Distinguishing early progression (EP) from pseudoprogression (PsP) following radiochemotherapy for glioblastoma is important but remains a significant clinical challenge. EP is more common in MGMT unmethylated (uMGMT) than methylated (mMGMT) cases. We aimed to develop a multiparametric scoring system that integrates advanced imaging techniques (magnetic resonance perfusion [MRP] and spectroscopy [MRS]) with MGMT status and time since radiochemotherapy to aid in the interpretation of anatomic imaging findings. We retrospectively reviewed records of glioblastoma patients diagnosed from 2016-2025 who had worsening contrast enhancement ≤ 6 months after radiochemotherapy and underwent MRS and MRP. Definitive distinction between EP and PsP was made by either histology from surgical resampling or subsequent clinico-radiographic evolution. Optimal cutoff for time since radiochemotherapy was calculated for uMGMT and mMGMT using receiver operating characteristic (ROC) curve; a time-MGMT prediction was determined for each case (PsP if time since radiochemotherapy was below the cutoff for their MGMT status, EP if above). Sensitivity, specificity and accuracy of MGMT status, of each imaging modality, and of a composite score (1 point for each of the following consistent with EP: MRP, MRS, and time-MGMT prediction) were calculated. 78 MRS/MRP instances from 65 patients (26 mMGMT, 39 uMGMT) were included; outcome (EP 49 [63%], PsP 29 [37%]) was determined histologically in 21 (27%) and clinically in 57 (73%) cases. Sensitivity/specificity/accuracy to predict EP were 65%/79%/71% for uMGMT, 82%/41%/67% for MRP, 78%/62%/72% for MRS, and 82%/69%/77% for composite score of 2-3 (vs. 0-1). In 21 cases with discordant MRS and MRP, the time-MGMT prediction correctly identified the outcome in 17 cases (81%). Incorporating clinical parameters—including MGMT status and time since radiochemotherapy—to imaging results improves accuracy in distinguishing EP from PsP, but challenges remain. Ongoing work will incorporate CSF cell-free DNA and nuclear imaging into the model.

Glioblastoma is the most common type of malignant primary brain tumor and a major cause of morbidity and mortality. In 2021, the World Health Organization updated the classification of Central Nervous System (CNS) tumors to restrict glioblastomas to isocitrate dehydrogenase-wildtype (IDHwt) tumors, improving understanding of the prognosis and optimal therapy for these tumors. This revision also enables more homogeneous populations of patients to be enrolled in clinical trials, facilitating the evaluation of novel therapies. In this updated consensus review from the Society for Neuro-Oncology (SNO) and the European Association of Neuro-Oncology (EANO), the current management of patients with glioblastoma is discussed. In addition, novel therapies such as immunotherapies, viral therapies, targeted molecular therapies, theranostics, and antibody-drug conjugates will be reviewed, as well as the current challenges and future directions for research.

Supplementary Figures S1-S6 and figure legends.

GBM AGILE (Glioblastoma Adaptive, Global, Innovative Learning Environment) is a multi-arm, international, seamless Phase 2/3 response adaptive randomization (RAR) platform trial designed to efficiently identify investigational therapies that improve overall survival and confirm efficacious therapies and biomarker signatures to support registration. GBM AGILE is a collaboration among academic investigators, patient organizations, and industry to support new drug applications for newly diagnosed and recurrent glioblastoma. The primary objective of GBM AGILE is to identify therapies that improve overall survival in patients with newly diagnosed or recurrent glioblastoma. Operating under a Master Protocol, GBM AGILE allows multiple drugs from different companies to be evaluated simultaneously and/or over time against a common control. Investigational therapies are added as information about promising drugs is identified, while other therapies are removed as they complete evaluation. RAR is used within subtypes of the disease to assign participants to investigational arms based on their performance. GBM AGILE has screened over 2300 patients and enrollment continues to be robust. In addition to the efficient evaluation of investigational arms, a goal of GBM AGILE is to expand knowledge of glioblastoma to support advancements in treatment using the data collected within the trial (learning environment). Over 7 million data points are currently available for inclusion in the development of a longitudinal model. Such a model may be able to inform randomization by providing earlier and continuous information regarding patient and arm performance. In addition, serial magnetic resonance imaging scans and biospecimens from baseline through patient progression are being collected for further analysis. An initial 500 baseline tissue samples are being characterized by genome sequencing and transcriptome analysis. NCT number: NCT03970447.

