Phase II Randomized Trial of Lenalidomide in Children With Pilocytic Astrocytomas and Optic Pathway Gliomas: A Report From the Children's Oncology Group

PURPOSE

Children with low-grade glioma often require long-term therapy and suffer from treatment morbidity. Although targeted agents are promising, tumor targets often encompass normal developmental pathways and long-term effects of inhibition are unknown. Lenalidomide is an immunomodulatory agent with wide-ranging properties. Phase I studies indicated greater tolerability of lenalidomide in children compared with adults and a potential dose-response effect.

PATIENTS AND METHODS

We performed a phase II trial of lenalidomide in children with pilocytic astrocytomas and optic pathway gliomas who failed initial therapy. Primary objectives included determination of objective response rate of children randomly assigned to regimen A, low-dose (20 mg/m²/dose), or regimen B, high-dose (115 mg/m²/dose) lenalidomide, and assessment for early progression. Secondary objectives included estimation of event-free survival, overall survival, incidence of toxic events, and assessment of plasma lenalidomide concentrations. Lenalidomide was administered once daily × 21 days of each 28-day cycle for each regimen.

RESULTS

Seventy-four eligible patients were enrolled (n = 37, each arm). The predefined activity level of interest was achieved for both arms. Four objective responses were observed in each arm, and the number of early progressors was low. Eighteen patients completed 26 cycles of therapy (regimen A, n = 12; regimen B, n = 6). The median number of cycles was 14 (range, 2-26) for regimen A and 11 for regimen B (range, 1-26). Of 74 eligible patients who received study drug, 30 required dose reduction for toxicity (regimen A, n = 6; regimen B, n = 24) and 16 discontinued because of toxicity (regimen A, n = 2; regimen B, n = 14).

CONCLUSION

Lenalidomide demonstrates a sufficient level of activity in children with low-grade glioma to warrant further exploration. Low-dose (20 mg/m²/dose administered once daily × 21 days of each 28-day cycle) lenalidomide appears to have better tolerability with comparable activity.

INTRODUCTION

Low-grade gliomas are the most common CNS tumors affecting children.¹ This group of tumors displays histologic and molecular heterogeneity and varied clinical behavior. Despite identification of tumor targets, treatment of unresectable lesions often entails long-term administration of cytotoxic chemotherapies and sometimes radiation therapy. Several chemotherapy regimens have demonstrated activity in pediatric low-grade gliomas, and the 5-year overall survival (OS) for this population exceeds 90%.² However, low-grade glioma is a chronic disease for many children, with approximately 50% of children treated with chemotherapy needing second-line therapy within 5 years.^3,4 As the majority of these children survive,⁵ there is a need to develop novel, tolerable, and effective therapy with minimal toxicity.

CONTEXT

• Key Objective

• Treatment selection for children with low-grade glioma focuses on tumor control while reducing long-term toxicities and preserving quality of life. Although targeted therapies are promising alternatives to cytotoxic agents, their efficacy, tolerability, and long-term consequences of inhibiting a normal developmental pathway are unknown. We describe the results of our phase II study of lenalidomide in children with recurrent, refractory, and progressive pilocytic astrocytomas and optic pathway gliomas.

• Knowledge Generated

• The threshold level of activity of interest was met. The standard, lower dosing of lenalidomide appeared as efficacious and better tolerated than high-dose lenalidomide. Long-term treatment (up to 2 years) was feasible and tolerable. Lenalidomide represents a potential additional treatment option to children with uncontrolled disease.

• Relevance (S. Bhatia)

• The limited range of acute toxicities and efficacy at the standard dose, make lenalidomide as a potential agent for managing low grade glioma using a chronic disease paradigm.*

  *Relevance section written by JCO Associate Editor Smita Bhatia, MD, MPH, FASCO.

