Unique brain injury patterns after proton vs photon radiotherapy for WHO grade 2-3 gliomas.

BACKGROUND: Central nervous system (CNS) injury following brain-directed radiotherapy remains a major challenge. Proton radiotherapy (PRT) minimizes radiation to healthy brain, potentially limiting sequelae. We characterized CNS radiotoxicity, including radiation-induced leukoencephalopathy (RIL), brain tissue necrosis (TN), and cerebral microbleeds (CMB), in glioma patients treated with PRT or photons (XRT). PATIENTS AND METHODS: Thirty-four patients (19 male; median age 39.6 years) with WHO grade 2-3 gliomas treated with partial cranial radiotherapy (XRT [n = 17] vs PRT[n = 17]) were identified and matched by demographic/clinical criteria. Radiotoxicity was assessed longitudinally for 3 years post-radiotherapy via serial analysis of T2/FLAIR- (for RIL), contrast-enhanced T1- (for TN), and susceptibility (for CMB)-weighted MRI sequences. RIL was rated at whole-brain and hemispheric levels using a novel Fazekas scale-informed scoring system. RESULTS: The scoring system proved reliable (ICC > 0.85). Both groups developed moderate-to-severe RIL (62%[XRT]; 71%[PRT]) within 3 years; however, XRT was associated with persistent RIL increases in the contralesional hemisphere, whereas contralesional hemispheric RIL plateaued with PRT at 1-year post-radiotherapy (t = 2.180; P = .037). TN rates were greater with PRT (6%[XRT] vs 18%[PRT]; P = ns). CMB prevalence (76%[XRT]; 71%[PRT]) and burden (mean #CMB: 4.0[XRT]; 4.2[PRT]) were similar; however, XRT correlated with greater contralesional hemispheric CMB burden (27%[XRT]; 17%[PRT]; X2 = 4.986; P = .026), whereas PRT-specific CMB clustered at the radiation field margin (X2 = 14.7; P = .002). CONCLUSIONS: CNS radiotoxicity is common and progressive in glioma patients. Injury patterns suggest radiation modality-specificity as RIL, TN, and CMB exhibit unique spatiotemporal differences following XRT vs PRT, likely reflecting underlying dosimetric and radiobiological differences. Familiarity with such injury patterns is essential to improve patient management. Prospective studies are needed to validate these findings and assess their impacts on neurocognitive function.

Impact of TP53 mutations on brain metastasis control in non-small cell lung cancer patients undergoing stereotactic radiosurgery.

Purpose: To investigate whether TP53 variants may be correlated with overall survival and local control following stereotactic radiosurgery (SRS) for brain metastases (BMs) from non-small cell lung cancer (NSCLC). Methods: Patients undergoing an initial course of SRS for NSCLC brain metastases between 1/2015 and 12/2020 were retrospectively identified. Overall survival and freedom from local intracranial progression (FFLIP) were estimated via Kaplan-Meier method. Cox models assessed TP53 variant status (pathogenic variant, PV; variant not detected, ND). Results: 255 patients underwent molecular profiling for TP53, among whom 144 (56%) had a TP53 PV. Median follow-up was 11.6 months. OS was not significantly different across TP53 status. A trend toward superior FFLIP was observed for PV (95% CI 62.9 months-NR) versus ND patients (95% CI 29.4 months-NR; p=0.06). Superior FFLIP was observed for patients with one TP53 variant versus those with TP53 ND. Conclusion: Among NSCLC patients with BMs, the potential association between TP53 status and post-SRS FFLIP warrants further investigation in a larger prospective cohort.

Dose-Incorporated Deep Ensemble Learning for Improving Brain Metastasis Stereotactic Radiosurgery Outcome Prediction.