ASCO Rapid Recommendation Updates highlight revisions to select ASCO guideline recommendations as a response to the emergence of new and practice-changing data. The rapid updates are supported by an evidence review and follow the guideline development processes outlined in the ASCO Guideline Methodology Manual. The goal of these articles is to disseminate updated recommendations, in a timely manner, to better inform health practitioners and the public on the best available cancer care options. Guidelines and updates are not intended to substitute for independent professional judgment of the treating clinician and do not account for individual variation among patients. See appendix for disclaimers and other important information (Appendix 1 and Appendix 2, online only).

PURPOSE:New therapies for glioblastoma are needed, especially MGMT-unmethylated (uMGMT) disease. NRG Oncology BN002 (phase I) demonstrated safety and suggested efficacy of ipilimumab (ipi) with nivolumab (nivo) in newly diagnosed glioblastoma, leading to this phase II/III trial. METHODS:Adults with newly diagnosed uMGMT glioblastoma and Karnofsky performance status (KPS) ≥70 were randomly assigned to radiotherapy with either immunotherapy (ipi and nivo) or temozolomide (TMZ), stratified by recursive partitioning analysis (RPA) class and intention to use tumor treating fields. With 95% power to detect a hazard ratio (HR) ≤0.58 for progression-free survival (PFS) at a one-sided significance level (P) of .15, superior PFS with immunotherapy in phase II would lead to phase III overall survival (OS) testing. Corticosteroids were disallowed when starting immunotherapy. Diagnosis, biomarkers, and PFS were centrally assessed. RESULTS:One hundred fifty-nine participants were randomly assigned (79 immunotherapy and 80 TMZ). Arms were well balanced for age (median 60 years, range, 28-79), sex (male n = 105, 66%), KPS (90-100 n = 97, 61%), resection extent (gross total, n = 103, 65%), and RPA class (III, n = 16, 10%; IV, n = 116, 73%; V, n = 27, 17%). A preplanned analysis of phase II data conducted after 100 centrally determined PFS events showed no significant PFS improvement for ipi and nivo versus TMZ (median 7.7 months v 8.5 months, HR, 1.47 [70% CI, 1.19 to 1.83]; one-sided P = .96 [95% CI, 0.98 to 2.2]). OS is immature (>50% alive) but with no observed difference between arms (median approximately 13 months each, HR, 0.95 [95% CI, 0.61 to 1.49]; P = .36). CONCLUSION:Ipi and nivo did not improve PFS among patients with newly diagnosed uMGMT glioblastoma versus TMZ. Accrual closed permanently; the trial will not proceed to phase III. No new safety signals were identified. Molecular correlative analyses and survival follow-up are ongoing.

STELLAR was a phase III randomized, open-label trial of eflornithine (ornithine decarboxylase inhibitor) with lomustine versus lomustine alone in patients with recurrent anaplastic astrocytoma (AA). We previously reported an overall survival (OS) and progression free survival (PFS) of the combination in the patients with grade 3 IDH-mutant astrocytoma. Here we report updated OS and PFS outcomes based on additional molecular classification and blinded independent central review (BICR). Eligibility included: age ≥ 18 years, AA (2016 WHO criteria) with first recurrence ≥ 6 months after radiation, KPS ≥ 70, and no imaging findings consistent with grade 4 glioblastoma. Patients received eflornithine (2.8 g/m2 TID Q 2/3 weeks) with lomustine (90 mg/m2 Q 6 weeks) (Arm A) or lomustine alone (110 mg/m2 Q 6 weeks) (Arm B). Primary endpoint was OS. Updated molecular data resulted in re-classification of the 343 AA patients enrolled into 3 WHO 2021 defined tumor subtypes, 196 tumors were IDHmut, CDKN2A/B intact, consistent with diagnosis of astrocytoma, IDHmut, WHO grade 3, 33 tumors had CDKN2A/B loss consistent with astrocytoma, IDHmut, WHO grade 4, 106 tumors were IDHwt consistent with diagnosis of glioblastoma, IDHwt, WHO grade 4. Updated results strengthen initial conclusion that eflornithine plus lomustine resulted in significant survival benefit in the astrocytoma, IDHmut, WHO grade 3 group (mOS 34.9 vs, 23.5 months, HR = 0.64, stratified log rank p =0.0136). PFS with eflornithine was also improved (15.8 vs, 7.2 months, HR = 0.57, stratified log rank p =0.0113). BICR demonstrated a high correlation between investigator and centrally determined PFS (p=0.0305), and high correlation between OS and PFS (p=.0001). The clinically meaningful OS and PFS benefits observed with eflornithine in the molecularly defined 2021 WHO CNS grade 3 astrocytomas is further confirmed and expanded based on additional molecular classification and BICR of PFS.

