Preoperative Single-Fraction Versus Fractionated Stereotactic Radiation Therapy: A Single-Institutional Analysis of 299 Resected Metastases Authors: Haley K. Perlow, Sydney Luu, Sohil Reddy, Jared Bradshaw, Sarah Hennings, Jennifer K. Matsui, Brett G. Klamer, Rituraj Upadhyay, Alexandria Oliver, John McGregor, Russell Lonser, Daniel Prevedello, James B. Elder, Kyle Wu, Roshan Prabhu, Simeng Zhu, Raj Singh, Sasha J. Beyer, John Grecula, Dukagjin M. Blakaj, Evan M. Thomas, Raju R. Raval, Joshua D. Palmer Published: 2026 ABSTRACT PurposeThe treatment standard for patients with large or symptomatic brain metastases and limited intracranial disease is surgical resection followed by postoperative stereotactic radiosurgery (SRS). Preoperative SRS may lead to a reduced incidence of radiation necrosis (RN), local failure (LF), and nodular meningeal disease (nMD). Fractionated treatments may reduce the incidence of LF and nMD. We hypothesize that preoperative fractionated stereotactic radiation therapy will reduce the incidence rate of RN, LD, and nMD compared to patients receiving preoperative SRS.Methods and MaterialsPatients who had surgical resection and preoperative 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) nMD; and/or (3) grade 2 or higher (symptomatic) RN.Results299 resected brain metastases from 260 patients were eligible for analysis. Two hundred thirty-five metastases received fractionated stereotactic radiation therapy (FSRT) and 64 received SRS. Thirty-eight patients had multiple metastases resected preoperatively. Overall, 4 (6.3%) SRS and 6 (2.6%) FSRT metastases experienced LF. Four (6.2%) SRS and 21 (9.1%) FSRT metastases experienced grade 2 or higher RN. Three (4.7%) SRS and 11 (4.7%) FSRT metastases were diagnosed with leptomeningeal disease. Additionally, 3 (4.6%) and 0 (0%) of SRS, and 8 (3.4%) and 4 (1.7%) of FSRT patients experienced classical or nodular meningeal disease, respectively. Eleven percent and 12% of SRS and FSRT metastases experienced the composite endpoint.ConclusionsIn our study, preoperative SRS and FSRT both appear to be safe and effective options to treat resectable brain metastases. It is important to prospectively compare preoperative SRS and FSRT in matched cohorts to assess any differences in treatment efficacy and toxicity. Introduction Adjuvant radiation after surgical resection of a large or symptomatic brain metastasis minimizes the risk of local recurrence. Traditionally this was delivered utilizing whole brain radiation (WBRT); Patchell et al1 showed how WBRT after resection can reduce local failure (LF) rates and incidence of neurologic death albeit without an overall survival (OS) benefit. Multiple subsequent clinical trials have shown improved cognitive preservation with postoperative stereotactic radiosurgery over WBRT.2,3 Mahajan et al4 showed improved LF rates with postoperative SRS over observation. However, postoperative SRS causes an increased rate of toxicity including LF, radiation necrosis (RN), and nodular meningeal disease (nMD). Preoperative SRS has been explored as an alternative treatment to minimize the risk of treatment toxicity. The seminal PROPS-BM showed a sizable decrease in preoperative SRS incidence of nMD and RN with 2 year rates of 7.6% and 6.8%, respectively, comparing favorably to data from historical postoperative SRS trials.5 Notably, the LF rates in this