PURPOSE: To develop a novel deep ensemble learning model for accurate prediction of brain metastasis (BM) local control outcomes after stereotactic radiosurgery (SRS). METHODS AND MATERIALS: A total of 114 brain metastases (BMs) from 82 patients were evaluated, including 26 BMs that developed biopsy-confirmed local failure post-SRS. The SRS spatial dose distribution (Dmap) of each BM was registered to the planning contrast-enhanced T1 (T1-CE) magnetic resonance imaging (MRI). Axial slices of the Dmap, T1-CE, and planning target volume (PTV) segmentation (PTVseg) intersecting the BM center were extracted within a fixed field of view determined by the 60% isodose volume in Dmap. A spherical projection was implemented to transform planar image content onto a spherical surface using multiple projection centers, and the resultant T1-CE/Dmap/PTVseg projections were stacked as a 3-channel variable. Four Visual Geometry Group (VGG-19) deep encoders were used in an ensemble design, with each submodel using a different spherical projection formula as input for BM outcome prediction. In each submodel, clinical features after positional encoding were fused with VGG-19 deep features to generate logit results. The ensemble's outcome was synthesized from the 4 submodel results via logistic regression. In total, 10 model versions with random validation sample assignments were trained to study model robustness. Performance was compared with (1) a single VGG-19 encoder, (2) an ensemble with a T1-CE MRI as the sole image input after projections, and (3) an ensemble with the same image input design without clinical feature inclusion. RESULTS: The ensemble model achieved an excellent area under the receiver operating characteristic curve (AUCROC: 0.89 ± 0.02) with high sensitivity (0.82 ± 0.05), specificity (0.84 ± 0.11), and accuracy (0.84 ± 0.08) results. This outperformed the MRI-only VGG-19 encoder (sensitivity: 0.35 ± 0.01, AUCROC: 0.64 ± 0.08), the MRI-only deep ensemble (sensitivity: 0.60 ± 0.09, AUCROC: 0.68 ± 0.06), and the 3-channel ensemble without clinical feature fusion (sensitivity: 0.78 ± 0.08, AUCROC: 0.84 ± 0.03). CONCLUSIONS: Facilitated by the spherical image projection method, a deep ensemble model incorporating Dmap and clinical variables demonstrated excellent performance in predicting BM post-SRS local failure. Our novel approach could improve other radiation therapy outcome models and warrants further evaluation.

Artificial intelligence in neurology: opportunities, challenges, and policy implications.

Neurological conditions are the leading cause of disability and mortality combined, demanding innovative, scalable, and sustainable solutions. Brain health has become a global priority with adoption of the World Health Organization's Intersectoral Global Action Plan in 2022. Simultaneously, rapid advancements in artificial intelligence (AI) are revolutionizing neurological research and practice. This scoping review of 66 original articles explores the value of AI in neurology and brain health, systematizing the landscape for emergent clinical opportunities and future trends across the care trajectory: prevention, risk stratification, early detection, diagnosis, management, and rehabilitation. AI's potential to advance personalized precision neurology and global brain health directives hinges on resolving core challenges across four pillars-models, data, feasibility/equity, and regulation/innovation-through concerted pursuit of targeted recommendations. Paramount actions include swift, ethical, equity-focused integration of novel technologies into clinical workflows, mitigating data-related issues, counteracting digital inequity gaps, and establishing robust governance frameworks balancing safety and innovation.

Effects of Ataxia-Telangiectasia Mutated Variants on Radionecrosis and Local Control After Stereotactic Radiation Surgery for Non-Small Cell Lung Cancer Brain Metastases.

