- Email: [email protected]
INVESTIGATION OF THE DIFFERENCE IN PB AND AAA CALCULATED DOSE DISTRIBUTIONS FOR DYNAMIC IMRT
IMRT: R OTATIONAL THERAPIES
519 poster (Physics Track) HIPPOCAMPAL AVOIDANCE WITH RAPID ARC AND HELICAL TOMOTHERAPY FOR BASE OF SKULL TUMORS E. Wiebe1 , L. Cozzi2 , S. Yartsev3 , A. Fogliata Cozzi2 , E. Vanetti de Palma2 , A. Clivio4 , G. Nicolini4 , J. Chen5 , A. Leung5 , G. Bauman5 1 L ONDON R EGIONAL C ANCER P ROGRAM, London, Canada 2 O NCOLOGY I NSTITUTE OF S OUTHERN S WITZERLAND, Bellinzona, Switzerland 3 T HE L ONDON R EGIONAL C ANCER C ENTRE, London, Ontario, Canada 4 O SPEDALE S AN G IOVANNI, Bellinzona, Switzerland 5 L ONDON R EGIONAL C ANCER C ENTRE, London, Canada
Purpose: The presence of radiosensitive neurogenic stem cells in the hippocampal area suggests identification and avoidance of the hippocampi may reduce the potential risks of radiation related cognitive and memory impairment. Materials: Six patients treated for base of skull tumors (4 pituitary adenoma, 2 meningioma) were re-planned with specific hippocampus sparing using coplanar helical tomotherapy (HT) as well as co-planar and non-coplanar volumetric arc techniques: Rapid Arc (RA). Two options (10 and 2 mm) of PTV margins were considered to evaluate the impact of on-board image guidance/stereotaxis enabling tighter margins. The PTV dose was set for all plans to 50 Gy. The hippocampal areas were identified and contoured as avoidance structures with the specific goal of minimizing the planned dose to the hippocampus while respecting other organ at risk dose limits. Results: Comparison of the hippocampus avoidance plans with HT and different number of arcs in RA approach demonstrated the importance of noncoplanar delivery when larger GTV to PTV margins were used (10mm). With smaller PTV margins (2mm) both co-planar and non-coplanar delivery provided similar degrees of hippocampal sparing although a benefit for brainstem and optic nerve sparing was still noted with non-coplanar delivery (although PTV and OAR constraints were met by all techniques). A similar benefit was noted with 2 versus 3 non-coplanar arc plans. Conclusions: Our comparisons suggest interventions to minimize GTV to PTV margins (accurate tumor volume delineation and image guidance) have a profound influence on the ability to spare intracranial OAR. Non coplanar techniques could be advisable when larger margins are used. 520 poster (Physics Track) INVESTIGATION OF THE DIFFERENCE IN PB AND AAA CALCULATED DOSE DISTRIBUTIONS FOR DYNAMIC IMRT A. Karlsson1 , C. Behrens1 , R. Ottosson1 , E. Samsoe1 , D. Sjöström1 1 C OPENHAGEN U NIVERSITY H OSPITAL H ERLEV, Department of Oncology, Division of Radiotherapy, Herlev, Denmark
Purpose: The anisotropic analytical algorithm (AAA) is the sole algorithm for dose calculation of RapidArcTM (Varian Medical Systems) treatment plans. For the sliding window IMRT (Varian Medical Systems), with static arcs, the dose is most commonly calculated using the pencil beam convolution algorithm (PB). Thus, introducing RapidArcTM , results in both new technology and a more sophisticated dose calculation algorithm. When comparing RapidArcTM and IMRT dose distributions it is important to be able to distinguish differences arising from the calculation algorithm and the new technology. The aim of this study was to investigate the effect on the dose distribution for clinical IMRT cases when using the AAA. Materials: Clinical pelvic and H&N IMRT plans, originally calculated using the PB algorithm were recalculated employing the AAA. The recalculation was undertaken importing the original MLC-files and using the same monitor unit settings as in the original IMRT treatment plans. The procedure was carried out using the EcipseTM treatment planning system (build 8.2.24.10720, Varian Medical Systems). Dose volume histogram data and slice based isodose line examination were used in the evaluation. Results: Generally, the dose distributions calculated using the AAA lead to lower mean dose and less dose coverage of the PTVs compared to dose distributions calculated using the PB. For the H&N cases the tumour PTV volume receiving more than 95% of the prescribed dose (PTV(95%)) were on average decreased with 2.7% units (±2.1), the mean dose were on average decreased with 0,9% units (±0,4). The corresponding results for the PTV(95%) for the associating node PTVs were approximately a