Comparison of dose calculation algorithms for treatment planning in external photon beam therapy for clinical situations

Abstract

A study of the performance of five commercial radiotherapy treatment planning systems (TPSs) for common treatment sites regarding their ability to model heterogeneities and scattered photons has been performed. The comparison was based on CT information for prostate, head and neck, breast and lung cancer cases. The TPSs were installed locally at different institutions and commissioned for clinical use based on local procedures. For the evaluation, beam qualities as identical as possible were used: low energy (6 MV) and high energy (15 or 18 MV) x-rays. All relevant anatomical structures were outlined and simple treatment plans were set up. Images, structures and plans were exported, anonymized and distributed to the participating institutions using the DICOM protocol. The plans were then re-calculated locally and exported back for evaluation. The TPSs cover dose calculation techniques from correction-based equivalent path length algorithms to model-based algorithms. These were divided into two groups based on how changes in electron transport are accounted for ((a) not considered and (b) considered). Increasing the complexity from the relatively homogeneous pelvic region to the very inhomogeneous lung region resulted in less accurate dose distributions. Improvements in the calculated dose have been shown when models consider volume scatter and changes in electron transport, especially when the extension of the irradiated volume was limited and when low densities were present in or adjacent to the fields. A Monte Carlo calculated algorithm input data set and a benchmark set for a virtual linear accelerator have been produced which have facilitated the analysis and interpretation of the results. The more sophisticated models in the type b group exhibit changes in both absorbed dose and its distribution which are congruent with the simulations performed by Monte Carlo-based virtual accelerator.

The photon dose calculation algorithm used in breast radiotherapy has significant impact on the parameters of radiobiological models

“…The photon dose calculation algorithms in commercial treatment planning systems are classified in two groups, depending on the method used for inhomogeneity approximation: type‐a algorithms and type‐b algorithms (1) . The inhomogeneity corrections used in the former are based on equivalent path length scaling — for example, the pencil beam convolution algorithm (2) with modified Batho heterogeneity correction (PBC‐MB) or with equivalent tissue‐air‐ratio correction.…”

“…The impact of the algorithm on the target coverage in lung cancer treatment has been described 1 , 7 , 8 , 10 . Two recent studies have quantified the difference between AAA and Acuros XB in stereotactic body radiation therapy of non‐small cell lung cancer 11 , 12 .…”

“…Two recent studies have quantified the difference between AAA and Acuros XB in stereotactic body radiation therapy of non‐small cell lung cancer 11 , 12 . There is some knowledge on the dose distributions in target volumes close to air cavities like, for example, in the head and neck region 1 , 8 , 13 . Recent planning studies discuss the impact of low‐density organs on the coverage of adjacent target volumes, as in breast cancer treatment 7 , 14 , 15 , 16 …”

A clinical database to assess action levels and tolerances for the ongoing use of Mobius3D

“…Eclipse utilizes a superposition/convolution type algorithm, iPlan a pencil beam convolution, and Mobius3D a collapsed cone. Generally speaking, model‐based calculation algorithms such as the AAA and CCCS have been shown to have a similar level of dose calculation accuracy, whereas PBC has severe limitations, especially when heterogeneous media (as is often found within a real patient) are involved 17, 18, 19…”

Dosimetry of oblique tangential photon beams calculated by superposition/convolution algorithms: a Monte Carlo evaluation

“…( 1 – 4 ) Semi‐analytical algorithms for dose calculation such as the superposition/convolution method ( 5 – 7 ) are currently used by most of commercial treatment planning systems (TPSs), and have proved to be efficient and accurate in the routine treatment planning practice. ( 8 – 12 ) …”