Older adults (OA) represent a functionally heterogeneous population. While aging increases vulnerability to disease, not all OA are frail. Cancer, particularly glioblastoma (GBM), is more frequent in OA, with a median age of cancer-related death exceeding 70 years in the U.S. However, they are often underrepresented in clinical trials, leading to challenges in optimizing treatment and patient outcomes. OA with cancer face aging-related biological changes and potentially higher risk for treatment-induced toxicities. The Eastern Cooperative Oncology Group (ECOG) Performance Status Scale and the Karnofsky Performance Status (KPS) score, fail to adequately predict treatment tolerance and functional decline inthis population. The Comprehensive Geriatric Assessment (CGA) evaluates multiple domains including nutrition, mobility, multimorbidity, polypharmacy, geriatric syndromes, cognition, mood, and patient-centered goals. However, CGAs are resource-intensive and require trained personnel. The G8 Geriatric Screening Tool offers a practical alternative—requiring only five minutes to administer without specialized training—while effectively identifying patients who may benefit from a CGA. The G8 tool has demonstrated prognostic value in hematologic malignancies, identifying geriatric impairments and serving as an independent predictor of one-year mortality (hazard ratio 3.93; 95% CI 1.67–9.22, p < 0.001). However, it has not yet been validated in primary brain tumors. This study aims to evaluate the G8’s prognostic value in OA with newly diagnosed GBM and compare its performance with KPS and ECOG scores. The primary objective is to determine whether the G8 screening tool is an independent predictor of 1-year mortality when adjusted for well-known prognostic factors such as age, performance status, extent of resection, and MGMT promoter methylation status. The study population consist of patients ≥70 years (n=100) with newly diagnosed glioblastoma, IDH wild-type from Columbia University Medical Center. Validating the G8 in this population could facilitate improved risk stratification and lay the foundation for a tailored assessment tool for primary brain tumors. Data is being analized and would be presented at the meeting.

2003 Background: FGFR genomic alterations occur in approximately 8% of gliomas. Inhibition of FGFR1-3 with pemigatinib showed antitumor activity in a multihistology basket trial (FIGHT-207) in which approximately 10% of participants (pts) had recurrent/progressive FGFR-altered glioblastoma (GBM). We further investigated pemigatinib activity in primary brain tumors by conducting an international, multicenter, single-arm, 2-cohort, phase 2 study specifically in adults with FGFR-altered pretreated gliomas (FIGHT-209; NCT05267106). Methods: Pts were enrolled in 2 cohorts: A, histologically or molecularly defined GBM; or B, other gliomas, glioneuronal tumors, and neuronal tumors. Eligible pts had tumors harboring a FGFR1-3 fusion/rearrangement or mutation detected by an accredited laboratory that had recurred/progressed after ≥1 prior therapy. Pemigatinib (oral, 13.5 mg on days 1-14/21) was intended to continue until progression by Response Assessment in Neuro-Oncology (RANO) criteria determined by an independent review committee (IRC) or unacceptable toxicity. Efficacy of each cohort was evaluated independently. The primary endpoint was objective response rate (ORR; partial plus complete) per RANO (cohort A), with a goal of > 28%. Key secondary and exploratory endpoints were ORR in cohort B, ORR by investigator assessment, progression-free survival (PFS) by IRC, overall survival (OS), safety, neurologic function by Neurologic Assessment in Neuro-Oncology (NANO), and efficacy correlations with diagnosis and specific FGFR-alterations. Results: Between May 2022 and December 2023, 74 pts were enrolled in cohort A and 9 in cohort B. FGFR1-3 fusions/rearrangements were the most common genomic alterations in cohort A (n = 65 [88%]) and in cohort B, FGFR1 mutations (n = 8 [89%]). Pts had a median (range) age of 56 (20-79) years; 60% were male. On September 27, 2024 (data cutoff), 16 pts remained on treatment (cohort A, n = 11 [15%]; cohort B, n = 5 [56%]); 67 discontinued, primarily due to progressive disease (n = 59 [71%]). In cohort A, ORR was 8% (6 partial responses [PR], 0 complete responses [CR]); 21 pts (28%) had stable disease (SD); estimated 6-month PFS rate was 17% (95% CI, 8.7-27.8) and 12-month OS rate 48% (95% CI, 35.6-60.2). In cohort B, the ORR was 22% (1 CR, 1 PR); 3 (33%) SD. Most treatment-emergent adverse events (AEs) were low grade in severity (grade ≥3, 36.1%). Hyperphosphatemia, a class effect of FGFR inhibitors, was the most common AE (75%); 6 pts (7%) required dose reduction and 4 pts (5%) discontinued due to AEs. Conclusions: ORR did not meet the pre-specified goal of > 28% among pts with GBM harboring pemigatinib-sensitizing FGFR alterations. However, durable disease stabilization was observed, notably in pts with CNS tumors other than GBM, and toxicities were manageable. More mature PFS and OS data will be presented with exploratory molecular correlations. Clinical trial information: NCT05267106 .

