A comparison of A comparison of heterogeneity correction heterogeneity correction algorithms within a lung PTV algorithms within a lung PTV Paola Alvarez, Andrea Molineu Molineu, Nadia , Nadia Hernandez Hernandez, , Paola Alvarez, Andrea David Followill Followill, Geoffrey , Geoffrey Ibbott Ibbott David Radiological Physics Center Radiological Physics Center July 2007 July 2007
RPC Lung Phantom RPC Lung Phantom - Plastic shell water Plastic shell water - fillable fillable - Designed based on Designed based on - patient anatomy patient anatomy - Imaging and Imaging and - dosimetric insert dosimetric insert
RPC Phantom RPC Phantom Target dimension Ovoid shape 3 cm diameter Tumor (TLD) 5 cm long Densities Lung = 0.33g/cm 3 Heart= 1.1 g/cm 3 Cord = 1.31 g/cm3 Film Slits Tumor = 1.04 g/cm 3 Dosimeters TLD and Gafchromic film
Phantom Process Phantom Process � Phantom is imaged Phantom is imaged � � Treatment plan developed by institution Treatment plan developed by institution � � Treatment is delivered to the phantom Treatment is delivered to the phantom � � Phantom is returned to the RPC for data analysis Phantom is returned to the RPC for data analysis � � Treatment plan is submitted electronically to the ITC Treatment plan is submitted electronically to the ITC � � The phantom is to be treated as if it were a patient The phantom is to be treated as if it were a patient �
Prescription Prescription � Energies: 4 � Energies: 4 – – 10 MV 10 MV � SBRT technique: SBRT technique: ≥ ≥ 7 non 7 non- -opposing static fields opposing static fields � ≥ 340 340° ° arc rotation technique arc rotation technique ≥ � Prescribed dose must cover 95% of the PTV Prescribed dose must cover 95% of the PTV � � Prescription isodose line between 60% to 90%. � Prescription isodose line between 60% to 90%. � Ignore lung heterogeneity for calculation of M.U. Ignore lung heterogeneity for calculation of M.U. � � Submit hetero. plan based on homo. M.U. set Submit hetero. plan based on homo. M.U. set �
Phantom Results Phantom Results � A total of 33 irradiations were processed � A total of 33 irradiations were processed � The 6 MV photon beam was used most often � The 6 MV photon beam was used most often � The TPSs TPSs used to plan the cases were: used to plan the cases were: � The Pinnacle, BrainLab, XiO, Precise, Eclipse Pinnacle, BrainLab, XiO, Precise, Eclipse Ergo and Hi- -ART. ART. Ergo and Hi � Superposition/Convolution algorithm was � Superposition/Convolution algorithm was used most often. used most often.
Phantom Results Phantom Results Center of Tumor Center of Tumor Measured Measured TPS Dose Calc. Algorithm # irradiation D TLD /D D hetero TPS Dose Calc. Algorithm # irradiation D TLD / hetero ± 3.1% 0.99 ± 3.1% Precise Precise Scatter Int. Clarkson Scatter Int. Clarkson 2 2 0.99 ± 2.4% 0.96 ± 2.4% BrainLab BrainLab Clarkson & Pencil beam Clarkson & Pencil beam 5 5 0.96 ± 1.8% 0.96 ± Eclipse Eclipse Pencil Beam Pencil Beam 5 5 0.96 1.8% ± 3.2% 0.98 ± Ergo 3D Convol Convol. Pencil Beam . Pencil Beam 2 0.98 3.2% Ergo 3D 2 Hi- -ART ART Superposition/Convolution 1 0.97 Hi Superposition/Convolution 1 0.97 ± 2.1% 0.99 ± Pinnacle Adaptive convolve 10 2.1% Pinnacle Adaptive convolve 10 0.99 ± 3.8% 0.98 ± Eclipse AAA 2 3.8% Eclipse AAA 2 0.98 ± 1.8% 0.96 ± XiO Superposition/Convolution 6 1.8% XiO Superposition/Convolution 6 0.96 0.97 ± 2.8%
Profile analysis Profile analysis Right Left Profile Convolution Superposition example 30 Left Right Prescribed D 20 D o s e (G y ) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution hetero Institution homo
