Conventional matching techniques of craniospinal irradiation (CSI) require patients to be in the prone position in order to use skin surface markers to align matched or gapped adjacent fields. Because of safety concerns regarding field overlap with a blind match, there has been reluctance to treat patients supine. Here, we present the clinical outcomes of patients treated with our supine CSI technique using a gradient match of brain and spinal fields.
Stephanie K. Childs, MD, Jon J. Kruse, PhD, Kevin L. Kisrow, CMD, RTT, Nadia N. Laack, MD; Mayo Clinic
Background: Conventional matching techniques of craniospinal irradiation (CSI) require patients to be in the prone position in order to use skin surface markers to align matched or gapped adjacent fields. Small setup misalignments of even 1 mm can lead to large deviations in delivered dose of up to 40%. In addition, the prone position is cumbersome for patient positioning, especially for pediatric patients requiring anesthesia for treatment. Because of safety concerns regarding field overlap with a blind match, there has been reluctance to treat patients supine. Here, we present the clinical outcomes of patients treated with our supine CSI technique using a gradient match of brain and spinal fields.
Methods: We retrospectively reviewed the charts of all patients who received CSI at Mayo Clinic from 2009–2013. All patients were simulated in the supine position, and all radiotherapy plans were created using a hybrid of forward and inverse planned intensity-modulated radiation therapy (IMRT). The inferior borders of the brain fields and superior border of the lower spine field are designed with long, gradual dose gradients created by sequential closing of the 5-mm multileaf collimator (MLC) leaves using forward planned, multiple static segment IMRT delivery. Next, a sliding window upper spine IMRT field is created by the inverse planning system to match the gradients of the brain and lower spine fields. Daily patient localization is confirmed by kV On-Board imaging of the cranial fields and a single posterior spine image. Subsequent shifts are calculated based on couch coordinates and planned isocenter shifts, not skin markings. To verify the robustness and safety of the CSI plans, dose was calculated, assuming a systematic setup error of 3 mm. Resulting dose deviations were +/− 15%.
Results: We identified 26 patients with a median age of 16 years (range: 3.4–65.5 y). Thirteen patients had medulloblastoma, nine had primitive neuroectodermal tumors (PNETs), two had pure germinomas, one had a mixed germ cell tumor, and one had an anaplastic glioneuronal tumor. The median dose delivered to the neuraxis was 36 Gy (range: 18–39.6 Gy), and the median boost dose to the tumor bed, posterior fossa, or ventricles was 55.8 Gy (range: 30.6–60 Gy). At a median follow-up of 2.8 years (range: 0.2–4.7 y), no patient had an isolated spinal recurrence and no patient had a clinically or radiographically apparent spinal myelopathy. One patient with medulloblastoma who presented with diffuse leptomeningeal disease and one patient with a supratentorial PNET recurred with diffuse leptomeningeal disease throughout the neuraxis. Seven of the nine patients with PNETs had an intracranial recurrence, five of which were in the primary tumor bed.
Conclusions: We saw no clinical evidence of isolated spinal recurrences or myelopathies at the field gradients using our hybrid forward and inverse planned IMRT supine CSI technique. Smooth field transitions obviate the need for match line shifts, allowing patients to be treated more comfortably in the supine position. In addition, the hybrid IMRT plans are less susceptible to geometric inaccuracies and the resulting dose deviations.