Purpose: Genetic variants affecting the radiation response protein ataxia-telangiectasia mutated (ATM) have been associated with increased adverse effects of radiation but also with improved local control after conventional radiation therapy. However, it is unknown whether ATM variants affect rates of radionecrosis (RN) and local intracranial progression (LIP) after stereotactic radiosurgery (SRS) for brain metastases. Methods and Materials: Patients undergoing an initial course of SRS for non-small cell lung cancer (NSCLC) brain metastases at a single institution were retrospectively identified. Kaplan-Meier estimates were calculated and Cox proportional hazards testing was performed based on ATM variant status. Results: A total of 541 patients completed SRS for brain metastasis secondary to NSCLC, of whom 260 completed molecular profiling. Variants of ATM were identified in 36 cases (13.8%). Among patients who completed molecular profiling, RN incidence was 4.9% (95% CI, 1.6%-8.2%) at 6 months and 9.9% (95% CI, 4.8%-15.0%) at 12 months. Incidence of RN was not significantly increased among patients with ATM variants, with an RN incidence of 5.3% (95% CI, 0.0%-15.3%) at both 6 and 12 months (P = .46). For all patients who completed genomic profiling, LIP was 5.4% (95% CI, 2.4%-8.4%) at 6 months and 9.8% (5.5%-14.1%) at 12 months. A significant improvement in LIP was not detected among patients with ATM variants, with an LIP incidence of 3.1% (0.0%-9.1%) at both 6 and 12 months (P = .26). Although differences according to ATM variant type (pathologic variant or variant of unknown significance) did not reach significance, no patients with ATM pathologic variants experienced LIP. Conclusions: We did not detect significant associations between ATM variant status and RN or LIP after SRS for NSCLC brain metastases. The current data set allows estimation of patient cohort sizes needed to power future investigations to identify genetic variants that associate with significant differences in outcomes after SRS.

Quantifying U-Net uncertainty in multi-parametric MRI-based glioma segmentation by spherical image projection.

BACKGROUND: Uncertainty quantification in deep learning is an important research topic. For medical image segmentation, the uncertainty measurements are usually reported as the likelihood that each pixel belongs to the predicted segmentation region. In potential clinical applications, the uncertainty result reflects the algorithm's robustness and supports the confidence and trust of the segmentation result when the ground-truth result is absent. For commonly studied deep learning models, novel methods for quantifying segmentation uncertainty are in demand. PURPOSE: To develop a U-Net segmentation uncertainty quantification method based on spherical image projection of multi-parametric MRI (MP-MRI) in glioma segmentation. METHODS: The projection of planar MRI data onto a spherical surface is equivalent to a nonlinear image transformation that retains global anatomical information. By incorporating this image transformation process in our proposed spherical projection-based U-Net (SPU-Net) segmentation model design, multiple independent segmentation predictions can be obtained from a single MRI. The final segmentation is the average of all available results, and the variation can be visualized as a pixel-wise uncertainty map. An uncertainty score was introduced to evaluate and compare the performance of uncertainty measurements. The proposed SPU-Net model was implemented on the basis of 369 glioma patients with MP-MRI scans (T1, T1-Ce, T2, and FLAIR). Three SPU-Net models were trained to segment enhancing tumor (ET), tumor core (TC), and whole tumor (WT), respectively. The SPU-Net model was compared with (1) the classic U-Net model with test-time augmentation (TTA) and (2) linear scaling-based U-Net (LSU-Net) segmentation models in terms of both segmentation accuracy (Dice coefficient, sensitivity, specificity, and accuracy) and segmentation uncertainty (uncertainty map and uncertainty score). RESULTS: The developed SPU-Net model successfully achieved low uncertainty for correct segmentation predictions (e.g., tumor interior or healthy tissue interior) and high uncertainty for incorrect results (e.g., tumor boundaries). This model could allow the identification of missed tumor targets or segmentation errors in U-Net. Quantitatively, the SPU-Net model achieved the highest uncertainty scores for three segmentation targets (ET/TC/WT): 0.826/0.848/0.936, compared to 0.784/0.643/0.872 using the U-Net with TTA and 0.743/0.702/0.876 with the LSU-Net (scaling factor = 2). The SPU-Net also achieved statistically significantly higher Dice coefficients, underscoring the improved segmentation accuracy. CONCLUSION: The SPU-Net model offers a powerful tool to quantify glioma segmentation uncertainty while improving segmentation accuracy. The proposed method can be generalized to other medical image-related deep-learning applications for uncertainty evaluation.