Abstract BACKGROUND Recurrent anaplastic astrocytoma (AA), a grade 3 brain cancer, remains an unmet medical need. STELLAR was a phase III randomized, open-label trial of eflornithine (ornithine decarboxylase inhibitor) with lomustine versus lomustine alone, originally for patients with recurrent AA. However, as the WHO definition of AA evolved during trial conduct (e.g., excluding IDH wild-type disease), we determined results in both the original ITT population and by revised molecularly defined WHO 2021 diagnoses in pre-planned analyses. METHOD Key eligibility: age ≥ 18, AA (2016 WHO criteria), first recurrence ≥ 6 months after radiation and temozolomide, KPS ≥ 70, no imaging findings consistent with grade 4 glioblastoma. Stratification: IDH, age, resection extent, and geography. Patients were randomized to Arm A: eflornithine (2.8 g/m2 TID Q 2/3 weeks) with lomustine (90 mg/m2 Q 6 weeks) or Arm B: lomustine (110 mg/m2 Q 6 weeks). Primary endpoint was OS. The pre-specified hazard ratio (HR) goal was 0.667 based on projected median OS (mOS) increase to 18 months from 12 months. RESULTS 343 pts (172/171 Arms A/B) were randomized (105 sites, 8 countries). ITT analysis found no difference between arms (mOS 23.4 vs. 20.3 months, HR = 0.94). However, a pre-specified subset analysis in IDH mutant, 2021 WHO-defined grade 3 astrocytoma (n=194), showed significant improvement with eflornithine (mOS 34.9 vs. 23.5 months, HR = 0.64, log rank p = 0.016) and PFS (15.8 vs. 7.2 months, HR = 0.58, log rank p = 0.015). No difference was observed in the IDH wild-type subset. AEs were consistent with historic data. CONCLUSION There was no difference in OS between arms in the ITT analysis. However, medically meaningful and statistically significant OS and PFS benefits were observed with eflornithine in the pre-planned analysis of patients with molecularly defined 2021 WHO CNS grade 3 astrocytoma.

