Results of a Phase I Trial of Dose Escalation for Preoperative Stereotactic Radiosurgery for Patients with Large Brain Metastases

PURPOSE/OBJECTIVE(S)

Single session stereotactic radiosurgery (SRS) alone for brain metastases larger than 2 cm in diameter results in unsatisfactory local control. Surgical resection alone also produces unreliable local control and perioperative radiation is required. We conducted a prospective phase I trial (NCT01891318) for brain metastases greater than 2 cm to determine the safety of preoperative SRS at escalating doses followed by surgical resection.

MATERIALS/METHODS

Radiosurgery dose started at RTOG 9005 dose levels for the 3 cohorts based on maximum tumor diameter of the index lesion: 18 Gy for >2-3 cm, 15 Gy for >3-4 cm, and 12 Gy >4-6 cm. Concurrent SRS alone to other smaller lesions was allowed using standard RTOG dose. Dose limiting toxicity (DLT) was defined as grade 3 or greater acute toxicity within 3 to 4 months after SRS. Patients underwent surgical resection within 2 weeks and were followed with imaging and neurological evaluations every 3 months.

RESULTS

A total of 35 patients were enrolled into the trial (see Table 1 below). The median age was 63, and median interval between SRS and surgery was 2 days. The most common histology was non-small cell lung cancer (57.1%), followed by breast cancer (14.3%). For tumor size >2-3 cm, patients were enrolled up to the 2nd dose level (21 Gy); for >3-4 cm and >4-6 cm cohorts the 3rd dose level (21 Gy and 18 Gy, respectively) was reached. There was a total of 3 DLTs: 2 in the >3-4 cm cohort and 1 in the >4-6 cm cohort (Table 1). The maximum tolerable dose (MTD) was 18 Gy (2nd dose level) for >3-4 cm, and 18 Gy (3rd dose level) for >4-6 cm. With a median follow-up of 64 months, the 6- and 12-month local control rates were 88.8% and 79.1%, respectively. The 6- and 12-month distant brain control was 63.1% and 55.3%, respectively. Overall survival at 6 and 12 months was 82.9% and 59.0%. The rate of leptomeningeal disease (LMD) at 2 years was 0%.

CONCLUSION

Preoperative SRS with dose escalation followed by surgical resection for brain metastases greater than 2 cm in size results in local control comparable to postoperative SRS or whole-brain radiation therapy and demonstrates acceptable acute toxicity. The Phase II portion of the trial will be conducted at the maximum tolerated SRS doses.

The impact of opioid administration for post gamma knife radiosurgery frame removal: a prospective quality-improvement study

PURPOSE

In our center, five Gamma Knife proceduralists differed in opioid administration practices prior to Leksell frame removal, providing the opportunity to improve the care of patients with brain metastases by studying whether opioid medications improve pain scores and patient satisfaction during Gamma Knife treatment in a prospective, pseudorandomized fashion.

METHODS

We prospectively administered a questionnaire to patients undergoing Gamma Knife Radiosurgery for metastases between November, 2017 and July, 2018. Using multivariable methods, we assessed whether opioid pain medication administration influenced the change in pain scores after frame removal, and whether they influenced patient satisfaction on how often their pain was controlled, and their overall satisfaction.

RESULTS

We included 142 patients. Mean age was 65.2 ± 10.8 years and 52.7% were female. Morphine was the most commonly administered medication. Pain increases were greater around frame removal than placement. Opioids were not associated with any difference in the change in pain scores before and after frame removal, or patient satisfaction. Patients with higher pre-removal pain scores had smaller increases in pain scores after removal; they also had worse pain control and overall satisfaction with their treatment.

CONCLUSION

Morphine administration prior to frame removal did not improve pain scores or pain control satisfaction. Absence of efficacy may be related to delayed onset of action, and stronger and faster-acting agents should be explored. Pre-removal pain scores were associated with decreased pain control and overall satisfaction, further identifying earlier and stronger pain treatment as a potential area for improvement.

Targeting the IL4 receptor with MDNA55 in patients with recurrent glioblastoma: Results of a phase IIb trial

BACKGROUND

MDNA55 is an interleukin 4 receptor (IL4R)-targeting toxin in development for recurrent GBM, a universally fatal disease. IL4R is overexpressed in GBM as well as cells of the tumor microenvironment. High expression of IL4R is associated with poor clinical outcomes.

METHODS

MDNA55-05 is an open-label, single-arm phase IIb study of MDNA55 in recurrent GBM (rGBM) patients with an aggressive form of GBM (de novo GBM, IDH wild-type, and nonresectable at recurrence) on their 1st or 2nd recurrence. MDNA55 was administered intratumorally as a single dose treatment (dose range of 18 to 240 ug) using convection-enhanced delivery (CED) with up to 4 stereo-tactically placed catheters. It was co-infused with a contrast agent (Gd-DTPA, Magnevist®) to assess distribution in and around the tumor margins. The flow rate of each catheter did not exceed 10μL/min to ensure that the infusion duration did not exceed 48 h. The primary endpoint was mOS, with secondary endpoints determining the effects of IL4R status on mOS and PFS.

RESULTS

MDNA55 showed an acceptable safety profile at doses up to 240 μg. In all evaluable patients (n = 44) mOS was 11.64 months (80% one-sided CI 8.62, 15.02) and OS-12 was 46%. A subgroup (n = 32) consisting of IL4R High and IL4R Low patients treated with high-dose MDNA55 (>180 ug) showed the best benefit with mOS of 15 months, OS-12 of 55%. Based on mRANO criteria, tumor control was observed in 81% (26/32), including those patients who exhibited pseudo-progression (15/26).

CONCLUSIONS

MDNA55 demonstrated tumor control and promising survival and may benefit rGBM patients when treated at high-dose irrespective of IL4R expression level.Trial Registration: Clinicaltrials.gov NCT02858895.

CTIM-01. TITLE: PHASE 1B STUDY OF AVELUMAB AND WHOLE BRAIN RADIOTHERAPY (WBRT) IN PATIENTS WITH LEPTOMENINGEAL DISEASE (LMD): PRELIMINARY RESULTS

BACKGROUND

LMD from systemic cancer has a dismal prognosis with median survivals of 8-10 weeks. A phase 2 trial of PD-1 inhibitor monotherapy in LMD showed median overall survival (OS) of 3.6 months (Brastianos P et al., 2020). We determined the safety/efficacy of avelumab, a PD-L1 inhibitor with WBRT in patients with LMD (NCT0371768). This combination can treat the tumor directly and increase BBB permeability (Li, 2003; Nordal, 2005) allowing the increased egress of activated T cells into the meninges/CSF.

METHODS

Patients received concurrent avelumab 800 mg IV q2 weeks for ≤ 5 cycles (unless PD or unacceptable toxicity) with WBRT 3000 cGy in 10 fractions. Primary endpoints are safety/DLTs and OS at 3 months. Secondary endpoints are CSF T-cell/cytokine profiles (scRNAseq, phosophoproteomics etc.).

RESULTS

A total of 15 patients (7 breast, 7 lung & 1 other) were enrolled (n = 13 F, ages 32-79). Pts receiving anti-PD-1/PD-1L/PD-L2/CD137,CTLA-4 therapy ≤ 6 months prior were excluded. Three of 15 patients had grade 3/4 AEs (diarrhea, lymphopenia, decreased WBC count in 3 patients). Seven patients (50%) were alive at 3 or 6 months. The estimated median follow up in 14 patients is 4.75 months (range, 0.92 – 30.05 months, 95% CI is 1.32 ~ 19.82). The median PFS is 3.75 months (95% CI = 0.85-15.16) and median OS is 6.89 months (95% CI = 1.18-14.7).

CONCLUSIONS

The combination of avelumab and WBRT is safe, well tolerated, and demonstrates encouraging activity in patients with LMD with an OS that is longer than other published series. Multiple platform interrogation of CSF (analysis underway) will determine mechanisms of LMD therapeutic/resistance effects.

CTNI-48. TRIAL UPDATE -RESULTS OF A PHASE I WINDOW OF OPPORTUNITY TRIAL EVALUATING A FIRST-IN-CLASS CD29 INHIBITOR, OS2966, IN RECURRENT HIGH-GRADE GLIOMA

BACKGROUND

OS2966 is a first-in-class, humanized, and deimmunized anti-CD29 (β1integrin) monoclonal antibody. CD29 is a mechanosignaling receptor prominently upregulated in glioblastoma. Preclinical studies in mice and non-human primates have demonstrated safety and encouraging efficacy. A phase I clinical trial was therefore initiated to evaluate the safety and feasibility of delivering OS2966 directly to the site of disease via convection-enhanced delivery (CED) in recurrent high-grade glioma patients.

METHODS

This study employs a 2-part design: In part 1, patients undergo stereotactic tumor biopsy followed by placement of a CED catheter(s) for delivery of OS2966 to the bulk contrast-enhancing tumor. Subsequently, in part 2, patients undergo a clinically-indicated tumor resection followed by placement of CED catheter(s) and delivery of OS2966 to the surrounding tumor-infiltrated brain. Co-convection of gadolinium enables real-time monitoring of therapeutic delivery. A concentration-based accelerated titration design is utilized for dose escalation. Given availability of pre and post-infusion samples, pharmacodynamic data will be analyzed to explore mechanism of action of OS2966.

RESULTS

Four patients have completed treatment in both study parts. No dose-limiting toxicity or adverse events related to treatment with OS2966 or CED catheter placement have been reported at the first 3 dose levels (ie, doses up to 0.6 mg/mL). For the four patients treated an average of 4.2 mL and 3.2 mL were infused in study parts 1 and 2 respectively. The Vd/Vi ratio ranged from 0.6 to 1.6 in part 1, and from 2.0 to 4.3 in Study Part 2. Analysis of tissue samples collected during the trial has demonstrated decreased VEGF expression post-treatment, with preinfusion samples showing > 50% VEGF expression and post-infusion samples showing 10% expression. Additional pharmacodynamic analysis via tissue-level biomarkers is ongoing and will be presented.

CONCLUSION

Initial data demonstrate the safety and feasibility of direct intracranial delivery of OS2966.

