Context.—

Leptomeningeal disease (LMD) is a clinical sequela of central nervous system metastasis involving the cerebrospinal fluid (CSF), often seen in late-stage solid tumors. It has a grave prognosis without urgent treatment. Standard of care methodologies to diagnose LMD include CSF cytology, magnetic resonance imaging, and clinical evaluation. These methods offer limited sensitivity and specificity for the evaluation of LMD. Here, we describe the analytic performance characteristics of a microfluidic-based tumor cell enrichment and detection assay optimized to detect epithelial cells in CSF using both contrived samples as well as CSF from patients having suspected or confirmed LMD from carcinomas.

Objective.—

To demonstrate the feasibility of using a microfluidic, multi-antibody cell capture assay to identify and quantify tumor cells in CSF.

Design.—

An artificial CSF solution was spiked with 34 different human carcinoma cell lines at different concentrations and assayed for the ability to detect tumor cells to assess analytic accuracy. Two cell lines were selected to assess linearity, intra-assay precision, interinstrument precision, and sample stability. Clinical verification was performed on 65 CSF specimens from patients. Parameters assessed included the number of tumor cells, coefficient of variation percentage, and percentage of tumor cell capture (TCC).

Results.—

Among contrived samples, average tumor cell capture ranged from 50% to 82% (261 of 522; 436 of 531), and coefficients of variation ranged from 7% to 67%. The cell capture assay demonstrated a sensitivity of 92% and a specificity of 95% among clinical samples.

Conclusions.—

This assay demonstrated the ability to detect and enumerate epithelial cells in contrived and clinical specimens in an accurate and reproducible fashion. The use of cell capture assays in CSF may be useful as a sensitive test for the diagnosis and longitudinal monitoring of LMD from solid tumors.

This content is only available as a PDF.

Author notes

This study was funded by Biocept Inc. The CNSide technology described here is the intellectual property of Biocept Inc, San Diego, California.

Sweed, Hsiao, Blouw, Pircher, Fisher, Naluz, Mayer, and Dugan are prior employees of Biocept Inc. Kesari was a consultant and advisor to Biocept Inc, and received compensation and research support. The other authors have no other relevant financial interest in the products or companies described in this article.

Sweed, Blouw, Dugan, Naluz, Mayer, Hsiao, Pircher, and Fisher are currently unaffiliated.