Heterogeneity of HER2 expression in circulating tumor cells of patients with breast cancer brain metastases and impact on brain disease control.

2029 Background: The HER2 expression switching in circulating tumor cells (CTC) of breast cancer is dynamic and may have prognostic and predictive clinical implications. This study aims to analyse the association between expression of HER2 in CTC of patients with breast cancer brain metastases (BCBM) and brain disease control. Methods: Exploratory analysis of a prospective assessment (NCT02941536) of CTC before (CTC1) and 4–5 weeks after (CTC2) stereotactic radiotherapy/radiosurgery (SRT). CTC were isolated and quantified by a method of isolation by size of tumors and analyzed by immunocytochemistry to evaluate the expression of HER2. Distant brain failure-free survival (DBFFS), the primary endpoint, and overall survival (OS) were estimated by Kaplan-Meier estimator. Log-rank tests were applied in order to compare the survival curves. For multivariate analysis of prognostic factors that affected DBFFS and OS, the Cox proportional model was adjusted. Results: The median age at SRT was 54 (34-70), the diagnosis-specific graded prognostic assessment (DS-GPA) was 1–2 in 17.5% and 2.5–4 in 82.5% and the primary immunophenotype (PIP) was HER2-enriched in 51%, luminal B (LB) in 31% and triple negative (TN) in 18% of the total of 39 patients. CTC were detected in all 39 patients before SRT and the median CTC1 was 2 CTC/mL. After SRT, CTC were detected in 34 of 35 patients (4 deaths between CTC1 and CTC2) and the median CTC2 was 2.33 CTC/mL. HER2 was expressed in CTC1 and/or CTC2 in 9 patients, of which only 2 patients had PIP HER2-enriched. After a median follow-up of 16.6 months, there were 15 patients with distant brain failure and 16 deaths. The median DBFFS and OS were 15.3 and 19.5 months, respectively. Median DBFFS was 7 months in patients with PIP TN and was not reached in PIP LB and HER2-enriched (p = 0.036); 14 months in patients with DS-GPA 1-2 and 7 months with DS-GPA 2.5-4 (p = 0.017); 10 months in patients without HER2 expressed in CTC and not reached in patients with HER2 expressed in CTC (p = 0.012). Median OS was 4.8 months in patients with PIP TN and was not reached in PIP LB and HER2-enriched (p = 0.0026); 19.54 months in patients with DS-GPA 1-2 and 7.6 months with DS-GPA 2.5-4 (p = 0.00088); 17 months in patients without HER2 expressed in CTC and not reached in patients with HER2 expressed in CTC (p = 0.104). On multivariate analysis, DBFFS was superior in patients with PIP HER2-enriched (HR 0.128, 95% CI 0.025–0.534; p = 0.013) and OS was superior in patients with PIP HER2-enriched (HR 0.073, 95% CI 0.018-0.288; p < 0.0001) and LB (HR 0.224, 95% CI 0.062–0.816; p = 0.023). The status of expression of HER2 in CTC was not included in Cox model for DBFFS due to lack of events in patients with HER2 expressed in CTC. Conclusions: The expression of HER2 in CTC was associated with a longer DBFFS and the switching of HER2 expression between PIP and CTC may have impact on prognosis and treatment selection of BCBM.

RADT-13. EARLY CONCURRENT IMMUNOTHERAPY WITH STEREOTACTIC RADIOSURGERY IS ASSOCIATED WITH PROLONGED SURVIVAL AND DECREASED DISTANT BRAIN FAILURE IN PATIENTS WITH NEWLY DIAGNOSED MELANOMA BRAIN METASTASES (MBM)

INTRODUCTION We evaluated outcomes of patients with newly diagnosed MBM treated with concurrent immune checkpoint inhibition (ICI) and stereotactic radiosurgery (SRS) (concurrentTx), defined as treatment delivery within 30 days of each other. METHODS Screening of 2,617 melanoma patients who received ICI (anti-CTLA4/anti-PD1/both) between 2011-2019 identified 151 pts who received concurrentTx for MBM. Among these, 51 had newly-diagnosed MBM and received no prior ICI or SRS, and were included in the current study. Overall survival (OS) and distant brain failure (DBF) were estimated using the Kaplan-Meier method. Incidence of radiation necrosis (RN) was captured. RESULTS Median follow up from treatment initiation (either ICI or SRS, whichever occurred first) was 37 months. Median OS was 30 months. Median interval between ICI/SRS was 12 days (range: 1-29). Twenty-two patients received ICI first and 29 received SRS first, without differences in OS (p=0.22), DBF (p=0.91), or development of RN (p=0.86). However, the interval between ICI and SRS was significant. Patients who received concurrentTx 1-11 days apart (n=25, “early”) experienced a significant improvement in OS and DBF compared to 12-29 days apart (n=26, “delayed”) (p=0.01, HR 2.8; 95%CI 1.3-6.2 for OS and p=0.02, HR 2.5; 95%CI 1.2-5.6 for DBF). OS and DBF at 36 months were 67% vs. 26% and 60% vs. 27%, respectively, for the early vs. delayed groups. Time to concurrentTx as a continuous variable was significantly associated with DBF (p=0.02), but not OS (p=0.06). Although not significant, more patients developed RN in the early (26.0%) versus delayed (3.8%) group (p=0.07). No additional patient or treatment differences were identified. CONCLUSIONS Early concurrentTx was associated with prolonged OS and improved DBF in newly diagnosed MBM patients who did not receive prior CNS-directed therapy. This finding suggests therapeutic synergism related to combined early treatment and should be validated in a prospective clinical trial.

