Open AccessMBRS-56. RE-EVALUATION OF LEPTOMENINGEAL METASTASIS IN MEDULLOBLASTOMA WITH MAGNETIC RESONANCE IMAGING, RELATED SYMPTOMS AND CEREBROSPINAL FLUID METABOLOMIC PROFILES
Author(s) -
Ji Hye Im,
Byong Chul Yoo,
Jun Hwa Lee,
KyungHee Kim,
Jong Heon Kim,
Hyeon Jin Park,
Meerim Park,
Ji Woong Kwon,
Sang Hoon Shin,
Heon Yoo,
Seung Ah Choi,
Jae Yeol Yang,
Ji Hoon Phi,
Seung-Ki Kim,
Ho-Shin Gwak
Publication year - 2020
Publication title -
neuro-oncology
Language(s) - English
Resource type - Journals
eISSN - 1523-5866
pISSN - 1522-8517
DOI - 10.1093/neuonc/noaa222.561
Subject(s) - cytology , medulloblastoma , magnetic resonance imaging , cerebrospinal fluid , medicine , pathology , radiology
BACKGROUND Diagnosis of leptomeningeal metastasis (LM) in medulloblastoma is made by positive findings in either MRI or CSF cytology. We studied if CSF metabolomics profile can differentiate the discordant results between MRI and CSF cytology and reflect the sampling time related to treatment. MATERIALS AND METHODS We prospectively collected 83 CSF samples from 45 medulloblastoma patients. A total of 6,527 low-mass ions (LMIs) were detected using liquid chromatography tandem mass spectrometry (LC-MS/MS). Discriminative low-mass ions (LMIs) between four different MRI and cytology results groups were evaluated and representative LMIs were identified. RESULTS CSF cytology and MRI finding were both positive for LM in 8 samples and both negative for 47 samples. Tests were cytology (-) and MRI (+) in 20 samples, whereas cytology (+) and MRI (-) status were in the remaining 8 samples. The diagnostic accuracy by area under the curve (AUC) was 0.722 for cytology and 0.888 for MRI each. Based on the exclusiveness of LMI between groups, we verified 27 discriminative LMIs in MRI (+)/ cytology (+), 9 LMIs in MRI (+)/ cytology (-), and 12 LMIs in MRI (-) and cytology (+) group, separately. Metabolic pathways involved in MRI (+)/ cytology (+) group were linoleic acid, phenylalanine, TCA cycle, retinol, arginine-ornithine, nicotinate-nicotinamide, etc. Low-mass-ion discriminant equation (LOME), which could differentiate both different MRI and cytology results and the sampling time or presence of LM-related symptoms was found. CONCLUSION Non-targeted MSanalysis CSF metabolite in medulloblastomas revealed significantly different profiles, and these results suggest LMI profiles might have a higher sensitivity for LM diagnosis than either MRI or cytology.