IMRT for Prostate Cancer
Robert A. Price Jr., Ph.D.
Philadelphia, PA
1
IMRT Patients at FCCC
Approximately 130150 patients per day
on 6 linacs
1481
Prostate
Breast
H&N
Other
1600
Number of Patients
1400
~ 55-60% are
treated via IMRT
1200
1000
>1900 IMRT
patients total
800
600
293
400
97
64
200
0
Treatment Site
All patients are simulated in the supine position. Reproducibility is
achieved using a custom alpha cradle cast that extends from the
mid-back to mid-thigh. The feet are positioned in a custom
plexiglas foot-holder. The patient is told to have a half-full bladder
because during treatment a full bladder is difficult to maintain.
foot
holder
alpha cradle
2
Simulation (Positioning and Immobilization )
Bad rectum
• The patient is asked to
empty the rectum using an
enema prior to simulation.
Also, a low residue diet the
night before simulation is
recommended to reduce gas.
If at simulation the rectum is
>3 cm in width due to gas or
stool, the patient is asked to
try to expel the rectal
contents.
6 cm
4.5 cm
Good rectum
3 cm
2.5 cm
CT Scans
• Scans are acquired from approximately
2 cm above the top of the iliac crest to
approximately mid-femur. This scan
length will facilitate the use of noncoplanar beams when necessary.
• Scans in the region beginning 2 cm
above the femoral heads to the bottom
of the ischial tuberosities are acquired
using a 3 mm slice thickness and 3 mm
table increment. All other regions are
scanned to result in a 1 cm slice
thickness.
3
MR Scans
0.23 T, Philips Medical Systems
•
All prostate patients also undergo MR
imaging within the department,
typically within one half hour before or
after the CT scan. Scans are obtained
without contrast media. The resultant
images are processed using a gradient
distortion correction (GDC) algorithm.
• CT and MR (after GDC) images
are fused according to bony
anatomy using either chamfer
matching or maximization of
mutual information methods. All
soft tissue structures are contoured
based on the MR information
while the external contour and
bony structures are based on CT.
MRI
Extracapsular extension
•Retrograde urethrograms are not performed.
CT
Seminal vesicles ?
Imaging modality may affect treatment regime
4
MRI
CT
Prostate (CT)
Prostate (MR)
Imaging artifacts may affect contouring
PTV growth = 8mm in all
directions except
posteriorly where a 5mm
margin is typically used
Rt FH
A
Bladder
S
I
CTV
P
Lt FH
Rectum
The “effective margin” is defined by the
distance between the posterior aspect of
the CTV and the prescription isodose line
and typically falls between 3 and 8 mm.
5
Acceptance Criteria
DVH Acceptance
Criteria
Good DVH
PTV95 % ≥ 100% Rx
R65 Gy ≤ 17%V
PTV95 = 100%
R40 Gy ≤ 35%V
B65 Gy ≤ 25%V
B40 Gy ≤ 50%V
FH50 Gy ≤ 10%V
R40 = 22.7%
R65 = 8.3%
R40 = 19%
B65 = 8.4%
6
Good plan example (axial)
100%
90%
80%
CTV
70%
60%
50%
“Effective margin”
The 50% isodose line should
fall within the rectal contour on
any individual CT slice
The 90% isodose line should
not exceed ½ the diameter of
the rectal contour on any
slice
Good plan example (sagital)
100%
90%
80%
70%
CTV
60%
50%
Attempting to get isodose
line “compression” (fast
dose fall-off) at the
prostate-rectum interface
7
DVH Acceptance
Criteria
DVH for bad plan
(meets DVH criteria)
PTV95 % ≥ 100% Rx
R65 Gy ≤ 17%V
PTV95 = 100%
R40 Gy ≤ 35%V
B65 Gy ≤ 25%V
B40 Gy ≤ 50%V
R40 = 31.5%
FH50 Gy ≤ 10%V
R65 = 13.4%
B40 = 21.3%
B65 = 9%
Bad plan example (axial)
100%
90%
80%
CTV
70%
60%
50%
The 50% isodose line falls
outside the rectal contour
8
Bad plan example (sagital)
100%
90%
80%
CTV
70%
60%
50%
The 50% isodose line falls
outside the rectal contour
Typical Dose
Routine treatments
• Prostate + proximal sv
(76 Gy @ 2.0 Gy/fx)
• Distal sv, lymphatics
(56 Gy @ ~1.5 Gy/fx)
• 38 fractions total
Post Prostatectomy
• Prostate bed
(64-66 Gy @ 2.0 Gy/fx)
9
Hypofractionation
• Prostate + proximal sv
(70.2 Gy @ 2.7 Gy/fx)
equivalent to 84.4 Gy
in 2 Gy fractions
assuming an α/β ratio
of 1.5.
