Proceedings of the 52nd Annual ASTRO Meeting S589

and December 2009. The median age at diagnosis of the 54 patients was 7 years (range, 3-21 years). The male/female patient ratiowas 3.5:1. Patients were treated with concurrent chemotherapy (vincristine) and reduced-dose cranial spinal irradiation (2340 cGy)followed by a conformal primary tumor bed boost (3060 cGy to 3600 cGy, 47 patients received 3240 cGy). Eight patients weretreated with IMRT boost plans and 46 patients were treated with 3D conformal boost plans. After radiation therapy patients were tobe treated with 8 cycles of chemotherapy (vincristine, cisplatin, and lomustine or cyclophosphamide). The median follow-up was68 months (range, 0-179 months).

Results: The 5-year Kaplan-Meier estimated overall survival rate was 85% (±10.8%, 95% confidence interval). The 5-year diseasefree survival and freedom from posterior fossa recurrence were 81% (±11.6%, 95% confidence interval) and 94% (±7.0%, 95%confidence interval), respectively. Ten patients had disease recurrences. The patterns of first site of failure included 1 patientwith distal central nervous system recurrence, 3 patients with recurrence in the tumor bed boost fields, 1 patient with simultaneousrecurrences in the tumor bed boost fields and distal central nervous system, and 5 patients with leptomeningeal disease. No failuresoccurred elsewhere in the posterior fossa.

Conclusions: The treatment of average-risk medulloblastoma with chemotherapy, reduced-dose cranial spinal irradiation, anda conformal tumor bed boost results in survival rates and primary site control rates comparable to those in contemporary studies.Patterns of recurrence suggest that a limited tumor bed boost and whole posterior fossa dose of 2340 cGy does not compromise OSor DFS. The presence of leptomeningeal recurrences with the reduced dose posterior fossa fields is similar to what has been ob-served in other studies and raises concern for further cranial spinal dose reductions.

Author Disclosure: R.P. Ermoian, None; J.L. Barker, None; S.R. Browd, None; J.G. Ojemann, None; J.M. Olson, None; J.R.Geyer, None; R.G. Ellenbogen, None; J.G. Douglas, None.

2863 Supine Craniospinal Irradiation (SCI) in Medulloblastoma (MBL): Long-term Outcomes

S. Firat, M. L. Siker, C. J. Schultz, S. S. Jogal, B. A. Kaufman, J. J. Stratton, K. J. Murray

Medical College of Wisconsin, Wauwatosa, WI

Purpose/Objective(s): To improve patient comfort, airway safety, and head immobilization compared to prone techniques, SCIwas implemented at Medical College of Wisconsin (MCW) during the 1990s. Although early results with SCI suggest equivalentefficacy and safety compared to prone techniques, long-term outcomes are unknown. This study reports long-term outcomes ofpatients with MBL treated with SCI at MCW.

Materials/Methods: All MBL patients treated with SCI at MCW between January 1990 and December 2007 were reviewed. SCIwas delivered with parallel opposed cranial fields collimated to match the divergence of the upper spine field. The cranial border ofthe upper spine field matched the caudal edge of the cranial fields on the spinal cord at midline, without table rotation but witha calculated gap at the skin surface. The cranial border of the lower spine field matched the caudal border of the upper field atthe posterior aspect of the vertebral body. Field junctions were moved by 1 cm cranially every 5 fractions. Appropriate gapswere calculated for all field junctions. Median dose to the craniospinal axis and posterior fossa were 35.2 Gy (18-40) and 54Gy (53.2 - 55.8) at 1.6-1.8 Gy/day. Posterior fossa boost technique was opposed lateral fields in 28 patients and a 5-field 3D con-formal boost in 10. Three patients received spine boosts to 44.8 Gy using a posterior field. Chemotherapy was used in 29 patientswho were enrolled (n = 18) or treated per trials (POG/COG: 9031, 9961, ACNS0331).

Results: Of the 38 patients reviewed, median age was 7 years (3-27). Twenty-five patients had average and 13 had high risk dis-ease. Gross or near total resection of the primary tumor was achieved in 29 patients. Five- and 10-year overall survival (OS) for allpatients was 69% and 65%. Average and high risk patients had a 10-year OS of 71% and 50%. Five- and 10-year OS of 17 averagerisk patients treated with chemotherapy and SCI was 81%. Of the 19 disease-free patients with at least 5 years follow-up, 11 werefollowed for over 10 years. In this group, endocrinopathy was reported in 15 patients, with growth hormone being the most com-mon deficiency (n = 13). Neurocognitive delay was seen in 12 patients with 8 needing individualized special education services.Sensorineural hearing loss was observed in 12 patients with 8 requiring hearing aids. Two patients had vascular headaches, and 2had CNS venous angiomas requiring no intervention. Two patients had cardiac valvular disease. There were no secondary malig-nancies, nor any radiation myelitis or disease recurrence observed at field junctions.