Although surgical resection is the mainstay of therapy and curative for some low-grade gliomas, many, particularly optic pathway gliomas, are not amenable to surgical resection and, historically, even biopsy has been avoided. Optic pathway gliomas, diagnosed on the basis of clinical and radiographic features, are assumed to be pilocytic astrocytomas. Most pediatric low-grade gliomas demonstrate activating genetic alterations in the mitogen-activated protein kinase (MAPK) pathway, with KIAA1549:BRAF fusion being the most common alteration in pilocytic astrocytomas.^6,7 Therapies targeting this pathway are currently in clinical trials. While demonstrating promising activity in early clinical trials,^8,9 it is concerning that the consequences of long-term inhibition of this normal developmental pathway are yet unknown as animal studies suggest that alterations or even transient interference in this pathway may be associated with neurotoxicity.^10-12

Lenalidomide (Revlimid, Celgene Corporation, Summit, NJ; Bristol Myers Squibb, Cambridge, MA), an oral, structural analog of thalidomide, is classified as an immunomodulatory agent (ImID). It is US Food and Drug Administration–approved for multiple myeloma, myelodysplastic syndrome with deletion of 5q, and mantle cell lymphoma. The potential antitumor mechanisms of action include immunomodulatory effects, antiangiogenic properties, and direct cytotoxic effects.^13-17 Recently, lenalidomide was identified as a protein degrader, recruiting specific substrates to CRL4^CRBN, inducing their ubiquitination and subsequent degradation.¹⁸ Early clinical trials of lenalidomide in children demonstrated tolerability of higher doses compared with adults and antitumor activity against low-grade glioma.^9-21 In the phase I clinical trial of lenalidomide for children with recurrent, refractory, or progressive CNS tumors, antitumor activity was observed, with long-term stable disease at all dose levels, yet objective responses were observed only at the higher dose levels of 88 mg/m²/d and 116 mg/m²/d, each administered once daily × 21 days of each 28-day cycle,¹⁹ raising the question of dose-related response.

We performed a randomized phase II study of standard-dose (20 mg/m²/dose administered once daily × 21 days of each 28-day cycle) and high-dose (115 mg/m²/dose administered once daily × 21 days of each 28-day cycle) lenalidomide in children with optic pathway gliomas and pilocytic astrocytomas who failed initial chemotherapy. The primary objectives were to determine objective response rates and assess the number of patients with early progression (ie, within the first 6 months of therapy). Secondary outcome measures included estimation of event-free survival (EFS), OS, assessment of toxicities, and pharmacokinetic parameters.

PATIENTS AND METHODS

Patients

Eligible patients were younger than 22 years with a pilocytic astrocytoma (histologic verification) or optic pathway glioma (histology not required) that relapsed, progressed, or became refractory to conventional therapy, including at least one prior regimen containing carboplatin. Patients were required to have body surface area ≥0.4 m² because of limitations of capsule size, measurable residual disease, and a Lansky/Karnofsky performance score ≥60 and be able to swallow capsules. Patients on corticosteroids needed to be on a stable or decreasing dose. Patients were required to have recovered from the toxic effects (<Common Toxicity Criteria v4.0 grade 2 toxicity) of previous therapy and received last dose of myelosuppressive chemotherapy ≥3 weeks before entry (6 weeks for nitrosourea), last dose of biologic agents ≥7 days, and last dose of immunotherapy ≥6 weeks; patients who received radiation therapy must have received the last fraction of craniospinal radiotherapy ≥6 months before study entry and ≥4 weeks after last fraction of focal radiotherapy. Patients were required to have adequate hematologic function (absolute neutrophil count [ANC] ≥1,000/μL, hemoglobin ≥8.0 g/dL, platelets ≥100,000/μL), an age-adjusted normal serum creatinine or creatinine clearance/radioisotope glomerular filtration rate of ≥70 mL/min/1.73 m², a serum glutamic-pyruvic transaminase level of ≤110 U/L, a total bilirubin level of ≤1.5 times the upper limit of normal for age, a serum albumen level of ≥2 g/dL, an O₂ saturation of >94%, and no dyspnea at rest. A codiagnosis of neurofibromatosis type 1 did not affect eligibility. Patients with a history of thromboembolism unrelated to central line or a known thromboembolic predisposition were ineligible. Pregnant or breastfeeding females were ineligible; sexually active females were required to use two methods of birth control beginning at least 28 days before initiating lenalidomide; all males were required to use latex condoms during intercourse with a woman of childbearing potential during the study and for 28 days after the study.