magnitude lower. The higher values for the tumour PTV, as well as the high standard deviation were due to artefacts from the dental implants sited close to the tumour PTV. For most H&N patients the tumour volume is sited close to the oral cavity. However, the associated nodes are typically located more caudally. For the spinal cord and the brainstem the calculated dose maximum were approximately 0,5 Gy lower for the AAA calculation compared with the original PB calculation. For the pelvic cases the difference in the calculated dose distributions were less pronounced. This was probably due to less inhomogeneities in and around the PTVs. Normally, none of the beams in the pelvic IMRT treatment plans pass entries through the femur heads. Conclusions: IMRT dose distributions calculated using the PB and the AAA were compared and investigated. In general, the differences between the PB and the AAA calculated dose distributions were relatively small. This was the
S 193
case in both the pelvic and the H&N region when the densities were close to water and/or the volumes with other densities were minor. However, the difference between the algorithms becomes evident when the volumes and the difference in density become larger e.g. dental implants. Verification measurements for the AAA will be the future prospective of this study. 521 poster (Physics Track) IS ONE SINGLE ARC SUFFICIENT IN VOLUMETRIC MODULATED ARC THERAPY FOR COMPLEX SHAPED TARGET VOLUMES? M. Guckenberger1 , A. Richter1 , T. Krieger1 , J. Wilbert1 , K. Baier1 , M. Flentje1 1 U NIVERSITAETSKLINIKUM W UERZBURG, Radiation Oncology Department, Wuerzburg, Germany
Purpose: To compare step-and-shoot intensity-modulated radiotherapy (sIMRT) with volumetric modulated arc therapy (VMAT) for complex shaped target volumes with a simultaneous integrated boost (SIB). Materials: This retrospective planning study was based on fifteen patients treated for prostate cancer (n=5), hypo-pharyngeal cancer (n=5) or cancer of the paranasal sinuses (n=5); a SIB boost technique was used in all patients. Doses of 60Gy and 74Gy in 33 fractions were prescribed to the initial planning target volume (PTV_1) and the boost target volume (PTV_2) for prostate patients. In postoperative treatment of head-and-neck patients after R1 resection, doses of 60Gy and 66Gy in 30 fractions were prescribed to PTV_1 and PTV_2, respectively. Treatment plans were generated in the Pinnacle planning system (research version v8.9) for an Elekta Synergy linear accelerator with 4mm (prostate) or 10mm leaf width (head-and-neck). S-IMRT plans were generated with 7 to 9 fields and 30 (prostate) or 50 (head-and-neck) segments using direct machine parameter optimization. VMAT plans were generated with one single arc (control points every 4◦ and every 2◦ , 120s per rotation), two arcs (control points every 2◦ , 120s per rotation) and three arcs (control points every 4◦ , 90s per rotation). S-IMRT and VMAT plans were planned using identical optimization objectives. Results: For prostate patients, single-arc VMAT plans were equivalent and two-arc plans were superior to s-IMRT regarding PTV dose homogeneity, coverage and conformity with similar doses in the organs-at-risks rectum and bladder. Monitor units were reduced from average 528 to 478 for s-IMRT and single-arc VMAT, respectively. For hypo-pharyngeal cancer, single-arc VMAT plans were inferior to s-IMRT in all patients: doses to organs-at-risks spinal cord and parotid glands were similar but dose coverage and dose homogeneity in both PTV_1 and PTV_2 were worse in single-arc VMAT plans. Two-arc VMAT plans were similar and three-arc plans were superior to s-IMRT; monitor units were identical for two-arcs while three-arcs increased monitor units by 15% on average compared to s-IMRT. For cancer of the paranasal sinuses, s-IMRT was superior to all VMAT plans in all patients. Especially target coverage in the region of the ethmoid sinus between the orbits was inferior in the VMAT plans; doses to lenses were increased in all VMAT plans with highest doses in multiple arc plans while doses to spinal cord and parotid glands were similar. The table shows composite objective values of the penalty function in the inverse treatment planning; smaller values indicate improved conformity for planning objectives. For cancer of paranasal sinuses, values after exclusion of lens objectives are shown in brackets.
