Radiation Necrosis Research Articles

Association between tumor location and toxicity outcomes after stereotactic radiosurgery for brain metastases.

Toxicities associated with stereotactic radiosurgery (SRS) are important when considering treatment and supportive management for patients with brain metastases. We herein assessed the association between brain metastasis location and risk of toxicity after SRS. We conducted a retrospective institutional review of patients treated with SRS for brain metastases between 2008 and 2023. Outcomes included radiation necrosis, seizure, local failure, and overall survival (OS). We reviewed 215 patients treated to 605 metastases (median diameter 10mm, IQR 5-17 mm), in the frontal (34%), cerebellar (19%), parietal (16%), temporal (13%), and occipital (13%) regions. Median follow-up was 16 months (IQR 7-36). New-onset seizures developed in 11% (19/174) of patients without prior seizure and was higher in patients with motor or sensory cortex lesions (12/48, 25%) on multivariate analysis (MVA, P = 0.02). SRS-related grade ≥ 2 symptomatic radionecrosis occurred in 6% (33/605) of lesions and correlated with larger metastasis volume (P < 0.001) and renal cell carcinoma histology (P < 0.05), while supratentorial location was nearly significant (MVA, P = 0.06). Median OS across all patients was 16 months (95% CI 12-20). Patients with symptomatic radiation necrosis had a longer median survival compared to those who did not (43 vs. 14 months, P = 0.002), which remained significant alongside Karnofsky performance status and extracranial disease on MVA. Brain metastasis location in the motor or sensory cortex is associated with increased risk of new-onset seizure following SRS and may warrant consideration of steroid and/or anti-epileptic prophylaxis. Symptomatic radiation necrosis is uncommon in the cerebellum and may be increasing with improvements in survival.

Outcomes of adjuvant radiation treatment following subtotal resection of world health organization grade II meningiomas.

Existing literature on adjuvant radiation after subtotal resection (STR) of WHO II meningiomas is limited by heterogenous patient cohorts, combining adjuvant and salvage radiation, gross total resection (GTR) and STR, primary radiation treatment vs. re-treatment, or grade II and III meningiomas, all of which have different expected outcomes. Tumor control estimates in a large homogenous patient cohort are needed to accurately counsel patients. A retrospective review of patients that had immediate post-operative imaging-confirmed residual WHO grade II meningioma followed by either adjuvant intensity-modulated radiation therapy (IMRT) or stereotactic radiosurgery (SRS) between 1996 and 2020 was conducted. Kaplan-Meier survival analysis and log-rank test was used to assess progression-free survival (PFS). Thirty-nine patients met inclusion criteria (IMRT = 32; SRS = 7). Overall, the 3-, 5-, and 10-year PFS was 81.1%, 61.2%, and 44.6%, respectively. Median follow-up time was 37 months. When comparing IMRT and SRS cohorts, baseline characteristics did not differ significantly between groups, but significantly larger residual tumor volumes were treated with IMRT (22.2 cm3 vs. 6.3 cm3, p = 0.004). PFS was not significantly different between IMRT and SRS at 3 years (81.1% vs. 80.0%) or 5 years (65.5% vs. 40%) (p = 0.19). There was no significant difference in radiation necrosis between groups (IMRT = 3/32 patients vs. SRS = 0/7 patients, p = 0.32). Our homogenous patient cohort displayed acceptable control rates at 3 years using SRS or IMRT as adjuvant therapy. No significant difference in PFS or radiation necrosis was noted between patients treated with adjuvant IMRT versus SRS.

BIOM-39. GENOMIC AND CLINICAL ASSOCIATIONS WITH RADIATION NECROSIS IN PATIENTS WITH DIFFUSE INFILTRATING GLIOMAS

Abstract BACKGROUND Radiation necrosis (RN) is a complication of radiotherapy (RT) that can significantly affect quality of life and outcomes in patients with gliomas. Clinical and molecular factors predisposing patients to RN are not clearly established. We hypothesize that clinicogenomic data can be leveraged to identify patients at high risk for RN. METHODS We identified patients from 2014-2024 with diffuse gliomas and RN diagnosed based on Advanced-Brain-Tumor-Imaging which includes conventional MRI, 3 perfusion parameters, spectroscopy and/or pathology at MD Anderson. Inclusion criteria included diagnosis of RN 12-weeks after RT and concordance of 4 ABTI parameters. For each case, we identified 2 controls who were gender-,diagnosis-, and age-matched using a nearest neighborhood algorithm. We evaluated the correlation between 15 most mutated genes in baseline tissue and occurance of RN. A logistic regression model was employed to assess the multivariable relationship between patient characteristics, gene mutations, and RN. RESULTS We identified 62 patients with RN; 50 glioblastoma (80.6%), 8 diffuse astrocytoma (12.9%) and 4 oligodendroglioma (6.5%). Time from RT to RN diagnosis was 254 days (IQR 154-483). 18(29%) of RN cases and 29(24.4%) of controls had received more than one course of RT. Patients treated with bevacizumab before RT had significantly lower odds of developing RN compared to those who did not receive bevacizumab (OR 0.103 [0.011 – 0.960], p=0.0459). BRAF mutations were more frequent in the RN cohort (8.1% vs. 2.4%) and patients with BRAF mutation were more likely to have RN than those without BRAF mutation (OR 6.842 [1.2, 39.01], p=0.0304). Age, gender, diagnosis, IDH, MGMT or 1p19q status were not associated with RN. CONCLUSION We found that bevacizumab administration was associated with a significant reduction in RN development. We also found a novel association between presence of BRAF mutations and increased risk of RN in glioma patients, which warrants further investigation.