Conclusions: SCI in MBL is not associated with an increased risk of myelitis or recurrence with long-term follow-up. Average riskpatients treated with chemotherapy and SCI had an excellent OS with SCI, similar to published reports using prone techniques.

Author Disclosure: S. Firat, None; M.L. Siker, None; C.J. Schultz, None; S.S. Jogal, None; B.A. Kaufman, None; J.J. Stratton,None; K.J. Murray, None.

2864 Time Course Quantification of Brainstem Necrosis by DTI in Medulloblastoma Patients following

Chemo-radiotherapy

C. Hua, Y. Zhang, R. J. Ogg, A. Broniscer, A. Gajjar, T. E. Merchant

St. Jude Children’s Research Hospital, Memphis, TN

Purpose/Objective(s): To quantify the evolution of therapy-induced brainstem necrosis and subsequent recovery using diffusiontensor imaging (DTI) in medulloblastoma patients treated with post-operative craniospinal irradiation and post-irradiation chemo-therapy.

Materials/Methods: Serial DTI studies were evaluated in five children (ages 5-11 years) with medulloblastoma who developedbrainstem necrosis after combined modality therapy that included CSI 36-39.6 Gy and primary site irradiation to 55.8 Gy followedby high-dose cisplatin, vincristine, and cyclophosphamide. Changes in DTI parameters were compared to 20 healthy children (6-24yrs) and 20 irradiated patients (10 medulloblastomas and 10 high-grade gliomas, 4-23 yrs) who did not develop necrosis. For eachsubject, fractional anisotropy (FA) and apparent diffusion coefficient (ADC) maps at different time points were co-registered withplanning computed tomography, T1-weighted magnetic resonance images and/or radiation dose maps. ADC and FA measure thediffusivity and the directionality of water diffusion, respectively, reflecting the integrity of white matter tracts.

S590 I. J. Radiation Oncology d Biology d Physics Volume 78, Number 3, Supplement, 2010

Results: The age-dependent increase in FA and decrease in ADC of healthy pediatric brainstems was modeled as logarithmic func-tions and served as a benchmark for comparison. For the 20 irradiated patients without brainstem necrosis, three distinctive lon-gitudinal patterns in FA were observed post irradiation: (1) rising or stable time trends similar to those observed in healthy children,(2) FA declining 10-20% during the first three years after treatment followed by a rebound to baseline values at year 4, (3) FAdeclining 15-35% with no evidence of recovery. Absolute FA values for all cases remained above 0.3 over the entire follow-upperiod. Though generally opposite to FA patterns, ADC patterns were less distinctive. Mean and maximum doses to the brainstemwere significantly different only between patients with patterns 1 and 3 (p = 0.01 and 0.03). For patients with brainstem necrosis,mean FA in the necrotic regions declined to 50% of the baseline values and below 0.3 within 5-12 months after the start of RT, butshowed signs of subsequent recovery in FA possibly due to medical intervention including corticosteroids and hyperbaric oxygentreatment.

Conclusions: Patients who developed brainstem necrosis showed early and more dramatic declines in FA compared to the late andprogressive changes in patients without necrosis. Longitudinal changes in DTI aligned with anatomical imaging findings in brain-stem necrosis patients and detected therapy effects in non-necrosis patients. Our inability to predict recovery of white matter injurybased on radiation dose alone implies that other clinical factors may be relevant.

Author Disclosure: C. Hua, None; Y. Zhang, None; R.J. Ogg, None; A. Broniscer, None; A. Gajjar, None; T.E. Merchant, None.

2865 Local Control of Metastatic Sites with Radiation Therapy in Metastatic Ewings and Rhabdomyosarcoma

A. K. Liu1, M. Stinauer1, T. Tello1, K. Maloney2

1University of Colorado Denver, Aurora, CO, 2The Children’s Hospital, Denver, Aurora, CO

Purpose/Objective(s): Ewing’s sarcoma (EWS) and rhabdomyosarcoma (RMS) are two of the most common sarcomas in chil-dren and outcomes with localized disease are good with overall survival rates of 70% - 80%. However, 20% of newly diagnosedchildren present with metastatic disease. Treatment for these children includes chemotherapy, local control of the primary site andradiation therapy to sites of metastasis. Even with intensive therapy, overall survival is only 30% at 5 years. There is little data onlocal control of metastatic sites in EWS and RMS, and as a result, it is unclear whether treatment of metastatic sites is beneficial. Weevaluated the role of radiation therapy in the control of metastatic sites in the treatment of metastatic EWS and RMS.