Long-Term Intracranial Outcomes With Combination Dual Immune-Checkpoint Blockade and Stereotactic Radiosurgery in Patients With Melanoma and Non-Small Cell Lung Cancer Brain Metastases.

PURPOSE: The intracranial benefit of offering dual immune-checkpoint inhibition (D-ICPI) with ipilimumab and nivolumab to patients with melanoma or non-small cell lung cancer (NSCLC) receiving stereotactic radiosurgery (SRS) for brain metastases (BMs) is unknown. We hypothesized that D-ICPI improves local control compared with SRS alone. METHODS AND MATERIALS: Patients with melanoma or NSCLC treated with SRS from 2014 to 2022 were evaluated. Patients were stratified by treatment with D-ICPI, single ICPI (S-ICPI), or SRS alone. Local recurrence, intracranial progression (IP), and overall survival were estimated using competing risk and Kaplan-Meier analyses. IP included both local and distant intracranial recurrence. RESULTS: Two hundred eighty-eight patients (44% melanoma, 56% NSCLC) with 1,704 BMs were included. Fifty-three percent of patients had symptomatic BMs. The median follow-up was 58.8 months. Twelve-month local control rates with D-ICPI, S-ICPI, and SRS alone were 94.73% (95% CI, 91.11%-96.90%), 91.74% (95% CI, 89.30%-93.64%), and 88.26% (95% CI, 84.07%-91.41%). On Kaplan-Meier analysis, only D-ICPI was significantly associated with reduced local recurrence (P = .0032). On multivariate Cox regression, D-ICPI (hazard ratio [HR], 0.4003; 95% CI, 0.1781-0.8728; P = .0239) and planning target volume (HR, 1.022; 95% CI, 1.004-1.035; P = .0059) correlated with local control. One hundred seventy-three (60%) patients developed IP. The 12-month cumulative incidence of IP was 41.27% (95% CI, 30.27%-51.92%), 51.86% (95% CI, 42.78%-60.19%), and 57.15% (95% CI, 44.98%-67.59%) after D-ICPI, S-ICPI, and SRS alone. On competing risk analysis, only D-ICPI was significantly associated with reduced IP (P = .0408). On multivariate Cox regression, D-ICPI (HR, 0.595; 95% CI, 0.373-0.951; P = .0300) and presentation with >10 BMs (HR, 2.492; 95% CI, 1.668-3.725; P < .0001) remained significantly correlated with IP. The median overall survival after D-ICPI, S-ICPI, and SRS alone was 26.1 (95% CI, 15.5-40.7), 21.5 (16.5-29.6), and 17.5 (11.3-23.8) months. S-ICPI, fractionation, and histology were not associated with clinical outcomes. There was no difference in hospitalizations or neurologic adverse events between cohorts. CONCLUSIONS: The addition of D-ICPI for patients with melanoma and NSCLC undergoing SRS is associated with improved local and intracranial control. This appears to be an effective strategy, including for patients with symptomatic or multiple BMs.

Long-term outcomes with reduced-dose whole-brain radiotherapy and a stereotactic radiosurgery boost for primary central nervous system lymphoma.

Background: Primary central nervous system lymphoma (PCNSL) is an aggressive diffuse large B-cell lymphoma. Treatment approaches are historically associated with neurotoxicity, particularly with high-dose whole-brain radiotherapy (WBRT). We hypothesized that reduced dose-WBRT (rd-WBRT) followed by a stereotactic radiosurgery (SRS) boost could provide durable disease control without significant adverse effects. Methods: We retrospectively reviewed PCNSL patients treated with rd-WBRT plus an SRS boost at Duke University between 2008 and 2021. Progression-free survival and overall survival (OS) were estimated using competing risk and Kaplan-Meier methods. Results: We identified 23 patients with pathologically confirmed PCNSL. Median age at diagnosis was 69 years (Q1Q3: 52-74) and median Karnofsky Performance Scale (KPS) was 80 (Q1Q3: 70-80). Median follow-up was 21 months. Median doses for rd-WBRT and SRS were 23.4 Gy (Q1Q3: 23.4-23.4) and 12 Gy (Q1Q3: 12-12.5), respectively. The cumulative incidence of intracranial progression at 2 years was 23% (95% CI: 8-42). Six patients (26%) developed distant radiographic progression while 2 patients (9%) developed both distant and local progression. Ten patients (44%) were alive without progression at last follow-up. By Kaplan-Meier estimate, the 2-year OS was 69% (95% CI: 46-84). There were no reported grade 3 + radiation-induced toxicities. Conclusions: The combination of rd-WBRT with an SRS boost appears well-tolerated with durable intracranial control. This approach may represent a treatment option for select patients, such as those with progressive or refractory disease. Further prospective studies are needed to validate these findings and determine whether this approach could be incorporated into consolidation strategies.