Role of Gamma Knife Radiosurgery in Small Cell Lung Cancer: A Multi-Institutional Retrospective Study of the International Radiosurgery Research Foundation (IRRF)

BACKGROUND Despite a high incidence of brain metastases in patients with small-cell lung cancer (SCLC), limited data exist on the use of stereotactic radiosurgery (SRS), specifically Gamma Knife™ radiosurgery (Elekta AB), for SCLC brain metastases. OBJECTIVE To provide a detailed analysis of SCLC patients treated with SRS, focusing on local failure, distant brain failure, and overall survival (OS). METHODS A multi-institutional retrospective review was performed on 293 patients undergoing SRS for SCLC brain metastases at 10 medical centers from 1991 to 2017. Data collection was performed according to individual institutional review boards, and analyses were performed using binary logistic regression, Cox-proportional hazard models, Kaplan-Meier survival analysis, and competing risks analysis. RESULTS Two hundred thirty-two (79%) patients received SRS as salvage following prior whole-brain irradiation (WBRT) or prophylactic cranial irradiation, with a median marginal dose of 18 Gy. At median follow-up after SRS of 6.4 and 18.0 mo for surviving patients, the 1-yr local failure, distant brain failure, and OS were 31%, 49%, and 28%. The interval between WBRT and SRS was predictive of improved OS for patients receiving SRS more than 1 yr after initial treatment (21%, &lt;1 yr vs 36%, &gt;1 yr, P = .01). On multivariate analysis, older age was the only significant predictor for OS (hazard ratio 1.63, 95% CI 1.16-2.29, P = .005). CONCLUSION SRS plays an important role in the management of brain metastases from SCLC, especially in salvage therapy following WBRT. Ongoing prospective trials will better assess the value of radiosurgery in the primary management of SCLC brain metastases and potentially challenge the standard application of WBRT in SCLC patients.

TRLS-10. MITIGATING NEUROCOGNITIVE DEFICITS FROM WHOLE-BRAIN RADIOTHERAPY IN PATIENTS WITH NUMEROUS BRAIN METASTASES VIA A NOVEL SUPEROXIDE DISMUTASE MIMETIC: RATIONALE & DESIGN OF A CLINICAL TRIAL

BACKGROUND: Patients with a large number of brain metastases (BM) and/or micrometastatic disease in the brain present a clinical challenge. While technical innovations in stereotactic radiosurgery (SRS) have extended the number of BM that can be effectively treated, SRS does not treat occult disease and distant brain failure (DBF) post-SRS remains high. Immuno- and targeted therapies show promise in treating metastatic disease to the brain, though response rates are variable. In contrast, whole-brain radiotherapy (WBRT) provides high rates of local control and, compared to SRS, reduces the risk of distant brain failure. Unfortunately, WBRT is also associated with substantial neurocognitive deficits and neither altered fractionation nor the use of available neuroprotectants has adequately addressed this issue. An agent that safely minimizes the adverse effects of WBRT while preserving or enhancing tumor control would provide meaningful clinical benefit. TRIAL DESIGN: BMX-001, a novel Mn-porphyrin superoxide dismutase mimetic, has been shown to protect normal tissues from ionizing radiation in preclinical trials, reducing neurocognitive adverse effects as well as enhancing tumor response. Based on the first-in-human trial of this agent in patients with high-grade gliomas, we have instituted a clinical trial of WBRT +/- BMX-001 in adult patients with more than 10 BM from melanoma, non-small-cell lung, breast and renal cancer. Following a safety lead-in of 5 patients, all of whom will receive WBRT and BMX-001, 69 patients will be randomized to WBRT (3Gy/fraction x 10 fractions) with or without BMX-001 administered subcutaneously before, twice weekly during and once after WBRT (6 injections total.) The primary endpoint is cognition, as measured by the Hopkins Verbal Learning, Trailmaking A/B and Controlled Oral Word Association tests. Secondary endpoints include health-related quality-of-life, overall and progression-free survival, rates of radiation necrosis, DBF and neurologic death. Enrollment began January 2019. (ClinicalTrials.gov Identifier: NCT03608020.)