NIMG-42. DURABLE OVERALL RESPONSE RATE (ORR) TARGETS FOR RECURRENT GLIOBLASTOMA (RGBM) CLINICAL TRIALS BASED ON THE HISTORIC ASSOCIATION BETWEEN ORR AND MEDIAN OVERALL SURVIVAL (MOS)

Durable overall response rate (ORR) remains a meaningful endpoint in recurrent GBM; however, the target ORR for single arm recurrent GBM trials has not been based on historic information or tied to patient outcomes. The current study reviewed past trials in recurrent GBM in order to judiciously define target ORRs for use in recurrent GBM trials. After reviewing 69 treatment arms comprising 4,971 patients, ORR was 6.1%[95% CI 4.23; 8.76%] for cytotoxic chemotherapies (ORR=7.59% for CCNU, 7.57% for TMZ, 0.64% for CPT-11, and 5.32% for other agents), 3.37% for biologic agents, 7.97% for immunotherapies (select trials), and 26.8% for anti-angiogenic agents. ORRs were significantly correlated with median overall survival (mOS) across chemotherapy (R2=0.4078, P< 0.0001), biologics (R2=0.4003, P=0.0003), and immunotherapy trials (R2=0.8994, P< 0.0001), but not anti-angiogenic agents (R2=0, P=0.8937). Pooling data from chemotherapy, biologics, and immunotherapy trials, a meta-analysis indicated a strong correlation between ORR and mOS (R2=0.3164, P< 0.0001; mOS[weeks]=0.6xORR+28.9), suggesting an ORR >20% results in an mOS of > 40.9 weeks, which is double the survival estimate of a treatment with ORR=0% and ≥ 2 months longer than treatments with ORR=5%. Assuming an ineffective therapy (control) has ORR=5%, a trial of 32 patients with a target ORR=20% leads to the 95% confidence interval higher than the control group. We conclude that single arm phase II studies in recurrent GBM with ≥ 32 patients should have a target ORR ≥ 20%. This was associated with a median OS of approximately 1 year. Importantly, durability of response should also be considered and was not assessed in the current meta-analysis.

TMIC-53. DEVELOPMENT OF A GEO-TAGGED TUMOR SAMPLE REGISTRY; LINKAGE OF TUMOR SAMPLE LOCATION TO IMAGING CHARACTERISTICS

Background

Significant intra- and inter-heterogeneity exists in gliomas. This provides clinical, radiological, diagnostic, and treatment challenges. To date, there have been few efforts to comprehensively catalog information obtained in the operating room that spatially links neuro-navigation localization to imaging characteristics, gross intraoperative visual findings, and pathological/molecular features. The value of this spatial localization is probably greatest in high grade gliomas, which have been demonstrated to have intra-tumoral histological and genetic/epigenetic heterogeneity.

METHODS

An IRB-approved institutional registry of patients undergoing clinically-indicated surgery with use of an image-guidance system (IGS) was launched in November 2019 and as of June 1, 2022 includes nearly 500 patients, of which 243 were diagnosed with gliomas. Intraoperatively, locations within the gross tumor or tumor-infiltrated brain were sampled at each surgeon’s discretion, and each sample was linked to their precise location with the IGS system (“geo-tagged”). The registry includes information regarding surgeon; anesthesia technique; use of intraoperative tumor fluorescence; tumor location and volume; pathologic diagnosis and molecular features, and sample number.

RESULTS

Of 243 gliomas, 26 were low grade and 217 were high grade with 174 being glioblastoma. For enhancing tumors, volume of enhancement ranged from 0.31 to 127.0 cm3 with an average of 22.9 cm3. Tumors were widely distributed throughout the cerebrum with 133 left-sided tumors, 110 right-sided and 32 spanning multiple lobes or deep subcortical structures including the brainstem. 51% of surgeries were under awake anesthesia, and 40% were performed using fluorescence guidance. The average number of navigation image-linked samples collected per tumor was 3.67; 3.48 in low grade gliomas, and 3.69 in high grade gliomas. Samples are archived in frozen and/or formalin-fixed, paraffin-embedded formats for future research.

CONCLUSION

This registry provides the foundation for correlation of imaging, intraoperative findings, and pathology in brain tumors, and it will support detailed laboratory/translational investigations addressing tumor heterogeneity.

IMMU-17. LEVERAGING CORE EPIGENETIC PROGRAMMING OF IMMUNOSUPPRESSIVE MYELOID CELLS FOR THERAPEUTIC TARGETING OF GLIOBLASTOMA

Accumulation of various immunosuppressive myeloid cells, including myeloid-derived suppressor cells (MDSCs), facilitates progression and treatment resistance of glioblastoma (GBM). We previous demonstrated that monocytic MDSCs (mMDSCs) but not granulocytic MDSCs (gMDSCs) infiltrate tumors in mouse models and patients, and that nonspecific targeting of mMDSCs with chemotherapies provided therapeutic benefit in preclinical models of GBM. To investigate the differential role and trafficking of mMDSCs versus gMDSCs in GBM, we adoptively transferred these cells into tumor-bearing mice. Mice that received mMDSCs succumbed to disease at an earlier point compared to control mice or mice that received gMDSCs. To delineate the cellular basis of this distinct behavior of MDSC subsets, we performed assay for transposase-accessible chromatin using sequencing (ATAC-seq) and observed that cell adhesion-related genes were significantly enriched in open chromatin regions of mMDSCs as opposed to gMDSCs. Aligned with this epigenetic profile, mMDSCs from blood and tumors had significantly higher surface integrin β1 expression compared to gMDSCs in both mouse models and GBM patients. To evaluate the functional role of these integrins, we pre-treated mMDSCs with anti-integrin β1 prior to adoptive transfer. Blockade of surface integrin β1 interfered with the pro-tumorigenic role of mMDSCs, as the survival span of mice receiving these cells was similar to that of vehicle controls. Further analysis of the ATAC-Seq data revealed that dipeptidyl peptidase-4 (Dpp4), an interacting partner of integrin β1, was more accessible in mMDSCs compared to gMDSC. Consistently, bone marrow-derived mMDSCs in mouse and tumor-infiltrating mMDSCs in GBM patients expressed high levels of Dpp4. Pharmacological inhibition of Dpp4 reduced chemotaxis of mMDSCs in vitro and extended survival duration of mice with tumors. The findings from this study have broad implications across cancer types to modulate immunosuppressive myeloid cells by leveraging these novel insights into their adhesion mechanisms.

RTID-01. ONC108: A RANDOMIZED PHASE 3 STUDY OF ONC201 IN PATIENTS WITH NEWLY DIAGNOSED H3 K27M-MUTANT DIFFUSE GLIOMA

BACKGROUND

H3 K27M-mutant diffuse midline glioma is a universally fatal malignancy primarily affecting children and young adults; while radiotherapy (RT) provides transient benefit, no effective systemic therapy is currently available. ONC201, a first-in-class imipridone, is an oral, blood-brain barrier penetrating, selective small molecule antagonist of dopamine receptor D2/3 (DRD2) and agonist of the mitochondrial protease ClpP. An integrated pooled analysis of objective response in ONC201-treated patients enrolled in one of five open-label trials has previously demonstrated efficacy in patients with recurrent disease. This phase 3 trial will be the first randomized, controlled study evaluating ONC201 in patients with H3 K27M-mutant disease.

METHODS

ONC108 is a randomized, double-blind, placebo-controlled, parallel-group, international Phase 3 study of ONC201 in patients with newly diagnosed H3 K27M-mutant diffuse glioma. Patients will be randomized to receive once-weekly ONC201 or placebo following standard frontline radiotherapy. Primary efficacy endpoints are overall survival (OS) and progression-free survival (PFS) in all participants; PFS will be assessed with the response assessment in neuro-oncology-high grade glioma by blind independent central review. Other objectives include assessments of safety, additional efficacy endpoints, clinical benefit, quality of life, pharmacokinetics, biomarkers, and healthcare resource utilization. Eligible patients will have histologically confirmed H3 K27M-mutant diffuse glioma, a Karnofsky/Lanksy performance status ≥ 70, and completed first-line radiotherapy. Eligibility will not be restricted based on age; however, patients must be ≥ 10 kg at time of randomization. Patients with a primary spinal tumor, diffuse intrinsic pontine glioma, leptomeningeal disease, or cerebrospinal fluid dissemination are not eligible.

RADT-11. CLINICAL OUTCOMES IN THE MANAGEMENT OF MELANOMA BRAIN METASTASES TREATED WITH STEREOTACTIC RADIOSURGERY AND ANTI-PD-1+CTLA-4

BACKGROUND

Anti-PD-1+CTLA-4 therapy has revolutionized melanoma brain metastases (MBM) treatment. Prospective trials show higher response in asymptomatic versus symptomatic patients. We evaluated clinical outcomes in MBM treated with stereotactic radiosurgery (SRS) and anti-PD-1+CTLA-4.

METHODS

Patients were included if MBM were diagnosed and treated with SRS within 3 months of anti-PD-1+CTLA-4, and this was their last course of systemic treatment. Endpoints of this study were distant MBM control, MBM local control (LC) defined as less than 20% volume increase on follow-up MRI, and overall survival (OS) from SRS. Adverse advents were evaluated using Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.

RESULTS

29 patients with 122 MBM treated over 40 SRS sessions between 2015-2020 were identified. Median SRS dose: 24 Gy (range: 15-24). Median MBM diameter: 0.6 cm (range: 0.3-2.9). Median follow-up using reverse Kaplan-Meier (KM) method: 19.3 months (interquartile range: 14.6-38.4).Six-, twelve-, and eighteen-month KM distant MBM control rates were 51%, 42%, 42%, respectively. LC rates: 90%, 86%, 85%. OS rates: 76%, 68%, 56%. 17 patients (59%) were asymptomatic and 12 (41%) symptomatic. KM distant MBM control and OS for asymptomatic and symptomatic patients were not significant; p=0.61 and p=0.67, respectively.On univariate analysis (UVA), Diagnosis-Specific Graded Prognostic Assessment (DS-GPA) 0-1 was associated with increased risk of distant MBM failure versus DS-GPA 3.5-4 (hazard ratio (HR): 9.8, 95% confidence interval (CI) 1.9-51.5, p=0.007). UVA showed decreased OS with increased number of organs with metastases at diagnosis (HR:12, 95% CI 2.0-83.1, p=0.0075).2 lesions (1.6%) developed symptomatic radiation necrosis requiring steroids; 10 lesions developed grade 3 edema (8%); 13 patients experienced grade 1-2 fatigue and/or headache (45%); no patients experienced grade >3 events.

CONCLUSION

Combination SRS and anti-PD-1+CTLA-4 in MBM shows durable intracranial control with similar outcomes between asymptomatic and symptomatic patients with acceptable toxicity. Further study is warranted.

SURG-19. PROGNOSTIC VALIDATION OF A NEW CLASSIFICATION SYSTEM FOR EXTENT OF RESECTION IN GLIOBLASTOMA: A REPORT OF THE RANO RESECT GROUP

BACKGROUND

Terminology to describe extent of resection in glioblastoma is inconsistent across clinical trials. A surgical classification system was previously proposed based upon residual contrast-enhancing (CE) tumor. We aimed to (I) explore the prognostic utility of the classification system and (II) define how much removed non-CE tumor translates into a survival benefit.