The Protocol (online only) was conducted under the oversight of the National Cancer Institute (NCI) pediatric central institutional review board and also obtained approval as needed by local institutions. Patients or their legal guardians provided written consent before enrollment; assent was obtained as appropriate.

Drug Administration

All patients received lenalidomide. Lenalidomide was supplied by the Clinical Trials Evaluation Program of the Division of Cancer Treatment, NCI as 2.5 mg, 5 mg, and 25 mg capsules. On eligibility confirmation, patients were randomly assigned to receive either regimen A (standard dose, 20 mg/m²/dose) or regimen B (high dose, 115 mg/m²/dose) lenalidomide; each dose was rounded to within 10% of prescribed dose for convenience of capsule size. Lenalidomide was administered daily × 21 days of each 28-day cycle; therapy was continued for up to 26 cycles in the absence of disease progression or off-protocol criteria. Before starting each 28-day cycle, ANC had to be ≥1,000/μL, platelets had to be ≥100,000/μL, and pregnancy test for females of childbearing potential had to be negative.

Dose Modifications for Toxicity

Toxicity was assessed using the NCI Common Terminology Criteria for Adverse Events Version 4. Patients experiencing significant toxicities were dose reduced by 25% on subsequent cycles; two dose reductions were allowed. Significant hematologic toxicities were defined as any grade 4 neutropenia of duration >3 days, any grade 4 thrombocytopenia or requirement for ≥2 platelet transfusions/cycle for platelet count <50,000/μL, and ≥14-day delay in starting subsequent cycles because of ANC <1,000/μL or platelets <100,000/μL; nonhematologic dose-limiting toxicities included any grade 3 or 4 nonhematologic toxicity with the exception of grade 3 nausea and vomiting controlled by antiemetics, grade 3 fever or infection, grade 3 electrolyte abnormalities responsive to oral supplements, any toxicity that persisted for >3 days and considered medically significant or intolerable so as to require treatment interruption or which recurred on rechallenge, and any toxicity of any grade that caused a ≥14-day delay between cycles.

Statistical Considerations

Any eligible patient who received at least one dose of lenalidomide was considered in each of the following outcome measures evaluation. Each patient's outcome was associated with the individual's randomized regimen.

Patient response evaluation.

Response was assessed at each time point according to the methodology described in the Response Assessment section. A patient who experienced progressive disease within 6 months of enrollment was considered to have early disease progression (EDP). A patient who did not have EDP and demonstrated a complete response (CR) or partial response (PR) before the termination of protocol therapy was considered a responder for the evaluation of the statistical rule. In all other cases, an evaluable patient was considered a nonresponder for the application of the statistical rule. On the basis of reported response rates in two clinical trials with similar, although not identical, populations and design,^22,23 a true objective response rate of at least 20% and an EDP of <25% were considered sufficient for further interest in this agent.

Statistical evaluation of treatment efficacy.

The two-stage statistical rule was applied to each dose assignment group independently. An initial stage of 20 evaluable patients was enrolled. If 0-1 patient responded or ≥10 patients had EDP, enrollment of that dose group was terminated with the conclusion that the dose did not provide sufficient efficacy for further evaluation. If that stopping rule was not met, an additional 17 evaluable patients were enrolled. If three or fewer were considered responders or ≥15 of the 37 patients experienced EDP, we concluded that the dose did not provide sufficient evidence for further evaluation. Otherwise, we concluded that the dose reached our activity level of interest and was sufficiently efficacious for further study. A summary of the operating characteristics of this rule is provided in the Data Supplement ([Table 1], online only). The 95% CI for the probability of response and probability of EDP were calculated using the method in the study by Jung and Kim.²⁴

Patient excessive toxicity evaluation.

