- Email: [email protected]
Dosimetric Accuracy of Ray-tracing Algorithm for Treatment of Thoracic Spine Lesions Using Robotic Radiosurgery
S280
International Journal of Radiation Oncology Biology Physics
displacements in the AP, LR, and SI directions were 0.44mm, 0.59mm and 0.55mm, and maximal displacements were 1.54mm, 2.56mm, and 2.21mm, respectively. In the AP, LR, and SI directions, bulk displacements were less than 0.25 mm in 8%, 0%, and 0%, respectively and more than 1.5mm in 8% of patients in the AP and LR directions, and 15% of patients in the SI direction. Conclusions: There is spinal cord motion; however, physiologic oscillatory motion was generally minor. In select patients, random bulk motion may be significant and more than 1.5mm, which reinforces the role of near-rigid body immobilization during spine SBRT. Author Disclosure: C. Tseng: None. M. Sussman: None. A. Simeonov: None. D. Letourneau: None. E. Yu: None. A. Sahgal: None.
Oncology, Providence, RI, 3Rhode Island Hospital, Department of Neurosurgery, Providence, RI
2162 Reirradiation With Stereotactic Body Radiation Therapy (SBRT): Analysis of Human Spinal Cord Tolerance Using the Generalized Linear-quadratic Model Z. Huang,1 N.A. Mayr,2 W.T. Yuh,2 A. Sahgal,3 J.Z. Wang,2 and S.S. Lo4; 1 East Carolina University, Greenville, NC, 2Ohio State University, Columbus, OH, 3University of Toronto, Toronto, ON, Canada, 4Case Western Reserve University, Cleveland, OH Purpose/Objective(s): Optimal dose delivery, while preventing devastating neurologic complications, is of utmost importance for spinal SBRT, particularly in the re-treatment setting; and defining spinal cord dose constraints is challenging. The recently described generalized linearquadratic (gLQ) model provides more consistent response modeling across conventional through ablative-dose ranges, but the gLQ model has not been used clinically in spinal SBRT. Using the gLQ model, we reanalyzed published dosimetric data of patients with radiation myelopathy (RM) after spinal SBRT, who had been previously treated with conventionally fractionated radiation (RTconv). Materials/Methods: Based on thecal sac (TS) contours for SBRT, spinal cord dose-volume histograms (DVHs) of 5 RM patients (5 spinal segments) and 14 no-RM patients (16 spinal segments) were reanalyzed. The maximum dose point (Pmax) within the TS and doses to 0.1-, 1.0-, and 2.0-cc TS volumes were obtained. The gLQ-based biologically effective doses were calculated and normalized to 2-Gy equivalent dose (nBEDgLQ Z Gy2/2_gLQ). The initial RTconv dose, converted to Gy2/2_gLQ, was added. Results: For the pre-treatment RTconv, nBEDgLQ was similar among noRM and RM patients (27.8-52.5, median: 38.6; and 20.0-50.8, median: 38.8 Gy2/2_gLQ, respectively).The mean Pmax nBEDgLQ to the TS from SBRT reirradiation was significantly lower in no-RM patients compared to RM patients: 18.0 vs 57.0 Gy2/2_gLQ (p Z 1E-6). Similarly, total (RTconv+ SBRT) mean Pmax nBEDgLQ was lower in no-RM than RM patients: 62.2 (range: 42.9-82.4) Gy2/2_gLQ vs. 94.8 (range: 70.2-133.4) Gy2/2_gLQ (p Z 0.005). The proportion of total Pmax nBEDgLQ accounted for by the hypofractionated SBRT Pmax nBEDgLQ was 0.52-0.74 (median: 0.58) for RM patients, compared to 0.04-0.49 (median: 0.28) for no-RM patients. In RM patients, mean nBEDgLQ to Pmax, 0.1-, 1.0- and 2.0-cc was 57.0, 48.3, 34.3, and 26.8 Gy2/2_gLQ. No RMs were seen below a total (RTconv+ SBRT) spinal cord nBEDgLQ of 70 Gy2/2_gLQ. Conclusions: Based on this clinical data, the gLQ-derived spinal cord dose threshold for total nBEDgLQ, incorporating both conventional-fraction pretreatment and ablative-fraction re-treatment, was 70 Gy2/2_gLQ. However, with the current scarcity of clinical data, this approach must be further clinically validated. Author Disclosure: Z. Huang: None. N.A. Mayr: None. W.T. Yuh: None. A. Sahgal: None. J.Z. Wang: None. S.S. Lo: None.
