Identifying Leptomeningeal Dissemination in Patients with Medulloblastomas: Cerebrospinal Fluid Cytology Compared to Magnetic Resonance Imaging

Sunday, May 5, 2024

4:31 PM – 4:33 PM CT

Location: S405b

Presenting Author(s)

Zain Peeran, BS

Medical Student
UCSF
San Francisco, CA, US

Co-Author(s)

Alexander A. Aabedi, MD

Resident Physician
Department of Neurological Surgery, University of California, San Francisco
San Francisco, California, United States
Manish Aghi, MD, PhD

PI
UCSF
San Francisco, California, United States
EF

Edward Fisher, n/a

Research Fellow
UCSF, United States
NG

Nalin Gupta, MD, PhD

PI
UCSF
San Francisco, California, United States
HK

Harmon Khela, BS

Medical Student
UPenn, United States
CK

Cassie Kline, MD, MAS

PI
UPenn, United States
RP

Ryan Phelps, MD

Research Fellow
UCSF
Palo Alto, California, United States
AR

Alyssa Reddy, MD

Physician
UCSF, United States
JR

Jennifer Rios, BS

Medical Student
UCSF, United States

Disclosure(s):

Zain Peeran, BS: No financial relationships to disclose

Introduction: For medulloblastomas, it is believed CSF cytology and MR imaging together are superior to either modality alone in identifying leptomeningeal dissemination (LMD). However, MRI has markedly improved since the initial assessment of this approach, and it remains unclear whether differences exist between medulloblastoma subtypes.

Methods: The records of 132 patients with medulloblastomas were retrospectively reviewed. CSF samples and MR images obtained within 6 weeks of each other from 2000-2023 were included. A total of 311 CSF-MRI pairs (range: 1-12 pairs per patient) were analyzed. A 3-tier system (negative, atypical/suspicious, positive) was utilized to account for equivocal findings.

Results: 62.9% of patients were male. Mean (SD) age at time of surgery was 13.2 (12.9) years. 71.2% received gross or near-total resection, 81.1% received radiotherapy, and 93.2% received chemotherapy. 18.9% were deceased, with the mean (SD) time till death being 4.3 (3.4) years following surgery. The mean (SD) follow-up time was 7.0 (5.1) years. The molecular subtypes were: SHH (18.0%), WNT (3.5%), non-WNT/non-SHH (51.8%), unknown (26.7%). There was a 59.8% concordance between CSF cytology and MRI. MRI had an 89.0% sensitivity and CSF cytology had a 23.1% sensitivity for detecting LMD that was positive on CSF and/or MRI. MRI demonstrated greater or equal ability to CSF cytology in 94.2% of cases. In all cases where CSF cytology was positive and MRI was negative (n=5, 1.6%), the CSF samples and MR images were obtained within 10 days of surgery. Concordance rates differed by subtype (75.0% SHH vs. 72.7% WNT vs. 50.9% NWNS; p< 0.01). The percentage of cases in which MRI exceeded CSF cytology in detecting LMD also differed by subtype (25% SHH vs. 18.2% WNT vs. 44.7% NWNS; p=0.01).

Conclusion : CSF cytology may not be routinely required for LMD surveillance as MRI detects most cases. However, this may vary by medulloblastoma subtype and time since surgery.