Any patient who experienced a toxicity requiring dose reduction as described in the Dose Modifications for Toxicity section after having their dose of lenalidomide reduced two times before that were considered to have experienced an excessive toxicity event. In all other cases, an evaluable patient was considered not to have experienced an excessive toxicity event.

Statistical evaluation of excessive toxicity.

A two-stage rule was applied to each dose regimen independently. If ≥5 of the first 18 evaluable patients experienced excessive toxicity, a safety data review was to occur. If not required, toxicity was assessed again after 36 patients were enrolled, and a review is required if ≥8 patients experienced excessive toxicity. A summary of the operating characteristics of this rule is provided in the Data Supplement (Fig 1).

Assessment of toxicity.

The maximum grade of each toxicity type experienced over all cycles of therapy was determined for each patient.

EFS.

EFS was calculated as the time from enrollment on study until the date of disease progression, date of death regardless of the cause or date of last follow-up, whichever occurred first. Patients who experienced disease progression or died were considered to have experienced an event; otherwise, the patient was censored at last contact.

OS.

OS was calculated as the time from enrollment on study until the date of death regardless of the cause or date of last follow-up, whichever occurred first. Patients who died were considered to have experienced an event; otherwise, the patient was censored at last contact.

Statistical analysis of EFS and OS.

The proportion of patients who were event-free or surviving as a function of time since enrollment was estimated by the Kaplan and Meier method.²⁵ CIs were calculated using the complementary log-log transformation of the survivor function estimate. Median potential follow-up for OS was calculated by the method in the study by Schemper and Kim.²⁶

Response Assessment

Magnetic resonance imaging (MRI) to assess response was obtained within 1 week preceding study entry, before every other cycle for the first 13 cycles and then every three cycles thereafter. Standard MRI obtained at each evaluation included T1 pre- and postcontrast, T2, and fluid-attenuated inversion recovery sequences. The institution radiologist selected the sequence best highlighting tumor, using the same sequence throughout each patient's study course to determine tumor response/progression for the purposes of maintaining or discontinuing patients on study; retrospective central radiology review of potential responses was performed to determine best response and overall response rate. Each patient was classified on the basis of their best objective response, with a responder defined as a patient whose best overall response was a CR or PR.

The definition of response was based on the measurement of the longest tumor dimension (excluding cyst) and its perpendicular for each target lesion, maintained for ≥4 weeks. Disease burden was calculated as the sum of the products of all measured lesions. CR was defined as complete disappearance of all known disease for ≥4 weeks, and PR was defined as a reduction of ≥50% in disease burden for ≥4 weeks and no new lesions or progression of any lesion. Stable disease was defined as a decrease of <50% or an increase of <25% in disease burden without the appearance of new lesions, and progressive disease was defined as a ≥25% increase in disease burden or appearance of new lesions.

RESULTS

Patient Characteristics

This study opened for accrual in March 2012 and closed to accrual January 2017. The data cutoff was December 31, 2020. Seventy-five patients were enrolled; 74 were evaluable for response (n = 1 ineligible because the on-study MRI was not performed within 7 days of enrollment) including n = 37 on regimen A and n = 37 on regimen B. Characteristics of evaluable patients at the time of enrollment are listed in Table 1. The number of administered cycles ranged from 1 to 26 (regimen A: median 14, range, 2-26; regimen B: median 11, range, 1-26).

TABLE 1.

Patient Characteristics

Efficacy/Responses and Progression-Free Survival

Both regimens reached the threshold for activity in the first stage and were expanded to stage 2. The number of patients with objective response rate (CR, PR) as determined by central review was n = 4 for regimen A and n = 4 for regimen B. The estimated objective response rate for each regimen is 12.9% with associated 95% CI, 4.09 to 22.6.

A total of 10 patients experienced EDP (range, 1.7-5.5 months) including n = 6 patients from regimen A and n = 4 patients from regimen B. The estimated probability of EDP for regimen A was 16.2% (95% CI, 9.2 to 29.5), and that for regimen B was 10.8% (95% CI, 5.48 to 23.1). Four additional patients from regimen B were removed from protocol therapy in the first 6 months of therapy for reasons other than disease progression. The level of efficacy, defined by both objective responses and lack of EDP, reached the predefined level of interest, that is, at least 20% for objective response and not more than 25% for early local progression. As specified in the statistical design, this fell within the 95% CIs for objective response and early local progression.