2163 Dosimetric Accuracy of Ray-tracing Algorithm for Treatment of Thoracic Spine Lesions Using Robotic Radiosurgery R.L. Ebeling III,1 J. Hiatt,2 O. Adetokunbo,3 M. Goldman,3 B. Curran,2 T. DiPetrillo,1 T. Kinsella,2 and J.T. Hepel2; 1Tufts University Affiliated Hospitals, Boston, MA, 2Rhode Island Hospital, Department of Radiation
Purpose/Objective(s): Frequently, dose adjustments or optimization using Monte Carlo are employed to correct for tissue homogeneity inaccuracies of the ray-tracing dose calculation algorithm used for robotic radiosurgery treatment planning of lung tumors. Spine radiosurgery patients are required to be treated in the supine position to allow for accurate intra-treatment spine tracking; yet posterior oriented beams are not possible with this system. As such, in the treatment of thoracic spine lesions the majority of treatment beams traverse the lungs. In addition, high-density surgical hardware is often present in and around the target lesion. The dosimetric impact of inaccuracies in correcting for these marked differences in tissue density for the treatment planning of thoracic spine lesions is not known. Materials/Methods: Thirty-one patients with 43 thoracic vertebral lesions treated in our department using robotic radiosurgery between 2009 and 2012 were identified. All patients were planned using the robotic radiosurgery Multiplan system employing ray-tracing algorithm (RTA) for tissue heterogeneity correction to a median dose of 24 Gy in 1-5 fractions. For each patient, a comparison plan was generated using Monte Carlo dose calculation algorithm (MCA) using identical beam arrangement and monitor units. Plans were compared for differences in GTV coverage by the prescription isodose line, GTV D90%, spinal cord maximum dose, and dose to 0.25cc of spinal cord. Results: Mean GTV coverage was 26.3% lower for the MCA as compared with RTA, 60.3% vs 86.6%, respectively (p < 0.0001). Similarly, mean GTV D90% was 2.7% lower for MCA (range: -7.34% to 9.26%, p Z 0.001). Conversely, mean doses to the spinal cord were significantly higher using MCA by 6.9% (range: -9.8% to 24.1%) for maximum dose and by 1.2% (range: -8.9% to 10.6%) for 0.25cc; p < 0.0001 and p < 0.01, respectively. Presence of surgical hardware (n Z 8) had only a small effect with mean differences of -5.4% for GTV coverage (p Z 0.6), -2.2% for GTV D90% (p Z 0.3), 4.3% for cord max (p Z 0.2), and 5.4% for cord 0.25cc (p Z 0.04). Conclusion: Ray-tracing dose calculation algorithm for the thoracic spine overestimates GTV coverage and underestimates spinal cord dose. This can result in significant impact on both disease control and potential for cord injury, especially in the retreatment setting where treatment doses and cord constraints are pushed to tolerance levels. Optimization of treatment plans using Monte Carlo dose calculation algorithm is, thus, warranted in these patients. Author Disclosure: R.L. Ebeling, III: None. J. Hiatt: None. O. Adetokunbo: None. M. Goldman: None. B. Curran: None. T. DiPetrillo: None. T. Kinsella: None. J.T. Hepel: None.
2164 Analysis of Robustness of a Combined Cranial Photon and MLCbased Spinal Proton Field Matching Technique for Delivery of Craniospinal Irradiation H. Zhai, C. Hill-Kayser, P. James, R. Lustig, H. Lin, J. Mcdonough, Z. Tochner, and S. Both; The Hospital of University of Pennsylvania, Philadelphia, PA Purpose/Objective(s): In general, many patients requiring craniospinal irradiation (CSI) are children, for whom radiation therapy is frequently associated with severe side effects, such as fertility dysfunction, cardiac morbidity, and secondary malignancies. Proton beam treatment (PBT) has increased our ability to maximize the dose to the tumor while sparing normal structures, reducing the risk of acute as well as late and long-term side effects. CSI patients are treated in a prone position and careful junctions are created between cranial photon and posterior spinal proton fields. In this study, we examine the robustness of this novel technique. Materials/Methods: Nine patients were treated with opposed lateral cranial photon beams (6 MV) and two to four posterior passive scattered spinal proton beams. The patients were positioned prone and immobilized using a thermoplastic frame and vacuum bag. Daily orthogonal kV imaging was used for alignment based on bony anatomy. Weekly verification CTs were acquired during the treatment. The verification CTs were
International Journal of Radiation Oncology Biology Physics
displacements in the AP, LR, and SI directions were 0.44mm, 0.59mm and 0.55mm, and maximal displacements were 1.54mm, 2.56mm, and 2.21mm, respectively. In the AP, LR, and SI directions, bulk displacements were less than 0.25 mm in 8%, 0%, and 0%, respectively and more than 1.5mm in 8% of patients in the AP and LR directions, and 15% of patients in the SI direction. Conclusions: There is spinal cord motion; however, physiologic oscillatory motion was generally minor. In select patients, random bulk motion may be significant and more than 1.5mm, which reinforces the role of near-rigid body immobilization during spine SBRT. Author Disclosure: C. Tseng: None. M. Sussman: None. A. Simeonov: None. D. Letourneau: None. E. Yu: None. A. Sahgal: None.