RADT-05. VASCULAR OVERLAP PROVIDES PROTECTION FROM GAMMA-KNIFE-INDUCED RADIATION NECROSIS IN A METASTATIC BRAIN TUMOR COHORT

Abstract INTRODUCTION Radiation necrosis (RN) is a long-term complication of Gamma-Knife (GK) Radiosurgery, an established treatment for metastatic brain tumors. Few dosimetric parameters have shown to predict RN, including tumor target volume, treatment dose, and V12 Gy (volume of normal brain receiving 12 Gy). Vascularity has yet to be explored as an independent predictor of RN. OBJECTIVE To evaluate the role of vascular overlap in GK-induced RN. METHODS All metastatic brain tumor patients treated with GK between 2014-2019 were retrospectively screened for post-treatment radiographic RN. Vascular overlap, defined by the MCA-ACA and MCA-PCA border regions, was determined by two blinded neurosurgeons and one radiation oncologist. Multivariate logistic regression analysis was performed to determine factors predictive of RN using variables determined significant by univariate analysis and clinical relevance. Covariates included V12 as a binary variable (≥8 vs. &amp;lt;8 cm3), vascular overlap, supratentorial location, and eloquent location. RESULTS 6% (20/321) of all metastatic brain tumor patients developed RN and were included for analysis, with the final cohort being 60% (12/20) female with a mean age of 59 years, comprising 77 lesions. V12 (≥8 cm3) showed a significant increase in risk of developing RN [OR = 8.18; (coef, 95% CI: 2.10 (0.49, 3.72), p=0.011)]. Conversely, vascular overlap showed a protective effect against RN [OR = 0.23 (-1.49 (-2.79, -0.19), p=0.024)], with a lower rate of RN in areas of vascular overlap (10%) than non-vascular overlap (26%). Neither supratentorial location [OR = 3.31 (1.20 (-0.76, 3.16), p=0.231)] nor location within eloquent brain [OR = 0.96 (-0.04 (-1.48, 1.39), p=0.95)] showed significance. CONCLUSION For the first time, we show that patients with metastatic brain lesions in anatomical regions of vascular overlap treated with GK are less likely to develop RN. Future work should validate these findings with the goal of optimizing radiation dosing to mitigate RN incidence and recurrence rates.

NCMP-14. RADIOGRAPHIC RESPONSE TO BEVACIZUMAB IN TREATING RADIATION NECROSIS AND ASSOCIATED EDEMA IN BRAIN TUMORS

Abstract OBJECTIVE This study evaluates the efficacy of bevacizumab in treating radiation necrosis (RN) and associated perilesional edema in brain tumors. We aimed to measure changes in radiologic volumes and clinical performance following treatment. MATERIALS AND METHODS We retrospectively reviewed patients treated with radiation therapy and/or stereotactic radiosurgery between September 2017 and August 2023. We identified 10 metastatic brain tumors and one glioblastoma with symptomatic RN. Diagnosis was primarily through MRI, with one case confirmed pathologically. All patients received bevacizumab (7.5 mg/kg every 3 weeks). Volumes of RN and perilesional edema were measured using gadolinium-enhanced T1-weighted and T2-weighted/FLAIR MRI, respectively, analyzed with Brainlab software. Volume changes and clinical symptoms, graded by the ECOG Performance Status (PS) scale, were recorded. RESULTS The minimal follow-up was 4 months. The 11 patients received a median of 2 bevacizumab cycles (range, 2-3). Pre-treatment median volumes (ranges) were: tumor/RN 5.45 cc (2.66-13.56), perilesional edema (T2, FLAIR) 53.93 cc (7.46-115.32), 58.01 cc (15.18-131.2). Post-treatment median volumes (ranges) were: tumor/RN 2.5 cc (2.23-7.23), perilesional edema (T2, FLAIR) 4.05 cc (3.08-11.18), 7.38 cc (2.887-17.647). Median percentage volume reductions were: tumor 39%, RN 38%, perilesional edema (T2, FLAIR) 80%, 79%. Significant reductions in edematous volumes were observed. Pre-treatment ECOG PS median was 2 (range 2-3); post-treatment ECOG PS median remained 2 (range 2-3). Clinical performance improved in most patients, with two exception where deterioration occurred after the third cycle despite initial improvement. CONCLUSION Bevacizumab is effective in reducing RN and perilesional edema, particularly within the first two cycles of treatment. Most patients showed clinical improvement, though further research is needed to optimize treatment and understand long-term outcomes.

