Supine Craniospinal Irradiation: Early Results on Patterns of Failure

Abstract

There has been a reluctance in the radiation oncology community to perform craniospinal irradiation (CSI) in the supine position as the prone position allows the physician and therapist to visualize junctions and skin gaps on the patient, minimizing the risk of underdosing the tumor or overdosing the spinal cord. Some of the advantages of the supine craniospinal set-up include patient comfort, easier access to maintaining an airway in anesthesized children and reduced variability of head tilt in the face mask. Here we describe our early experience with supine CSI. From June 2004 to December 2005, 23 patients (pts) were treated at 1 facility using the supine CSI approach. There were 14 male and 9 female pts with a median age of 8.9 years (range, 2.7 to 23 years). Diagnosis included medulloblastoma in 15, germ cell tumor in 5, pineoblastoma in 1 and rhabdoid tumor in 2. Anesthesia was given to 11 pts. During simulation an Orfit mask and body cast were made for each child. A CT scan of the brain down to the inferior pelvis was performed using 2–5 mm slice thickness with a Picker PQ5000 scanner. These images were transferred to the ADAC planning system. Cranial fields were treated with a parallel-opposed lateral field set-up with the table kicked to match the spinal field. Cranial fields were also collimated to match the divergence from the proximal spinal field. Spine fields were treated using an SSD technique with all isocenters set to the couch top. All superior spine fields were set to a gantry of 180 degrees with a SSD of approximately 106 cm. When a second field was required, the inferior spine field was set to match the divergence of the superior spine field below the level of L2 vertebra when possible. To achieve this, the table and collimator were rotated 90 degrees to allow for the use of the multileaf collimator and to rotate gantry to match the divergence of the superior spinal field. In 6 patients, 2 spinal fields were employed. Junction shifts were performed in 2 ways. The first 12 pts' shifts at the cranial-spinal and spinal-spinal fields, when applicable, were performed every 5 treatments. In the remaining 11 pts, daily dynamic feathering of the junctions were completed along with field-in-field dose control. Dose to the craniospinal axis varied from 1800 to 3960 cGy. Wires were placed at the lower most portion of the cranial field and were visually matched to the upper spine field using port films. With a median follow-up of 17 months, there have been no cranial-spinal or spinal-spinal junction failures. There have been no cases of radiation myelitis. There have been 5 documented failures: 2 primary site alone, 2 spinal axis alone and 1 primary site and distant metastasis. Median time to recurrence was 17 months with a range of 1 to 23 months. The supine approach in delivering CSI has not been associated with increased relapses at field junctions or radiation myelitis. Longer follow-up is needed to determine if there will be future junctional failures or cases of myelitis.