We retrospectively evaluated rates of central nervous system toxicity following proton therapy in potentially high-risk patients in order to assess the incidence, clinical and dosimetric risk factors, and natural course of radiation treatment effects.
Nicholas S. Boehling, MD, David R. Grosshans, MD, PhD, Nikhil G. Thaker, MD, Mark A. Edson, MD, PhD, Anita Mahajan, MD; Department of Radiation Oncology, UT MD Anderson Cancer Center
Background: Recently reported studies have indicated low rates of radiation-induced central nervous system (CNS) toxicity following proton therapy in patients ≤ 18 years of age. However, subgroup analysis revealed potentially higher rates of toxicity in younger patients or those receiving methotrexate. We retrospectively evaluated rates of CNS toxicity following proton therapy in potentially high-risk patients in order to assess the incidence, clinical and dosimetric risk factors, and natural course of radiation treatment (RT) effects.
Methods: Patients who were treated for brain tumors at the MD Anderson Proton Therapy Center from 2006–2013, ≤ 9 years of age at the time of radiation, and receiving local-field RT only with or without chemotherapy were included. Eighty patients, all of whom were treated as part of a prospective data collection protocol, were identified. Clinical records as well as serial magnetic resonance imaging (MRI) were reviewed, and dosimetric factors were recorded. Patients with less than 1 month of imaging follow-up were excluded.
Results: A total of 60 patients met the follow-up criteria; 30 males and 30 females were included, with a median age of 2.6 years (range: 10 mo–9 yr). Location was posterior fossa in 39 patients and supratentorial in 21 patients. Ten patients were treated for recurrent disease but had not received prior irradiation. One patient received reirradiation. The most common histologies were anaplastic ependymoma (24), ependymoma (11), medulloblastoma (10), and atypical teratoid rhabdoid tumor (ATRT) (9). Forty-seven patients underwent a gross total resection prior to radiotherapy, with the remainder having subtotal of biopsy only. Patients had received chemotherapy before, during, and after radiation in 28, 6, and 17 cases, respectively. Fourteen patients received methotrexate. Median total dose and fraction size prescribed was 54 Gy at 1.8 Gy.
With a median imaging follow-up of 13.7 months (range: 1–80 mo), a total of seven patients (12%) developed RT effect consistent with radiation necrosis, as determined on serial MRI. Six of the seven (10%) were symptomatic from the treatment effect. Of the symptomatic patients, one had treatment effect in the supratentorial brain; the others developed treatment effect in the brainstem. The median time to development of was 3.9 months (range: 2.6–4.5 mo). The median age of those who developed radiation necrosis was 1.9 years vs 2.6 for those who did not (t-test P = .4). Two patients were treated with methotrexate before radiation for ATRT. Five of the seven patients were treated for ependymoma.
Conclusions: In this small retrospective series, proton radiation for pediatric brain malignancies was associated with a 12% crude risk of radiation-induced imaging changes and a 10% risk of symptomatic necrosis. The incidence of MRI changes was similar to reported photon series. The relatively high incidence of symptomatic necrosis likely reflects the young age at treatment and common use of chemotherapy in the patient population that is included. Changes tend to occur early within 3–5 months after completion of radiation. Due to the low incidence, no statistically significant predictive factors were found. However, further study that includes dosimetric modeling is warranted.