Mitigating Radiotoxicity in the Central Nervous System: Role of Proton Therapy.

OPINION STATEMENT: Central nervous system (CNS) radiotoxicity remains a challenge in neuro-oncology. Dose distribution advantages of protons over photons have prompted increased use of brain-directed proton therapy. While well-recognized among pediatric populations, the benefit of proton therapy among adults with CNS malignancies remains controversial. We herein discuss the role of protons in mitigating late CNS radiotoxicities in adult patients. Despite limited clinical trials, evidence suggests toxicity profile advantages of protons over conventional radiotherapy, including retention of neurocognitive function and brain volume. Modelling studies predict superior dose conformality of protons versus state-of-the-art photon techniques reduces late radiogenic vasculopathies, endocrinopathies, and malignancies. Conversely, potentially higher brain tissue necrosis rates following proton therapy highlight a need to resolve uncertainties surrounding the impact of variable biological effectiveness of protons on dose distribution. Clinical trials comparing best photon and particle-based therapy are underway to establish whether protons substantially improve long-term treatment-related outcomes in adults with CNS malignancies.

Novel Mechanisms and Future Opportunities for the Management of Radiation Necrosis in Patients Treated for Brain Metastases in the Era of Immunotherapy.

Radiation necrosis, also known as treatment-induced necrosis, has emerged as an important adverse effect following stereotactic radiotherapy (SRS) for brain metastases. The improved survival of patients with brain metastases and increased use of combined systemic therapy and SRS have contributed to a growing incidence of necrosis. The cyclic GMP-AMP (cGAMP) synthase (cGAS) and stimulator of interferon genes (STING) pathway (cGAS-STING) represents a key biological mechanism linking radiation-induced DNA damage to pro-inflammatory effects and innate immunity. By recognizing cytosolic double-stranded DNA, cGAS induces a signaling cascade that results in the upregulation of type 1 interferons and dendritic cell activation. This pathway could play a key role in the pathogenesis of necrosis and provides attractive targets for therapeutic development. Immunotherapy and other novel systemic agents may potentiate activation of cGAS-STING signaling following radiotherapy and increase necrosis risk. Advancements in dosimetric strategies, novel imaging modalities, artificial intelligence, and circulating biomarkers could improve the management of necrosis. This review provides new insights into the pathophysiology of necrosis and synthesizes our current understanding regarding the diagnosis, risk factors, and management options of necrosis while highlighting novel avenues for discovery.

Erratum to "E.C. Lerner, E.S. Srinivasan, G. Broadwater, et al. Factors Associated With New-Onset Seizures Following Stereotactic Radiosurgery for Newly Diagnosed Brain Metastases. Adv Radiat Oncol. 2022;7:101054."

[This corrects the article DOI: 10.1016/j.adro.2022.101054.].

A neural ordinary differential equation model for visualizing deep neural network behaviors in multi-parametric MRI-based glioma segmentation.