TPS2100 Background: Cytomegalovirus (CMV) antigens have been reported in over 90% of GBMs. CD4+ and CD8+ T cells are most frequently directed against the gB and pp65 antigens, respectively, which are immunogenic targets in a CMV-based GBM immunotherapeutic. A total of 28 first-recurrent GBM subjects were enrolled in a trial, designed as a single arm, 3-dose escalation phase (n=18) followed by a dose-expansion phase (n=10), to receive VBI-1901 (a gB/pp65 enveloped virus-like particle vaccine immunotherapeutic adjuvanted with GM-CSF) given intradermally every 4 weeks until clinical disease progression (NCT03382977). In Phase I (Part A) of the study, 3 vaccine doses (0.4µg, 2µg, and 10µg pp65) were evaluated. No DLTs or safety concerns with any of the doses. The 10µg dose was chosen for the dose-expansion phase. Injection site erythema was the most common adverse event. Vaccination with VB-1901 led to both humoral response with increase in CMV gB antibodies, and cellular response with increase in CD4+ effector memory T cells against CMV pp65 antigen; immunological responses were associated with MRI/clinical response. Among 16 subjects receiving the highest dose of VBI-1901, the disease control rate was 44%, including 2 durable partial responses, which translated into a mOS of 12.9 months. Based on these results, the study was amended into a randomized study. Methods: Approximately 60 adult subjects (18 years or older) with first recurrence of WHO 2016 grade IV Glioblastoma, IDH-wildtype will be randomized at a 1:1 ratio to two open-label cohorts to receive: VBI-1901 or standard of care (SOC) treatment with lomustine or carmustine. The contrast-enhancing lesion may not have an area measuring greater than 600 mm2 (subjects with a resected first recurrence tumor, which may not be measurable after surgery, are eligible). Subjects must have a KPS ≥70, corticosteroid (dexamethasone or equivalent) dosage must be ≤ 2 mg daily that has been stable for at least 5 days before randomization into the study, and only subjects with a CD4/CD8 ratio ≥1 OR a CD4 count of ≥ 400/uL at screening are eligible, as people with a preserved immune system are more likely to respond to immunization with VBI-1901 based on the initial phase study. The primary endpoints are hierarchical. If the safety endpoint is achieved, then efficacy (overall survival) will be considered as a co-primary endpoint. Tumor response rate is a secondary endpoint with planned interim analyses. For a one-sided log rank test with an equal allocation of subjects in the two groups, 80% power and a 0.20 significance level, a total number of 47 events are needed to detect a hazard ratio for mortality of 0.6145 in the VBI-1901 cohort relative to the SOC cohort. This hazard ratio corresponds to a difference in the mOS observed previously with VBI-1901 and a historical mOS of 8 months in the SOC group. As of January 31st 2024 , 12 of 60 patients have been randomized and dosed across 8 US sites. Clinical trial information: NCT03382977 .

Abstract Cytomegalovirus (CMV) antigens have been reported in over 90% of GBMs. CD4+ and CD8+ T cells are most frequently directed against the gB and pp65 antigens, respectively, which are immunogenic targets in a CMV-based GBM immunotherapeutic. First-recurrent GBM patients were enrolled, with Karnofsky Performance Status of at least 70, to receive VBI-1901 (a gB/pp65 enveloped virus-like particle [eVLP]) adjuvanted with GM-CSF and given intradermally (NCT03382977). Patients were vaccinated with VBI-1901 every 4 weeks, with surveillance brain MRI scans every 6 weeks. Among 16 subjects receiving the highest dose of VBI-1901 in the Phase I/IIa portion of the study, the disease control rate was 44%, including 2 durable partial responses, which translated into a mOS of 12.9 months. Based on these results, the study was amended into a randomized study. Approximately 60 adult subjects (18 years or older) with first recurrence of WHO 2016 grade IV Glioblastoma, IDH-wildtype will be randomized at a 1:1 ratio to two open-label cohorts to receive: VBI-1901 or standard of care (SOC) treatment with lomustine or carmustine. Only subjects with a CD4/CD8 ratio ≥1 or a CD4 count of ≥ 400/uL at screening are eligible, as people with a preserved immune system are more likely to respond to immunization with VBI-1901 based on the initial phases of the study. As of May 15, 2024, 23 patients have been randomized across 6 sites. Among evaluable patients (n=13), disease control rates of 43% and 0% have been observed in the VBI-1901 and SOC arms, respectively, including 1 partial response in the VBI-1901 arm. Full enrollment (n=60) is anticipated by the end of the year across 11 sites, and updated tumor response data and select biomarkers associated previously with tumor responses will be presented.

Abstract BACKGROUND Decompressive hemicraniectomy (DHC) is an uncommonly needed treatment in high-grade gliomas (HGG), most used for refractory increased intracranial pressure (ICP) from vasogenic edema, mass effect, or related complications. Outcomes and imaging following DHC are not well described. METHODS We retrospectively identified cases of DHC for HGG at our institution since 2019. RESULTS Three patients with HGG underwent DHC (2 females, ages 39-70) for newly diagnosed (2 patients) or recurrent tumor (1 patient). Causes of increased ICP were tumor/vasogenic edema (2 patients) or abscess (1 patient). Subsequent cranioplasty was performed in 2 patients: at 1.8 months after DHC in one patient, and at 8.6 months in another patient who developed sinking skin flap syndrome (SSFS). SSFS presented as progressive worsening of tumor-related neurologic deficits after completion of radiation therapy with no change in imaging, followed by frank signs of increased ICP 8 months after DHC. Two patients died of disease progression (5.3 months and 9.1 months after DHC), and one patient remains alive at 13.5 months (patient with SSFS; neurologic deficits virtually resolved after cranioplasty repair). Imaging will be described for all patients. CONCLUSION DHC can be a life-saving intervention in HGG. When feasible, early cranioplasty should be considered to avoid complications, including SSFS, which may present with neurological deficits without corresponding imaging change.

