BIOM-05. CASE SERIES OF MULTI-INSTITUTIONAL UTILITY OF CNSide TO MANAGE LEPTOMENINGEAL DISEASE IN PATIENTS WITH METASTATIC BREAST CANCER

INTRODUCTION Leptomeningeal Disease (LMD) occurs in 5% of breast cancer patients. Diagnosing LMD remains challenging. Current standard of care has limited sensitivity and is inadequate for monitoring treatment response. Biocept’s CNSide™ is a proprietary assay utilizing a 10-antibody capture cocktail with microfluidic chamber that quantitatively detects tumor cells in the cerebrospinal fluid (CSF). We present a case series using CNSide to manage LMD of 4 unique breast cancer patients treated at three different institutions and demonstrate its impact on clinical management. METHODS Patients were treated at Smilow Cancer Hospital at Yale-New Haven (1 patient), Northwestern Medicine Lou and Jean Malnati Brain Tumor Institute (1 patient) and Barrow Neurological Institute (2 patients). All patients received intrathecal treatment (IT) via an Ommaya Reservoir. CSF tumor cells were detected via cytology and CNSide at diagnosis (3 patients) and throughout treatment (4 patients). RESULTS At diagnosis, CNSide detected tumor cells in 3/3 patients, vs 2/3 patients for cytology. The fourth patient was diagnosed with LMD before CNSide was available. CNSide detected CSF tumor cells in 9/9 (100%) of measurements, vs 4/9 (44%) for cytology for samples analyzed in parallel. Throughout treatment, CNSide was able to track the quantitative LMD response and showed a decrease in CSF tumor cells in all four patients, ranging from 99.7% (from 773 to 2 cells, 1 patient) to 100% (from 4447 to 0 cells; and from 33 to 0 cells, 2 patients). CONCLUSION Intrathecal treatment of LMD via Ommaya reservoir is not widely adopted across the US. Our experience suggests that using CNSide for quantitative CSF tumor cell detection may aid in diagnosing LMD, as well as in quantifying response to treatment particularly in the setting of intrathecal therapy. However, larger prospective clinical trials are needed to establish the role of CNSide in the diagnosis and management of LMD.

LMD-18. Detection and serial monitoring of CSF ctDNA in breast cancer leptomeningeal disease (BCLM)

Background CSF cytology is the gold standard diagnostic test for BCLM, but is hampered by a low sensitivity, often necessitating repeated lumbar puncture to confirm or refute the diagnosis. Furthermore, during the treatment of BCLM, there is no robust quantitative response tool to guide treatment decisions. Material and Methods cfDNA was obtained from CSF and plasma in patients with breast cancer undergoing investigation for BCLM (n = 28) and during subsequent intrathecal treatment (n = 13). Ultra low pass whole genome sequencing (ulpWGS) and estimation of the ctDNA fraction was performed. Results were validated by mutation-specific digital droplet PCR (ddPCR). Results 22/28 cases had confirmed BCLM by positive MRI and/or CSF cytology. The remaining 6/28 had suspected but non-confirmed BCLM, and at median 20 months follow up, these patients were BCLM-free. CSF ctDNA fraction was significantly elevated (median 57.5, IQR 38.3 - 84.9%) in confirmed BCLM compared to 6 non-confirmed BCLM (median 5.0, IQR 0.0 - 6.7%) (p <0.0001). ctDNA fraction was detected in BCLM confirmed cases regardless of negative cytology or MRI. Plasma ctDNA fraction was only detected in extra-cranial disease progression. ctDNA fraction was concordant with mutant allele fraction measured by ddPCR (n = 118 samples). Serial CSF ctDNA fraction during intrathecal treatment showed dynamic changes, while CSF cytology and MRI were often unchanged or equivocal. Early reduction in CSF ctDNA fraction was associated with longer responses to intrathecal therapy. Further, rising ctDNA fraction during intrathecal chemotherapy could be detected up to 6 weeks before relapse in neurological symptoms, cytology or MRI. Conclusion Measuring CSF ctDNA fraction is a sensitive diagnostic test for BCLM and could lead to more timely and accurate diagnosis. During intrathecal chemotherapy, CSF ctDNA also provides a quantitative response biomarker to help guide clinical management in this difficult treatment scenario.

CMET-02. ESTABLISHING THE SAFETY AND EFFICACY OF A NEW MULTI-AGENT INTRATHECAL TREATMENT PROTOCOL FOR PATIENTS WITH NEOPLASTIC MENINGITIS

BACKGROUND The prognosis for neoplastic meningitis (NM) remains dismal and single-agent (rather than multi-agent) chemotherapy remains the standard of care. Objective: To analyze survival of patients with NM receiving multi-agent vs. single-agent intrathecal chemotherapy and evaluate the safety of antitumor intrathecal biological agents. METHODS We compared a standardized multi-agent intrathecal treatment cohort to the corresponding patient level data acquired from 6 of the 7 randomized controlled trials (RCTs) conducted in patients with NM, all of which used single-agent intraventricular therapies. The toxicity of intrathecal biological agents was studied by analyzing the outcome data form patients with NM treated with at least 1 intrathecal biological drug. RESULTS 283 patients receiving multi-agent therapy were compared to 290 patients from the RCTs. Patients and tumor characteristics did not differ between groups. All patients from the multi-agent patient group were included in an intent-to-treat model. For all solid tumors, median survival (multi-agent vs. single-agent treatment) was 211 vs. 97 days, hazard ratio (HR) 3.39, p< 0.001. For lymphoma, survival was 304 vs. 81 days, HR 2.10, p< 0.001. Stratified by tumor histology, median survival for breast cancer NM was 315 days [95% CI 248–449], for lung cancer 193 days [127- 200]; for melanoma 307 days [65–1492], and for primary brain tumors 253 days [187- 348]. No additional toxicity was seen in the multi-agent compared to the single-agent treatment groups. In 110 patients, grade III toxicity occurred in 5.3%, 3.1%, 0% and 0% of patients receiving rituximab, trastuzumab, panitumumab, and alpha interferon. No difference in survival was seen when comparing patients with and without treatment related toxicity in any histological groups. CONCLUSION Multi-agent intraventricular chemotherapy was associated with dramatically increased survival in patients with solid tumor and lymphomatous NM. Incorporating biologic agents as part of a multi-agent intrathecal treatment regimen is very safe.