study remained over 15%. Lehrer et al6 published a meta-analysis of 1887 brain metastasis and showed a reduced incidence of RN and LF when delivering fractionated stereotactic radiation therapy (FSRT) compared to SRS. Subsequent publications by the PROPS-BM group show that fractionated preoperative treatments may reduce the risk of LF to under 5%.7,8 We hypothesize that preoperative FSRT can reduce the incidence of RN and LF compared to preoperative SRS while maintaining the low rates of nMD seen in historical trials. Methods and Materials This retrospective single institution study was approved by our institutional review board. Patients were eligible for preoperative radiosurgery if they had a large dominant brain metastasis, neurologic symptoms that stabilized on steroids, and after a multidisciplinary discussion involving radiation oncology, neurosurgery, neuroradiology, and neuro-oncology. There were no specific size criteria used for resection, nor was there a maximum total number of brain metastasis cutoff to be eligible for surgical resection. Surgery was typically performed within a few days after the last radiation fraction. Pertinent demographic, clinical, radiation, surgical, radiologic, pathologic, and follow-up data were collected for each patient. Karnofsky Performance Status (KPS) was documented for each patient at the time of their radiation oncology consult. RN was diagnosed utilizing a multidisciplinary team including radiation oncology, neuro-surgery, neuro-oncology, neuro-radiology, and also pathology if needed. RN was graded according to NCI Common Terminology Criteria for Adverse Events (CTCAE) v5.0 (grade 1: asymptomatic; grade 2: moderate symptoms, corticosteroids indicated; grade 3: severe symptoms, medical intervention indicated; grade 4: life-threatening; urgent intervention; and grade 5: death). Patients were eligible for analysis if they had more than 1 brain metastasis resected, either at the same time point or a new metastasis at a different time in the disease course. Adverse events including RN, LF, and nMD were evaluated on a per-lesion basis. A per-lesion composite endpoint was measured and was defined as metastases with either (1) grade 2 or higher (G2+, symptomatic) RN; (2) LF; or (3) nMD. This composite endpoint is similar to the composite endpoint being used on NRG BN-012 (NCT05438212), a clinical trial evaluating preoperative versus postoperative SRS. The treatment planning for this patient population has been previously published.9, 10, 11 Patients were simulated supine with a thermoplastic mask for linear-accelerator treatments, or immobilized with a headframe for gamma knife treatments. T1 postcontrast volumetric imaging was utilized to delineate the gross tumor volume for each patient. A clinical target volume of 2 mm and an optional 1-mm planning target volume (PTV) was used for all metastases. Linear-accelerator based radiation therapy was planned and delivered uniformly using intensity modulated radiation therapy or volumetric modulated arc therapy-based planning with 2 to 3 noncoplanar 6 megavoltage arcs with daily image guidance using cone beam computed tomography. Gamma Knife-based SRS was prescribed to the 50% isodose line. Risk-adjusted follow-up was utilized; patients were generally followed with a detailed history and physical and brain magnetic resonance imaging every 2 to 3 months for