RADT-02. PRE-OPERATIVE VERSUS POST-OPERATIVE FRACTIONATED STEREOTACTIC RADIOTHERAPY: A SINGLE-INSTITUTION ANALYSIS OF 534 RESECTED METASTASES

Abstract Patients with large or symptomatic brain metastases and limited intracranial disease typically have surgery followed by post-operative (post-op) stereotactic radiosurgery (SRS). However, SRS can lead to elevated rates of radiation necrosis (RN), meningeal disease (MD), and local failure (LF). Fractionated treatments can deliver a higher biological effective dose and may reduce the risk of LF, and pre-operative (pre-op) treatments may reduce the risk of RN and MD through treating smaller volumes and tumor sterilization. We hypothesize that pre-op fractionated stereotactic radiation therapy (FSRT) will reduce the incidence rate of RN, MD, and LF when compared to patients who receive post-op FSRT. A retrospective analysis was performed at a single institution and included patients who had surgical resection and radiation to at least one brain metastasis. Patients with multiple metastases resected, either during the same surgery or at different times in the disease course, were eligible for inclusion. Outcomes were evaluated on a per-lesion basis. The primary outcome was a composite endpoint defined by 1) LF, 2) MD, and/or 3) Grade 2 or higher (symptomatic) RN. 458 patients with 534 resected brain metastases were eligible for analysis. 235 metastases received pre-op FSRT, and 299 metastases received post-op FSRT. Overall, 15% of patients had multiple brain metastases resected. Overall, 6 (2.6%) pre-op and 13 (4.3%) post-op patients experienced LF. 21 (8.9%) pre-op and 38 (12.7%) post-op patients experienced symptomatic RN. 11 (4.7%) pre-op and 29 (9.7%) post-op patients were diagnosed with MD (p=0.031). 14% and 24% of metastases that received pre-op and post-op FSRT, respectively, experienced the composite endpoint (p=0.005). In our study, pre-op FSRT compares favorably to post-op FSRT primarily due to a 50% reduction in the incidence of MD. Differences in symptomatic RN or LF were small on adjusted analyses. Prospective validation of pre-op FSRT is needed.

RADT-45. INTEGRATION OF FUNCTIONAL MAGNETIC RESONANCE IMAGING INTO STEREOTACTIC RADIOSURGERY PLANNING TO REDUCE SYMPTOMATIC RADIATION NECROSIS

Abstract BACKGROUND Radiation necrosis (RN) is a common complication of stereotactic radiosurgery (SRS). Functional magnetic resonance imaging (fMRI) localizes eloquent cortical areas for neurosurgical planning but is seldom applied to SRS. We explored the relationship between symptomatic RN and functional neuroanatomy on fMRI. We evaluated the potential benefit of reducing radiation dose to functional areas. METHODS From 2011 to 2022, patients with an fMRI within 3 months of SRS for resected or unresected brain metastases (BM) were included. The BM nearest to the primary motor, language, or visual cortices was designated as the index BM. The functional area on fMRI nearest to the index BM was designated as the fOAR. Focally symptomatic RN (FSRN) was defined as radiographic RN alongside new or worsening focal neurologic symptoms consistent with the fOAR. Patients treated with 5-fraction SRS and modern planning techniques after 2013 were included in the dosimetric analysis. Associations between fOAR dose and FSRN were evaluated with logistic regression. SRS replanning was performed to determine if fOAR dose could be reduced without sacrificing target coverage. RESULTS Among 94 patients, 82% received SRS postoperatively, and 71% received 5-fraction SRS. fOARs included the primary motor (77%), language (19%), and visual (2%) cortices. Of 20 patients with radiographic RN, 20% were neurologically asymptomatic, 20% had anatomically unrelated neurologic symptoms, and 60% had FSRN. The dosimetric analysis included 39 patients. The fOAR volume receiving a single-fraction equivalent dose ≥14 Gy (V14Gy) was associated with FSRN at 18 months post-SRS (OR=6.8/mL, p=0.05). Replanning reduced fOAR V14Gy in all cases. The mean reduction was 28%. CONCLUSION Neurologic symptoms of SRS-induced RN reflect nearby functional neuroanatomy on fMRI. Dose reduction to functional areas while maintaining target coverage is consistently achievable and may mitigate symptomatic RN. Integrated fMRI-based SRS planning may protect neurologic function when treating tumors near eloquent areas.

NIMG-63. LEVERAGING LLMS FOR ACCURATE DIFFERENTIATION OF RADIATION NECROSIS AND TUMOR PROGRESSION IN BRAIN MRI REPORTS: A STUDY ON AUTOMATED SCORING AND CLINICAL IMPLICATIONS