Abstract BACKGROUND GBM AGILE (Glioblastoma Adaptive, Global, Innovative Learning Environment) is a biomarker based, multi-arm, international, seamless Phase 2/3 platform trial designed to rapidly identify and confirm experimental therapies that improve overall survival for approval along with their associated biomarker signatures. GBM AGILE is a collaboration between academic investigators, patient organizations, and industry to support new drug applications for newly diagnosed (ND) and recurrent glioblastoma. METHODS The primary objective of GBM AGILE is to identify therapies that improve the overall survival in patients with ND or recurrent glioblastoma. Bayesian response adaptive randomization is used within subtypes of the disease to assign participants to investigational arms based on their performance. GBM AGILE operates under a Master Protocol allowing multiple drugs to be evaluated simultaneously and over time against a common control. New experimental therapies are added as information about promising new drugs is identified, while therapies are removed as they complete evaluation. Six investigational arms have been included to date. Along with an adaptive trial design, shared control arm and operational processes, GBM AGILE continues to incorporate new design and operational elements. One recent design element added to the Master Protocol for consideration for future arms is an extended evaluation period, of up to 11 months, for arms that reach maximum sample size in Stage 1 (Stage 1: learn most responsive signature) without having reached a decision for futility or graduation. Allowing data to mature and additional events to accumulate may increase the ability of the trial to detect efficacy, and if appropriate, for an arm to graduate to Stage 2 (Stage 2: confirm activity in graduating signature). Using improved and flexible processes, GBM AGILE continues to serve as a global trial that supports the efficient and rapid incorporation and evaluation of new experimental therapies for patients with glioblastoma. NCT number: NCT03970447.

Radiographic assessment plays a crucial role in the management of patients with central nervous system (CNS) tumors, aiding in treatment planning and evaluation of therapeutic efficacy by quantifying response. Recently, an updated version of the Response Assessment in Neuro-Oncology (RANO) criteria (RANO 2.0) was developed to improve upon prior criteria and provide an updated, standardized framework for assessing treatment response in clinical trials for gliomas in adults. This article provides an overview of significant updates to the criteria including (1) the use of a unified set of criteria for high and low grade gliomas in adults; (2) the use of the post-radiotherapy MRI scan as the baseline for evaluation in newly diagnosed high-grade gliomas; (3) the option for the trial to mandate a confirmation scan to more reliably distinguish pseudoprogression from tumor progression; (4) the option of using volumetric tumor measurements; and (5) the removal of subjective non-enhancing tumor evaluations in predominantly enhancing gliomas (except for specific therapeutic modalities). Step-by-step pragmatic guidance is hereby provided for the neuroradiologist and imaging core lab involved in operationalization and technical execution of RANO 2.0 in clinical trials, including the display of representative cases and in-depth discussion of challenging scenarios.