Materials/Methods: We reviewed our experience in children with metastatic EWS and RMS treated at the Children’s Hospital inDenver and at the Department of Radiation Oncology at University of Colorado Denver from 1995 to 2008. We included childrenwho presented with metastatic disease and received radiation therapy to curative doses (greater than 40 Gy in fractionated treat-ments) to each site of metastasis. Children who only received whole lung radiation or palliative doses of radiation therapy wereexcluded. Clinical information including surgery, chemotherapy and radiation therapy was reviewed.

Results: Thirteen children with metastatic disease were treated at our institutions. Five children had metastatic EWS and 8 hadmetastatic RMS. The median age at diagnosis was 13.6 yrs (range, 11.2 to 18.1 yrs) with 6 males and 7 females. The most commonsite of primary disease was the trunk followed by the extremities. Bone was the most common metastatic site. The majority ofpatients were treated with vincristine, doxorubicin, cyclophosphamide, ifosfamide and etoposide. Median radiation dose was50.4 Gy (range, of 41.4 to 59.4 Gy). With median follow-up of 14.4 months (range, 5.2 to 69.4 months) local failure at metastaticsites was seen in 2 patients for an overall local control rate of 76.9% at 5 years. The median time to local failure was not reached, butin patients who failed it was 5.5 months. Overall survival was 40.1% at 5 years with a median time of 38.8 months.

Conclusions: Current literature provides little data regarding local control of metastatic sites in EWS and RMS. We had someconcern that metastatic disease would correlate with poor local control. However, our data would suggest that despite the high bur-den of disease, local control of metastatic sites is just as effective as in the treatment of localized disease. This suggests thataggressive local treatment of metastatic sites is an important component in the management of metastatic EWS and RMS.

Author Disclosure: A.K. Liu, None; M. Stinauer, None; T. Tello, None; K. Maloney, None.

2866 Impact of Kilo-voltage Imaging Doses to the Radiotherapy of Pediatric Cancer Patients

J. Deng, Z. Chen, K. Roberts, R. Nath

Yale University School of Medicine, New Haven, CT

Purpose/Objective(s): In image-guided radiation therapy, frequent use of kilo-voltage cone beam computed tomography(kVCBCT) can potentially add a significant amount of dose to the patients and the treatment plan. Yet, this portion ofkVCBCT-induced radiation dose has so far not been accounted for, due largely to the inability of dose calculation algorithmsin commercial treatment planning systems (TPS) for kilo-voltage photon beams. This portion of doses from imaging studies shouldbe incorporated into treatment planning for more accurate dose accumulation and treatment design, especially in pediatric cancerpatients. Children are far more susceptible to radiation-induced second malignancies than adults and the radiation-induced adverseeffects on growth and development of normal tissues can cause extra morbidity and late side effects in the children. The purpose ofthis work is to investigate kVCBCT-induced doses to pediatric cancer patients undergoing radiotherapy and strategies for optimaldose reduction.

Materials/Methods: Four pediatric cases were selected with patient CTs and contours exported from Eclipse TPS via DICOM andconverted into EGS4 patient phantoms. An EGS4 Monte Carlo code was employed to calculate dose distributions on four patientsscanned by kVCBCT. 3D dose distributions and DVHs were analyzed. In addition, distance from organ-at-risk (OAR) to CBCTfield border, kilo-voltage peak energy and testicular shielding have been studied for strategies to reduce doses to OARs.

Results: Due to circular gantry rotations of CBCT and smaller dimensions in pediatric patients compared to adults, kVCBCT-induced doses were quite symmetrically deposited around patient anatomy, with much higher doses to bony structures due to en-hanced photoelectric effect. DVH analyses indicated mean doses induced by kVCBCT ranged from 2.9 cGy to testes to 10.5 cGy tofemur heads. Increasing distances from OARs to CBCT field border would greatly cut doses to OARs, ranging from 50% reductionfor spinal cord to 2300% reduction for testes. Reducing beam energy from 125 to 60 kV generally increased doses to OARs, but