Abstract BACKGROUND Accurate differentiation between radiation necrosis and tumor progression or recurrence in patients treated with stereotactic radiosurgery for brain metastases is critical for guiding clinical management. This study leverages the advanced natural language processing capabilities of Meta Llama3, an artificial intelligence (AI) large language model (LLM), combined with prompt engineering, to rapidly categorize brain magnetic resonance imaging (MRI) radiology reports. Our objective was to develop an automated scoring system to classify concern for radiation necrosis, tumor progression, equivocal findings, or stable exams. METHODS Using a comprehensive dataset of reports annotated by expert radiologists, we ran inference on a 70-billion parameter Llama3 model (temperature 0.2, top_p 0.9) with specific prompts designed to capture the nuanced language and diagnostic criteria related to radiation necrosis or tumor progression. RESULTS The first pass was performed on a training dataset of 107 reports and did not predefine the clinical conditions. This demonstrated 43.4% accuracy in scoring when compared to a human user’s classification of each report. Agreement between the human reader and Llama3 was assessed using the Gwet agreement coefficient, AC1=0.411 (99%CI 0.396-0.426). Multiple iterations of targeted prompt engineering were then employed to narrow the definition of radiation necrosis and tumor progression, with specific examples and nuanced language used to achieve a higher degree of accuracy at 72.0%, AC1=0.719 (99%CI 0.717-0.722). DISCUSSION/FUTURE DIRECTIONS This surpasses recent demonstration of lower human-LLM agreement in radiographic score assignment, with further room for calibration on a dataset of several thousand reports. Next, we will correlate automated interpretations with actual clinical management decisions for radiation necrosis (e.g., initiation of steroids, bevacizumab, Laser Interstitial Thermal Therapy, and/or repeat imaging). This automated scoring system holds significant potential for LLMs in clinical applications. Future work will focus on integrating this model into clinical workflows and expanding its capabilities to include longitudinal monitoring of patient outcomes.

RADT-01. PRE-OPERATIVE SINGLE FRACTION VERSUS FRACTIONATED STEREOTACTIC RADIOTHERAPY: A SINGLE-INSTITUTION ANALYSIS OF 299 RESECTED METASTASES

Abstract The treatment standard for patients with large or symptomatic brain metastases and limited intracranial disease is surgical resection followed by post-operative (post-op) stereotactic radiosurgery (SRS). The multicenter PROPS-BM cohort showed how pre-operative (pre-op) SRS may lead to a reduced incidence of radiation necrosis (RN), local failure (LF), and meningeal disease (MD) compared to historical controls. Fractionated treatments can deliver a higher biological effective dose and may reduce the incidence of LF and MD. We hypothesize that pre-op fractionated stereotactic radiation therapy (FSRT) will reduce the incidence rate of RN, MD, and LF when compared to patients who receive pre-op SRS. Patients who had surgical resection and pre-operative radiation to at least one brain metastasis at a single institution were retrospectively analyzed. Outcomes were evaluated on a per-lesion basis. The primary outcome was a composite endpoint defined by 1) LF, 2) MD, and/or 3) Grade 2 or higher (symptomatic) RN. 260 patients with 299 resected brain metastases were eligible for analysis. 64 metastases received SRS and 235 metastases received FSRT. 38 patients had multiple metastases resected pre-operatively. The median gross tumor volume was 4 ccs for SRS and 10 ccs for FSRT (p&amp;lt;0.001). Overall, 4 (6.3%) SRS and 6 (2.6%) FSRT patients experienced LF. 4 (6%) SRS and 21 (8.9%) FSRT patients experienced Grade 2 or higher RN. 3 (4.7%) SRS and 11 (4.7%) FSRT patients were diagnosed with MD. 14% of both SRS and FSRT patients experienced the composite endpoint. There were no statistically significant differences in outcome between these two treatment groups. In our study, pre-op SRS and FSRT both appear to be safe and effective options to treat resectable brain metastases. It is important to prospectively compare pre-op SRS and FSRT in matched cohorts to assess any differences in treatment efficacy and toxicity.

RADT-17. FRACTIONATED REIRRADIATION OF RECURRENT IDH WILDTYPE GLIOBLASTOMA: HIGHER REIRRADIAITON DOSE AND SAFETY

Abstract The optimal regimen, normal tissue tolerances, and appropriate indications for reirradiation for recurrent glioblastoma (GBM) are uncertain. We evaluate outcomes and toxicity after fractionated reirradiation. Patients with recurrent IDH wildtype GBM after prior standard radiotherapy during 2007-2022 were retrospectively reviewed. All patients initially received conventionally fractionated radiotherapy. Radiation necrosis (RN) was determined by imaging or pathology. Overall survival (OS) was estimated using Kaplan-Meier methodology. Univariate and multivariate Cox proportional hazard models for OS were constructed. Bivariable associations with RN were estimated using fisher exact tests. 99 patients (median age 56 yrs, median KPS 80) were treated with reirradiation (median followup 7.4 mo). Median time between initial and reirradiation was 20 months (range, 2 – 214 mo). Median received dose at reirradiation was 41.4 Gy (30.6 – 60 Gy) to a median planning target volume (PTV) of 183 cc (16 – 825 cc). Median cumulative maximum dose to the optics and brainstem were 54 Gy (4 – 103 Gy) and 76 Gy (5 – 146 Gy), respectively. Median OS was 9 mo (95% CI 6.2 – 10.6 mo). On univariate analysis, time between initial and reirradiation ≥24 mo (HR 0.58, CI 0.36 – 0.94) and reirradiation dose ≥41.4 Gy (HR 0.56, CI 0.35 – 0.88) were associated with longer OS. On multivariate analysis, interval between radiation sessions (HR 0.56, CI 0.34 – 0.91) and higher reirradiation dose (HR 0.55, CI 0.35 – 0.87) remained associated with OS. No cases of radiation injury to the optics or brainstem were reported. 10% of patients experienced RN. Reirradiation dose ≥41.4 Gy, reirradiation PTV ≥200 cc and interval between radiation sessions &amp;lt; 24 mo were not significantly associated with RN. These data support the safety and efficacy of fractionated reirradiation for recurrent GBM including for large tumors. This study suggests that higher reirradiation dose may be feasible, including for large treatment volumes.