Abstract BACKGROUND GBM AGILE(NCT03970447) is a phase 2/3 Bayesian adaptive registration platform trial testing multiple therapies efficiently against a common control (C) with a primary endpoint of overall survival (OS). VAL-083 (VAL) is a DNA targeting agent that, independent of O6-methylguanine DNA methyltransferase promoter methylation status, targets the N7 position of guanine residues and facilitates inter-strand DNA crosslinks, leading to DNA double-strand breaks and cell death. It entered the trial in January 2021, and it is the 2nd arm (of 6) to complete its evaluation. MATERIAL AND METHODS Patient subtypes considered in GBM AGILE are newly diagnosed methylated (NDM), ND unmethylated (NDU), & recurrent disease (RD). C is temozolomide (in ND) & lomustine (in RD). Arms open to all 3 subtypes are evaluated in 5 prospectively defined signatures (sig): NDU, NDM, RD, all ND and All. Randomization (rand.) to C is 20% in each subtype. Exp arms in GBM AGILE have 1 or 2 stages. Efficacy is based on OS hazard ratio (HR) of arm/C. Efficacy goal is a final Bayesian probability ≥ 98% for HR <1.00 in combined Stages 1 & 2. Arms stop accruing in Stage 1 if they reach max sample size (N) or drop for futility or safety. Exp arms in Stage 1 are adaptively rand. with allocation being proportional to an arm’s current probability of having ≥ 30% benefit in OS, P(HR <0.70). In stage 1, exp arms are evaluated monthly, and arms showing Bayesian predictive power (PP) ≥ 0.8, graduate into Stage 2 with fixed rand. in one sig. For all exp arms, follow up continues for 12 mos after accrual stops. Arms are declared futile at any monthly analysis when PP is <0.25 for all sigs. Open to all 3 subtypes, VAL entered as the 1st arm in NDM and was rand. 1:1 to C in this subtype until additional arms entered. The target max N for VAL in its Stages 1 & 2 were 150 and 50, resp. RESULTS At the interim after VAL reached max sample size in Stage 1, the PP for all sigs was <0.8 and >0.25 for at least one sig. Thus, VAL did not graduate nor drop for futility, but accrual stopped for max N in Stage 1 VAL/C sample sizes were 39/92, 28/30, 87/225, 67/122, 154/347 for sigs NDU, NDM, RD, ND, and all. Resp. PPs: 0.086, 0.286, 0.005, 0.148, and 0.014. Resp. mean HRs: 1.36, 1.07, 1.69, 1.18, 1.4. At the interim after which VAL stopped for max N, mean HRs were 1.36, 1.07, 1.69, 1.18, and 1.40 with final probabilities of benefit (HR< 1.00) equal to 0.794, 0.564, 0.345, 0.708, and 0.564. Final results will be presented at the meeting. CONCLUSION GBM AGILE is an efficient & effective model for phase 3 drug development. VAL did not increase OS compared to C in any glioblastoma subtype. GBM AGILE evaluated this agent in less time, at lower cost, & with fewer patients than typical registration trials & is currently evaluating several other arms. © 2024 American Society of Clinical Oncology, Inc. Reused with permission. This abstract was accepted and previously presented at the 2024 ASCO Annual Meeting. All rights reserved.

Abstract BACKGROUND GBM AGILE (NCT03970447;https://www.gcaresearch.org/research/gbm-agile) is a phase 3 Bayesian adaptive platform trial that efficiently tests multiple arms against common control, with 6 arms included to date. Primary endpoint is overall survival (OS). Stage 1 experimental arms are adaptively randomized against other arms. Demonstrated efficacy in stage 1 leads to fixed randomization stage 2. Stages 1 and 2 are combined for registration. Control randomization is fixed. Regorafenib, a multikinase-inhibitor, entered into GBM AGILE as the first arm and therefore was equally randomized against control. Regorafenib showed OS benefit in recurrent disease (RD) in randomized phase 2 REGOMA trial. MATERIAL AND METHODS Patient subtypes in GBM AGILE are newly diagnosed unmethylated (NDU), RD, and—not considered for regorafenib—ND methylated (NDM). Arm indications (signatures) are combinations of subtypes. Control is temozolomide (ND) and lomustine (RD). Efficacy is assessed by OS hazard ratio(HR), arm/control. Efficacy is demonstrated when Bayesian probability of benefit (HR< 1.00) ≥ 98% (roughly analogous P-value: 0.02). Futility occurs at any monthly analysis when Bayesian predictive power (PP) is < 25% for all signatures. Follow-up continues for 12 months after arm’s accrual stops. RESULTS When PP for all 3 pre-defined signatures was < 25%, regorafenib’s accrual was stopped for futility. Regorafenib/control sample sizes were 49/51, 127/128, 176/179 for signatures NDU, RD, and both. Respective PPs: 0.138, 0.030, 0.025—none close to 0.25. Respective mean HRs: 1.26, 1.25, 1.23. Probabilities of benefit (HR< 1.00): 0.35, 0.18, 0.17. At final analysis, mean HRs were 1.07, 1.08, 1.08 with final probabilities of benefit (HR< 1.00) equal to 0.421, 0.312, 0.296—none close to 0.98. CONCLUSION GBM AGILE efficiently and compellingly addressed regorafenib’s role in GBM, in RD and NDU. These findings are germane as they fail to confirm the REGOMA results in RD. GBM AGILE continues to efficiently assess other therapies, including utilizing concurrent and previously accrued controls. This abstract was accepted and previously presented at the 2023 SNO Annual Meeting and published in Neuro-Oncology, Volume 25, Issue Supplement_5, November 2023, Pages v97-v98.

