Session Item

November 28
10:30 - 11:30
Clinical Stream 2
Proffered papers 4: CNS
Proffered Papers
11:00 - 11:10
Patterns of Failure in Pediatric Medulloblastoma and Implications for Hippocampal Sparing


Patterns of Failure in Pediatric Medulloblastoma and Implications for Hippocampal Sparing
Authors: BALIGA|, Sujith(1)*[];Bajaj|, Benjamin (2);Vanbenthuysen|, Liam(2);Adams|, Judith(2);Gallotto|, Sarah(2);Weyman|, Elizabeth A. (2);Lawell|, Miranda P. (2);Tarbell|, Nancy J(2);MacDonald|, Shannon M. (2);Yock|, Torunn I. (2);
(1)The Ohio State University Comprehensive Cancer Center, Radiation Oncology, Columbus, USA;(2)Massachusetts General Hospital/Harvard Medical School, Radiation Oncology, Boston, USA;
Show Affiliations
Purpose or Objective

Hippocampal Sparing (HS) has been associated with preservation of memory in adult patients with brain metastases. This approach has not yet been prospectively evaluated in the pediatric population due to concerns of increased disease relapse in the peri-hippocampal region. We aim to determine the patterns of local, distant, and peri-hippocampal failures in a cohort of patients treated with proton beam therapy (PBT) for pediatric medulloblastoma (MB).

Material and Methods

We performed a retrospective review and identified patients with histologically confirmed standard-risk (SR) or high-risk (HR) MB, who were treated with PBT at the Massachusetts General Hospital from January 2000-December 2016. Patients were treated with surgery, radiation (CSI and boost) and chemotherapy.  Overall Survival (OS) and Event-free Survival (EFS) were evaluated using the Kaplan Meier method. Univariate and multivariate analysis were performed using a Cox proportional hazards model. The covariates included were age, histology, extent of resection, RT duration, and sex.  Local, distant, and peri-hippocampal failures (defined as a failure within 15 mm of the hippocampus) were tabulated.


There were a total of 144 patients including in the analysis, (SR: 102 (70.8%) HR:  42 (29.1%)). The median follow-up was 4.8 years (Range: 1-14.7). 90% of patients had a GTR/near GTR. The 5-year EFS was 80.4% for SR and 64.6% for HR patients. The 5-year OS was 85.3% for SR and 68.2% for HR patients. On univariate analysis, anaplastic histology (HR: 2.45, p=0.02) and subtotal resection (HR: 2.89, p=0.009) were associated with inferior EFS. Multivariate analysis confirmed that anaplastic histology (HR: 2.89, p=0.008) and STR (HR: 3.57, p=0.0037) remained significant for inferior EFS. 34 patients progressed, with a 5-year cumulative incidence of local and distant recurrence for the entire cohort of 16.0% and 13.3%, respectively. In 25 patients (74%), detailed imaging of the recurrence was available. No patient failed within the contoured hippocampus. The incidence of peri-hippocampal metastases in standard risk (n=13) and high risk (n=12) patients who failed was 7.7% in SR (1/13) and 8.8% in HR (1/12).


This is the first study to demonstrate the low incidence of peri-hippocampal failures in a proton-radiotherapy treated medulloblastoma cohort and provides the rationale to study hippocampal sparing in a prospective study of standard risk MB patients.  This approach may improve neuro-cognitive outcomes in the standard risk MB cohort.