Conclusions: Based on these results, single arc VMAT was not equivalent to s-IMRT for highly complex target volumes but will be a viable option for semi-complex target volumes allowing reduced treatment time and reduced number of monitor units. Multiple arcs improved dose distributions in all patients. 522 poster (Physics Track) QA STRATEGY FOR ELEKTA VMAT AT JUNTENDO UNIVERSITY HOSPITAL S. Ozawa1 , M. Kawashima1 , T. Furuya1 , Y. Tsutsumi1 , A. Isobe1 , T. Fujita1 , C. Kurokawa1 , S. Sugimoto1 , C. Toramatsu1 , K. Itoh1 , K. Karasawa1 1 J UNTENDO U NIV. S CHOOL OF M EDICINE, Department of Medical Physics for Advanced Radiotherapy, Tokyo, Japan 2 U NIVERSITY OF F LORIDA, Department of Radiation Oncology, Gainesville, FL, USA 3 U NIVERSITY OF F LORIDA P ROTON T HERAPY I NSTITUTE, Department of Radiation Oncology, Jacksonville, Florida, USA
519 poster (Physics Track) HIPPOCAMPAL AVOIDANCE WITH RAPID ARC AND HELICAL TOMOTHERAPY FOR BASE OF SKULL TUMORS E. Wiebe1 , L. Cozzi2 , S. Yartsev3 , A. Fogliata Cozzi2 , E. Vanetti de Palma2 , A. Clivio4 , G. Nicolini4 , J. Chen5 , A. Leung5 , G. Bauman5 1 L ONDON R EGIONAL C ANCER P ROGRAM, London, Canada 2 O NCOLOGY I NSTITUTE OF S OUTHERN S WITZERLAND, Bellinzona, Switzerland 3 T HE L ONDON R EGIONAL C ANCER C ENTRE, London, Ontario, Canada 4 O SPEDALE S AN G IOVANNI, Bellinzona, Switzerland 5 L ONDON R EGIONAL C ANCER C ENTRE, London, Canada
Purpose: The presence of radiosensitive neurogenic stem cells in the hippocampal area suggests identification and avoidance of the hippocampi may reduce the potential risks of radiation related cognitive and memory impairment. Materials: Six patients treated for base of skull tumors (4 pituitary adenoma, 2 meningioma) were re-planned with specific hippocampus sparing using coplanar helical tomotherapy (HT) as well as co-planar and non-coplanar volumetric arc techniques: Rapid Arc (RA). Two options (10 and 2 mm) of PTV margins were considered to evaluate the impact of on-board image guidance/stereotaxis enabling tighter margins. The PTV dose was set for all plans to 50 Gy. The hippocampal areas were identified and contoured as avoidance structures with the specific goal of minimizing the planned dose to the hippocampus while respecting other organ at risk dose limits. Results: Comparison of the hippocampus avoidance plans with HT and different number of arcs in RA approach demonstrated the importance of noncoplanar delivery when larger GTV to PTV margins were used (10mm). With smaller PTV margins (2mm) both co-planar and non-coplanar delivery provided similar degrees of hippocampal sparing although a benefit for brainstem and optic nerve sparing was still noted with non-coplanar delivery (although PTV and OAR constraints were met by all techniques). A similar benefit was noted with 2 versus 3 non-coplanar arc plans. Conclusions: Our comparisons suggest interventions to minimize GTV to PTV margins (accurate tumor volume delineation and image guidance) have a profound influence on the ability to spare intracranial OAR. Non coplanar techniques could be advisable when larger margins are used. 520 poster (Physics Track) INVESTIGATION OF THE DIFFERENCE IN PB AND AAA CALCULATED DOSE DISTRIBUTIONS FOR DYNAMIC IMRT A. Karlsson1 , C. Behrens1 , R. Ottosson1 , E. Samsoe1 , D. Sjöström1 1 C OPENHAGEN U NIVERSITY H OSPITAL H ERLEV, Department of Oncology, Division of Radiotherapy, Herlev, Denmark