SURG-44. TO PROCEED OR NOT TO PROCEED: THE IMPACT OF FROZEN SECTION PATHOLOGY ON INTRAOPERATIVE BRAIN BRACHYTHERAPY IMPLANTATION FOR RECURRENT GLIOBLASTOMA

Abstract BACKGROUND Glioblastoma (GBM) is a highly aggressive malignancy that recurs in up to 90% of patients following standard-of-care therapy, including surgical resection, concurrent chemoradiotherapy, and subsequent adjuvant chemotherapy. In cases of prior radiation therapy (RT) use, one option for localized, recurrent GBM (RGBM) is STaRT (Surgically Targeted Radiation Therapy) (GammaTile, GT Medical Technologies, Tempe AZ USA), a form of brachytherapy utilizing cesium-131 seeds implanted at the time of re-resection. One of the potential concerns with STaRT implantation for RGBM is that intraoperative frozen section pathology may demonstrate only necrosis, in which case further RT may not be advisable. This concern is often cited as a deterrent to offering an intraoperative brachytherapy treatment in RGBM cases. Herein we review the frequency of intraoperative frozen section diagnosis returning radiation necrosis leading to aborted GammaTile placement. METHODS The GammaTile Registry, a prospectively enrolling multi-center phase 4 study (NCT04427384), was assessed for all patients receiving intraoperative brachytherapy for RGBM from 10/2020 to 3/2024. Data regarding patient demographics, procedure details, and pathology were extracted and analyzed. RESULTS A total of 122 RGBM patients with a median age of 59 years (range 28-86) were scheduled for STaRT. Of these, 109 (89.3%) patients had confirmed prior radiation with a median RT dose of 60.0Gy (range 16-75). Only two (1.6%) planned cases of GammaTile placement were aborted after intraoperative frozen section pathology demonstrated extensive radiation necrosis without viable tumor. The patients had received fractionated RT to 55.8Gy and 60.0Gy, respectively. While both frozen samples were predominantly necrosis, the final pathology demonstrated foci of residual glioblastoma in both cases. CONCLUSIONS While it is uncommon in RGBM to abort STaRT due to intraoperative pathology indicating only radiation necrosis, careful consideration should be given when aborting the procedure as in this series final pathology ultimately demonstrated viable tumor cells.

SURG-07. QUANTIFYING THE RISK OF SHORT TERM SEIZURES FOLLOWING LASER INTERSTITIAL THERMAL THERAPY

Abstract BACKGROUND Tumor-related epilepsy is a common sequela of both primary and metastatic brain tumors, as well as radiation necrosis (RN). Laser interstitial thermal therapy (LITT) is an effective option for lesional cytoreduction but induces transient peritumoral edema that typically resolves within 3 months of the procedure. Accordingly, we quantified the risk of seizure development early post-LITT in patients with tumors or RN of diverse histologies. METHODS Patients treated with LITT for brain tumors or radiation necrosis at a high-volume center from 2015 – 2021 with at least 3 months of clinical follow-up were evaluated for seizure risk and time to discontinuation of anti-epileptic drugs (AEDs) for at least 4 weeks, with censoring at death or last follow-up. RESULTS One-hundred and seventeen patients met inclusion criteria, and 22 (19%) experienced a post-LITT seizure within 3 months of surgery. Those with pre-LITT seizures (P = 0.03), prior whole-brain radiation (WBRT) (P = 0.0099), or lower pre-operative KPS (P = 0.04) were more likely to have a post-LITT seizure. Frontal lobe tumors appeared to be more frequently associated with seizures (68% vs 45%, P = 0.06). Prior WBRT, odds ratio (OR) 4.27 (95% CI 1.01 – 18.55), P = 0.047 and pre-LITT seizure, OR 5.39 (95% CI 1.87 – 16.31), P = 0.002 remained predictive on multivariate analyses. Successful AED weaning occurred at a median of 1.68 months (no seizure within 3-months group) versus undefined for patients experiencing seizures (P = 0.015). Post-LITT seizures within 3 months were associated with worsened post-LITT overall survival, 5.95 vs 17.2 months, P &amp;lt; 0.0001. DISCUSSION Prior seizure history or previous WBRT pose a risk of breakthrough seizures following laser ablation. Neurosurgeons should consider extended AED treatment or referral for management of tumor-related epilepsy in patients with these neurologic risk factors.

