Quantitative validation of a new semi-automatic inverse planning software in Leksell GammaPlan® version 10.1 by comparison with clinically applied manual dose plansTobias Nelder1, Thomas Kraepelien2, Bodo Lippitz31Stockholm, Sweden 2Department of Medical Physics, Karolinska University Hospital 3Gamma Knife Center, Department of Neurosurgery, Karolinska University Hospital Keywords: gamma knife, Imaging, dose planning, Dose distribution, quality assuranceInteractive Manuscript
Ask Questions of this Manuscript:
What is the background behind your study?
A new automatized dose planning tool Inverse Planning IP has been implemented in a recent version of the treatment planning system Leksell GammaPlan® v.10 for Gamma Knife radiosurgery.
What is the purpose of your study?
The current study aims to quantify performance of the new dose planning tool compared to the previously used manually optimized dose planning system.
Describe your patient group.
Clinically applied manual dose plans of 50 consecutive patients (63 targets) treated for intracranial metastasis at the Karolinska University Hospital Gamma Knife Center between 2004-07-27 to 2005-01-18 were recreated using the new automatized IP system in Leksell GammaPlan® version 10.1 and were quantitatively compared with the manual treatment plans according to a standardized quantitative criteria.
Describe what you did.
Target coverage, selectivity, gradient index, conformity index and Paddick''s conformity index, together with the KARE-factor was calculated.
Describe your main findings.
Each new recreated plan with IP was pairwise compared to their manually planned MP counterpart with regard to these parameters. It was shown that a significant improvement in all of the parameters except for target coverage could be achieved. A mean percentage difference introduced with IP was calculated, selectivity improved with 9.6% (IP: median 0.76: range 0.40-0.90 vs. MP: median 0.69: range 0.32-0.89 (P<0.05)), gradient index with 10.2% (IP: median 2.65: 2.54-3.51 vs. MP: median 2.99: 2.44-4.00 (P<0.05)), Paddick''s conformity index with 9.0% (IP: median 0.74: 0.40-0.87 vs. MP: median 0.68: 0.32-0.88 (P<0.05)), conformity index with 8.2% (IP: median 1.30: 1.01-2.51 vs. MP: median 1.44: 1.11-3.17 (P<0.05)) and the KARE-factor with 17.4% (IP: median 3.87: 2.59-7.60 vs. MP: median 4.73: 2.45-11.29 (P<0.05)) while the target coverage deteriorated with 0.5% (IP: median 0.99: 0.77-1.00 vs. MP: median 1.00: 0.86-1.00 (P>0.05)). The most important findings in this paper were the average decrease of the prescribed isodose volume together with the more rapid dose fall-off outside of the target surface introduced with the use of IP while still maintaining the same clinical goal as the manually optimized plans.
Describe the main limitation of this study.
This was a retrospective study.
Describe your main conclusion.
Due to the known correlations of high peripheral doses in adjacent healthy brain tissue and adverse radiation effects, this rapid dose fall-off could minimize untoward radiation effects in patients receiving Gamma Knife treatments dose planned with IP.
Describe the importance of your findings and how they can be used by others.
The value of inverse planning will require further validation.
A new automatized dose planning tool Inverse Planning IP has been implemented in a recent version of the treatment planning system Leksell GammaPlan® v.10 for Gamma Knife radiosurgery.
The current study aims to quantify performance of the new dose planning tool compared to the previously used manually optimized dose planning system.
Clinically applied manual dose plans of 50 consecutive patients (63 targets) treated for intracranial metastasis at the Karolinska University Hospital Gamma Knife Center between 2004-07-27 to 2005-01-18 were recreated using the new automatized IP system in Leksell GammaPlan® version 10.1 and were quantitatively compared with the manual treatment plans according to a standardized quantitative criteria.
Target coverage, selectivity, gradient index, conformity index and Paddick''s conformity index, together with the KARE-factor was calculated.
Each new recreated plan with IP was pairwise compared to their manually planned MP counterpart with regard to these parameters. It was shown that a significant improvement in all of the parameters except for target coverage could be achieved. A mean percentage difference introduced with IP was calculated, selectivity improved with 9.6% (IP: median 0.76: range 0.40-0.90 vs. MP: median 0.69: range 0.32-0.89 (P<0.05)), gradient index with 10.2% (IP: median 2.65: 2.54-3.51 vs. MP: median 2.99: 2.44-4.00 (P<0.05)), Paddick''s conformity index with 9.0% (IP: median 0.74: 0.40-0.87 vs. MP: median 0.68: 0.32-0.88 (P<0.05)), conformity index with 8.2% (IP: median 1.30: 1.01-2.51 vs. MP: median 1.44: 1.11-3.17 (P<0.05)) and the KARE-factor with 17.4% (IP: median 3.87: 2.59-7.60 vs. MP: median 4.73: 2.45-11.29 (P<0.05)) while the target coverage deteriorated with 0.5% (IP: median 0.99: 0.77-1.00 vs. MP: median 1.00: 0.86-1.00 (P>0.05)). The most important findings in this paper were the average decrease of the prescribed isodose volume together with the more rapid dose fall-off outside of the target surface introduced with the use of IP while still maintaining the same clinical goal as the manually optimized plans.
This was a retrospective study.
Due to the known correlations of high peripheral doses in adjacent healthy brain tissue and adverse radiation effects, this rapid dose fall-off could minimize untoward radiation effects in patients receiving Gamma Knife treatments dose planned with IP.
The value of inverse planning will require further validation.
Project Roles:
T. Nelder (), T. Kraepelien (), B. Lippitz ()