EFS and OS are demonstrated in Figure 1. The median potential follow-up was 60.2 months. The 2-year EFS and OS were 46% (95% CI, 34 to 57) and 93% (95% CI, 84 to 97), respectively. For regimen A, the 2-year EFS and OS were 43% (95% CI, 27 to 58) and 95% (95% CI, 80 to 99), and for regimen B, the 2-year EFS and OS were 48% (95% CI, 31 to 63) and 92% (95% CI, 77 to 97), respectively. The response rates did not differ on the basis of the number of previous treatment regimens (Data Supplement [Table 2]). The time to achieve an objective response varied (cycles 6, 8, 11, and 13 in regimen A and cycles 4, 6, 6, and 13 in regimen B; Data Supplement [Table 3]). Duration of response is described in the Data Supplement (Fig 2).

FIG 1.

Kaplan-Meier survival curve: EFS and OS, including all eligible patients from enrollment. EFS, event-free survival; OS, overall survival.

Tolerability and Toxicity

Table 2 shows the incidence of significant toxicities. As expected, myelosuppression was the most common toxicity, with grade 4 neutropenia occurring in n = 2 patients on regimen A and n = 12 patients on regimen B. Fifteen patients discontinued therapy for reasons other than protocol-defined toxicity including patient's best interest (n = 1), not otherwise specified (n = 3), vision loss (n = 1), alternative therapy (n = 2), persistent diarrhea (n = 1), compliance (n = 1), and parent refusal (n = 6).

TABLE 2.

Incidence of Significant (Grade 3, 4) Toxicities

Pharmacokinetics

Cycle 1 pharmacokinetics data are available for 49 enrolled patients (regimen A, n = 20; regimen B, n = 28). One patient is not included in this analysis (missing identifier, n = 1). Lenalidomide concentrations (ng/mL) ranged from 0.49 to 2,401.24. The median serum lenalidomide concentration was 15.14 (range, 0.49-732.22) for regimen A and 160.09 (range, 2.69-2,401.24) for regimen B.

DISCUSSION

In this randomized phase II trial, we demonstrate that lenalidomide is a safe and effective therapy for children with pilocytic astrocytomas and optic pathway gliomas. Higher dosing does not offer additional benefit when compared with standard dosing, and standard dosing (regimen A) is better tolerated than the higher dosing (regimen B).

Following the phase I studies of lenalidomide in children, two questions remained, namely, were responses dose-related and was toxicity dose-related. A maximum tolerated dose was not defined in the initial phase I pediatric brain tumor study as doses much higher than the adult maximum tolerated dose were tolerated and dose escalation was halted after 10 dose levels (range, 15-116 mg/m² once daily × 21 days of each 28-day cycle) were evaluated. Long-term stable disease or objective responses were observed at each of the dose levels. We now demonstrate that drug efficacy, most often expressed in this study as long-term stable disease, was not dose-related. In addition, treatment-limiting toxicities did appear to be dose-related, suggesting unnecessary additional risk without additional benefit.

The level of efficacy, defined by both objective responses and lack of early progression, reached the predefined threshold of interest. Defining appropriate cohorts for comparison is complicated by the historical incorporation of all low-grade gliomas into study cohorts, frequent absence of biopsy, variable response assessment methods and definition, heterogeneous study end points, and the absence of long-term tolerability data. The threshold for this study was based on trials with similar but not identical eligibility and response criteria^22,23; we also took into consideration additional trials that incorporated cytotoxic agents although the response assessment and criteria differed significantly.^27-30 This included, for example, measurement of change in contrast enhancement on computed tomography or MRI to define response or no requirement for sustained response. We used a more conservative estimate, requiring both an objective response threshold and a threshold not to be exceeded for the number of early progressors. We recognize that objective response assessment in children with low-grade gliomas is particularly difficult; clinical benefit in this population may also be demonstrated by long-term stability of disease. We incorporated the standard Children's Oncology Group method of tumor measurements, with confirmation by central review, and assessment of the number of early progressors. This study met the preset threshold of interest for both. Notably, nearly 25% of patients (n = 18 of 74) on our study completed 26 courses of study therapy.

