similar response rates. While 41% (9/22) of patients responded, non-responders quickly progressed, which resulted in relatively short PFS in the entire cohort (median, 2.1 months). For responders, however, the efficacy lasted unexpectedly longer as the median DoR was 20.9 months which has rarely been observed in other reports dealing with RR PCNSL. Nivolumab treatment did not require dose reduction, and only one patient experienced delay of treatment due to a drug-related adverse event.
As PCNSL also frequently harbors mutations involving in BTK pathways including MYD88, CD79B, and/or CARD11 [5], and antitumor immunity of ICIs may be enhanced by BTK inhibitors [26], incorporating BTK inhibitors and ICIs seems a reasonable option. In a case report by Feng et al. [27], patients who had been treated with zanubrutinib plus tislelizumab (a PD-1 inhibitor) showed CR after two cycles of treatment and the patient remained in CR for 20 months. In a single-arm phase II study, 18 patients (16 PCNSL, 2 secondary CNS involvement) were treated with ibrutinib plus nivolumab [28]. The ORR was 78% and 50% achieved CR. The median PFS and OS were 6.6 months and 25.4 months, respectively. Eight patients (44%) developed grade 3 or 4 adverse events and four and two patients experienced treatment interruption and discontinuation, respectively. Along with the current study, details these trials are provided in Table 2. Considering that most PCNSLs are non-GCB subtype, IMiDs may be effective in disease, and a number of small studies incorporating IMiDs demonstrated efficacy [29,30]. There is an ongoing trial which includes lenalidomide and nivolumab for PCNSL patients (NCT04609046).
Studies including ICI-based treatment in PCNSL treatment
Other than its small size and retrospective nature, the current study has several limitations. Although the response rate (41%) of our study is quite similar to reported incidence of chromosome 9p24.1 aberrancy in PCNSL, genomic profiling to define predictive role of ICIs has not been performed. Therefore, we cannot rule out the possibility that this number is a coincidence. Another limitation is that the responses were examined by individual investigators. It is well-known that distinguishing CR and PR with brain MRI is challenging in daily practice. However, as progression of PCNSL inevitably accompanied by neurologic deterioration, and we also assessed OS, we believe the conclusive data have validity.
In conclusion, we have analyzed 22 patients with RR PCNSL who had been treated with nivolumab monotherapy. Nine patients responded (41%) and their DoR exceeded 20 months. Given that only restricted treatment strategies are available for RR PCNSL patients, our study suggests that nivolumab can be a suitable option for the disease. In addition, as there are no widely accepted ‘standard-of-care’ for RR PCNSL, genomics-guided treatment approaches will be necessary for this disease, and our study provided one of the largest outcomes for ICI treatment in PCNSL.
Notes
Ethical Statement
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the Declaration of Helsinki or comparable ethical standards. Waiving any informed consent was approved by the institutional review boards of the respective sites (Samsung Medical Center, SMC2023-03-076-001; National Cancer Center, NCC2022-0350; Seoul National University Bundang Hospital, 2023-0039; Asam Medical Center, 2023-0945).
Author Contributions
Conceived and designed the analysis: Yi JH, Yoon DH.
Collected the data: Kim SJ, Kim SA, Jung J, Yoon DH.
Contributed data or analysis tools: Kim SJ, Kim SA, Jung J, Yoon DH.
Performed the analysis: Yi JH, Yoon DH.
Wrote the paper: Yi JH, Yoon DH.
Conflicts of Interest
Conflict of interest relevant to this article was not reported.