Oncology, Providence, RI, 3Rhode Island Hospital, Department of Neurosurgery, Providence, RI
2162 Reirradiation With Stereotactic Body Radiation Therapy (SBRT): Analysis of Human Spinal Cord Tolerance Using the Generalized Linear-quadratic Model Z. Huang,1 N.A. Mayr,2 W.T. Yuh,2 A. Sahgal,3 J.Z. Wang,2 and S.S. Lo4; 1 East Carolina University, Greenville, NC, 2Ohio State University, Columbus, OH, 3University of Toronto, Toronto, ON, Canada, 4Case Western Reserve University, Cleveland, OH Purpose/Objective(s): Optimal dose delivery, while preventing devastating neurologic complications, is of utmost importance for spinal SBRT, particularly in the re-treatment setting; and defining spinal cord dose constraints is challenging. The recently described generalized linearquadratic (gLQ) model provides more consistent response modeling across conventional through ablative-dose ranges, but the gLQ model has not been used clinically in spinal SBRT. Using the gLQ model, we reanalyzed published dosimetric data of patients with radiation myelopathy (RM) after spinal SBRT, who had been previously treated with conventionally fractionated radiation (RTconv). Materials/Methods: Based on thecal sac (TS) contours for SBRT, spinal cord dose-volume histograms (DVHs) of 5 RM patients (5 spinal segments) and 14 no-RM patients (16 spinal segments) were reanalyzed. The maximum dose point (Pmax) within the TS and doses to 0.1-, 1.0-, and 2.0-cc TS volumes were obtained. The gLQ-based biologically effective doses were calculated and normalized to 2-Gy equivalent dose (nBEDgLQ Z Gy2/2_gLQ). The initial RTconv dose, converted to Gy2/2_gLQ, was added. Results: For the pre-treatment RTconv, nBEDgLQ was similar among noRM and RM patients (27.8-52.5, median: 38.6; and 20.0-50.8, median: 38.8 Gy2/2_gLQ, respectively).The mean Pmax nBEDgLQ to the TS from SBRT reirradiation was significantly lower in no-RM patients compared to RM patients: 18.0 vs 57.0 Gy2/2_gLQ (p Z 1E-6). Similarly, total (RTconv+ SBRT) mean Pmax nBEDgLQ was lower in no-RM than RM patients: 62.2 (range: 42.9-82.4) Gy2/2_gLQ vs. 94.8 (range: 70.2-133.4) Gy2/2_gLQ (p Z 0.005). The proportion of total Pmax nBEDgLQ accounted for by the hypofractionated SBRT Pmax nBEDgLQ was 0.52-0.74 (median: 0.58) for RM patients, compared to 0.04-0.49 (median: 0.28) for no-RM patients. In RM patients, mean nBEDgLQ to Pmax, 0.1-, 1.0- and 2.0-cc was 57.0, 48.3, 34.3, and 26.8 Gy2/2_gLQ. No RMs were seen below a total (RTconv+ SBRT) spinal cord nBEDgLQ of 70 Gy2/2_gLQ. Conclusions: Based on this clinical data, the gLQ-derived spinal cord dose threshold for total nBEDgLQ, incorporating both conventional-fraction pretreatment and ablative-fraction re-treatment, was 70 Gy2/2_gLQ. However, with the current scarcity of clinical data, this approach must be further clinically validated. Author Disclosure: Z. Huang: None. N.A. Mayr: None. W.T. Yuh: None. A. Sahgal: None. J.Z. Wang: None. S.S. Lo: None.