The question remains regarding the clinical role of lenalidomide and comparison with current upfront therapy options, which involve primarily cytotoxic agents. Similar to many of these agents, myelosuppression is the main toxicity. Unlike carboplatin compounds, allergic reactions do not appear to be an issue for lenalidomide. Although alkylating agents are associated with an overall increased risk of second neoplasms,^31,32 this is less clear for lenalidomide. Other toxicities associated with standard agents used for pediatric low-grade glioma, including peripheral neuropathy and ototoxicity, were not observed with lenalidomide.

The exact antitumor mechanism of action of lenalidomide against pediatric pilocytic astrocytoma and optic pathway gliomas is not understood, but warrants further study. In view of lenalidomide's ImID properties, it is plausible that this is a primary antitumor mechanism, as a neuroimmune axis and nuclear factor kappa-light-chain-enhancer of activated B cells pathways have been implicated in growth of pediatric low-grade gliomas.^33,34 Pilocytic astrocytomas are highly vascular and display angiogenic profiles^35,36; therefore, lenalidomide's antiangiogenic and anti-inflammatory activity may be contributing.^37,38 The unique pharmacologic and pharmacodynamic properties of lenalidomide as a protein degrader enable targeting of previously undruggable proteins for destruction.³⁹ The role of this potential antitumor mechanism in pediatric low-grade gliomas is intriguing and currently being explored. It is also feasible that lenalidomide incorporates multiple simultaneous mechanisms of attack to interfere with cancer stem-cell stemness, effects on the tumor microenvironment, and direct effects against key functions of tumor cells.⁴⁰

As with other pediatric low-grade glioma studies, our study is limited by the lack of a specific comparable historical cohort. Additional limitations include the lack of tissue obtained from most patients, limiting our understanding of possible biologic explanations for tumor response or lack thereof. We did not stratify for patients with and without neurofibromatosis. At the time of study initiation, we were just beginning to understand pediatric low-grade glioma pathophysiology and routine biopsy of optic pathway gliomas was not established. We have since learned much about pediatric low-grade glioma, including involvement and targeting of the MAPK pathway, with clinical trials inhibiting this pathway underway. How efficacy, tolerability, and safety compare with our trial remain to be determined.

In conclusion, lenalidomide has antitumor activity in children with low-grade gliomas and is well tolerated with similar efficacy at standard doses compared with higher dosing. It may be a treatment option in this population, particularly since low-grade glioma is treated as a chronic disease, lenalidomide has a limited range of acute toxicities, and does not have the known risk of second malignancies associated with alkylator therapies. Further evaluation of its activity and underlying mechanisms of action in this population is warranted to best determine its role in the treatment of these patients.

AUTHOR CONTRIBUTIONS

Conception and design: Katherine E. Warren, Gilbert Vezina, Mark Krailo, Linda Springer, Cody J. Peer, Chris William-Hughes, Amar Gajjar, Daniel Bowers

Provision of study materials or patients: Katherine E. Warren, Sandy Kessel, Amar Gajjar, Daniel Bowers

Collection and assembly of data: Katherine E. Warren, Gilbert Vezina, Mark Krailo, Linda Springer, Allen Buxton, Cody J. Peer, William D. Figg, Sandy Kessel, Maryam Fouladi, Daniel Bowers

Data analysis and interpretation: Katherine E. Warren, Gilbert Vezina, Mark Krailo, Allen Buxton, Cody J. Peer, William D. Figg, Daniel Bowers

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Phase II Randomized Trial of Lenalidomide in Children With Pilocytic Astrocytomas and Optic Pathway Gliomas: A Report From the Children's Oncology Group

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).