Abstract GBM AGILE(NCT03970447) is a phase 2/3 Bayesian adaptive registration platform trial testing multiple therapies against a common control. Paxalisib, a PI3K/mTOR inhibitor, is the 3rd arm in the trial to conclude evaluation. METHODS Paxalisib was open to patients with newly diagnosed unmethylated (NDU) and recurrent (RD) glioblastoma, with three possible signatures: NDU, RD, and All(NDU+RD). Arm enrollment occurred Dec2020 through May2022. Control patients were enrolled from study initiation (July2019) and were treated with temozolomide(NDU) or lomustine(RD). GBM AGILE investigational arms have 1 or 2 stages, with adaptive randomization in stage 1 and fixed randomization if arms continue to stage 2. Efficacy is based on OS hazard ratio(HR) of Arm/Control. Efficacy goal is final Bayesian probability ≥ 98% for HR<1.00 in combined Stages. An Arm continues to Stage 2 if Bayesian predictive power (PP) ≥ 0.8. An Arm stops accruing in Stage 1 if it reaches maximal sample size(N) or does not meet a minimum efficacy threshold (PP<0.25 for all signatures when N>50). Clinical cut-off is 12 months after accrual stops. The maximum N for paxalisib was approximately 150(Stage 1) and 50(Stage 2). RESULTS After paxalisib reached >150 patients in Stage 1, accrual stopped [Paxalisib/control N 54/75(NDU), 100/188(RD)]. Neither PP thresholds for moving to Stage 2 nor stopping for minimum efficacy were met. Following an unplanned public disclosure that the arm did not continue to stage 2, clinical cut-off for final analysis (planned May2023) was updated to public disclosure date (August2022). At final analysis mean HRs were 0.89(NDU), 1.25(RD), 1.05(All), probabilities of HR<1.00 of 0.72(NDU), 0.076(RD), 0.398(All). Model estimated median OS paxalisib/control (months) were 14.77/13.84(NDU) and 8.58/10.06(RD). CONCLUSION Paxalisib did not show survival benefit over cumulative control in final primary analysis; additional secondary analyses are being considered. GBM AGILE continues to rapidly and efficiently assess therapies in ND and RD glioblastoma.

Abstract BACKGROUND GBM AGILE (Glioblastoma Adaptive, Global, Innovative Learning Environment) is a biomarker based, multi-arm, international, seamless Phase 2/3 platform trial designed to continuously add new arms and to rapidly identify experimental therapies that improve overall survival and confirm efficacious experimental therapies and associated biomarker signatures to support new drug approvals and registration. It is estimated that about 25% of glioblastoma patients that enroll in clinical trials in the US participate in GBM AGILE. Given the evergreen nature of GBM AGILE, with arms entering and exiting over time, combined with the high enrollment rate, there is commitment to providing regular updates on the advancement of the trial’s progress to the broader glioblastoma community. MATERIAL AND METHODS GBM AGILE (Trial sponsor: Global Coalition for Adaptive Research, GCAReseach.org) is a collaboration between academic investigators, patient organizations, and industry to support new drug applications for newly diagnosed (ND) and recurrent glioblastoma. The primary objective of GBM AGILE is to identify therapies that improve the overall survival in patients with ND or recurrent glioblastoma. Bayesian response adaptive randomization is used within subtypes of the disease to assign participants to investigational arms based on their performance. GBM AGILE operates under a Master Protocol, which allows multiple drugs from different companies to be evaluated simultaneously and/or over time against a common control. New investigational arms are added as potential promising new drugs are identified, while therapies are removed as they complete their evaluation. Six investigational arms have been included to date since 2019, with arm 6 entering in 2023. Evaluation of the first three arms in the trial has been completed. During 2023, GBM AGILE continued its global expansion and opened in Australia and Germany. As of April 2024, there are 41 active sites in the US, 4 active sites in Canada, 3 active sites in France, 5 active sites in Germany, 2 active sites in Switzerland, and 4 active sites in Australia, with additional sites in start-up. Along with a global reach, adaptive trial design, shared control arm and operational processes to serve the goal of optimizing patient care, GBM AGILE also incorporates design and operational elements to enhance efficiencies, cost-savings, and allow for arm-specific customizations. Regional arm activation is one such customization. Although the goal is for all agents in the trial to be available in all regions, arms may activate only in certain regions based on their clinical development plan. Through the use of improved and flexible processes, GBM AGILE continues to serve as a global trial that supports the efficient and rapid incorporation and evaluation of new experimental therapies for patients with GBM. Clinical trial information: NCT03970447.