NIMG-05. ARTIFICIAL INTELLIGENCE-BASED RESPONSE ASSESSMENT FOR PATIENTS RECEIVING LASER INTERSTITIAL THERMAL THERAPY

Abstract BACKGROUND Laser interstitial thermal therapy (LITT) is a minimally invasive option for tissue diagnosis, cytoreduction, and rapid post-operative return to systemic therapies for patients with brain tumors or radiation necrosis. As ablated tissue remains in situ, post-LITT edema may be associated with apparent increased lesion size and transient clinical worsening, complicating assessment of progression. METHODS All patients receiving LITT at a single center for tumors or radiation necrosis from 2015 – 2023 with ≥ 9 months of MRI follow-up were retrospectively reviewed. A 3D U-net segmentation model implemented in nnU-Net was developed for automated segmentation of Contrast-enhancing Lesion Volume (CeLV) of LITT-treated lesions on T1-weighted post-contrast MR images. CeLVs were analyzed to establish volumetric post-LITT response assessments. RESULTS Across 384 unique MRI exams, 61LITT-treated lesions and 6 control cases of medically-managed radiation necrosis were analyzed. Automated segmentation was qualitatively accurate in 367/384 (95.6%) images. CeLV increased to a median of 68.3% (IQR 35.1 – 109.2%) from baseline at 1 – 3 months from LITT (P = 0.0012) and returned to baseline thereafter. Based on established criteria, volumetric disease progression was defined as lesion expansion 40% from volumetric nadir or baseline. Twenty-one of 56 (37.5%) patients experienced volumetric progression with a progression-free survival of 21.4 (6.0 – 93.4) months. Patients experiencing volumetric progression had lower mOS (17.3 vs 62.1 months, P = 0.0015). CONCLUSIONS We observed a stereotyped increase in CeLV at 1 – 3 months post-LITT, which resolved within 6 months of the procedure. Post-LITT disease response can be reliably and feasibly assessed with criteria that account for this transient CeLV expansion. Automated lesion segmentation may speed adoption of volumetric response assessment criteria to clinical practice.

NIMG-03. USE OF ARTERIAL SPIN LABELING TO DIFFERENTIATE ICTAL PHENOMENON FROM TUMOR PROGRESSION IN HIGH GRADE ASTROCYTOMA

Abstract A 39-year-old female with grade-3 left parietal IDH-mutant astrocytoma following resection presented with convulsive status. EEG showed electrographic seizures arising from the left parietal region. MRI-brain showed the left temporoparietal post-treatment bed without any contrast enhancement (Fig-1A). Arterial spin labeling (ASL) – MR perfusion showed markedly increased cerebral blood flow (CBF) in the left temporoparietal lobe (Fig-1 C, D) with corresponding diffusion restriction (Fig-1B). Repeat imaging two weeks after the seizure control showed complete resolution (Fig-2) suggesting the imaging abnormalities to be a peri-ictal phenomenon.1 ASL-MR perfusion is a highly sensitive marker of increased CBF with its emerging application in distinguishing tumor progression (increased CBF) from post-treatment changes such as radiation necrosis (reduced CBF).1,2 This case highlights the utility of ASL-MR perfusion in differentiating peri/post-ictal phenomenon (increased CBF without T1-contrast enhancement) from tumor progression (increased CBF with T1-contrast enhancement) when employed routinely in clinical oncology practice.

RADT-50. CLINICAL OUTCOME OF DAILY 5-FRACTIONATED GAMMA KNIFE RADIOSURGERY FOR LARGE BRAIN METASTASES

Abstract INTRODUCTION With the introduction of the Gamma Knife Icon, hypo-fractionated Gamma Knife radiosurgery (GKRS) for large brain metastases (LBMs) is beginning to be performed. This study aimed to assess the clinical outcome and risk of the daily 5-fractionated GKRS for LBMs. METHODS A total of 100 patients who underwent 5-fractionated GKRS for LBMs (&amp;gt; 14cm3) were enrolled. Forty-six were male patients, and the median age was 60 years. Most common primary cancer sites were lung (41 non-small cell cancers), breast (24), and kidney (14). The median tumor volume was 22cm3 (14 – 69cm3) and marginal dose was 35Gy (50% isodose line) in 5 fractions. Median follow-up duration was 18 months (3 – 72 months). RESULTS The last follow-up showed local tumor control in 74 cases. The cumulative 1, 2, and 3-year local tumor control rates were 73, 65, and 60%, respectively. Eighty-six tumors showed best MRI response in a year, and median tumor volume decrease rate was 80% (22 – 100%). Thirty patients experienced dramatically tumor volume decrease (&amp;gt; 95%). Median progression free and overall survival were 7 and 16 months, respectively. Radiation necrosis occurred in 24 patients, of which only 9 had neurological symptoms and treated with steroid and bevacizumab. CONCLUSIONS The daily 5-fractionated GKRS for very LBMs showed favorable local tumor control and acceptable radiation necrosis, but questionable survival benefit. In the future, prospective multi-center study is needed.