• Distal sv, lymphatics
(50 Gy @ ~1.5 Gy/fx)
• 26 fractions total
BED for rectum & bladder
R50 Gy ≤ 17%V
R31 Gy ≤ 35%V
B50 Gy ≤ 25%V
B31 Gy ≤ 50%V
FH40 Gy ≤ 10%V
Number of Beam Directions
In the interest of
delivery time we
typically begin with 6
and progress to ≤ 9
Simpler plans such as
prostate only or prostate
+ seminal vesicles
typically result in fewer
beam directions than
with the addition of
lymphatics
10
Localization
BAT Alignment
Separation of seminal vesicles
into proximal and distal
This has allowed for increased accuracy. Patient
scans randomly evaluated; 303 prior to and 310 after
technique adopted. Evaluated by same physician.
Substandard alignments dropped from 15.1% to
3.5% (p=0.006)
-McNeeley et al. AAPM 2004
11
“CT-on-rails”
Prostate
20
15
LAT
CT shifts (mm)
10
1σ =2.16mm
AP
LONG
5
2σ =4.32mm
0
Ave=-.62mm
-5
-10
-15
-20
-20
-15
-10
-5
0
5
BAT shifts (mm)
10
15
20
BAT vs. Primatom shifts. Data for 218
alignments are presented (differences between
the 2 sets of shifts). The solid line is the line of
perfect agreement between the two systems.
1σ =2.14mm
1σ =2.36mm
2σ =4.28mm
2σ =4.72mm
Ave=-.2mm
Ave=-.32mm
Feigenberg et al. (submitted)
Prostate Bed
Note lack of physical
structures for
alignment with BAT
BAT first, CT second
(Mon, Thurs, Mon…)
BAT first, CT second
Patient #2 (with no template)
P a t ie n t # 7 ( w it h te m p la t e )
1 5 .0
AP
10.0
Lat
Long
5.0
0.0
1
2
3
4
5
6
7
8
9
-5.0
-10.0
10
11
12
13
Residual shift with sign (mm)
Residual shift with sign (mm)
15.0
AP
Initial Template (CT first and
“forced” BAT shift)
Template
1 0 .0
5 .0
Lat
Long
0 .0
1
2
3
4
5
6
7
8
9
10
11
- 5 .0
- 1 0 .0
- 1 5 .0
-15.0
Observation number
O b s e r v a t io n n u m b e r
Paskalev et al. (In Press)
12
Regions for dose constraint
Region
1
2
3
4
5
6
Limit
90% of target goal
80%
70%
50%
30%
20%
Minimum
Maximum
% volume ↑ limit
20
45% of target goal
Target Max
20
40%
90% of target goal
20
35%
75%
1
25%
55%
1
15%
35%
1
10%
25%
Price et al. IJROBP 2003
Region 1
CTV
Region 6
Attempt to assure that “PTV”
will be encompassed by the
1st region
Bladder
Rectum
Region 5
Region 2
Region 3
Region 4
Bladder
CTV
Region 1
Region 6
Region 2
Region 5
Region 3
Region 4
13
Regions
• 26 previously treated
patients (6 and 10 MV)
• The average number of
beam directions decreased
by 1.62 with a standard
error (S.E.) of 0.12.
• The average time for
delivery decreased by
28.6% with a S.E. of 2.0%
decreasing from 17.5 to 12.3
minutes
• The amount of nontarget tissue receiving
D100 decreased by
15.7% with a S.E. of
2.4%
• Non-target tissue
receiving D95, D90, D50
decreased by 16.3, 15.1,
and 19.5%, respectively,
with S.E. values of
approximately 2%
Price et al. IJROBP 2002
14
Bladder
Bladder
PTV
PTV
S.V.
S.V.