2163 Dosimetric Accuracy of Ray-tracing Algorithm for Treatment of Thoracic Spine Lesions Using Robotic Radiosurgery R.L. Ebeling III,1 J. Hiatt,2 O. Adetokunbo,3 M. Goldman,3 B. Curran,2 T. DiPetrillo,1 T. Kinsella,2 and J.T. Hepel2; 1Tufts University Affiliated Hospitals, Boston, MA, 2Rhode Island Hospital, Department of Radiation
Purpose/Objective(s): Frequently, dose adjustments or optimization using Monte Carlo are employed to correct for tissue homogeneity inaccuracies of the ray-tracing dose calculation algorithm used for robotic radiosurgery treatment planning of lung tumors. Spine radiosurgery patients are required to be treated in the supine position to allow for accurate intra-treatment spine tracking; yet posterior oriented beams are not possible with this system. As such, in the treatment of thoracic spine lesions the majority of treatment beams traverse the lungs. In addition, high-density surgical hardware is often present in and around the target lesion. The dosimetric impact of inaccuracies in correcting for these marked differences in tissue density for the treatment planning of thoracic spine lesions is not known. Materials/Methods: Thirty-one patients with 43 thoracic vertebral lesions treated in our department using robotic radiosurgery between 2009 and 2012 were identified. All patients were planned using the robotic radiosurgery Multiplan system employing ray-tracing algorithm (RTA) for tissue heterogeneity correction to a median dose of 24 Gy in 1-5 fractions. For each patient, a comparison plan was generated using Monte Carlo dose calculation algorithm (MCA) using identical beam arrangement and monitor units. Plans were compared for differences in GTV coverage by the prescription isodose line, GTV D90%, spinal cord maximum dose, and dose to 0.25cc of spinal cord. Results: Mean GTV coverage was 26.3% lower for the MCA as compared with RTA, 60.3% vs 86.6%, respectively (p < 0.0001). Similarly, mean GTV D90% was 2.7% lower for MCA (range: -7.34% to 9.26%, p Z 0.001). Conversely, mean doses to the spinal cord were significantly higher using MCA by 6.9% (range: -9.8% to 24.1%) for maximum dose and by 1.2% (range: -8.9% to 10.6%) for 0.25cc; p < 0.0001 and p < 0.01, respectively. Presence of surgical hardware (n Z 8) had only a small effect with mean differences of -5.4% for GTV coverage (p Z 0.6), -2.2% for GTV D90% (p Z 0.3), 4.3% for cord max (p Z 0.2), and 5.4% for cord 0.25cc (p Z 0.04). Conclusion: Ray-tracing dose calculation algorithm for the thoracic spine overestimates GTV coverage and underestimates spinal cord dose. This can result in significant impact on both disease control and potential for cord injury, especially in the retreatment setting where treatment doses and cord constraints are pushed to tolerance levels. Optimization of treatment plans using Monte Carlo dose calculation algorithm is, thus, warranted in these patients. Author Disclosure: R.L. Ebeling, III: None. J. Hiatt: None. O. Adetokunbo: None. M. Goldman: None. B. Curran: None. T. DiPetrillo: None. T. Kinsella: None. J.T. Hepel: None.
2164 Analysis of Robustness of a Combined Cranial Photon and MLCbased Spinal Proton Field Matching Technique for Delivery of Craniospinal Irradiation H. Zhai, C. Hill-Kayser, P. James, R. Lustig, H. Lin, J. Mcdonough, Z. Tochner, and S. Both; The Hospital of University of Pennsylvania, Philadelphia, PA Purpose/Objective(s): In general, many patients requiring craniospinal irradiation (CSI) are children, for whom radiation therapy is frequently associated with severe side effects, such as fertility dysfunction, cardiac morbidity, and secondary malignancies. Proton beam treatment (PBT) has increased our ability to maximize the dose to the tumor while sparing normal structures, reducing the risk of acute as well as late and long-term side effects. CSI patients are treated in a prone position and careful junctions are created between cranial photon and posterior spinal proton fields. In this study, we examine the robustness of this novel technique. Materials/Methods: Nine patients were treated with opposed lateral cranial photon beams (6 MV) and two to four posterior passive scattered spinal proton beams. The patients were positioned prone and immobilized using a thermoplastic frame and vacuum bag. Daily orthogonal kV imaging was used for alignment based on bony anatomy. Weekly verification CTs were acquired during the treatment. The verification CTs were