Premium
Performance evaluation of a collapsed cone dose calculation algorithm for HDR Ir‐192 of APBI treatments
Author(s) -
Terribilini Dario,
Vitzthum Veronika,
Volken Werner,
Frei Daniel,
Loessl Kristina,
Veelen Bob,
Manser Peter,
Fix Michael K.
Publication year - 2017
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1002/mp.12490
Subject(s) - brachytherapy , imaging phantom , monte carlo method , dosimetry , nuclear medicine , formalism (music) , physics , biomedical engineering , mathematics , radiation therapy , medicine , radiology , statistics , art , musical , visual arts
Purpose Most dose calculations for HDR brachytherapy treatments are based on the AAPM ‐ TG 43 formalism. Because patient's anatomy, heterogeneities, and applicator shielding are not considered, the dose calculation based on this formalism is inaccurate in some cases. Alternatively, collapsed cone ( CC ) methods as well as Monte Carlo ( MC ) algorithms belong to the model‐based dose calculation algorithms, which are expected to improve the accuracy of calculated dose distributions. In this work, the performance of a CC algorithm, ACE in Oncentra Brachy 4.5 ( ACE 4.5), has been investigated by comparing the calculated dose distributions to the AAPM ‐ TG 43 and MC calculations for 10 HDR brachytherapy accelerated partial breast irradiation treatments ( APBI ). Comparisons were also performed with a corrected version of ACE 4.5 ( ACE 4.5/corr). Methods The brachytherapy source microSelectron mHDR ‐v2 (Elekta Brachytherapy) has been implemented in a MC environment and validated by comparing MC dose distributions simulated in a water phantom of 80 cm in diameter with dose distributions calculated with the AAPM ‐ TG 43 algorithm. Dose distributions calculated with ACE 4.5, ACE 4.5/corr, AAPM ‐ TG 43 formalism, and MC for 10 APBI patients plans have then been computed and compared using HU scaled densities. In addition, individual dose components have been computed using ACE 4.5, ACE 4.5/corr, and MC , and compared individually. Results Local differences between MC and AAPM ‐ TG 43 calculated dose distributions in a large water phantom are < 1%. When using HU s scaled densities for the breast cancer patients, both accuracy levels of ACE 4.5 overestimate the MC calculated dose distributions for all analyzed dosimetric parameters. In the planning target volume ( PTV ), ACE 4.5 ( ACE 4.5/corr) overestimates on average V 100%, PTV by 3% ± 1% (1% ± 1%) and D 50, PTV by 3% ± 1% (1% ± 1%) and in the organs at risk D 1cc, skin by 4% ± 2% (1% ± 1%), D 0.5cc, ribs by 4% ± 2% (0% ± 1%), and D 1cc, heart by 8% ± 2% (3% ± 1%) compared to MC . Comparisons of the individual dose components reveals an agreement for the primary component of < 2% local differences for both ACE 4.5 and ACE 4.5/corr. Local differences of about 40% (20%) for the first and residual scatter components where observed when using ACE 4.5 ( ACE 4.5/corr). Using uniform densities for one case shows a better agreement between ACE 4.5 and MC for all dosimetric parameters considered in this work. Conclusions In general, on the 10 APBI patients the ACE 4.5/corr algorithm results in similar dose distributions as the commonly used AAPM ‐ TG 43 within the PTV . However, the accuracy of the ACE 4.5/corr calculated dose distribution is closer to MC than to AAPM ‐ TG 43. The differences between commercial version ACE 4.5 and MC dose distributions are mainly located in the first and residual scatter components. In ACE 4.5/corr, the changes done in the algorithm for the scatter components substantially reduce these differences.