BIOM-72. INTEGRATION OF ULTRASENSITIVE ELECTROLUMINESCENT IMMUNOASSAY AND CELL-FREE DNAMETHYLATION ANALYSIS FOR NON-INVASIVE MONITORING OF ADULT DIFFUSE GLIOMAS

Abstract Gliomas are highly aggressive brain tumors with nearly universal recurrence rate. Despite this, the ability to accurately predict and identify tumor recurrence relies solely on serial MRI imaging, which is marred by treatment effects such as radiation necrosis, and necessarily identifies progression retrospectively. It is believe that tumor undergo molecular changes upon tumor progression, however the resampling of tumors upon recurrence imposes significant risks. This highlights the need for novel non-invasive biomarkers capable of identifying tumor recurrence. Due to the low accuracies of individual biomarkers, we have proposed the use of an integrated, multi-platform approach to biomarker discovery. A cohort of 107 glioma plasma samples, including 30 pairs, underwent plasma proteomic analysis, consisting of a panel of serum proteins (FABP4, GFAP, NFL, Tau and MMP3,4 &amp;7) quantified through ultrasensitive electrochemiluminescence multiplexed immunoassays, and plasma DNA methylation analysis, captured through cell-free methylated DNA immunoprecipitation and high-throughput sequencing. Unsupervised hierarchal clustering revealed robust separation of primary and recurrent tumors through plasma proteomics, associated with a distinct plasma methylation signature. NFL, Tau and MMP3 levels differed between primary and recurrent samples; pair-wise analysis revealed increased in NFL and Tau concentrations upon recurrence. Tau levels predicted outcome independent of WHO Grade and IDH status. A predictive generalized linear regression model created through the integration of the proteomic and methylation signatures allowed for the discrimination of primary and recurrent samples in 83% of cases. This work suggests that the combination of DNA methylation and plasma proteomics may improve the ability of these techniques for the serial monitoring of gliomas patients.

SURG-54. RECURRENT BRAIN METASTASES (RBM) TREATED WITH SURGICALLY TARGETED RADIATION THERAPY (START): PRELIMINARY SAFETY EVALUATION FROM A MULTI-INSTITUTIONAL PROSPECTIVE REGISTRY

Abstract BACKGROUND Resection and intraoperative brachytherapy for operable recurrent brain metastasis allows for pathologic confirmation of recurrent disease, mass effect relief, and immediate initiation of radiotherapy (RT). In this analysis, we report patterns-of-use and treatment-related adverse events (AEs) for rBM patients treated with Cs-131 collagen tiles, an FDA-cleared intracranial brachytherapy device. METHODS Patients with rBM who underwent resection and surgically-targeted radiation therapy (GammaTile, GT Medical Technologies Inc., Tempe, AZ USA) on a prospectively enrolling phase 4 registry study (NCT04427384) were analyzed. AEs were graded per CTCAE v5.0. RESULTS Between 11/2020 and 2/2024, 56 rBM in 51 consecutive patients underwent STaRT at 19 centers, with 5 patients having 2 metastases implanted concurrently. 44 patients (86%) had prior same-site RT (median interval 14.5 mo, range 3-56). Primary tumor histologies were lung (27), melanoma (8), breast (7), renal (4), colon (2), and “other” (3). Median pre-operative maximum diameter was 3.0 cm (range 1.4-5.7); age 63 (range 28-81); 53% females; KPS median 90 (range 40-100); and median implantation time 3 minutes. 26 patients were implanted at a 1st, 15 at a 2nd, and 10 at ≥ 3rd same-site recurrence (range 1-9). At a median follow-up of 6.2 months (range &amp;lt;1-35.1), 6/51 patients (11.8%) experienced ≥Gr 3 AEs at a median of 12 (range 1-69) days postoperatively (POD). No radiation necrosis (RN) events were observed, and no AEs occurred in multi-implant cases or where STaRT was the initial form of RT. CONCLUSIONS In this prospective multi-institutional study, STaRT demonstrated an excellent safety profile in a cohort of larger rBM, even in the setting of multi-recurrent disease. Accrual and follow-up are on-going and will provide data on tumor control and long-term RN rates.

SURG-17. SALVAGE SURGERY FOR LOCAL PROGRESSION AFTER RADIATION THERAPY OF BRAIN METASTASES

Abstract Recent advances in cancer treatments have prolonged survival after the onset of brain metastases (BM), and this trend increases the risk of local progression (LP) after radiotherapy (RT)of BM. Salvage options include surgical resection, re-irradiation and medical therapies, but there is no standard management for this clinical setting. This study aimed to evaluate the outcome of salvage surgery. We performed surgical resection for intracranial metastases in 221 cases in a single institution between Apr 2014 and Mar 2024. Of those, the data of the cases underwent initial surgery after irradiation for BM were extracted. After excluding duplicated cases and those underwent surgery within one month after irradiation, we finally analyzed 49 cases. Overall survival (OS) and local progression free survival (LPFS) were evaluated by Kaplan-Meier method. All but two cases had a history of stereotactic radiosurgery/radiotherapy for BM. Histopathological diagnosis of surgical specimens was tumor recurrence in 33 cases and radiation necrosis (RN) in 16 cases. RN cases showed longer duration between radiation and LP, when comparing with recurrent cases (median 42.3. vs. 9.3 months). OS was longer in RN cases than in tumor recurrent cases (median 68.5 vs. 26.9 months). In the recurrent cases, preoperative poor KPS (under 70), active state of extracranial lesions and RPA class were associated with shorter OS. Extent of resection, postoperative chemotherapy and local irradiation did not have significant impact on OS. Twenty cases experienced local recurrence after salvage surgery, and LPFS was 7.0 months. Any of extent of removal, postoperative chemotherapy and RT were not associated with LPFS. We reported relatively long survival after salvage surgery for local failure of BM after irradiation. Salvage surgery should be considered especially in patients with good extracranial control and performance status. High recurrence rate after salvage treatment urges the development of additional treatment approaches.