PURPOSE: To develop a neural ordinary differential equation (ODE) model for visualizing deep neural network behavior during multi-parametric MRI-based glioma segmentation as a method to enhance deep learning explainability. METHODS: By hypothesizing that deep feature extraction can be modeled as a spatiotemporally continuous process, we implemented a novel deep learning model, Neural ODE, in which deep feature extraction was governed by an ODE parameterized by a neural network. The dynamics of (1) MR images after interactions with the deep neural network and (2) segmentation formation can thus be visualized after solving the ODE. An accumulative contribution curve (ACC) was designed to quantitatively evaluate each MR image's utilization by the deep neural network toward the final segmentation results. The proposed Neural ODE model was demonstrated using 369 glioma patients with a 4-modality multi-parametric MRI protocol: T1, contrast-enhanced T1 (T1-Ce), T2, and FLAIR. Three Neural ODE models were trained to segment enhancing tumor (ET), tumor core (TC), and whole tumor (WT), respectively. The key MRI modalities with significant utilization by deep neural networks were identified based on ACC analysis. Segmentation results by deep neural networks using only the key MRI modalities were compared to those using all four MRI modalities in terms of Dice coefficient, accuracy, sensitivity, and specificity. RESULTS: All Neural ODE models successfully illustrated image dynamics as expected. ACC analysis identified T1-Ce as the only key modality in ET and TC segmentations, while both FLAIR and T2 were key modalities in WT segmentation. Compared to the U-Net results using all four MRI modalities, the Dice coefficient of ET (0.784→0.775), TC (0.760→0.758), and WT (0.841→0.837) using the key modalities only had minimal differences without significance. Accuracy, sensitivity, and specificity results demonstrated the same patterns. CONCLUSION: The Neural ODE model offers a new tool for optimizing the deep learning model inputs with enhanced explainability. The presented methodology can be generalized to other medical image-related deep-learning applications.

Risk of Tract Seeding Following Laser Interstitial Thermal Therapy for Brain Tumors.

BACKGROUND: The management of intracranial oncological disease remains a significant challenge despite advances in systemic cancer therapy. Laser interstitial thermal therapy (LITT) represents a novel treatment for local control of brain tumors through photocoagulation with a stereotactically implanted laser fiber. Because the use of laser interstitial thermal therapy continues to increase within neurosurgery, characterization of LITT is necessary to improve outcomes. OBJECTIVE: To quantify the risk of tumor seeding along the laser fiber tract in patients receiving LITT for primary or metastatic brain tumors at a high-volume treatment center. METHODS: We retrospectively reviewed all patients receiving LITT from 2015 to 2021 at our medical center. Patients with biopsy-confirmed tumors were included in this study. Tract seeding was identified as discontinuous, newly enhancing tumor along the LITT tract. RESULTS: Fifty-six patients received LITT for biopsy-confirmed tumors from 2015 to 2021, with tract seeding identified in 3 (5.4%). Twenty-nine (51.8%) patients had gliomas, while the remainder had metastases, of which lung was the most common histology (20 patients, 74%). Tract seeding was associated with ablation proceeding inward from superficial tumor margin closest to the cranial entry point ( P = .03). Patients with tract seeding had a shorter median time to progression of 1.1 (0.1-1.3) months vs 4.2 (2.2-8.6) months ( P = .03). CONCLUSION: Although the risk of tract seeding after LITT is reassuringly low, it is associated with decreased progression-free survival. This risk may be related to surgical technique or experience. Follow-up radiosurgery to the LITT tract has the potential to prevent this complication.

Factors Associated With New-Onset Seizures Following Stereotactic Radiosurgery for Newly Diagnosed Brain Metastases.