Purpose: The anisotropic analytical algorithm (AAA) is the sole algorithm for dose calculation of RapidArcTM (Varian Medical Systems) treatment plans. For the sliding window IMRT (Varian Medical Systems), with static arcs, the dose is most commonly calculated using the pencil beam convolution algorithm (PB). Thus, introducing RapidArcTM , results in both new technology and a more sophisticated dose calculation algorithm. When comparing RapidArcTM and IMRT dose distributions it is important to be able to distinguish differences arising from the calculation algorithm and the new technology. The aim of this study was to investigate the effect on the dose distribution for clinical IMRT cases when using the AAA. Materials: Clinical pelvic and H&N IMRT plans, originally calculated using the PB algorithm were recalculated employing the AAA. The recalculation was undertaken importing the original MLC-files and using the same monitor unit settings as in the original IMRT treatment plans. The procedure was carried out using the EcipseTM treatment planning system (build 8.2.24.10720, Varian Medical Systems). Dose volume histogram data and slice based isodose line examination were used in the evaluation. Results: Generally, the dose distributions calculated using the AAA lead to lower mean dose and less dose coverage of the PTVs compared to dose distributions calculated using the PB. For the H&N cases the tumour PTV volume receiving more than 95% of the prescribed dose (PTV(95%)) were on average decreased with 2.7% units (±2.1), the mean dose were on average decreased with 0,9% units (±0,4). The corresponding results for the PTV(95%) for the associating node PTVs were approximately a magnitude lower. The higher values for the tumour PTV, as well as the high standard deviation were due to artefacts from the dental implants sited close to the tumour PTV. For most H&N patients the tumour volume is sited close to the oral cavity. However, the associated nodes are typically located more caudally. For the spinal cord and the brainstem the calculated dose maximum were approximately 0,5 Gy lower for the AAA calculation compared with the original PB calculation. For the pelvic cases the difference in the calculated dose distributions were less pronounced. This was probably due to less inhomogeneities in and around the PTVs. Normally, none of the beams in the pelvic IMRT treatment plans pass entries through the femur heads. Conclusions: IMRT dose distributions calculated using the PB and the AAA were compared and investigated. In general, the differences between the PB and the AAA calculated dose distributions were relatively small. This was the
S 193
case in both the pelvic and the H&N region when the densities were close to water and/or the volumes with other densities were minor. However, the difference between the algorithms becomes evident when the volumes and the difference in density become larger e.g. dental implants. Verification measurements for the AAA will be the future prospective of this study. 521 poster (Physics Track) IS ONE SINGLE ARC SUFFICIENT IN VOLUMETRIC MODULATED ARC THERAPY FOR COMPLEX SHAPED TARGET VOLUMES? M. Guckenberger1 , A. Richter1 , T. Krieger1 , J. Wilbert1 , K. Baier1 , M. Flentje1 1 U NIVERSITAETSKLINIKUM W UERZBURG, Radiation Oncology Department, Wuerzburg, Germany
Purpose: To compare step-and-shoot intensity-modulated radiotherapy (sIMRT) with volumetric modulated arc therapy (VMAT) for complex shaped target volumes with a simultaneous integrated boost (SIB). Materials: This retrospective planning study was based on fifteen patients treated for prostate cancer (n=5), hypo-pharyngeal cancer (n=5) or cancer of the paranasal sinuses (n=5); a SIB boost technique was used in all patients. Doses of 60Gy and 74Gy in 33 fractions were prescribed to the initial planning target volume (PTV_1) and the boost target volume (PTV_2) for prostate patients. In postoperative treatment of head-and-neck patients after R1 resection, doses of 60Gy and 66Gy in 30 fractions were prescribed to PTV_1 and PTV_2, respectively. Treatment plans were generated in the Pinnacle planning system (research version v8.9) for an Elekta Synergy linear accelerator with 4mm (prostate) or 10mm leaf width (head-and-neck). S-IMRT plans were generated with 7 to 9 fields and 30 (prostate) or 50 (head-and-neck) segments using direct machine parameter optimization. VMAT plans were generated with one single arc (control points every 4◦ and every 2◦ , 120s per rotation), two arcs (control points every 2◦ , 120s per rotation) and three arcs (control points every 4◦ , 90s per rotation). S-IMRT and VMAT plans were planned using identical optimization objectives. Results: For prostate patients, single-arc VMAT plans were equivalent and two-arc plans were superior to s-IMRT regarding PTV dose homogeneity, coverage and conformity with similar doses in the organs-at-risks rectum and bladder. Monitor units were reduced from average 528 to 478 for s-IMRT and single-arc VMAT, respectively. For hypo-pharyngeal cancer, single-arc VMAT plans were inferior to s-IMRT in all patients: doses to organs-at-risks spinal cord and parotid glands were similar but dose coverage and dose homogeneity in both PTV_1 and PTV_2 were worse in single-arc VMAT plans. Two-arc VMAT plans were similar and three-arc plans were superior to s-IMRT; monitor units were identical for two-arcs while three-arcs increased monitor units by 15% on average compared to s-IMRT. For cancer of the paranasal sinuses, s-IMRT was superior to all VMAT plans in all patients. Especially target coverage in the region of the ethmoid sinus between the orbits was inferior in the VMAT plans; doses to lenses were increased in all VMAT plans with highest doses in multiple arc plans while doses to spinal cord and parotid glands were similar. The table shows composite objective values of the penalty function in the inverse treatment planning; smaller values indicate improved conformity for planning objectives. For cancer of paranasal sinuses, values after exclusion of lens objectives are shown in brackets.
Conclusions: Based on these results, single arc VMAT was not equivalent to s-IMRT for highly complex target volumes but will be a viable option for semi-complex target volumes allowing reduced treatment time and reduced number of monitor units. Multiple arcs improved dose distributions in all patients. 522 poster (Physics Track) QA STRATEGY FOR ELEKTA VMAT AT JUNTENDO UNIVERSITY HOSPITAL S. Ozawa1 , M. Kawashima1 , T. Furuya1 , Y. Tsutsumi1 , A. Isobe1 , T. Fujita1 , C. Kurokawa1 , S. Sugimoto1 , C. Toramatsu1 , K. Itoh1 , K. Karasawa1 1 J UNTENDO U NIV. S CHOOL OF M EDICINE, Department of Medical Physics for Advanced Radiotherapy, Tokyo, Japan 2 U NIVERSITY OF F LORIDA, Department of Radiation Oncology, Gainesville, FL, USA 3 U NIVERSITY OF F LORIDA P ROTON T HERAPY I NSTITUTE, Department of Radiation Oncology, Jacksonville, Florida, USA