TMR 10: A comparison with TMR classic for clinical treatment plans





Keywords: dosimetry, physics, gamma knife, radiosurgery, dose planning

Interactive Manuscript

Ask Questions of this Manuscript:

   



Abstract

     Leksell GammaPlan 10.0 introduced an improved dose calculation algorithm, TMR 10, which entails new beam profiles, depth doses and output factors. This enhanced beam modelling has made inaccuracies of the previous algorithm, TMR Classic, apparent. Furthermore, multi-isocentric planning techniques can amplify subtle differences in single shot profiles, which in turn can affect the dose calculated, normalisation and beam on time.
     We aimed to quantify dosimetrical differences between the two algorithms. In a non-clinical investigation, the greatest dose discrepancy possible between the two algorithms was also explored.
     
     30 targets planned with TMR Classic for the Leksell Gamma Knife® Perfexion™ (PFX) were retrospectively re-calculated using the TMR 10 algorithm. This method was repeated for 37 targets planned for the Model B/C. The maximum dose, the mean dose to the target, and the maximum dose to the organs at risk (OARs) were compared.
     Assuming that TMR 10 correctly calculates dose to the skull, TMR Classic generally underestimated maximum Perfexion doses by 2.0% (range: -3.3% to 6.7%), mean doses 1.5% (range: -2.0% to 5.3%) and OARs 3.6% (range: -3.9 to 7.1%). Model B/C data differed by less. TMR Classic underestimated maximum doses by an average of 1.2% (range: -0.7% to 4.2%), mean doses 0.9% (range: -1.3% to 2.7%) and overestimated OARs by 1.8% (range: -0.9% to 4.5%). In a non-clinical setting, the largest discrepancy was for a plan with 8mm and 16mm shots with identical x and y coordinates, but with the 8mm shot placed 13.0mm in the superior direction. TMR Classic recorded a maximum dose 18.1% lower than that of TMR 10.
     This was a single center validation of this method.
     Dosimetrical differences between the two algorithms can be significant in certain circumstances. Fortunately, dose differences between TMR Classic for Model B/C (where most of our empirical data lies) and TMR 10 is minimal. Doses to OARs were typically underestimated by TMR Classic for PFX, but overestimated for the Model B/C. As a result, some users may be unintentionally giving doses to OARs that are several percent higher with PFX.
     When quoting doses to critical structures with clinical results, users should recalculate the doses delivered using TMR 10.


Acknowledgements

Project Roles: