Impact of systemic therapy regimen on survival of PCNSL

2070

Background: Primary CNS Lymphoma (PCNSL) is a rare and often fatal disease. Treatment includes multi-agent systemic therapy with a backbone of high-dose methotrexate (HD-MTX). Despite multiple drug and radiotherapy combinations for induction and consolidation treatment there remains no clear standard of care. The purpose of this analysis is to evaluate how varying treatment approaches impacted clinical outcomes at our institution. Methods: Data retrospectively collected for 95 consecutive patients with PCNSL pathologically confirmed from 2002 to 2021. Primary endpoint was OS with secondary endpoints of PFS and LC. Progression based on RANO criteria. Kaplan-Meier analyses, Log-rank test and Cox proportional hazard models used for time to event endpoints. MVA by backward selection applying an alpha of 0.2 for associations with 1st line chemo agents, number of cycles of HD-MTX (>6 or 0-5), size of enhancing tumor at presentation, CSF cytology, type of surgery (biopsy, STR, or GTR), and use of WBRT. Results: Most patients had KPS >70 (64.2%), were HIV negative (89.5%), and had no history of solid organ transplant (95.8%). Diagnosis was made by biopsy (73.7%) or resection (GTR 13.7%, STR 12.6%). 54.3% had <14 cc contrast-enhancing tumor volume (median 12.6 cc, range 0.5 - 67.8 cc) and 48.6% had single enhancing lesion. Of the 62 patients treated first line with at least 1 cycle of HD-MTX, 61.3% were treated with HD-MTX + Rituximab (R) and 33.9% with HD-MTX + R + temozolomide (TMZ). With or after induction HD-MTX, 1-3 patients received one or a combination of cytarabine, thiotepa, procarbazine, vincristine, carmustine, or ASCT. Of the 60 patients with evaluable CSF, 30.0% had positive cytology. IT chemotherapy (ITc) was administered to 12 patients (5 with + cytology, 4 with - cytology, 3 with unknown cytology). WBRT for consolidation after chemotherapy used for 3 patients and as monotherapy for 9 patients. 2-year OS and PFS rate was 50.1% (95% CI 38.6%-60.5%) and 38.5% (95% CI 27.9%-49.0%). On MVA, > 6 cycles of MTX was associated with superior OS, PFS, and LC. For patients receiving any chemotherapy, addition of R was associated with inferior OS while ITc was associated with improved OS, PFS, and LC (Table). There was no OS association on MVA with TMZ, GTR, consolidation WBRT, or size or number of initial lesions (p>0.05). Conclusions: Completion of induction HD-MTX and use of ITc was associated with better outcomes in this population. Incorporation of R into 1st line therapy was associated with worse OS. Survival remained poor throughout the study period, underscoring importance of further innovation. [Table: see text]

The timing of chemoradiotherapy after surgical resection and its impact on overall survival in glioblastoma

2051

Background: Prior studies examining time to initiate chemoradiotherapy (CRT) after surgical resection (S) in glioblastoma (GBM) have not provided clear consensus on its clinical impact. We sought to evaluate the effect that differential timing of adjuvant therapy may have on overall survival (OS). Methods: With the National Cancer Database (NCDB), patients (pts) with GBM who underwent S and adjuvant CRT from 2004-2013 were analyzed. Analysis was performed for the entire cohort as well as by Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis (RPA) classes (i.e. I, II, and III). Time from S to CRT was grouped weekly (i.e. 0-1, 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-8, and > 8 weeks). Pts were excluded if they died within the first 8 weeks to account for immortal time bias. Kaplan-Meier analysis, log-rank testing, and multivariate (MVA) Cox proportional hazards regression were performed with OS as the primary outcome. Results: A total of 30,414 pts were included for analysis. RPA class I, II, and III contained 903, 4,347, and 25,164 pts, respectively. The most common time to initiate CRT was week 4-5 (n = 7389), and this group served as reference for survival analysis. On MVA, weeks 0-1 (hazard ratio [HR] 1.18, 95% confidence interval [CI] 1.02-1.35), 1-2 (HR 1.24, CI 1.17-1.32), and 2-3 (HR 1.11, CI 1.07-1.15) demonstrated worse OS (all p < 0.03). For RPA class I pts, week 1-2 (HR 2.07, CI 1.08-3.95) was associated with worse OS (p = 0.028). For RPA class II pts, weeks 1-2 (HR 1.34, CI 1.14-1.57), 2-3 (HR 1.18, CI 1.07-1.31), and 3-4 (HR 1.10, CI 1.0-1.21) were associated with worse OS (all p < 0.05). For RPA class III pts, weeks 0-1 (HR 1.18, CI 1.02-1.38), 1-2 (HR 1.22, CI 1.14-1.3), and 2-3 (HR 1.09, CI 1.05-1.14) were associated with worse OS (all p < 0.03). No time point after week 5 was associated with change in OS for the overall cohort or any RPA class subgroup. Conclusions: These data provide insight into the optimal timing of CRT in GBM and describe RPA-class specific outcomes. In general, OS was negatively impacted if CRT started less than 3 weeks from S. Waiting up to 8 weeks, however, was not detrimental to OS and suggests delaying CRT beyond week 4-5 should be considered if clinically indicated without undue concern.

Is less more? Comparing chemotherapy alone to chemotherapy and radiation for high risk, grade II glioma—An analysis of the National Cancer Data Base

2031

Background: Grade II glioma patients with subtotal resection (STR) or age ≥ 40 are considered high risk. RTOG 9802 demonstrated that for these high-risk patients, chemotherapy and radiation therapy improved overall survival (OS) compared to radiation alone. The purpose of this study is to compare the OS of high risk, grade II glioma patients treated with adjuvant chemotherapy alone (CA) to chemotherapy and radiation therapy (CRT). Methods: Using the National Cancer Data Base (NCDB), high risk (age ≥ 40 or STR) grade II glioma patients with oligodendroglioma, astrocytoma, or mixed tumors were identified. Patients receiving CA were compared to patients receiving CRT. Univariate and multivariable analyses (MVA) were performed. Propensity score (PS) matching was utilized to account for difference in patient characteristics. Kaplan Meier statistics were utilized to compare OS. Results: 1054 high risk, grade II glioma patients were identified, 47.1% receiving CA and 52.9% receiving CRT. Median follow up time was 55.1 months. Patients treated with CA were statistically more likely (all p < 0.05) to be oligodendroglioma histology (65.5% vs. 34.2%), 1p/19 co-deleted (22.8% vs. 7.5%), younger median age (47 vs. 48 years) and treated at an academic program (65.2% vs. 50.3%). MVA demonstrated treatment type was not a significant predictor for OS (p = 0.125), while tumor size > 6cm, astrocytoma histology, and older age were predictors for worse survival (all p < 0.05). Utilizing 1:1 PS matching, with 662 total patients, OS was statistically similar (p = 0.919) for CA and CRT at 5 years (69.1% vs. 68.5%, respectively) and 7 years (55.5% vs. 60.0%, respectively). Conclusions: In this retrospective analysis of the NCDB, long term OS for high-risk, grade II glioma patients treated with CA appears similar to CRT. These findings are hypothesis generating, with the standard of care still remaining CRT as established by RTOG 9802. Prospective clinical trials comparing CA and CRT are warranted.

Using proton MRSI to predict response to vorinostat treatment in recurrent GBM

3055

Background: A major impediment to the development of new therapies for glioblastoma (GBM) is a lack of biomarkers indicating response. Epigenetic modifications are now recognized as a frequent occurrence in the early phases of tumorigenesis, playing a central role in tumor development. Epigenetic alterations differ significantly from genetic modifications in that they may be reversed by ‘‘epigenetic drugs’’ such as histone deacetylase inhibitors (HDACis). As a promising new modality for cancer therapy, the first generation of HDACi is currently being tested in phase I/II clinical trials. Methods: GBM alterations from therapy with HDACis, such as vorinostat (SAHA), include tumor redifferentiation/cytostasis rather than tumor size reduction limits the utility of traditional imaging methods such as MRI. Magnetic resonance spectroscopic imaging (MRSI) quantitates various metabolite levels in tumor and normal brain, allowing characterization of metabolic processes in live tissue. Results: In our preclinical model, MRS detected metabolic response to SAHA after only 3 days of treatment: reduced alanine and lactate and elevated myo-inositol, N-acetyl aspartate and creatine; each returning toward normal brain levels. This led to our clinical study of MRSI to evaluate the metabolic response of recurrent GBMs to SAHA + temozolomide. After only 7 days of SAHA treatment, MRSI can distinguish metabolic responders (normalization/restoration of tumor metabolites towards normal brain-like metabolism) from non-responders (no significant change in tumor metabolites). Our initial cohort (n=6) consists of 3 responders and 3 non-responders with highly significant differences in their change in metabolite levels (p < 0.001). Conclusions: Our results provide exciting insights into the mechanisms by which HDACi exerts its effect on GBMs. Tumor cells have increased biosynthetic needs requiring reprogramming of cellular metabolism. This creates increased energy demands, making tumor cells even more vulnerable to interventions targeting their metabolism. HDACi may induce redifferentiation in tumors by targeting tumor metabolism. Thus, MRSI provides a novel modality to predict response to HDACi-containing combination therapy in GBM.