METHODS

The international RANO resect group retrospectively searched the databases from seven neuro-oncological centers in the USA and Europe for patients with newly diagnosed glioblastoma per WHO 2021 classification. Clinical and volumetric information from pre- and post-operative MRI were collected.

RESULTS

We collected 1021 patients with newly diagnosed glioblastoma, including 1008 IDHwt patients. 744 IDHwt glioblastomas were treated with radiochemotherapy per EORTC 26981/22981 (TMZ/RT→TMZ) following surgery. Among such homogenously treated patients, lower absolute residual tumor volumes (in cm3) were favorably associated with outcome: patients with ‘maximal CE resection’ (class 2) had superior outcome compared to patients with ‘submaximal CE resection’ (class 3) or ‘biopsy’ (class 4) (median OS: 19 versus 15 versus 10 months; p=0.001). Extensive resection of non-CE tumor (≤ 5 cm3 residual non-CE tumor) provided an additional survival benefit in patients with complete CE resection, thus defining class 1 (‘supramaximal CE resection’) (median OS: 24 versus 19 months; p=0.008). The prognostic value of the resection classes was retained on multivariate analysis when adjusting for molecular and clinical markers including MGMT promotor status. Relative tumor reduction (in percentage) was not prognostic for outcome on multivariate analysis, and inter-rater agreement for CE and non-CE tumor on post-operative MRI was sufficient.

CONCLUSION

The proposed “RANO categories for extent of resection in glioblastoma” are highly prognostic and may serve for stratification of clinical trials. Removal of non-CE tumor beyond the CE tumor borders translates into additional survival benefit, providing a rationale to explicitly denominate such a ‘supramaximal CE resection’.

A preclinical model of patient-derived cerebrospinal fluid circulating tumor cells for experimental therapeutics in leptomeningeal disease from melanoma

BACKGROUND

Leptomeningeal disease (LMD) occurs as a late complication of several human cancers and has no rationally designed treatment options. A major barrier to developing effective therapies for LMD is the lack of cell-based or preclinical models that recapitulate human disease. Here, we describe the development of in vitro and in vivo cultures of patient-derived cerebrospinal fluid circulating tumor cells (PD-CSF-CTCs) from patients with melanoma as a preclinical model to identify exploitable vulnerabilities in melanoma LMD.

METHODS

CSF-CTCs were collected from melanoma patients with melanoma-derived LMD and cultured ex vivo using human meningeal cell-conditioned media. Using immunoassays and RNA-sequencing analyses of PD-CSF-CTCs, molecular signaling pathways were examined and new therapeutic targets were tested for efficacy in PD-CSF-CTCs preclinical models.

RESULTS

PD-CSF-CTCs were successfully established both in vitro and in vivo. Global RNA analyses of PD-CSF-CTCs revealed several therapeutically tractable targets. These studies complimented our prior proteomic studies highlighting IGF1 signaling as a potential target in LMD. As a proof of concept, combining treatment of ceritinib and trametinib in vitro and in vivo demonstrated synergistic antitumor activity in PD-CSF-CTCs and BRAF inhibitor-resistant melanoma cells.

CONCLUSIONS

This study demonstrates that CSF-CTCs can be grown in vitro and in vivo from some melanoma patients with LMD and used as preclinical models. These models retained melanoma expression patterns and had signaling pathways that are therapeutically targetable. These novel models/reagents may be useful in developing rationally designed treatments for LMD.

BIOM-02. LEPTOMENINGEAL DISEASE SECONDARY TO MELANOMA: UPDATES ON THE VALIDITY OF THE VERIDEX CELLSEARCH SYSTEM

BACKGROUND

Leptomeningeal disease (LMD) is devastating with a median survival of 8-10 weeks without treatment. LMD affects approximately 5% to 25% of melanoma patients. Its pathophysiology remains unknown and effective treatments are virtually non-existent. The primary aim of this study was to evaluate the validity of Veridex CellSearch® System (VCS) compared to Gold Standard test (i.e., CSF cytology).

MATERIALS AND METHODS

A retrospective chart review was performed of subjects with suspected LMD from melanoma enrolled in the MCC 19332/19648 at Moffitt Cancer Center. Patients underwent standard of care with different treatments as deemed appropriate by treating physician. CSF samples were obtained from lumbar punctures, surgeries, and Ommaya reservoir. CSF was evaluated for quantification of CSF circulating tumor cells (CTCs) with the Veridex CellSearch® System (VCS).

RESULTS

Forty-eight patients were identified with melanoma as primary tumor, ages 29-80. Twenty-seven had LMD (median age 62) with median KPS 70. N=19 (70%) were diagnosed radiographically and n=5 (19%) with CSF cytology; n=14 (54%) had positive cytology on first LP. From 24 patients with LMD who underwent VCS, n=22 (92% patients had positive CSF CTCs. Number of CTCs/mL CSF was significantly higher in patients with LMD versus in patients without LMD (mean SD 227.6 vs. 0.07, p < 0.001). VCS sensitivity and specificity was analyzed. AUC was 0.515, with TPR 0.250 and FPR 0.286. CSF analysis and treatments were described. The median survival of those with LMD was 2.7 months.

CONCLUSION

These results indicate the potential value of the VCS as an additional tool to the gold standard in the diagnosis of LMD in patients with high suspicion of the disease. Future directions involve doing prospective studies to further validate this method, and to better understand this patient population to enhance diagnostic tools and management of LMD.

CTNI-16. TROTABRESIB (CC-90010, BMS-986378), A REVERSIBLE, POTENT ORAL BROMODOMAIN AND EXTRATERMINAL INHIBITOR (BETi) IN PATIENTS WITH HIGH-GRADE GLIOMAS: A PHASE 1 OPEN-LABEL ‘WINDOW OF OPPORTUNITY’ STUDY

Trotabresib is a potent, reversible oral BETi with antitumor activity in patients with advanced malignancies (Moreno et al. ESMO 2020. Abstract 5270). The CC-90010-GBM-001 study (NCT04047303) enrolled patients with progressive or recurrent astrocytoma or recurrent glioblastoma scheduled for salvage resection. Patients were treated with trotabresib 30 mg daily for 4 days before surgery, then trotabresib 45 mg daily 4 days on/24 days off after recovery. Primary objectives were trotabresib tumor tissue concentration and plasma pharmacokinetics (PK). Secondary and exploratory objectives included safety, antitumor activity, cerebrospinal fluid concentration, and pharmacodynamics (PD). Twenty patients were enrolled; blood PK, blood PD, and tumor PD data were available for 14, 12, and 11 patients, respectively. Geometric mean peak trotabresib plasma concentration on day 4 was 1.92 μM; median time to peak concentration was 1.5 hours. At the time of resection, geometric mean trotabresib concentrations in plasma and brain tumor tissue were 1.01 and 0.68 μM, respectively. Blood CCR1 mRNA was reduced ≥ 50% from baseline after dose 4. Blood HEXIM1 mRNA increased at 72–96 hours following first dose, and at the time of surgery the percentage increase was related to plasma trotabresib concentration. Tumor HEXIM1 RNA increased in 10 of 11 patients. Eighteen patients (90%) had ≥ 1 treatment-related adverse event (TRAE). Nine patients (45%) had grade 3/4 TRAEs, most frequently thrombocytopenia (5 patients [25%]). Only 1 patient had serious TRAEs (hemiparesis and lethargy). Two patients died of intracranial hemorrhage unrelated to study drug. Of 16 patients evaluable for antitumor response, 7 had stable disease per RANO criteria, with 3 ongoing beyond data cutoff at cycles 4–11. Median progression-free survival was 1.9 months (95% CI, 1.4–3.3). Overall, trotabresib showed good tumor tissue penetration, with PD signals of response, and was well tolerated. A study of trotabresib + temozolomide in first-line glioblastoma is ongoing (NCT04324840).

NCMP-03. ANATOMIC AND SURGICAL FACTORS PREDICT DEVELOPMENT OF LEPTOMENINGEAL DISEASE IN PATIENTS WITH METASTATIC MELANOMA

Leptomeningeal disease (LMD) is a devastating complication of systemic malignancy, portending a poor prognosis with an estimated median survival of 4-6 weeks if left untreated. Several reports have suggested surgical resection as a potential causative factor. Herein, we explore if surgical and anatomical factors are correlated with development of LMD in patients with melanoma brain metastases. METHODS: Patients treated at our institution between 1999-2019 for primary melanoma with brain metastasis were compiled into a database based on ICD9/10 coding. 1,079 patients with melanoma brain metastases and appropriate imaging were identified, and 834 patients with a minimum of 3 months’ follow up were included. Patients were dichotomized by development of LMD or lack thereof, and categorized into an overall cohort, and surgical and non-surgical cohorts. Anatomic factors and ventricular access during surgery were investigated as possible correlative factors for the development of LMD. RESULTS: In the overall cohort, female gender(p=0.033), presence of dural metastasis(p=0.018), presence of periventricular lesions(p< .001), presence of intraventricular lesions(p< .001), and ventricular access during surgery(p< .001) were significantly associated with LMD. Patients undergoing surgery, or those undergoing surgery without ventricular access, were not at higher risk of LMD. On multivariate analysis, female gender(p=.033), presence of periventricular lesions (p< .001), presence of intraventricular lesions(p< .002), and presence of dural metastasis(p=0.032) were significantly associated with development of LMD. In patients who had surgery, iatrogenic ventricular access(p< .001) was significantly correlated with LMD. In the group of patients without surgery, those with periventricular lesions had significantly higher odds of LMD(p< .001). CONCLUSIONS: In a retrospective cohort of patients with melanoma metastatic to the brain, surgical intervention does not increase odds of LMD; however, iatrogenic access to the CSF space during surgery is highly correlated with LMD development. Anatomic contact with the CSF space predicts LMD regardless of surgical status.

NCMP-21. IMMUNE-RELATED ACUTE MOTOR AXONAL NEUROPATHY: A SMALL CASE SERIES AND REVIEW OF THE LITERATURE

BACKGROUND

Immunotherapy revolutionized cancer treatment in the past decade, with a significant increased survival in patients with solid tumors. However, immune checkpoint inhibitors (ICIs) have been associated with a growing number of neurotoxicities, some of which can be fatal if not recognized and treated promptly. Some of these neurotoxicities include very uncommon syndromes like Acute Motor Axonal Neuropathy (AMAN). Herein we present four oncological cases of patients who underwent immunotherapy and developed AMAN.

METHODS

Four patients were diagnosed with immune-related AMAN between 2017 and 2000 at H. Lee Moffitt Cancer Center. Patients were treated with standard of care.

RESULTS

We describe four oncological patients who developed a motor axonal neuropathy (i.e., AMAN) confirmed on nerve conduction studies following 2 cycles of immunotherapy, including a 28 year old woman with melanoma brain metastasis and a 50 year old woman with renal cell carcinoma both treated with ipilimumab and nivolumab, a 32 year old man with Hodgkin lymphoma who was treated with nivolumab and brentuximab, and a 77 year old woman with renal urothelial cancer who was treated with pembrolizumab and cabozantinib. All four patients were promptly recognized as having immune-related neurotoxicity (irNs), were promptly treated (i.e., high dose steroids +/- IVIG +/- other immunomodulators), and significantly improved and have remained stable.

CONCLUSION

This is the first case series of patients with AMAN following two cycles of immunotherapy, who were successfully treated. It is crucial to develop a better understanding of irNs, including those rare conditions that are difficult to diagnose and treat, as the utilization of these immunomodulating therapies continues to expand to include other solid malignancies. Neurologists should be involved early on in any case of suspected irN to assist in the management of these complicated patients and a swift work up should be initiated for timely diagnosis and treatment.

SURG-10. The evolving role of neurosurgery for central nervous system metastases in the era of personalized medicine

Background

Novel therapies translating into improved survival of patients with advanced cancer have emerged. The number of metastases in the central nervous system is therefore seen to increase. Neurosurgery assumes an expanding role within multi-disciplinary care structures for such patients.

Methods

We performed a comprehensive literature review on the current status of neurosurgery for brain metastases patients. Based on the extracted data, we developed a review from experts in the field on the role of brain metastasis surgery in the era of personalized medicine.

Results

Traditionally, three metastases were considered the cutoff to offer surgical resection. With respect to the clinical status, the resection of a symptomatic mass may nowadays be considered even in presence of multiple tumors in a multimodal setting: surgical resection of brain metastasis provides immediate relief from mass effect-related symptoms and histology in case of unknown primary tumor; surgery may help stabilizing the disease, thus enabling further therapy; and in situations where immunotherapy is considered and non-surgical management would require long-term steroid administration, surgery may also provide expeditious relief of edema and reduction of needs for steroids. In patients with multiple brain metastasis and mixed response to non-surgical therapy, tumor resampling may allow tissue analysis for expression of molecular tumor targets. In patients with leptomeningeal dissemination and consecutive hydrocephalus, ventriculo-peritoneal shunting improves quality of life but also allows for time to administer more therapy thus prolonging survival. Addressing the limited efficacy of many oncological drugs for brain metastases, clinical trial protocols in which surgical specimens are analyzed for pre-surgically administered agents may offer pharmacodynamic insights.

Conclusion

Comprehensive neurosurgical care will have to be an integral element of multi-disciplinary oncological centres providing care to patients with brain metastases to improve on therapy and tumour biology research.

LMD-05. Phase 1B Study of Avelumab and Whole Brain Radiotherapy (WBRT) in Patients with Leptomeningeal Disease (LMD): Preliminary Results

Background

LMD has a dismal prognosis with median survivals of 8–10 weeks. Recently the first phase 2 trial of PD-1 inhibitor monotherapy in solid tumor LMD showed median overall survival (OS) 3.6 months. We aimed to determine the safety/efficacy of avelumab with WBRT in patients with LMD from solid malignancies (NCT0371768). This combination can treat tumor directly and increase the permeability of the blood-brain-barrier with increased egress of activated T cells into the meninges/CSF and facilitated Avelumab entry into the CSF.

Hypothesis

Combination radioimmunotherapy will produce an activated immunocyte/cytokine profile in CSF.

Methods

Patients received concurrent Avelumab 800mg IV q2weeks x≤5 cycles with WBRT 3000cGy, 10 fractions. Primary endpoints: Safety/DLTs and OS at 3 months. Secondary endpoints: CSF T-cell/cytokine profiles (scRNAseq/phosophoproteomics) and clinical outcomes, to be performed when all 15 patients are accrued to minimize batch effects.

Results

Ten patients (5 breast, 4 lung & 1 undifferentiated sinonasal carcinoma) were enrolled (n=8 females, n=2 males, ages 32–79); n=1 patient did not complete WBRT. Patients who received anti-PD-1/PD-1L/PD-L2/CD137/CTLA-4 therapy within 6 months prior to enrollment were excluded. 30% had grade 3 AEs at least possibly related to treatment (n=3 diarrhea, lymphopenia, decreased WBC count). There were no grade 4–5 toxicities. Six patients (66.7%) were alive at 3 months. The estimated median follow up in 9 patients (regardless whether patients failed or not) is 10.49 months (range, 0.95–19.82 months, 95% CI) and the estimated median follow up survival was 19.8 months assessed using the reverse Kaplan-Meier method. Median PFS is 4.27 months (range, 0.30–16.73 months, 95% CI).

Conclusions

In this pilot study, combination of Avelumab and WBRT is safe, and demonstrates encouraging activity in patients with solid tumor LMD. Multiple platform interrogation of CSF may determine mechanisms of LMD therapeutic effects and differentiate responders from non-responders.

IMMU-10. GENOMIC DIFFERENCES UNDERLIE MYELOID-DERIVED SUPPRESSOR CELL SEXUAL DIMORPHISM IN GLIOBLASTOMA

A potently immunosuppressive tumor microenvironment facilitates progression of glioblastoma (GBM). We previously demonstrated that myeloid-derived suppressor cell (MDSC) subsets promote tumorigenesis in a sex-specific manner, contributing to sexual dimorphism in GBM incidence and prognosis. Our findings indicated that proliferating monocytic MDSCs (mMDSCs) accumulate in tumors of male mice and patients, while female tumor-bearing mice had an increase in circulating granulocytic MDSC (gMDSC) frequency, and a high gMDSC gene signature correlated with worse outcome of female patients. However, the mechanisms underlying sexual dimorphism of MDSC heterogeneity remain understudied and can provide insights for improved immunotherapy response. Using syngeneic mouse glioma models and sequencing approaches, we show that expression of Y-chromosome-linked genes correlates with upregulation of multiple RNA transcription-related pathways specifically in male mMDSCs. Consistently, adoptive transfer of male mMDSCs but not gMDSCs worsened GBM outcome in male recipients, while the transfer of sex-matched mMDSCs did not impact survival of female mice. In contrast to this cell-intrinsic regulatory pathway, sex steroids had no impact on MDSC profile, as castration or ovariectomy failed to alter MDSC subset accumulation patterns in GBM-bearing mice. Correspondingly, IL-1β, which we had identified as a female-specific drug target, was highly expressed in female but not male gMDSCs. Single-cell sequencing revealed that circulating but not tumor-infiltrating gMDSCs were the primary source of IL-1β and that its neutralization provided a female-specific survival advantage by reducing circulating gMDSCs. This was accompanied by declines in tumor infiltration of microglia, microglia activation status and tumor cell proliferation. In vitro, IL-1β inhibition reduced viability and expression of activation markers by primary microglia. These findings highlight a peripheral gMDSC-microglia communication axis mediated by IL-1β signaling in females with GBM and indicate that expression differences in MDSC subsets represent opportunities for improved immunotherapy efficacy while accounting for sex as a biological variable.

RADT-02. CLINICAL OUTCOMES OF BREAST BRAIN METASTASES BY SUBTYPE FOLLOWING LINAC BASED STEREOTACTIC RADIATION

BACKGROUND

Management of breast cancer brain metastases has become an increasing concern due to improved systemic control. Failure patterns in the brain may vary by breast cancer subtype.

OBJECTIVE

In this study, we sought to distinguish our institutional clinical outcomes following stereotactic radiation by breast cancer subtype.

METHODS

A total of 180 breast cancer patients treated over 279 stereotactic sessions to 646 brain metastases were identified from our LINAC based stereotactic radiation institutional registry. Patients were treated between August 2004 and May 2019. Outcomes including distant brain metastases control (DC) as well as overall survival (OS) following stereotactic radiation were assessed from review of the clinical chart and radiologic examinations.

RESULTS

The median age of patients was 55 (range: 28-86 years). Subtypes in order of decreasing frequency were hormone receptor (HR)+ (n=64; 36%), HR+/HER2+ (n=47; 26%), triple negative (TN) (n=43; 24%), and HR-/HER2+ (n=26; 14%). TN patients had the shortest interval from systemic metastases to brain metastases diagnosis; HR-/HER2 + 16 months, HR+ 13.3 months, HR+/HER2 + 11 months, and TN 1.4 months (p=0.02). Median follow-up from brain metastases diagnosis was 21.2 months (range: 0.9-135.4 months). Twenty-four month Kaplan-Meier (KM) DC rates varied by subtype and were 49% (HR+/HER2+), 38% (HR+), 33% (HR-/HER2+) and 21% (TN) (p=0.0004), respectively. Similar differences were noted in OS with 24 month KM rates of 58% (HR+/HER2+), 51% (HR-/HER2+), 27% (HR+), and 14% (TN), p< 0.0001. A total of 26 patients (14%) were noted to undergo leptomeningeal disease (LMD) progression. No differences were noted by subtype and LMD progression (p=0.88).

CONCLUSIONS

In our institutional series of breast cancer brain metastases treated with stereotactic radiation, significant differences were noted in clinical outcomes by breast cancer subtype. HR+/HER2+ patients had the best DC and OS rates while outcomes were poorest for TN patients.