Profile analysis Profile analysis Right Left Profile Pencil Beam example 30 Left Right Prescribed D 20 Dose (Gy) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution hetero Institution homo
Phantom analysis analysis Phantom Criteria on on heterogeneous heterogeneous case case Criteria DTLD/DInst DInst : 0.97 +/ : 0.97 +/- - 5% 5% DTLD/ DTA ≤ ≤ 5mm 5mm at at all all side side of of PTV PTV DTA 23 irradiations irradiations passed passed the the test test. . 23 An analysis analysis of of the the dose dose distribution distribution was was done done over over An the central 80% central 80% of of the the PTV PTV for for these these 23 23 the irradiations. . irradiations
Superposition/Convolution R- -L Profile L Profile Superposition/Convolution R Right Left Profile Axial plane 30 Left Righ Average Average displacement displacement Left side: Rigth side: on: 3 mm Prescribed D on: 1 mm off: 1 mm 20 off: 5 mm Dose (Gy) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution hetero Institution homo RPC Regression Institution Regression
Superposition/Convolution R- -L Profile L Profile Superposition/Convolution R Right Left Profile Axial plane 30 Left Righ 25 Average Average displacement displacement Left side: 23 Rigth side: on: 3 mm Prescribed D on: 1 mm off: 1 mm 20 off: 5 mm 21 -1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6 Dose (Gy) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution hetero Institution homo RPC Regression Institution Regression
Superposition/Convolution D RPC /D D Inst Superposition/Convolution D RPC / Inst D RPC / D Inst over 80% of PTV on Rt Lt profile 1.10 1.05 5% 1.00 D RPC /D Inst 0.95 0.90 0.85 0.80 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 distance (cm)
Pencil- -Beam profile Beam profile Pencil Right Left Profile Axial plane 30 Left Right Average Average displacement displacement Left side: Right side: on: -2 mm on: -4 mm off: -4 mm Prescribed D off: 0 mm 20 Dose (Gy) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution homo Institution hetero RPC Regression Institution Regression
Pencil- -Beam profile Beam profile Pencil Right Left Profile Axial plane 30 Left Right 28 26 Average Average displacement 24 displacement Left side: Right side: 22 on: -2 mm on: -4 mm off: -4 mm Prescribed D off: 0 mm 20 20 -1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6 Dose (Gy) D 2cm 10 PTV 0 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Distance (cm) RPC Film Institution homo Institution hetero RPC Regression Institution Regression
Pencil Beam D RPC /D D Inst Pencil Beam D RPC / Inst D RPC /D Inst over 80% of PTV on Rt Lt profile 1.10 10% 1.05 7% 5% 1.00 D RPC /D Inst 0.95 0.90 0.85 0.80 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 distance (cm)
Summary of Systems Passing Summary of Systems Passing Existing Criteria Existing Criteria Percent of Points Within: Percent of Points Within: System/Algorithm 5% 7% 10% System/Algorithm 5% 7% 10% Pencil Beam- - Pencil Beam Clarkson 69 ± ±27% 27% 83 ± ±14% 14% 92 ± ±8% 8% Clarkson 69 83 92 (n=9) (n=9) S uperposition C onvolution / S uperposition C onvolution / AAA AAA 87 ± ±20% 20% 95 ± ±13% 13% 99 ± ±5% 5% 87 95 99 (n=14) (n=14)
Conclusions Conclusions � The average target TLD/Inst ratio is 0.97 (range 0.96 to 0.99). � The average target TLD/Inst ratio is 0.97 (range 0.96 to 0.99). � The calculation from Superposition Convolution and AAA � The calculation from Superposition Convolution and AAA algorithms agree well with the measurements. algorithms agree well with the measurements. � New evaluation methods needed to assess each algorithm’ ’s s � New evaluation methods needed to assess each algorithm accuracy. accuracy.
Thank you Thank you
Recommend
More recommend