Purpose: Stereotactic radiosurgery (SRS) is a highly effective therapy for newly diagnosed brain metastases. Prophylactic antiepileptic drugs are no longer routinely used in current SRS practice, owing to a perceived low overall frequency of new-onset seizures and potential side effects of medications. It is nonetheless desirable to prevent unwanted side effects following SRS. Risk factors for new-onset seizures after SRS have not been well established. As such, we aimed to characterize variables associated with increased seizure risk. Methods and Materials: Patients treated with SRS for newly diagnosed brain metastases between 2013 and 2016 were retrospectively reviewed at a single institution. Data on baseline demographics, radiation parameters, and clinical courses were collected. Results: The cohort consisted of 305 patients treated with SRS without prior seizure history. Median age and baseline Karnofsky Performance Scale score were 64 years (interquartile range, 55-70) and 80 (interquartile range, 80-90), respectively. Twenty-six (8.5%) patients developed new-onset seizures within 3 months of SRS. There was no association between new-onset seizures and median baseline Karnofsky Performance Scale score, prior resection, or prior whole brain radiation therapy. There were significant differences in the combined total irradiated volume (12.5 vs 3.7 cm3, P < .001), maximum single lesion volume (8.8 vs 2.8 cm3, P = .003), lesion diameter (3.2 vs 2.0 cm, P = .003), and number of lesions treated (3 vs 1, P = .018) between patients with and without new-onset seizures, respectively. On multivariate logistic regression, total irradiated volume (odds ratio, 1.09 for every 1-cm1 increase in total volume; confidence interval, 1.02-1.17; P = .016) and pre-SRS neurologic symptoms (odds ratio, 3.08; 95% confidence interval, 1.19-7.99; P = .020) were both significantly correlated with odds of seizures following SRS. Conclusions: Our data suggest that larger total treatment volume and the presence of focal neurologic deficits at presentation are associated with new-onset seizures within 3 months of SRS. High-risk patients undergoing SRS may benefit from counseling or prophylactic antiseizure therapy.

Combination laser interstitial thermal therapy plus stereotactic radiotherapy increases time to progression for biopsy-proven recurrent brain metastases.

Background: Improved survival for patients with brain metastases has been accompanied by a rise in tumor recurrence after stereotactic radiotherapy (SRT). Laser interstitial thermal therapy (LITT) has emerged as an effective treatment for SRT failures as an alternative to open resection or repeat SRT. We aimed to evaluate the efficacy of LITT followed by SRT (LITT+SRT) in recurrent brain metastases. Methods: A multicenter, retrospective study was performed of patients who underwent treatment for biopsy-proven brain metastasis recurrence after SRT at an academic medical center. Patients were stratified by "planned LITT+SRT" versus "LITT alone" versus "repeat SRT alone." Index lesion progression was determined by modified Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) criteria. Results: Fifty-five patients met inclusion criteria, with a median follow-up of 7.3 months (range: 1.0-30.5), age of 60 years (range: 37-86), Karnofsky Performance Status (KPS) of 80 (range: 60-100), and pre-LITT/biopsy contrast-enhancing volume of 5.7 cc (range: 0.7-19.4). Thirty-eight percent of patients underwent LITT+SRT, 45% LITT alone, and 16% SRT alone. Median time to index lesion progression (29.8, 7.5, and 3.7 months [P = .022]) was significantly improved with LITT+SRT. When controlling for age in a multivariate analysis, patients treated with LITT+SRT remained significantly less likely to have index lesion progression (P = .004). Conclusions: These data suggest that LITT+SRT is superior to LITT or repeat SRT alone for treatment of biopsy-proven brain metastasis recurrence after SRT failure. Prospective trials are warranted to validate the efficacy of using combination LITT+SRT for treatment of recurrent brain metastases.

Resolution of radiation necrosis with bevacizumab following radiation therapy for primary CNS lymphoma.