CTNI-51. NEUROCOGNITIVE FUNCTION (NCF) OUTCOMES OF RTOG FOUNDATION 3508: A PHASE 3 TRIAL OF ABT-414 WITH CONCURRENT CHEMORADIATION AND ADJUVANT TEMOZOLOMIDE IN PATIENTS WITH EGFR-AMPLIFIED NEWLY DIAGNOSED GBM

RTOG 3508/AbbVie M13-813/INTELLANCE-1 was a phase 3 trial of depatuximab-mafodotin (depatux-m, formerly ABT-414) that accrued 639 patients with EGFR-amplified newly diagnosed GBM. At the pre-specified interim OS analysis, the futility criteria were met and there was no survival benefit from adding depatux-m to SOC. Pre-specified secondary NCF analyses included time to decline in verbal learning and memory as assessed by the HVLT-R Total Recall based on the reliable change index. Exploratory NCF analyses examined changes in other HVLT-R outcomes over time. As corneal epitheliopathy causing visual impairment is a known toxicity of depatux-m, NCF tests that did not depend on visual acuity were employed. NCF testing occurred at baseline, day 1 of the first cycle of adjuvant depatux-m, every other cycle (i.e., 8 weeks) thereafter, and at progression. Compliance with test completion was 95% at screening and 80%, 70%, 58%, 51%, 47% thereafter through cycle 9. The most common reasons for missing data was site error. Time to HVLT-R Total Recall decline trended worse in the depatux-m arm compared to placebo but the difference was not significant (12 month deterioration: 41.2%, 95% CI: 3.50–47.2 vs 32.4%, 95% CI: 26.6- 38.4, p=0.052). The depatux-m arm, in comparison to the placebo arm, showed greater decline from baseline on the HVLT-R at the following time points: cycle 3 (Total Recall: mean= -1.8, SD=5.7 vs mean= -0.5, SD=5.5, respectively, p=0.046; Delayed Recall: mean= -1.1, SD=3.0 vs. mean= -0.2, SD=2.7, respectively, p=0.01), cycle 7 (Total Recall: mean= -0.6, SD=5.1 vs mean= 1.4, SD=5.0, respectively, p=0.009; Delayed Recall: mean -0.6, SD=3.0 vs. mean= 0.5, SD=2.7, respectively, p=0.01), and cycle 9 (Delayed Recall: mean=-0.4, SD=2.7 vs. mean= 0.8, SD=2.4, respectively, p=0.003). Depatux-m added to concurrent chemoradiation and adjuvant temozolomide was associated with faster time to deterioration and worse episodic learning and memory over time than placebo.

NCMP-27. THE USE OF INTRAVENOUS IMMUNOGLOBULIN (IVIG) DURING SEVERE NEUROTOXICITY AMONG THE RECIPIENTS OF CHIMERIC ANTIGEN RECEPTOR T-CELLS (CAR-T) THERAPY

INTRODUCTION

Severe Immune effector Cell-Associated Neurotoxicity Syndrome (ICANS) occurs in ~ 30% of Diffuse Large B Cell Lymphoma (DLBCL) patients treated with CAR-T cell therapy. The current treatment for severe ICANS is glucocorticoids, and this may be combined with tocilizumab for concurrent cytokine release syndrome. Even with these treatments, neurotoxicity can persist. It is essential, therefore, to find additional treatments to more effectively reverse CAR-T neurotoxicity.

OBJECTIVE

We summarize our institutional experience using IVIG to treat severe or refractory ICANS after glucocorticoids alone.

METHOD

This is a single center retrospective cohort study of neurologic and oncologic outcomes among patients who received axicabtagene ciloleucel (axi-cel) for DLBCL or its variants from May 2015 to February 2019. We identified those patients who developed severe ICANS subsequently treated with glucocorticoids alone (n = 10) or glucocorticoids plus IVIG (n = 9).

RESULTS

The median age of all 19 patients was 63 (range 47-75, and 68% were males). All IVIG recipients received glucocorticoids prior to IVIG administration. There was no significant difference in time to resolution (TTR) of severe ICANS between both groups (median 3 vs. 3 days, Log-rank p = 0.331). However, we found that all IVIG recipients had experienced either escalation of their neurotoxicity grade or persistent severe neurotoxicity after initiation of steroids (p = 0.001). Moreover, IVIG recipients had worse ECOG performance status prior to axi-cel therapy (p = 0.03).

CONCLUSION

Although we found no difference in TTR of severe ICANS with addition of IVIG to glucocorticoids, our analysis suggests that the addition of IVIG blunted the effects of steroid-refractory neurotoxicity. Patients who received IVIG had worsening ICANS despite administration of steroids and therefore might have benefited from the earlier addition of IVIG. Controlled studies may clarify the potential efficacy of IVIG in severe neurotoxicity after CAR-T therapy.

CTNI-23. IDH1/2wt ANAPLASTIC GLIOMAS OF THE EORTC RANDOMIZED PHASE III INTERGROUP CATNON TRIAL: OVERALL SURVIVAL RELATED TO TREATMENT, MGMT STATUS AND MOLECULAR FEATURES OF GLIOBLASTOMA

BACKGROUND

The phase III CATNON trial randomized 751 adult patients with newly diagnosed 1p/19q non-codeleted anaplastic glioma to 59.4 Gy radiotherapy +/- concurrent and/or adjuvant TMZ. Here, we present the molecular analysis of the IDH1/2wt subgroup, and associations between molecular characteristics and patient outcomes.

METHODS

CNV data and MGMT promoter methylation status were assessed from EPIC methylation array data. IDH1/2 and H3F3A mutation status were determined with a glioma tailored next-generation sequencing panel and TERT promoter mutation status using a SNaPshot assay. Overall survival (OS) was measured from date of randomization.

RESULTS

Of 654 assessed tumors, 211 (32%) were IDH1/2wt. An H3F3A mutation was present in 14 tumors (K27M: n=10; G34R: n=4). Of the remaining 197 patients, 154 tumors had molecular features of glioblastoma according to cIMPACT-NOW 3 criteria (‘IDH1/2wt astrocytomas WHO IV’), 39 tumors did not (‘IDH1/2wt astrocytomas WHO III’), and 4 had inconclusive molecular data. IDH1/2wt astrocytomas WHO III patients had significantly better survival than WHO IV patients: median OS of 2.83 yrs vs 1.43 yrs respectively [log-rank test: p< 0.001]. Of the 154 IDH1/2wt astrocytoma WHO IV, 55 (36%) were found MGMT promoter methylated. MGMT promoter methylation was prognostic in IDH1/2wt astrocytomas WHO IV patients, with a median OS of 1.86 yrs for methylated vs 1.34 yrs for unmethylated [HR 1.62, p=0.006]. In the IDH1/2wt astrocytomas WHO IV, no effect of concurrent and/or adjuvant TMZ was observed; the HR for OS after RT with any TMZ vs RT alone was 1.31 [95% CI 0.73–2.36, p=0.4] for MGMT promoter methylated and 0.90 [95% CI 0.55–1.45, p=0.7] for unmethylated glioma patients.

CONCLUSIONS

Our study validated the prognostic value of the cIMPACT-NOW 3 criteria. MGMT promoter methylation is prognostic but not predictive for outcome to TMZ treatment in this cohort of IDH1/2wt anaplastic gliomas with molecular features of glioblastoma.

PATH-11. PROGNOSTIC SIGNIFICANCE OF EPIGENETIC SUBTYPES AND CpGs ASSOCIATED WITH PROGRESSION TO G-CIMP LOW IN THE EORTC RANDOMIZED PHASE III INTERGROUP CATNON

BACKGROUND

Uncontrolled studies have suggested that methylation-based epigenetic subtypes can be used for prognostication of glioma. We used the prospective randomized CATNON trial to validate the clinical relevance of these epigenetic subtypes.

METHODS

The phase III CATNON trial randomized 751 adult patients with newly diagnosed 1p/19q non-codeleted anaplastic glioma to 59.4 Gy radiotherapy +/- concurrent and/or adjuvant TMZ. CNV data and methylation data were derived from Infinium MethylationEPIC arrays. Epigenetic subtyping and risk of progression to G-CIMP low were determined from random forest models and 7 specific CpGs (PMID: 29642018). IDH1/2 status was determined with a glioma-tailored NGS panel. Overall survival (OS) was measured from date of randomization.

RESULTS

Methylation analysis was performed on 654 tumors: 440 were IDH1/2mt, 204 IDH1/2wt and of 10 IDH1/2 status was unknown; 8 IDH1/2mt were 1p/19q codeleted. Based on methylation, tumors were classified as G-CIMP high (n=409), G-CIMP low (n=19), codel-like (n=18), mesenchymal-like (n=107), classic-like (n=48), and PA-like tumors (n=53). Median OS between these epigenetic subtypes varied considerably: codel-like 9.1 yrs, G-CIMP high 9.5 yrs, G-CIMP low 2.8 yrs, mesenchymal-like 1.3 yrs, classic-like 1.6 yrs, and PA-like 2.8 yrs. The difference in OS of the IDH1/2mt astrocytoma subgroup patients was prominent [G-CIMP low vs G-CIMP high: HR 4.12, 95% CI 2.37-7.19, p < 0.001]. Within the IDH1/2mt G-CIMP high astrocytoma patients, 115 tumors were predicted to have risk of progression to G-CIMP low and patients with such tumors indeed had poorer survival [risk vs no-risk: HR 1.59, 95% CI 1.10-2.31, p = 0.02]. Median OS in G-CIMP high tumors with (n=37) and without (n=366) CDKN2A/B HD was 3.3 yrs versus not reached [p< 0.001], in G-CIMP low tumors it was 1.2 yrs (n=6) versus 4.4 yrs (n=12) [p=0.008].

CONCLUSIONS

In IDH1/2mt anaplastic astrocytoma, G-CIMP status and CDKN2A/B HD are of independent prognostic value.

CTIM-13. CLINICAL EFFICACY OF MDNA55, AN INTERLEUKIN-4 RECEPTOR TARGETED IMMUNOTHERAPY, IN RECURRENT GBM DELIVERED BY CONVECTION ENHANCED DELIVERY (CED)

BACKGROUND

MDNA55 is an IL4R-targeted toxin in development for GBM, a universally fatal disease. IL4R is over-expressed in GBM, the tumor microenvironment, and high expression is associated with poor outcomes in GBM.

METHOD

MDNA55-05 is an open-label, single-arm study of MDNA55 delivered by CED as a single treatment in rGBM patients at 1st or 2nd recurrence with an aggressive form of GBM (de novo GBM, IDH wild-type, not-resectable at recurrence, ~ 50% expressing high levels of IL4R). Primary endpoint is mOS, secondary endpoint includes impact of IL4R status on mOS. A Synthetic Control Arm (SCA) served as an external comparator constructed from patient registries at neurosurgery centers under IRB-approved protocols. Propensity score weighting corrected for imbalances in baseline characteristics between the two groups.

RESULTS

MDNA55 showed an acceptable safety profile at doses up to 240μg. In all evaluable subjects (n=44) mOS was 11.6 and OS-12 was 46%. A sub-population (n=32) consisting of all IL4RHigh subjects + only IL4RLow subjects treated with high dose (median 180µg) showed most benefit: mOS is 15 months, OS-12 is 55%. Tumor control, assessed by mRANO criteria, was seen in 81% (26/32) of this sub-population, including those that had pseudo-progression (15/26). In these subjects, tumor control was associated with longer mPFS (4.7 months) and mOS (15.0 months) than those with progressive disease (mPFS 1.0 month, mOS 7.7 months). Comparison against the SCA demonstrated > 100% increase in mOS: 15.7 months vs 7.2 months (HR 0.52, 95% CI 0.30, 0.91).

