Eclipse TP algorithm guide (non eMC)

Question 1

does Acuros use relative electron density curve or relative mass density curve

Answer 1

relative mass density

Question 2

how are HU values outside of the curve handled in the different algorithms?

Answer 2

AAA-assigns max value. Warning
Acuros- if small enough volume, converts to a material. Otherwise, prevents calc until high density is assigned to a structure
-MRDC and FTDC- aaigns max value. No warning
-eMC- assigns max value. Warning

Question 3

calc options for AAA and Acuros

Answer 3

● Calculation resolution
● Calculation resolution for SRS and HyperArc plans
● Heterogeneity correction
● Field normalization
● Angular resolution in conformal arc and VMAT calculations
● Dose reporting mode (Acuros XB only)
● Plan dose calculation (Acuros XB only)
● GPU calculation (Acuros XB only)
● Automatic high-density material (Acuros XB only)
● Maximum automatic high-density volume in cm3
(Acuros XB only)

Question 4

dose matrix for AAA vs Acruros

Answer 4

Acuros is cartesian grid
AAA is divergent

Question 5

grid resolution in axis perpendicular to image slices

Answer 5

AAA and Acuros XB adapt the grid resolution to
ensure that the dose is always calculated exactly on the image slices. If the slice thickness is
larger than the defined resolution, AAA and Acuros XB may calculate the dose on dose planes
between the image slices. If the slice spacing is smaller than the defined resolution, AAA and
Acuros XB may skip calculating the dose on some slices

Question 6

modeled photon sources

Answer 6

: primary photon source, second photon source, electron contamination source, and
photons scattered from the hard wedge (wedge scatter source)

Question 7

parameters required for modelling of primary source

Answer 7

-photon energy spectrum
-mean radial energy
-intensity profile (ie. profile)

Question 8

limitations of AAA in lung

Answer 8

-For 4 MV to 6 MV energies and field sizes larger than or equal to 5×5 cm2
, AAA tends to
underestimate the dose in lung and overestimate the dose in water-equivalent tissue after the
lung. F

-For 10 MV to 20 MV energy modes and field sizes smaller than or equal to 5×5 cm2
, AAA tends
to overestimate the dose in lung.

Question 9

accuracy of AAA and acuros for static MLC fields

Answer 9

-For 18 MV, AAA and Acuros XB tend to underestimate the dose at shallow depths for certain
static MLC shapes

-For 6 MV, AAA and Acuros XB may underestimate the dose at large depths (larger than or equal
to 20 cm) for certain static MLC shapes.

Question 10

do you measure output factors for small fields in eclipse?

Answer 10

Output factors from 3×3 cm2
up to the maximum field size deliverable with the machine.
Output factors for field sizes 1×1 cm2
and 2×2 cm2
can be included, if desired. However,
these will not affect the calculation results for small MLC collimated fields in treatment
units, where MLC is located below the jaws (for example Varian). This is because the
backscatter in these cases is determined from the size of the jaw opening. If small jawcollimated fields are used in the treatments, the inclusion of output factor measurements
for these field sizes may improve the accuracy.

Note also that depth dose curve and profile measurements for
field sizes smaller than 2×2 cm2
are ignored by the configuration program.

Question 11

objecrive function for photons

Answer 11

2 terms:
-total gamma error
-f penalties for noise, an increasing mean energy curve and an
increasing intensity profile (outside the field edge), and for unphysical second source
parameters

Question 12

sigma parameters

Answer 12

smoothing factors for the electron contamination

Question 13

scatter kernels in AAA

Answer 13

-fromEGS MC
- A polyenergetic scatter kernel is constructed as a weighted sum of the monoenergetic
scatter kernels. During the 3D dose calculation these kernels are scaled according to the
densities of the actual patient tissues determined from the CT images.

Question 14

how does AAA convolution work?

Answer 14

The 3D dose distribution is calculated from separate convolutions for the primary photon source,
second photon source, wedge scatter source and contaminating electron source. The
convolutions are performed for all finite-sized beamlets that comprise the clinical broad beam.
The final dose distribution is obtained by a simple superposition of the individual beamlet
contributions

Question 15

heterogeneity corrections in AAA?

