A Method to Reduce the Statistical Uncertainty Caused by High Energy Cutoffs in Monte Carlo Treatmen

1
A Method to Reduce the Statistical Uncertainty
Caused by High Energy Cutoffs in Monte Carlo
Treatment Planning Dose Calculation

• Jinsheng Li, Ph.D. and C-M Charlie Ma, Ph.D.
• Fox Chase Cancer Center, Philadelphia, PA 19111

2
Outline

• Brief introduction of FCCC
• Effects of energy cutoffs on Monte Carlo dose
  accuracy
• Effects on CPU time
• Effects on statistical uncertainty and thus
  simulation efficiency
• Methods to remove the effects
• Some examples

3
IMRT Patients at FCCC by 3/2007
Approximately 140-180 patients per day on 6 linacs
2250
Prostate
1600
Breast
55-60 are treated via IMRT
HN
1400
Other
1200
1000
gt3200 IMRT patients total
Number of Patients
800
600
500
400
300
150
200
0
Treatment Site
4
Features of MCSIM

• An EGS4 user code for dose calculation in a 3D
  rectilinear phantom
• Various source configurations
• Various beam modifiers
• CT phantom (DICOM-RT,RTOG,RTP files)
• Dose calculation for conventional treatments
• IMRT dose verification
• 3-D dose and DVH data output

5
Simulation Geometry
Source plane
IREGFLAG
Zero dose outside
6
The IREGFLAG Parameter
Every bit is used separately in MCSIM to mark a
structure
0
5
1
3
Bit 0 set if the voxel is inside ext contour
Bits 1-31 set if the voxel is in a structure
marked by the bit number
For example IREGFLAG 12832 43 The voxel
is inside the external contour and in structures
marked by bits 1, 3 and 5
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
Energy Cutoff Effects on Monte Carlo Dose
Distribution
AE0.7MeV, ECUT 2.5MeV (thin lines) vs. AE0.7Me
V, ECUT 0.7MeV (thick lines)
12
Energy Cutoff Effects on Monte Carlo Dose
Distribution
AE0.7MeV, ECUT 1.0MeV (thin lines) vs. AE0.7Me
V, ECUT 0.7MeV (thick lines)
13
Energy Cutoff Effects on Monte Carlo Dose
Distribution
AE0.7MeV, ECUT 1.5MeV (thin lines) vs. AE0.7Me
V, ECUT 0.7MeV (thick lines)
14
Energy Cutoff Effects on Monte Carlo Dose
Distribution
6MeV electron beam 4,10,30,50 and 90 Isodose
lines.
AE0.521MeV, ECUT 0.521MeV (thin
lines) vs. AE0.7MeV, ECUT 1.0MeV (thick lines)
15
Summary of Energy Cutoff Effects on Dose
Distribution

• 1.0MeV ECUT can be used for headneck and lung
  patient dose calculation
• 1.5MeV ECUT can be used for other part of body

Q How much can we gain from using high ECUT?
16
Effects of High Energy Cutoff on Monte Carlo Dose
Calculation Efficiency
For a prostate patient treated with 10MV photon
IMRT
17
Effect of High Energy Cutoff on Statistical
Uncertainty in Air Cavity
6MV photon IMRT plan
AE0.7MeV, ECUT 0.7MeV
18
A Method to Reduce the Statistical Uncertainty
Caused by High ECUT

• Continue Transport the electron after its energy
  is lower than ECUT
• Transport the electron along a straight line
• The continue energy loss is calculated based on
  the electron stopping power.
• Increase energy loss by 70 to account for the
  approximations made in transporting the electron
  in a straight line rather than a curved path.

19
Results of the new method
For a prostate patient treated with 10MV photon
IMRT
20
Results of the new method
Before
After
21
Results of the new method
Dose profiles along the line depicted in the
previous slice. The red line is the result of the
new method.
22
Conclusion

• High electron energy cutoff reduces the
  simulation time but increases the statistical
  uncertainty, especially in low density regions
  such as an air cavity.
• A simple method was developed to reduce the
  statistical uncertainty caused by high ECUT that
  improved the EGS4 Monte Carlo dose calculation
  efficiency.
• A significantly improved dose distribution can be
  achieved by using the new method for patients
  with low-density regions in the treatment volume
  without losing simulation efficiency and dose
  accuracy.

23
Thank You