CONCLUSIONS

MDNA55 demonstrates improved survival and tumor control in this study design and has potential to benefit all rGBM patients treated at high dose irrespective of IL4R expression. There are no approved therapies for rGBM that can extend survival by 50%; the potential for MDNA55 to extend survival by > 100% is promising for this devastating disease.

IMMU-08. MODELING UPFRONT GLIOBLASTOMA SURGICAL RESECTION AND STEROID USE REVEALS IMMUNOSUPPRESSIVE CHANGES AND SUGGESTS THAT PERIPHERAL LYMPHOCYTE COUNTS ARE ASSOCIATED WITH TUMOR VOLUME AND PROGNOSIS

Glioblastoma (GBM) and its treatment produces systemic immunosuppression, which is being targeted by immunotherapies. However, it remains unclear how surgical resection and steroids specifically in GBM alter the immune system. To further explore this issue, immunocompetent C57Bl/6 mice were intracranially inoculated with syngeneic glioma cells (GL261 and CT-2A) and growth of tumors was evaluated by MRI. Host immune cell populations were analyzed during surgical resection and steroid administration. Mice with surgically resected tumors had a longer median survival compared to mice subjected to tumor biopsies, and had increased bone marrow sequestration of both CD4 and CD8 T cells with corresponding decreased blood lymphocytes. Furthermore, physiologic doses of dexamethasone administered perioperatively decreased tumor edema, but increased the number and proliferative capacity of both marrow and circulating MDSCs while generating no survival benefit. Independent of therapy or dexamethasone, intracranial tumor volume correlated linearly with decreased CD4 and CD8 T cells in peripheral blood, and increased T cell sequestration within the bone marrow. We validated these parameters in steroid-naïve newly diagnosed GBM patients and observed decreased lymphocytes correlated linearly with increased tumor volume. When initial lymphocyte counts in both steroid-naïve and steroid-administered patients were used in univariate and multivariate models predicting progression-free survival and overall survival, decreased initial lymphocyte counts were an independent predictor of decreased progression free survival and decreased overall survival, with steroid use and initial tumor size falling out of significance during stepwise selection. Taken together, tumor volume is linearly correlated with marrow sequestration of lymphoid cells, but both surgery and steroid administration further suppress active immune responses along lymphoid and myeloid lineages. Furthermore, decreasing peripheral lymphocyte counts at diagnosis of GBM indicate an immune system less able to mount responses to the tumor and portent a worse progression free and overall survival.

CTNI-29. CODEL: PHASE III TRIAL OF RT ALONE, RT PLUS TMZ, OR TMZ ALONE FOR NEWLY-DIAGNOSED, 1p/19q CODELETED ANAPLASTIC OLIGODENDROGLIOMA. ANALYSIS FROM THE INITIAL STUDY DESIGN. (NCCTG N0577, ALLIANCE)

BACKGROUND

The original 3-arm CODEL design included a radiotherapy (RT)-alone control arm, an RT plus temozolomide (TMZ) arm, and an exploratory TMZ-alone arm. We report the analysis involving patients treated per the initial design.

METHODS

Adults (18+ years) with newly-diagnosed 1p/19q codeleted WHO grade III oligodendroglioma were randomized to RT (5940 cGy) alone (Arm A); RT with concomitant and adjuvant TMZ (Arm B); or TMZ alone (Arm C). Primary endpoint was OS, Arm A vs. B. Secondary comparisons were performed for OS and PFS, comparing pooled RT arms with the TMZ-alone arm.

RESULTS

36 patients were randomized equally to the three arms. At median follow-up of 7.5 years, 83.3% (10/12) TMZ-alone patients had progressed, versus 37.5% (9/24) patients on the RT arms. PFS was shorter in TMZ-alone patients compared to RT-treated patients (HR=3.12; 95% CI: 1.26, 7.69; p=0.014). Death from disease progression occurred in 3/12 (25%) of TMZ-alone patients and 4/24 (16.7%) of RT-treated patients. OS did not statistically differ between arms, although this comparison was underpowered. After adjustment for IDH status (mutated vs. wildtype) in a Cox regression model, with IDH status and RT treatment status as co-variables (Arm C vs pooled A and B), PFS remained shorter for patients not receiving RT (HR= 3.33; 95% CI: 1.31, 8.45; p=0.011), and OS differences remained non-significant ((HR = 2.78; 95% CI 0.58, 13.22, p=0.20). Grade 3+ adverse events occurred in 25%, 42% and 33% patients (Arms A, B and C, respectively). Neurocognitive assessments, comparing baseline and 3 month timepoints, showed no significant differences between arms.

CONCLUSIONS

TMZ-alone treated patients experienced significantly shorter PFS than patients treated on the pooled RT arms, which remained significant when adjusting for IDH status. CODEL has been redesigned to compare the efficacy and toxicity of RT+PCV versus RT+TMZ. Clinicaltrials.gov Identifier: NCT00887146. Support: U10CA180821, U10CA180882, https://acknowledgments.alliancefound.org.

Proposed response assessment and endpoints for meningioma clinical trials: report from the Response Assessment in Neuro-Oncology Working Group

No standard criteria exist for assessing response and progression in clinical trials involving patients with meningioma, and there is no consensus on the optimal endpoints for trials currently under way. As a result, there is substantial variation in the design and response criteria of meningioma trials, making comparison between trials difficult. In addition, future trials should be designed with accepted standardized endpoints. The Response Assessment in Neuro-Oncology Meningioma Working Group is an international effort to develop standardized radiologic criteria for treatment response for meningioma clinical trials. In this proposal, we present the recommendations for response criteria and endpoints for clinical trials involving patients with meningiomas.

RBTT-11. NRG ONCOLOGY NRG-BN006: A PHASE II/III RANDOMIZED, OPEN-LABEL STUDY OF TOCA 511 AND TOCA FC WITH STANDARD OF CARE COMPARED TO STANDARD OF CARE IN PATIENTS WITH NEWLY DIAGNOSED GLIOBLASTOMA

The standard of care (SOC) in newly diagnosed glioblastoma (nGBM) includes resection and chemoradiotherapy. With a median overall survival (OS) of only 16–18 months for well-selected patients in clinical trials, better therapeutic options are needed. Toca 511 (vocimagene amiretrorepvec) is a retroviral replicating vector encoding a codon optimized, heat stabilized cytosine deaminase that converts Toca FC (extended-release 5- fluorocytosine,5-FC) into 5-fluorouracil. Preclinical evidence demonstrates that Toca 511 & 5-FC kill cancer cells and immunosuppressive myeloid cells in the tumor microenvironment, leading to durable antitumor immune responses. Three phase (P) 1 studies in patients with recurrent high grade glioma have demonstrated tolerable safety profile and encouraging efficacy. NRG-BN006 is a randomized P2/3 trial of Toca 511 & Toca FC with SOC versus SOC for patients with nGBM. Optune use is allowed on the SOC arm, but not on the experimental arm. Patients will be stratified by age and KPS score for 1:1 randomization. The primary endpoint is progression-free survival for P2 and OS for P3. The secondary endpoints include objective response rate in patients with measurable disease and safety. Key inclusion criteria include presumptive diagnosis of glioblastoma with anticipated 80% resection, unifocal tumor, and KPS≥70. Immune monitoring and molecular profiling will be performed. P2 has 90% power to detect a hazard ratio (HR)=0.67 in 250 nGBM patients. P3 has 85% power to detect a HR=0.75 in 720 nGBM patients. Since patients are enrolled prior to surgery and confirmatory diagnosis of GBM, approximately 900 patients will be enrolled, and two interim analyses are planned for OS. In addition, two interim safety analyses will be conducted for the experimental treatment, with the first 15 and 30 eligible and analyzable patients randomized to the experimental arm. NRG-BN006 is anticipated to start enrollment in Q4 2019. Supported by grants U10CA180868, U10CA180822 from NCI and Tocagen.

ACTR-21. A RANDOMIZED, DOUBLE-BLIND, PLACEBO-CONTROLLED PHASE 3 TRIAL OF DEPATUXIZUMAB MAFODOTIN (ABT-414) IN EPIDERMAL GROWTH FACTOR RECEPTOR (EGFR) AMPLIFIED (AMP) NEWLY DIAGNOSED GLIOBLASTOMA (nGBM)

BACKGROUND

Approximately 50% of nGBMs harbor EGFR-amp. Depatuxizumab mafodotin (depatux-m) is an antibody drug conjugate: a monoclonal antibody that binds activated EGFR (wild-type and EGFRvIII mutant) linked to a microtubule-inhibitor toxin. Pre-clinical and earlier clinical trials suggested efficacy.

METHODS

RTOGF 3508/AbbVie M13-813 (INTELLANCE-1, NCT02573324) was a phase 3 academic-industry collaboration (RTOG-Foundation, AbbVie). Eligible adults (KPS ≥ 70, EGFR-amp nGBM, centrally confirmed histology and biomarkers) were randomized 1:1 to radiotherapy (RT) and temozolomide and either depatux-m (2.0 mg/kg during RT, 1.25 mg/kg thereafter, q 14 days) or placebo, stratified by region of world, RPA class, MGMT methylation, and EGFRvIII mutation. Primary endpoint was overall survival (OS), with 640 patients planned for randomization; 441 events yielded 85% power to detect 25% reduction in hazard of death (HR 0.75), one-sided 2.5% level of significance by stratified weighted log-rank.

RESULTS

2229 patients were screened and 639 (median age 60, range 22–84; 394 men, 62%) randomized. Pre-specified interim analysis after 346 events (≥ 75% required) found no OS improvement for depatux-m over placebo (median 18.9 vs. 18.7 months, HR 1.01, 95% CI 0.82–1.25, one-sided p= 0.63). Progression-free survival (PFS) trended toward depatux-m (median 8.0 vs. 6.3 months; HR 0.84, 95% CI 0.70–1.02), particularly among the ~50% with EGFRvIII mutation (median 8.3 vs. 5.9 months, HR 0.72, 95% CI 0.56–0.93) but without an OS improvement (median 19.8 vs. 18.2, HR=0.95, 95% CI 0.71–1.27). Ocular side effects (grade ≥ 1) occurred in 95% of depatux-m treated patients, 61% grade 3–4, causing 12% to discontinue, and were the most common treatment related adverse events.

CONCLUSION

Interim analysis demonstrated no OS benefit for treating EGFR-amp nGBM with depatux-m. PFS trended toward favoring depatux-m, particularly in the EGFRvIII harboring subgroup. No new important safety risks were identified. The trial was stopped for futility. Active patients are permitted to continue treatment.

CMET-17. RENAL CELL CARCINOMA BRAIN METASTASES TREATED WITH STEREOTACTIC RADIATION THERAPY AND NIVOLUMAB DOES NOT ALTER LESIONAL OR CLINICAL OUTCOMES

Metastases (BM) carries the risk of hemorrhaging lesions and can be effectively treated using stereotactic radiotherapy (SRT). Nivolumab is a recently approved immunotherapy for stage IV RCC. We evaluate whether patients with RCC BM treated with SRT overlapping with nivolumab have altered clinical and BM outcomes. METHODS: 38 consecutive patients were identified in our retrospective database from 1/2011 to 6/2018. Analyses were performed on a per-lesional basis (n=170), per-treatment session basis (n=79), and per-patient basis (n=38). Patients who received nivolumab within 6 months of SRT were considered to have overlapping treatments. ROC curve, chi-squared, Kaplan-Meier, log-rank, and Cox regression model were employed for statistical analyses. RESULTS: A total of 7 (18.4%) patients received overlapping treatments for 64 (37.6%) eligible lesions. Median follow-up was 15.4 months and median overall survival from first BM treatment was 14.8 months (0.5 – 98.4). Median time to subsequent distant brain and non-brain failures were 3.4 and 2.2 months, respectively. Median time to local failure was 20.2 months (two lesions). There were 11 hemorrhagic toxicities (7 in the nivolumab group) and 17 radionecrosis toxicities (4 in the nivolumab group) with no significant difference amongst the groups. Lesions receiving nivolumab within 6 months of SRT did not exhibit a higher rate of toxicity (p=0.521) but had a shortened time to hemorrhage (p< 0.001). Patients who received SRT > 1 month after nivolumab had a prolonged time to subsequent distant brain failure (median 11.1 months) than patients who received SRT > 1 month before (median 3.1 months) or within 1 month (median 1.6 months) of nivolumab, p=0.014. CONCLUSIONS: Overlapping nivolumab with SRT is safe with no increased risk of hemorrhagic lesions. An optimal treatment sequence of nivolumab administration followed by SRT prolongs the time to subsequent BM and warrants further investigation.

IMMU-28. TARGETING IMMUNOSUPPRESSIVE MYELOID DERIVED SUPPRESSOR CELLS VIA MIF/CD74 SIGNALING AXIS TO ATTENUATE GBM GROWTH

The immunosuppressive microenvironment in glioblastoma (GBM) enables persistent tumor growth and evasion from tumoricidal immune cell recognition. Despite a large accumulation of immune cells in the GBM microenvironment, tumor growth continues, and evidence for potent immunosuppression via myeloid derived suppressor cells (MDSCs) is now emerging. In agreement with these observations, we have recently established that increased MDSCs over time correlates with poor prognosis in GBM, making these cells of interest for therapeutic targeting. In seeking to reduce MDSCs in GBM, we previously identified the cytokine macrophage migration inhibitory factor (MIF) as a possible activator of MDSC function in GBM. Here, using a novel in vitro co-culture system to reproducibly and rapidly create GBM-educated MDSCs, we observed that MIF was essential in the generation of MDSCs and that MDSCs generated via this approach express a repertoire of MIF receptors. CD74 was the primary MIF receptor in monocytic MDSCs (M-MDSC), which penetrate the tumor microenvironment in preclinical models and patient samples. A screen of MIF/CD74 interaction inhibitors revealed that MN-166, a clinically relevant blood brain barrier penetrant drug, which is currently fast tracked for FDA approval, reduced MDSC generation and function in vitro. This effect was specific to M-MDSC subsets expressing CD74, and appeared as reduced downstream pERK signaling and MCP-1 secretion. In vivo, MN-166 was able reduce tumor-infiltrating MDSCs, while conferring a significant increase in survival in the syngeneic glioma model GL261. These data provide proof of concept that M-MDSCs can be targeted in the tumor microenvironment via MN-166 to reduce tumor growth and provide a rationale for future clinical assessment of MN-166 to reduce M-MDSCs in the tumor microenvironment. Ongoing studies are assessing the effects of MDSC inhibition in combination with immune activating approaches, in order to inhibit immune suppression while simultaneously activating the immune system.

CMET-06. AN ANALYSIS OF RESPONSE ENDPOINTS FOR BRAIN METASTASIS PATIENTS TREATED WITH STEREOTACTIC RADIOSURGERY AND PD(L)-1 INHIBITORS

PURPOSE

Recent trials have evaluated the role of anti-PD(L)-1 inhibitors alone in patients with newly diagnosed brain metastasis, resulting in a suggested paradigm shift. In the absence of randomized comparisons, an evaluation of outcomes for similar patients treated with radiotherapy are needed to provide baseline comparative data.

METHODS

This retrospective cohort study included patients diagnosed with brain metastasis from 2010–2017 at a single tertiary care institution who received SRS and at least one anti-PD(L)-1 inhibitor. Primary endpoints included the rate of best intracranial response (either a complete [CR] or partial response [PR]), rate of intracranial benefit (percentage of patients with stable disease [SD] for at least 6 months), and overall objective response rates.

RESULTS

150 patients met the eligibility criteria for this study and the median follow-up time was 10 months (Range: 1–130 months). Only a minority of patients from this cohort would have been eligible for comparative systemic therapy alone trials: 8 (CheckMate 204, CM), 8 (ABC), and 50 patients (Yale Institutional trial, YI). Best intracranial objective response rates for these trial eligible patients were 64% (CM), 64% (ABC), and 72% (YI), respectively. The 6-month rates of intracranial benefit across patients were 75% (CM), 75% (ABC), and 93% (YI), respectively. The rates of either a CR, PR, or SD across trial eligible patients were 100% (CM), 100% (ABC), and 94% (YI), respectively. More than 50% of patients experienced out-of-field progression in each of the patient subgroups.

CONCLUSIONS

Over a seven year period, very few patients treated with SRS and anti-PD(L)-1 inhibitor therapy in a busy academic practice would have been eligible for systemic therapy alone per trial eligibility criteria. In these highly favorable subgroups, patients experience high rates of objective response, best intracranial response, and intracranial benefit after SRS and anti-PD(L)-1 therapy, representing optimal brain metastasis management.

ATIM-30. COMBATING RECURRENT GLIOBLASTOMA WITH MDNA55, AN INTERLEUKIN-4 RECEPTOR TARGETED IMMUNOTHERAPY, THROUGH MRI-GUIDED CONVECTIVE DELIVERY

MDNA55, an IL4R-directed toxin, is being studied in a Phase 2b trial in recurrent GBM (rGBM) patients at first or second relapse. MDNA55 is co-infused with Gadolinium-based contrast agent and delivered as a single intratumoral infusion using Convection Enhanced Delivery (CED). Primary endpoint is median Overall Survival (mOS) and secondary endpoint is objective response rate (ORR) assessed by mRANO-based criteria incorporating advanced imaging modalities. Enrollment is complete (n=46). Current safety data show similar profile to previous MDNA55 trials with no systemic toxicities or drug related deaths. Current mOS in subjects treated with low doses of MDNA55 (median 63µg; n=21) is 11.8 months. When stratified by IL4R expression, a biomarker for more aggressive GBM, IL4R+ve subjects (mOS 15.2 months; n=8) show a survival advantage of 7 months compared to IL4R-ve subjects (mOS 8.1 months; n=10). Updated survival and response outcomes including subjects receiving the high dose (median 180µg; n=25) and stratification by IL4R expression will be reported. Review of serial imaging within 90 days following MDNA55 treatment demonstrated tumor shrinkage or stabilization from baseline in 19/42 evaluable subjects (disease control rate of 45%). To account for initial pseudo-progression in some subjects, tumor response was also assessed from nadir: 83% (35/42) showed disease control. Multi-parametric MRI biomarkers including relative cerebral blood volume (rCBV) and apparent diffusion coefficient (ADC) measurements demonstrated distinct imaging phenotypes among different disease states (pseudo-progression vs true-progression, pseudo-response vs true-response) and improved response staging. This trial is advancing neurosurgical methods for CED, potential of IL4R expression as a biomarker to select GBM patients most likely to benefit from MDNA55 treatment, and optimal use of multi-parametric MRI as an adjunct to clinical decision making. The improved survival and disease control seen after only a single infusion of MDNA55, especially in IL4R+ve subjects, may provide promising clinical benefit for rGBM patients.

STEM-10. BIDIRECTIONAL INTERACTION BETWEEN TUMOR-ASSOCIATED PLATELETS AND GLIOMA STEM CELLS IN GLIOBLASTOMA MULTIFORME

Glioblastoma Multiforme (GBM) is the most common and lethal malignant primary adult brain tumor. Therapy resistance and tumor recurrence in GBM have been attributed to glioma stem cells (GSCs), which are found in hypoxic but perivascular niches. We hypothesized that the clinically documented and prognostically relevant increase in platelets in GBM promotes formation of functional hypoxic but perivascular niches that support GSCs due to formation of intravascular thromboses, a hallmark of GBM. Our preliminary in silico analysis from TCGA GBM samples indicates significant correlation of platelet and stemness signature expression (P < 0.001). High expression of platelet gene signatures also inversely correlates with overall survival in GBM patients (P < 0.02). Furthermore, we found significant co-localization of known platelet and stemness markers in primary GBM specimens, supporting our hypothesis and motivating further interrogation of GSC-platelet crosstalk. Our preliminary data suggest that GSCs exposed to either tumor or healthy platelets demonstrate increased self-renewal and stemness, as determined by a significant increase in OCT4, NANOG, and OLIG2 expression compared to unexposed GSCs. This increase in stemness and self-renewal markers was not seen in platelet-exposed differentiated glioma cell (DGC) counterparts. Conversely, platelets are stimulated by GSCs and GSC-conditioned media while DGCs elicit no stimulatory effects as measured by ATP release and aggregation assays. Our results implicate a functional role for platelet-GSC interactions in the maintenance of tumor cellular hierarchy that ultimately contributes to poor clinical outcomes in GBM. RNA-Seq of platelet-privileged GSCs is currently underway to elucidate the mechanism by which platelets impact GSC self-renewal. Successful characterization of potential crosstalk between platelets and GSCs may offer new clinical perspectives into GBM and inform development of a novel treatment paradigm to target these specific cell-to-cell interactions.

SURG-26. READMISSION FOLLOWING RESECTION FOR PATIENTS WITH BRAIN METASTASES IN THE UNITED STATES

OBJECTIVE

Up to 30% of cancer patients will develop brain metastasis during the course of their systemic disease with a significant proportion undergoing resection of at least one lesion. The objective of the present study was to characterize the rates, predictors, and costs of 30-day readmissions following craniotomy for brain metastases using a nationally representative database.

