Cerebrospinal fluid circulating tumor cells: a novel tool to diagnose leptomeningeal metastases from epithelial tumors

Background: Diagnosis of leptomeningeal metastasis (LM) remains challenging due to low sensitivity of CSF cytology and infrequent unequivocal MRI findings. In a previous pilot study, we showed that rare cell capture technology (RCCT) could be used to detect circulating tumor cells (CTC) in the CSF of patients with LM from epithelial tumors. To establish the diagnostic accuracy of CSF-CTC in the diagnosis of LM, we applied this technique in a distinct, larger cohort of patients. Methods: In this institutional review board-approved prospective study, patients with epithelial tumors and clinical suspicion of LM underwent CSF-CTC evaluation and standard MRI and CSF cytology examination. CSF-CTC enumeration was performed through an FDA-approved epithelial cell adhesion molecule-based RCCT immunomagnetic platform. LM was defined by either positive CSF cytology or imaging positive for LM. ROC analysis was utilized to define an optimal cutoff for CSF-CTC enumeration. Results: Ninety-five patients were enrolled (36 breast, 31 lung, 28 others). LM was diagnosed in 30 patients (32%) based on CSF cytology (n = 12), MRI findings (n = 2), or both (n = 16). CSF-CTC were detected in 43/95 samples (median 19.3 CSF-CTC/mL, range 0.3 to 66.7). Based on ROC analysis, 1 CSF-CTC/mL provided the best threshold to diagnose LM, achieving a sensitivity of 93%, specificity of 95%, positive predictive value 90%, and negative predictive value 97%. Conclusions: We defined >= 1 CSF-CTC/mL as the optimal cutoff for diagnosis of LM. CSF-CTC enumeration through RCCT is a robust tool to diagnose LM and should be considered in the routine LM workup in solid tumor patients.