the first year, 3 to 4 months for the second year, and every 6 months after 2 years. Statistics Descriptive statistics were used to summarize baseline demographic, clinical, and treatment variables. The median and interquartile range (IQR) were used for continuous variables and frequency counts and proportions for categorical variables. Patients were stratified by method of treatment (FSRT vs SRS). Differences between treatment groups were summarized using the standardized mean difference. OS was defined as the date of onset of treatment to the date of death and censored at the date of last follow-up for those still alive. The Kaplan–Meier estimator was used to calculate 12-month probability of OS. Cox proportional hazard and the Fine–Gray proportional subdistribution hazard models were used for OS and cumulative incidence, respectively. Each model adjusted for treatment method, sex, age, number of lesions, and log-transformed gross tumor volume (GTV). The primary event in cumulative incidence analysis was the per-lesion composite endpoint, whereas death without experiencing the composite endpoint was a competing event. A sensitivity analysis was performed by replicating analyses on the subgroup of patients whose GTV were at least the 15th percentile in the SRS group and at most the 85th percentile in the FSRT group. Complete case analysis was used for all summaries and reported P values and CI were unadjusted for multiplicity. Statistical analyses were performed with R version 4.4.3 using the survival (version 3.8-3) and tidycmprsk (version 1.1.0) packages. Results Between 2016 and 2023, 299 metastases from 260 patients received preoperative radiosurgery and were included in the final analyses. The median follow-up time either to censoring or time of death was 10.4 months (IQR, 4.5-23.4). During the follow-up period, 141 deaths were observed (54%). Fifty-one percent of resectable metastases were from male patients (Table 1). A plurality of metastases (48%) were from a lung primary tumor, with genitourinary (14%) and melanoma (13%) forming a smaller proportion. Most patients (91%) had a KPS of ≥70.Table 1Patient and tumor characteristics per index lesion (n = 299)Table 1 dummy alt textVariablePreoperative SRS, n = 641Preoperative FSRT, n = 2351P value†Difference95% CI‡OverallSex.40.13−0.15, 0.41 Men36 (56%)117 (50%)153 (51%) Women28 (44%)118 (50%)146 (49%)Age (y, IQR)62 (54, 66)62 (55,69).60.07−0.20, 0.3562 (55,68)Primary tumor0.740.46, 1.0 Lung41 (64%)102 (39%)143 (48%) Genitourinary13 (20%)30 (15%)43 (14%) Melanoma5 (7.8%)34 (13%)39 (13%) Breast1 (1.6%)26 (11%)27 (9.0%) Gastrointestinal2 (3.1%)18 (7.4%)20 (6.7%) Sarcoma0 (0%)10 (7.0%)10 (3.3%) Gynecologic1 (1.6%)7 (3.7%)8 (2.7%) Head and neck0 (0%)6 (2.0%)6 (2.0%) Other1 (1.6%)2 (1.0%)3 (1.0%)Extracranial disease control.0110.360.08, 0.64 Controlled35 (55%)87 (37%)122 (41%) Uncontrolled29 (45%)148 (63%)177 (59%)Extracranial metastases.0130.350.07, 0.63 Absent35 (55%)88 (37%)238 (41%) Present29 (45%)147 (63%)296 (59%)KPS.0110.470.19, 0.75 ≤601 (1.6%)24 (10%)25 (8.4%) 70-8042 (66%)111 (47%)153 (51%) 90-10021 (33%)100 (43%)121 (40%)SIMT<.0010.760.47, 1.0 No53 (83%)116 (49%)169 (57%) Yes11 (17%)119 (51%)130 (43%)Number of fractions<.001 164 (100%)0 (0%)64 (21%) 30 (0%)217 (92%)217 (73%) 50 (0%)18 (7.7%)18 (6.0%)GTV (cm3, IQR)4 (2,7)10 (5,17)<.0010.470.19, 0.768 (4,15)PTV (cm3, IQR)6 (4,10)16 (10,26)<.0010.670.38, 0.9614 (7,25)Location of resected lesion.0250.600.32, 0.88 Frontal30 (47%)76 (32%)106 (35%) Parietal18 (28%)51 (22%)69 (23%) Cerebellar6 (9.4%)42 (18%)48 (16%) Occipital3 (4.7%)40 (17%)43 (14%) Temporal7 (11%)19 (8.1%)26 (8.7%) Other0 (0%)7 (3.0%)7 (2.4%)Extent of resection>.90.19−0.09, 0.46 Gross total resection64 (100%)231 (98%)295 (99%) Near total resection0 (0%)3 (1.3%)3 (1.0%) Subtotal resection0 (0%)1 (0.4%)1 (0.3%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multi target; SRS = stereotactic radiosurgery.*n (%); median (quartile1, quartile3).†Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test.‡Standardized mean difference. Table 1Patient and tumor characteristics per index lesion (n = 299)Table 1 dummy alt textVariablePreoperative SRS, n = 641Preoperative FSRT, n = 2351P value†Difference95% CI‡OverallSex.40.13−0.15, 0.41 Men36 (56%)117 (50%)153 (51%) Women28 (44%)118 (50%)146 (49%)Age (y, IQR)62 (54, 66)62 (55,69).60.07−0.20, 0.3562 (55,68)Primary tumor0.740.46, 1.0 Lung41 (64%)102 (39%)143 (48%) Genitourinary13 (20%)30 (15%)43 (14%) Melanoma5 (7.8%)34 (13%)39 (13%) Breast1 (1.6%)26 (11%)27 (9.0%) Gastrointestinal2 (3.1%)18 (7.4%)20 (6.7%) Sarcoma0 (0%)10 (7.0%)10 (3.3%) Gynecologic1 (1.6%)7 (3.7%)8 (2.7%) Head and neck0 (0%)6 (2.0%)6 (2.0%) Other1 (1.6%)2 (1.0%)3 (1.0%)Extracranial disease control.0110.360.08, 0.64 Controlled35 (55%)87 (37%)122 (41%) Uncontrolled29 (45%)148 (63%)177 (59%)Extracranial metastases.0130.350.07, 0.63 Absent35 (55%)88 (37%)238 (41%) Present29 (45%)147 (63%)296 (59%)KPS.0110.470.19, 0.75 ≤601 (1.6%)24 (10%)25 (8.4%) 70-8042 (66%)111 (47%)153 (51%) 90-10021 (33%)100 (43%)121 (40%)SIMT<.0010.760.47, 1.0 No53 (83%)116 (49%)169 (57%) Yes11 (17%)119 (51%)130 (43%)Number of fractions<.001 164 (100%)0 (0%)64 (21%) 30 (0%)217 (92%)217 (73%) 50 (0%)18 (7.7%)18 (6.0%)GTV (cm3, IQR)4 (2,7)10 (5,17)<.0010.470.19, 0.768 (4,15)PTV (cm3, IQR)6 (4,10)16 (10,26)<.0010.670.38, 0.9614 (7,25)Location of resected lesion.0250.600.32, 0.88 Frontal30 (47%)76 (32%)106 (35%) Parietal18 (28%)51 (22%)69 (23%) Cerebellar6 (9.4%)42 (18%)48 (16%) Occipital3 (4.7%)40 (17%)43 (14%) Temporal7 (11%)19 (8.1%)26 (8.7%) Other0 (0%)7 (3.0%)7 (2.4%)Extent of resection>.90.19−0.09, 0.46 Gross total resection64 (100%)231 (98%)295 (99%) Near total resection0 (0%)3 (1.3%)3 (1.0%) Subtotal resection0 (0%)1 (0.4%)1 (0.3%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multi target; SRS = stereotactic radiosurgery.*n (%); median (quartile1, quartile3).†Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test.‡Standardized mean difference. Patient and tumor characteristics per index lesion (n = 299) Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multi target; SRS = stereotactic radiosurgery. *n (%); median (quartile1, quartile3). Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test. Standardized mean difference. Nearly all lesions had a gross total resection (99%). The 2 most common locations for resected lesions were the frontal (35%) and parietal (23%) lobes. Almost half of patients (43%) had single isocenter multitarget (SIMT) treatments. The median GTV for metastases receiving SRS and FSRT were 4 cm3 (IQR, 2-7) and 10 cm3 (IQR, 5-17), respectively. The median PTV for metastases receiving SRS and FSRT were 6 cm3 (IQR: 4, 10) and 16 cm3 (IQR, 10-26), respectively. Ninety-two percent of FSRT metastases received 3 fractions, and 8% of FSRT patients received 5 fractions. All SRS patients received 20 Gy in 1 fraction. All patients receiving FSRT received either 24 Gy in 3 fractions or 25 Gy in 5 fractions. The 12-month OS probability for all patients was 0.57 (95% CI, 0.51-0.64). Patients receiving FSRT had increased hazard of death compared to SRS (hazard ratio, 1.23; 95% CI, 0.79-1.93; Fig. 1).Figure 1Overall survival probability.Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery.Figure 1 dummy alt text Overall survival probability. Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery. Overall, 6.3% of SRS and 2.6% of FSRT metastases experienced LF (Table 2). The proportion of G2+ RN for SRS and FSRT was 6.3% and 8.9%, respectively. Three (4.7%) SRS and 11 (4.7%) FSRT metastases were diagnosed with leptomeningeal disease. Additionally, 3 (4.6%) and 0 (0%) of SRS, and 8 (3.4%) and 4 (1.7%) of FSRT metastases experienced classical or nodular meningeal disease, respectively. Eleven percent and 12% of SRS and FSRT metastases experienced the composite endpoint, respectively. At 12 months, the cumulative incidence of the composite endpoint was 0.088 (95% CI, 0.032-0.18) for SRS, 0.11 (95% CI, 0.25-1.71) and 0.10 (95% CI, 0.07-0.14) overall (Fig. 2). The adjusted hazard ratio for the composite endpoint was 0.65 (95% CI, 0.25-1.71), indicating reduced hazard for FSRT compared to SRS. A similar result was observed in the subgroup sensitivity analysis (Table 3 and Fig. 3; HR, 0.65; 95% CI, 0.12-3.57).Table 2Adverse outcomes per index lesion (n = 299)Table 2 dummy alt textVariablePreoperative SRS, n = 64Preoperative FSRTn = 235Adjusted HR (95% CI)P valueOverallLocal progression0.17 (0.03, 0.98).2 No60 (94%)229 (97%)289 (97%) Yes4 (6.3%)6 (2.6%)10 (3.3%)Radiation necrosis1.25 (0.38, 4.10)>.9 Grade 22 (3.1%)12 (5.2%)14 (4.7%) Grade 32 (3.1%)9 (3.9%)11 (3.7%) Grade 4 or higher00Leptomeningeal disease->.9 No61 (95%)224 (95%)285 (95%) Yes3 (4.7%)11 (4.7%)14 (4.7%)Type of leptomeningeal disease-.6 Classic3 (100%)7 (64%)10 (71%) Nodular0 (0%)3 (27%)3 (21%) Both0 (0%)1 (9.1%)1 (7.1%)Composite endpoint0.65 (0.25, 1.71)>.9 No57 (89%)207 (88%)264 (88%) Yes7 (11%)28 (12%)35 (12%)Follow-up.7 Censored30 (47%)103 (44%)133 (44%) Death34 (53%)132 (56%)166 (56%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery.Figure 2Cumulative incidence curve.Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery.Figure 2 dummy alt textTable 3Patient characteristics for subset analysis of GTVs ≥3.8 cm3 and ≤12.3 cm3Table 3 dummy alt textVariablePreoperative SRS, n = 19*Preoperative FSRT, n = 104*P value†Difference‡95% CI‡Overall123*Sex.20.30−0.19, 0.79 Men13 (68%)56 (54%)69 (56%) Women6 (32%)48 (46%)54 (44%)Age (y, IQR)62 (54, 66)62 (57, 68).60.20−0.29, 0.6962 (56, 68)Primary Tumor.80.760.27, 1.3 Lung11 (58%)47 (45%)58 (47%) Genitourinary3 (16%)15 (14%)18 (15%) Melanoma4 (21%)12 (12%)16 (13%) Breast0 (0%)12 (12%)12 (9.8%) Gastrointestinal1 (5.3%)9 (8.7%)10 (8.1%) Sarcoma0 (0%)3 (2.9%)3 (2.4%) GYN0 (0%)3 (2.9%)3 (2.4%) Head and neck0 (0%)2 (1.9%)2 (1.6%) Other0 (0%)1 (1.0%)1 (0.8%)Extracranial disease control.0450.500.01, 1.0 Controlled11 (58%)35 (34%)46 (37%) Uncontrolled8 (42%)69 (66%)77 (63%)Extracranial metastases.90.03−0.46, 0.52 Absent7 (37%)40 (38%)47 (38%) Present12 (63%)64 (62%)76 (62%)KPS>.90.17−0.32, 0.66 ≤601 (5.3%)10 (9.6%)11 (8.9%) 70-809 (47%)49 (47%)58 (47%) 90-1009 (47%)45 (43%)54 (44%)SIMT.0030.850.35, 1.3 No16 (84%)49 (47%)65 (53%) Yes3 (16%)55 (53%)58 (47%)Number of fractions<.001 119 (100%)0 (0%)19 (15%) 30 (0%)100 (96%)100 (81%) 50 (0%)4 (3.8%)4 (3.3%)GTV (cm3, IQR)6.40 (4.40, 7.58)7.28 (5.50, 9.56).0770.45−0.04, 0.957.00 (5.30, 9.50)PTV (cm3, IQR)7.6 (5.3, 11.2)12.8 (9.9, 16.6)<.0010.970.46, 1.512.4 (9.3, 16.2)Location of resected lesion.80.550.06, 1.0 Frontal8 (42%)35 (34%)43 (35%) Parietal6 (32%)21 (20%)27 (22%) Cerebellar3 (16%)17 (16%)20 (16%) Occipital1 (5.3%)16 (15%)17 (14%) Temporal1 (5.3%)11 (11%)12 (9.8%) Other0 (0%)2 (1.9%)2 (1.6%)Extent of resection>.90.14−0.35, 0.632 (1.6%) Gross total resection19 (100%)103 (99%) Near total resection0 (0%)1 (1.0%)122 (99%) Subtotal resection0 (0%)0 (0%)1 (0.8%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multitarget; SRS = stereotactic radiosurgery.⁎n (%); median (quartile1, quartile3).†Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test.‡Standardized mean difference.Figure 3Cumulative incidence curve in subgroup of GTVs ≥3.8 cm3 and ≤12.3 cm3.Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; HR = hazard ratio; SRS = stereotactic radiosurgery.Figure 3 dummy alt text Table 2Adverse outcomes per index lesion (n = 299)Table 2 dummy alt textVariablePreoperative SRS, n = 64Preoperative FSRTn = 235Adjusted HR (95% CI)P valueOverallLocal progression0.17 (0.03, 0.98).2 No60 (94%)229 (97%)289 (97%) Yes4 (6.3%)6 (2.6%)10 (3.3%)Radiation necrosis1.25 (0.38, 4.10)>.9 Grade 22 (3.1%)12 (5.2%)14 (4.7%) Grade 32 (3.1%)9 (3.9%)11 (3.7%) Grade 4 or higher00Leptomeningeal disease->.9 No61 (95%)224 (95%)285 (95%) Yes3 (4.7%)11 (4.7%)14 (4.7%)Type of leptomeningeal disease-.6 Classic3 (100%)7 (64%)10 (71%) Nodular0 (0%)3 (27%)3 (21%) Both0 (0%)1 (9.1%)1 (7.1%)Composite endpoint0.65 (0.25, 1.71)>.9 No57 (89%)207 (88%)264 (88%) Yes7 (11%)28 (12%)35 (12%)Follow-up.7 Censored30 (47%)103 (44%)133 (44%) Death34 (53%)132 (56%)166 (56%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery. Adverse outcomes per index lesion (n = 299) Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery. Cumulative incidence curve. Abbreviations: FSRT = fractionated stereotactic radiation therapy; HR = hazard ratio; SRS = stereotactic radiosurgery. Table 3Patient characteristics for subset analysis of GTVs ≥3.8 cm3 and ≤12.3 cm3Table 3 dummy alt textVariablePreoperative SRS, n = 19*Preoperative FSRT, n = 104*P value†Difference‡95% CI‡Overall123*Sex.20.30−0.19, 0.79 Men13 (68%)56 (54%)69 (56%) Women6 (32%)48 (46%)54 (44%)Age (y, IQR)62 (54, 66)62 (57, 68).60.20−0.29, 0.6962 (56, 68)Primary Tumor.80.760.27, 1.3 Lung11 (58%)47 (45%)58 (47%) Genitourinary3 (16%)15 (14%)18 (15%) Melanoma4 (21%)12 (12%)16 (13%) Breast0 (0%)12 (12%)12 (9.8%) Gastrointestinal1 (5.3%)9 (8.7%)10 (8.1%) Sarcoma0 (0%)3 (2.9%)3 (2.4%) GYN0 (0%)3 (2.9%)3 (2.4%) Head and neck0 (0%)2 (1.9%)2 (1.6%) Other0 (0%)1 (1.0%)1 (0.8%)Extracranial disease control.0450.500.01, 1.0 Controlled11 (58%)35 (34%)46 (37%) Uncontrolled8 (42%)69 (66%)77 (63%)Extracranial metastases.90.03−0.46, 0.52 Absent7 (37%)40 (38%)47 (38%) Present12 (63%)64 (62%)76 (62%)KPS>.90.17−0.32, 0.66 ≤601 (5.3%)10 (9.6%)11 (8.9%) 70-809 (47%)49 (47%)58 (47%) 90-1009 (47%)45 (43%)54 (44%)SIMT.0030.850.35, 1.3 No16 (84%)49 (47%)65 (53%) Yes3 (16%)55 (53%)58 (47%)Number of fractions<.001 119 (100%)0 (0%)19 (15%) 30 (0%)100 (96%)100 (81%) 50 (0%)4 (3.8%)4 (3.3%)GTV (cm3, IQR)6.40 (4.40, 7.58)7.28 (5.50, 9.56).0770.45−0.04, 0.957.00 (5.30, 9.50)PTV (cm3, IQR)7.6 (5.3, 11.2)12.8 (9.9, 16.6)<.0010.970.46, 1.512.4 (9.3, 16.2)Location of resected lesion.80.550.06, 1.0 Frontal8 (42%)35 (34%)43 (35%) Parietal6 (32%)21 (20%)27 (22%) Cerebellar3 (16%)17 (16%)20 (16%) Occipital1 (5.3%)16 (15%)17 (14%) Temporal1 (5.3%)11 (11%)12 (9.8%) Other0 (0%)2 (1.9%)2 (1.6%)Extent of resection>.90.14−0.35, 0.632 (1.6%) Gross total resection19 (100%)103 (99%) Near total resection0 (0%)1 (1.0%)122 (99%) Subtotal resection0 (0%)0 (0%)1 (0.8%)Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multitarget; SRS = stereotactic radiosurgery.⁎n (%); median (quartile1, quartile3).†Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test.‡Standardized mean difference. Patient characteristics for subset analysis of GTVs ≥3.8 cm3 and ≤12.3 cm3 Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; KPS = Karnofsky Performance Status; PTV = planning target volume; SIMT = single isocenter multitarget; SRS = stereotactic radiosurgery. n (%); median (quartile1, quartile3). Pearsons’s χ2 test, Wilcoxon rank sum test, Fisher’s exact test. Standardized mean difference. Cumulative incidence curve in subgroup of GTVs ≥3.8 cm3 and ≤12.3 cm3. Abbreviations: FSRT = fractionated stereotactic radiation therapy; GTV = gross tumor volume; HR = hazard ratio; SRS = stereotactic radiosurgery. Discussion This study represents the largest single-institutional comparison of preoperative SRS and FSRT to treat brain metastases. These data suggest that preoperative SRS and FSRT are both safe and effective options with 11% and 12% of metastases experiencing a composite endpoint, respectively. GTVs and PTVs were significantly larger for metastases in the FSRT arm. However, estimates adjusting for GTV and subgroup analyses with overlapping GTV produced similar results. Prospective validation in matched cohorts is necessary to determine differences in efficacy with both treatment approaches. Historically, postoperative radiosurgery has increased the rates of RN, LF, and MD compared to WBRT. Specific rates of RN, MD, and LF in historical studies have ranged from 0% to 3%, 7% to 28%, and 25% to 50%, respectively.2, 3, 4 Given the cognitive benefits of stereotactic radiosurgery over WBRT for patients with brain metastases, more work is needed to deliver a cognitive sparing treatment that minimizes the incidence of adverse events. Preoperative stereotactic radiosurgery treatments for patients with brain metastases were initially introduced by Asher et al12 as a technique to improve historical rates of LF, RN, and MD. Preoperative SRS can minimize iatrogenic leptomeningeal spread through tumor sterilization prior to surgery. Preoperative treatments, through treatment of an intact tumor versus a larger postoperative cavity, can also reduce the risk of RN. An early study from Patel et al13 showed significantly reduced 2-year rates of leptomeningeal disease (3.2% vs 16.6%) and symptomatic RN (4.9% vs 16.4%). Multiple additional studies have shown reduced rates of RN and MD compared to historical postoperative SRS data.9,10,14,15 In our study, our composite endpoint of 11% to 12% inclusive of RN, nMD, and LF compares favorably to previously referenced historical data treating patients with postoperative SRS. One historical challenge with delivering stereotactic radiosurgery for brain metastases is balancing the competing outcomes of RN and LF. The historical data guiding single-fraction dosing were based on tumors receiving reirradiation treatments in RTOG 900516 which led to maximum tolerated doses of 24 Gy for tumors <2 cm and 18 Gy for tumors 2 to 3 cm. These doses were similarly adopted for N0574 with the exception of lesions 2 to 3 cm receiving 20 Gy for patients receiving SRS alone.17 However, in the SRS alone arm, the 12-month local control rate was 72.8%. For patients with larger metastases, the rates of local control with SRS can even dip below 50%.18 Delivering fractionated treatments may increase the biological effective dose to the tumor cells which can in turn reduce the rates of LF. Lehrer et al6 show how delivering fractionated radiation treatments can improve the historical rates of LF while also minimizing the rate of subsequent RN. In our study, there was a reduced incidence of LF in patients receiving FSRT although it did not reach statistical significance due to a low event rate. There were similar rates of RN, although patients receiving FSRT had a larger mean tumor volume treated which makes a direct comparison challenging. To our knowledge, this is the largest single-institutional study comparing preoperative SRS versus FSRT. These data show that preoperative treatments are well tolerated and can decrease toxicity compared to historical postoperative SRS data. Our composite endpoint has been previously published in studies comparing pre- and postoperative SRS and is the endpoint for a national cooperative group trial. A limitation of this study is the low incidence rate of RN, nMD, and LF. Although there was a 50% reduction in the rate of LF for patients receiving FSRT, the difference was not statistically significant due to an overall low event rate. A composite endpoint can be challenging to interpret given the variable severity of individual adverse events. RN and LF can be challenging to diagnose prospectively or retrospectively and oftentimes requires pathologic confirmation for diagnostic certainty. The median follow-up was 10.4 months and longer follow-up may capture more events. Neuro-cognitive and quality of life data were not collected for patients in this study. There was selection bias in this cohort with larger tumors receiving FSRT; this could underestimate the incidence of adverse events for patients receiving SRS. Given the heterogeneity of histology and systemic therapy options at appearance of the resectable brain metastasis, systemic therapy administration before and after radiosurgery was not evaluated. A large randomized control trial is needed to definitively compare outcomes between preoperative SRS and FSRT, as the large reported confidence intervals do not definitively support equivalence. Conclusions In this cohort, both preoperative SRS and FSRT are associated with a low rates of RN, LF, and nMD, confirming the efficacy of preoperative radiosurgery regardless of fractionation approach Prospective evaluation in matched cohorts can effectively determine differences in efficacy with both treatment approaches. Disclosures Haley K. Perlow is an associate editor for Advances in Radiation Oncology. Haley K. Perlow also reports honoraria from Novocure and grant funding from Varian. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.