METHODS

This study was a retrospective analysis of data from the Nationwide Readmissions Database (NRD) from 2010–2014. We included patients who underwent craniotomy for brain metastasis, identified using ICD-9-CM diagnosis (198.3) and procedure (01.59) codes. The primary outcome of the study was unplanned 30-day all-cause readmission rates. Secondary outcomes included predictors and costs of readmissions.

RESULTS

During the study period, there were 44,846 index hospitalizations for patients who underwent resection of brain metastasis. Among this cohort, 17.8% (n=7,965) had unplanned readmissions within the first 30 days after discharge from the index hospitalization. The readmission rate did not change significantly during the study period (P=0.286). The odds of unplanned readmission were significantly greater in patients with thromboembolic complications (aOR, 1.53; 95% CI: 1.18–2.01), patients with Elixhauser comorbidities >3 (aOR, 1.35; 95% CI: 1.22–1.50), male patients (adjusted odds ratio [aOR], 1.29; 95% CI: 1.17–1.42), patients with an initial length of stay ≥5 days (aOR, 1.02; 95% CI: 1.01–1.03). The median per-patient cost for 30-day unplanned readmission was $11,109 and this accounted for a total cost of $132.1 million during the study period.

CONCLUSIONS

Unplanned readmissions after resection for brain metastases involve substantial healthcare expenditures. Though there have been many interventions for improving surgical quality, post-operative care, and cost metrics, unplanned readmission rates have not changed. Key patient-specific variables and high rates of comorbidities should be considered to focus our efforts on patient selection for resection, and for strengthening existing interventions for high-risk patients.

GENE-26. HOST GENETIC VARIATIONS IN MACROPHAGE MIGRATION INHIBITOR FACTOR CONFER WORSE PROGNOSIS IN GLIOBLASTOMA

Genetic and systemic prognostic factors have been identified in glioblastoma (GBM) that correlate with survival, but patient-specific genomic risk factors that impact prognosis or response to immunotherapies have not been elucidated. GBM promotes an immune suppressive microenvironment via many mechanisms including the production of macrophage migration inhibitory factor (MIF) by GBM cancer stem cells, which activates myeloid derived suppressor cells (MDSCs). Increased circulating and intracranial MDSCs as well as intratumoral MIF expression in GBM patients portend a worse overall prognosis. Endogenous MIF expression is dependent on two genetic microsatellite loci: single nucleotide polymorphisms (SNPs) and CATT repeats within the promoter. To determine whether these genetic variations were linked to GBM development and/or prognosis, we assessed peripheral nucleated blood from 520 GBM patients for the MIF SNPs and CATT repeats. MIF microsatellite frequencies were similar between the normal population and GBM patients indicating loci variability was not a risk factor for GBM development. However, newly diagnosed IDH wild-type GBM patients with a minor allele SNP who received standard of care therapy had a 3.1 month shorter progression free survival (PFS), and a 4.3 month shorter overall survival (OS) when compared to patients with two major allele SNPs. This association was seen also with the CATT-repeat analyses. Furthermore, in a multivariate analysis for PFS that included age, sex, Karnofsky performance status, MGMT methylation status, 1p/19q co-deletion, and SNP status as covariates, only age and SNP status were independently associated with shorter PFS. Taken together, patients with variant MIF microsatellite loci experienced shorter PFS and decreased OS compared to those with the most common loci. These results are currently being validated in a separate 1700 patient cohort with the intent of understanding how MIF expression influences myeloid populations within the GBM microenvironment and then developing novel peripheral GBM screening markers for aggressiveness.

Long-term trends in glioblastoma survival: implications for historical control groups in clinical trials

Background

Historical controls continue to be used in early-phase brain tumor trials. We aim to show that historical changes in survival trends for glioblastoma (GBM) call into question the use of noncontemporary controls.

Methods

We analyzed data from 46 106 primary GBM cases from the SEER database (1998-2016). We performed trend analysis on survival outcomes (2-year survival probability, median survival, and hazard ratios) and patient characteristics (age, sex, resection extent, and treatment type).

Results

In 2005-2016 (ie, the post-Stupp protocol era), fitting a parameter independently to each year, there was a demonstrable increase in median survival (R² = 0.81, P < .001) and 2-year survival probability (R² = 0.55, P = .006) for GBM. Trend analysis of the hazard ratio showed a significant time-dependent downward trend (R² = 0.62, P = .002). When controlling, via multivariable Cox regression modeling, for age, sex, resection extent, and treatment type, there was a persistent downward trend in hazard ratios with increases in calendar time, especially in the most recent data.

Conclusion

Contemporary GBM patients face a different overall hazard profile from their historical counterparts, which is evident in changes in measures of patient survival and parametric hazard modeling. Though there was a plateau in these measures before 2005 (pre-Stupp protocol), there is no evidence of a new plateau in recent years even when controlling for known prognostic factors (age, sex, resection extent, and treatment type), suggesting that it may be insufficient to match contemporary patients and noncontemporary controls on the basis of these factors.

TRLS-05. EARLY RESULTS FROM A PROSPECTIVE PHASE I/II DOSE ESCALATION STUDY OF NEOADJUVANT RADIOSURGERY FOR BRAIN METASTASES

OBJECTIVES: Single-session stereotacic radiosurgery (SRS) alone for brain metastases larger than 2cm in maximal dimension results in local control of only 50%. Surgical resection followed by SRS to the resection cavity can result in leptomeningeal failure (LMD). This Phase I/II study aims to determine the safety and local control of neoadjuvant SRS at escalating doses followed by surgical resection of brain metastases greater than 2 cm. METHODS: Radiosurgery dose was escalated at 3 Gy increments from currently accepted RTOG standard. If no dose-limiting toxicities (DLT) were observed, the dose was escalated. Patients underwent surgical resection of brain metastases within 2 weeks and were followed with brain MRIs and neurologic evaluations every 3 months. RESULTS: 27 patients were enrolled. For tumor size &gt;2.0–3.0 cm, 2 patients completed treatment at 18 Gy and 3 patients at 21Gy. For tumor size &gt;3.0–4.0 cm, 4 patients were treated at 15 Gy and 9 patients were treated at 18 Gy and 1 patient at 21 Gy. For tumor size &gt;4.0–5.0 cm, 1 patient was treated at 12 Gy and 7 patients at 15 Gy. No DLT have occurred. With a mean follow-up of 13.1 months, the 6 and 12 month local control was 93.8% and 72.3%, respectively. Six and 12 month distant brain control was 38.6% and 25.8%. Overall survival at 12 months was 53.5%. One patient developed LMD 5 months following SRS. 4 patients (15%) had acute grade 1/2 toxicity, and no grade 3/4 toxicity was observed. CONCLUSIONS: Neoadjuvant SRS with dose escalation followed by surgical resection for brain metastases greater than 2 cm results in local control comparable to postoperative SRS or WBRT, and demonstrates acceptable acute toxicity. A low rate of LMD failure was found. The Phase II portion of the trial will be conducted at the maximum tolerated SRS doses.

RADI-35. AN ANALYSIS OF RESPONSE ENDPOINTS FOR BRAIN METASTASIS PATIENTS TREATED WITH STEREOTACTIC RADIOSURGERY AND PD(L)-1 INHIBITORS

PURPOSE: Recent trials have evaluated the role of anti-PD(L)-1 inhibitors alone in patients with newly diagnosed brain metastasis, resulting in a suggested paradigm shift. In the absence of randomized comparisons, an evaluation of outcomes for similar patients treated with radiotherapy are needed to provide baseline comparative data. METHODS: This retrospective cohort study included patients diagnosed with brain metastasis from 2010–2017 at a single tertiary care institution who received SRS and at least one anti-PD(L)-1 inhibitor. Primary endpoints included the rate of best intracranial response (either a complete [CR] or partial response [PR]), rate of intracranial benefit (percentage of patients with stable disease [SD] for at least 6 months), and overall objective response rates. RESULTS: 150 patients met the eligibility criteria for this study and the median follow-up time was 10 months (Range: 1–130 months). Only a minority of patients from this cohort would have been eligible for comparative systemic therapy alone trials: 8 (CheckMate 204, CM), 8 (ABC), and 50 patients (Yale Institutional trial, YI). Best intracranial objective response rates for these trial eligible patients were 64% (CM), 64% (ABC), and 72% (YI), respectively. The 6-month rates of intracranial benefit across patients were 75% (CM), 75% (ABC), and 93% (YI), respectively. The rates of either a CR, PR, or SD across trial eligible patients were 100% (CM), 100% (ABC), and 94% (YI), respectively. More than 50% of patients experienced out-of-field progression in each of the patient subgroups. CONCLUSIONS: Over a seven year period, very few patients treated with SRS and anti-PD(L)-1 inhibitor therapy in a busy academic practice would have been eligible for systemic therapy alone per trial eligibility criteria. In these highly favorable subgroups, patients experience high rates of objective response, best intracranial response, and intracranial benefit after SRS and anti-PD(L)-1 therapy, representing optimal brain metastasis management.

RADI-25. REPEAT STEREOTACTIC RADIOSURGERY FOR RECURRENT BRAIN METASTASES

INTRODUCTION: Mainstream modality of treatment of oligo-metastatic disease is stereotactic radiosurgery (SRS). While the local control rate is nearing 90%, 7% of lesions and 10 to 15% of patients develop radiation necrosis post treatment. In the face of increasing lesion size and evidence of recurrence, re-treatment of the enlarging lesion with radiosurgery can be attempted. The aim of the project is to evaluate outcomes of lesions treated with repeat SRS. METHODS: We conducted a retrospective review of all patients that were treated with Gamma Knife radiosurgery at our institution from 2000 to 2018. Fifty-one lesions in 39 patients were identified that had recurrence during follow-up period and were treated with a second single-fraction SRS. RESULTS: A combination of imaging studies, such as PET and/or perfusion studies, lesion biopsy, and clinical course were used to make the diagnosis of lesion recurrence. The average radiation dose at first treatment was 21 Gy, and the average dose at second treatment was 19 Gy. The median time between treatments was 16.8 months, ranging from 2.5 to 75.3 months, and the median follow-up after second treatment was 10.2 months. Of 51 lesions that received two SRS treatments, 49% (25 lesions) continued to progress at a median interval of 4.8 months post treatment, of which 35% (18 lesions) were diagnosed as radiation necrosis based on biopsy results or advanced brain imaging. The overall rate of radiation necrosis post second SRS treatment was determined to be 16% per lesion and 21% per patient. CONCLUSION: Recurrent brain metastases that are re-treated with single fraction SRS are associated with a higher risk of radiation necrosis. Alternative treatment strategies, including fractionation of subsequent SRS treatments, radiation dose reduction, and combination with laser ablation could be considered to ensure symptom and disease control to reduce the rate of subsequent radiation necrosis.