IMPORTANCE: Radiation necrosis (RN) is a rare but serious adverse effect following treatment with radiation therapy. No standard of care exists for the management of RN, and efforts to prevent and treat RN are limited by a lack of insight into the pathomechanics and molecular drivers of RN. This case series describes the outcomes of treatment with bevacizumab (BV) in two primary CNS lymphoma (PCNSL) patients who developed symptomatic biopsy-proven RN after whole brain radiation (WBRT) with a stereotactic radiosurgery (SRS) boost. OBSERVATIONS: Patient 1 is a 52 year-old female with PCNSL treated with WBRT followed by an SRS boost. She developed symptomatic biopsy-proven RN, and initial treatment with tocopherol and pentoxifylline was unsuccessful. A 100% clinical and radiographic response was achieved with 4 cycles of BV. Patient 2, a 48 year-old male with PCNSL, presented with seizures and biopsy-proven RN after radiation therapy. Initial empiric treatment with tocopherol and pentoxifylline was unsuccessful. A 100% clinical and radiographic response was achieved with 3 cycles of BV. CONCLUSIONS AND RELEVANCE: Monitoring for RN in patients with PCNSL treated with radiation therapy is warranted. BV is an efficacious treatment and a viable alternative to corticosteroids or surgical intervention.

Central nervous system injury from novel cancer immunotherapies.

PURPOSE OF REVIEW: Neurotoxicity from antineoplastic treatment remains a challenge in oncology. Cancer treatment-induced central nervous system (CNS) injury can be therapy-limiting, severely disabling, and even fatal. While emerging cancer immunotherapies have revolutionized oncology during the past decade, their immunomodulatory properties can cause immune-related adverse effects (IRAE) across organ systems, including the nervous system. Central neurologic IRAEs from chimeric antigen receptor T cells (CAR-T) and immune checkpoint inhibitors (ICPI) are challenging complications of such therapies.We aim to provide clinicians with a comprehensive review of the relevant forms of CAR-T and ICPI-associated CNS toxicity, focusing on clinical features of such complications, diagnostic workup, predictive biomarkers, and management considerations in affected patients. RECENT FINDINGS: Unique forms of CAR-T and ICPI-related CNS toxicity have been characterized in the recent literature. CAR-T-related neurotoxicity is common and clinically well delineated. ICPI-related CNS toxicity is relatively rare but includes a heterogenous spectrum of severe and diagnostically challenging conditions. While putative risk factors, neurotoxicity biomarkers, imaging correlates and treatment strategies have been put forward, development of tailored diagnostic and management consensus guidelines awaits further clinical investigation. SUMMARY: As CAR-T and ICPI become more widely adopted, early recognition, documentation, and management of immunotherapy-related CNS toxicity are of paramount importance in the clinical setting.

Eosinophil and lymphocyte counts predict bevacizumab response and survival in recurrent glioblastoma.

BACKGROUND: There is a lack of biomarkers to identify glioblastoma (GBM) patients who may benefit from specific salvage therapies, such as the anti-angiogenic agent bevacizumab. We hypothesized that circulating blood counts may serve as biomarkers for treatment response and clinical outcomes. METHODS: Complete blood counts, clinical data, and radiographic information were collected retrospectively from 84 recurrent GBM patients receiving bevacizumab (10 mg/kg every 2 weeks). Significant biomarkers were categorized into quartiles and the association with clinical outcomes was assessed using the Kaplan-Meier method. RESULTS: The median treatment duration and survival on bevacizumab (OS-A) was 88 and 192 days, respectively. On multivariate analysis, MGMT promoter methylation (hazard ratio [HR] 0.504, P = .031), increases in red blood cells (HR 0.496, P = .035), and increases in eosinophils (HR 0.048, P = .054) during treatment predicted improved OS-A. Patients in the first and fourth quartiles of eosinophil changes had a 12-month survival probability of 5.6% and 41.2% (P < .0001), respectively. Treatment response was associated with increases in eosinophil counts (P = .009) and improved progression-free survival (P = .013). On multivariate analysis, increases in lymphocyte counts among responders predicted improved OS-A (HR 0.389, P = .044). Responders in the first and fourth quartiles of lymphocyte changes had a 12-month survival probability of 0% and 44.4% (P = .019), respectively. Changes in platelet counts differed before and after radiographic response (P = .014). CONCLUSIONS: Changes in circulating eosinophil, lymphocyte, and platelet counts may predict treatment response and clinical outcomes in patients with recurrent GBM receiving bevacizumab.