Answer 15

-radiological depth is scaled
-also lateral density scaling

Question 16

Acuros calculation steps

Answer 16

1. photon beam source model
2. ray tracing of beams into the patient
3. scattered photon fluence calculation
4. collided electron fluence calculation
5. dose calculation

Question 17

what does acuros XB do?

Answer 17

Acuros XB solves the time-independent
three-dimensional system of coupled Boltzmann transport equations (LBTE)

Question 18

assumptions in acuros Botlzmann equation

Answer 18

. Both charged pair production secondary particles are assumed to be electrons instead
of one electron and one positron. Also, the partial coupling technique is assumed, whereby
photons can produce electrons, but electrons do not produce photons. Regarding the latter, the
energy from photons produced by the electrons is accounted for, but assumed to be deposited
locally.

Question 19

computational grid in acuros

Answer 19

spatially variable; the local
element size is adapted to achieve a higher spatial resolution inside the beam, with reduced
resolution in lower dose and lower gradient regions outside the beam penumbra.

Question 20

discretization in Acuros

Answer 20

-angular, energy, spatial

Question 21

output dose grid control in accuros

Answer 21

Regardless of the output grid size, Acuros XB will account for the effects of photon and electron
transport throughout the full CT extents on the dose within the output dose grid. However,
Acuros XB will reduce the spatial resolution external to the output dose grid to reduce the
calculation time

Question 22

acuros discretization errors

Answer 22

energy= solution bias
angular= ray effects
spatial= local solution over/under shooting

Question 23

MRDC algorithm

Answer 23

-multi-resolution dose calculation
-fast dose estimation inside PO
-convolution superposition
-Finer resolution is used close to the location of the primary interaction, while much lower
resolution is used to compute the scatter component for larger distances

Question 24

FTDC algorithm

Answer 24

-fourier transform dose calculation
-convolution/superposition dose
calculation algorithm optimized for speed.
-uses fast fourier transform to calc convolution

Question 25

objectives in optimizer

Answer 25

dose-volume objectives
smoothing objectives
MU objective
normal tissue objective

Question 26

aperture shape controller

Answer 26

ASC favors apertures of minimal local curvature

Question 27

what method does PO use?

Answer 27

gradient search method. The gradient search is divided into two phases:
gradient evaluation and line search. Gradient evaluation generates the gradient direction and the
gradient length, and line search evaluates the objectives on a line segment along the gradient
and finds the minimum along the line segment

Question 28

progressive resolution

Answer 28

The angle resolution of the dose calculation segments gets more accurate as the optimization
progresses, and in consequence, the dose also gets more accurate. The number of control points
remains the same during the whole optimization.

Question 29

jaw tracking

Answer 29

dynamically moves the collimator jaws during beam-on to keep them as close to the
actual MLC aperture as possible. This reduces leakage between the MLC leaves. T

Question 30

explain rapid plan

Answer 30

DVH estimation algorithm is used for creating and applying DVH estimation models to be used in
treatment planning with RapidPlan. The main purpose of the DVH estimation algorithm is to
estimate what would be an achievable Dose-Volume Histogram (DVH) for various critical organs.
The estimations are based on the actual DVHs achieved in earlier planned patient cases. These
estimated DVHs can be translated into optimization objectives. The optimization objectives are
set so that the optimizer tries to achieve the estimated DVHs

Question 31

Multi-Criteria Optimization-Based Trade-Off
Exploration

Answer 31

a vector of objective functions is optimized instead of a
single objective function, as described in the equation below. Contrary to optimization of a single
objective function, no single best objective function value exists, but a set of best-compromise
points which constitute the Pareto surface of the problem.

Question 32

why measure cutout factors?

Answer 32

to validate your model since cutout factors are not modelled in eMC

Question 33

metrics of quality for electron beams

Answer 33

DTA
R50 (mm)

Question 34

Bremsstrahlung contribution to electron treatment

Answer 34

< 5 %

Question 35

where do the photons from electron interactions come from?

Answer 35

Bremss

Question 36

how decent is it to approximate secondary photons as going in same direction as electrons in eMC?

Answer 36

Bremss < 5 % of dose, seems reasonable

Question 37

why use higher resolution dose grid for low energies?

Answer 37

smaller scatter interactions- harder to model
-need more histories to bring uncertainty down

Question 38

remember difference between systemic bias vs uncertainty

Answer 38

uncertainty can be improved with more histories
bias is an inaccuracy (from approximations)