Impact of Convolution Algorithm Based Treatment Planning in Gamma Knife Radiosurgery
gopishankar natanasabapathi1*, Raj Kishor Bisht2, All India Institute of Medical Sciences, New Delhi, India.SU-E-T-542 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
To evaluate the influence of Convolution algorithm based calculation in Leksell Gamma Knife treatment planning.
Leksell Gamma Plan (LGP version 10.1) was upgraded with convolution algorithm option for planning purpose. An electron density phantom (CIRS, VA, USA) with different insert materials was CT (Siemens Emotion6) scanned for calibration purpose. Hounsfield units corresponding to different insert materials of the phantom were measured by ImageJ software. A graphical plot between Hounsfield units and Electron density of different materials was fed in LGP TPS for activation of Convolution Algorithm based planning (CABP). Following scanning parameters were used: Tube voltage-110kV, Slice thickness-1mm and FOV-240mm. Images were transferred to the TPS for planning. The CT images were defined and converted to electron density map for CABP. Study1: A dummy patient CT image was acquired and a tumor was simulated adjacent to an air cavity. Study2: A patient image with multiple tumors on either side of brainstem was planned. A dose of 25Gy and 12Gy was prescribed at 50% isodose level respectively for both studies. Both these plans were compared with CABP and TMR10 algorithm.
Study1 showed that lower level isodose lines (8Gy) shifted 10 mm near tumor and air interface and 11 mm near tumor and normal tissue junction contributing additional dose to normal tissue. In Study2 integral dose to brainstem was five times more with CABP than TMR10 based planning. The treatment time calculated by CABP was two to three times more than TMR10 algorithm based calculations in both studies.
The inclusion of heterogeneity corrections through CABP gives a new perspective to the Gamma Plan which may be beneficial for patients as it gives a more realistic picture to treatment. At present treatment time generated by CABP appears to be one of the limiting factors in implementing this algorithm.