NIMG-38. DIAGNOSIS AND TREATMENT OF BRAIN TUMORS USING METHIONINE PET

Abstract INTRODUCTION The usefulness of methionine PET examination in diagnosing brain tumors is widely recognized, and it is a frequently used examination method for distinguishing primary tumors from non-neoplastic lesions and recurrent lesions from radiation necrosis. In addition, during surgery, it has the property of being taken up by actively dividing tumor cells, which is useful for identifying the extent of resection. Method: Methionine PET examination was performed in 228 consecutive cases in our department from January 2021 to April 2023, and medical records and genetic analysis data were retrospectively analyzed. RESULTS Of the 228 cases, 146 cases provided useful data for clinical practice, such as the diagnosis of tumors and the differentiation of recurrent tumors from necrosis. On the other hand, there were 9 cases in which the pathological diagnosis of the resected specimen was “tumor” despite the lack of significant methionine accumulation. The details were astrocytoma, cavernous hemangioma, spinal ependymoma, etc. On the other hand, 26 cases were finally diagnosed as “non-neoplastic disease”. Of these, 13 cases had methionine accumulation. The details of the cases that were finally diagnosed as “non-neoplastic disease” were multiple sclerosis, cerebral infarction, dural arteriovenous fistula, inflammatory disease, etc. There were 12 cases in which a diagnosis of “neoplastic disease” was not revealed after methionine PET examination, and the patient was continuously observed. None of these 12 cases had an acute or malignant course. CONCLUSION Although methionine PET is considered useful for diagnosing neoplastic disease, there are cases in which there is no uptake, especially in astrocytoma, and a certain number of cases in which uptake was observed even in non-neoplastic diseases, so caution is required. In addition, in terms of treatment, it is possible to aim to improve surgical accuracy by integrating methionine PET images with intraoperative MRI and intraoperative molecular analysis techniques, and we report an example of this.

NIMG-56. MAGNETIC RESONANCE SPECTROSCOPY IN THE POST-TREATMENT METABOLIC CHARACTERIZATION OF BRAIN TUMORS OF GLIAL ORIGIN OF CENTRAL NERVOUS SYSTEM

Abstract BACKGROUND and Purpose Central Nervous System tumors have a very low incidence worldwide, however they constitute an important source of mortality and morbidity. Treatment for these types of tumors in adults may include surgery, radiosurgery, radiation therapy, chemotherapy, and targeted treatment. These treatments in brain tumors of glial origen in adults can produce changes in the pathophysiology of the tumor. The purpose of this study was to characterize the different metabolic patterns as effect of treatment and identify tumor recurrences. Materials and METHODS A prospective study was conducted with patients with low- and high-grade tumors treated, subjecting them to the single-voxel MRS and multivoxel MRSI protocol at 1.5 T. RESULTS Of the 44 patients included, 21 were histopathologically diagnosed as low-grade tumors and 23 as high-grade tumors. Of the low-grade tumors, 11 were caracterized as low-grade (mean Cho/NAA and NAA/Cr ratios 1.49 ((p=0.036) and 0.92 (p=0.038) respectively), 8 cases presented gliosis (mean Cho/NAA and NAA/Cr ratios 1.57 (p=0.028) and 1.36 (p=0.026) respectively), and 4 cases presented radiation necrosis (mean Cho/Cr ratios 2.14 (p=0.03)), (mean Cho/Cr and NAA/Cr ratios 1.25 (p=0.01) and 1.25 (p=0.00)) respectively). Furthermore, tumor recurrence that presented high Cho/NAA ratio were observed in 11 cases. Of the high-grade tumors, 13 were caracterized as high-grade (mean Cho/NAA and Cho/Cr ratios 2.24 (p=0.026) and 4.48 (p=0.026) respectively), 13 cases presented gliosis (mean Cho/NAA and NAA/Cr ratios 1.82 (p=0.028) and 0.64 (p=0.026)) respectively) and radiation necrosis was detected in 10 cases (mean Cho/Cr ratio 1.47 (p=0.03), (mean Cho/Cr and NAA/Cr ratios 2.25 (p=0.01) and 0.97 (p=0.00) respectively). Moreover, tumor recurrence that presented high Cho/NAA ratio were observed in 9 cases. CONCLUSION MRS and MRSI contribute to identifying the recurrences, to characterizing the different metabolic patterns as an effect of the treatment and differentiate between low and high grade tumors.