Rectum
Direct right lateral
Rectum
Anterior 45 degree oblique
IMRT Plan Comparison (Rectum)
110.0
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Percent Volum e Covered
5fd ax
7fd ax (non eq)
7fd ax (eq)
9fd ax
5fd xfire
7fd xfire
0.0
10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
Percent of Delivered Dose
15
Nodal Irradiation
JCO 21:1904-1911, 2003
The inclusion of pelvic lymphatic irradiation in
the treatment of prostate cancer for some
patients has been suggested in RTOG 9413.
16
WP followed by
prostate boost
Hormonal therapy
and WP
PSA control
Roach et al, RTOG 94-13
Targeting Progression
Intermediate risk (group 1)
High risk (group 2)
PTV = prostate + proximal sv
PTV1 = prostate + proximal sv
PTV2 = distal sv (no lymph nodes)
High risk (group 3)
PTV1 = prostate + proximal sv
PTV2 = distal sv
PTV3 = periprostatic + peri sv LNs
High risk (group 5)
High risk (group 4)
PTV1 = prostate + proximal sv
PTV1 = prostate + proximal sv
PTV2 = distal sv
PTV2 = distal sv
PTV3 = periprostatic + peri sv LNs
PTV3 = periprostatic + peri sv LNs + LN ext
+ LN ext
+ presacral LN
LN ext = external iliac, proximal obturator and proximal internal iliac
17
Prostate
Proximal SVs
Prostate
Distal SVs
Proximal SVs
18
Prostate
Regional
Lymphatics
Distal SVs
Proximal SVs
Extended
Lymphatics
Prostate
Regional
Lymphatics
Distal SVs
Proximal SVs
19
Extended
Lymphatics
Prostate
Regional
Lymphatics
Distal SVs
Proximal SVs
Rectum
Bladder
Extended
Lymphatics
Prostate
No longer
a geometry
problem;
avoidance
is only
minimally
useful
Regional
Lymphatics
Distal SVs
Proximal SVs
Rectum
20
Bladder
Extended
Lymphatics
Prostate
Regional
Lymphatics
Distal SVs
Proximal SVs
Rectum
Presacral Nodes
Lymphatic irradiation study
• 10 patient data sets
• Generate plans for each stage in targeting progression
• Evaluate effect of nodal irradiation on our routine prostate
IMRT plan acceptance criteria
• Evaluate effect on bowel
• Evaluate effect on erectile tissues
• Treatment time (logistical concerns as well as patient comfort)
• Physics concerns (dose per fraction vs. “cone downs”,
increased “hot spots”, PTV growth and localization technique,
rectal expansion and inclusion of presacral nodes, etc.)
21
Extended Lymphatics (good)
100%
90%
80%
70%
60%
50%
56 Gy
76 Gy
100%
90%
80%
70%
60%
50%
56 Gy
Extended Lymphatics (good)
B40 = 40.5%
B65 = 8.2%
R40 = 29.1%
R65 = 10.2%
22
Extended Lymphatics (bad)
100%
90%
80%
70%
60%
50%
56 Gy
76 Gy
100%
90%
80%
70%
60%
50%
56 Gy
Extended Lymphatics (bad)
B40 = 74.3%
B65 = 35.3%
R40 = 34.4%
R65 = 12.4%
23
R65
R40
R65 limit
R40 limit
1
2
3
4
Proximal SVs
5
6
7
8
9
Percent Volume
R65
R40
R65 limit
R40 limit
2
3
4
5
6
7
8
9
B65 limit
B40 limit
2
3
4
5
7
8
9
10
Bladder Dose
Segments = 55.1 (σ = 14.2)
80
70
60
50
40
30
20
10
0
B65
B40
B65 limit
B40 limit
1
2
3
4
5
6
7
8
9
10
Patient Index
Tx Time ~ 9.1 min
Beams = 8.2 (σ = 0.9)
Rectal Dose
6
Patient Index
10
Patient Index
Proximal SVs
B40
Beams = 7.0 (σ = 1.2)
40
35
30
25
20
15
10
5
0
1
B65
Tx Time ~ 9.0 min
Rectal Dose
Prostate
Distal SVs
Segments = 54.4 (σ = 13.1)
1
Patient Index
Bladder Dose
70
60
50
40
30
20
10
0
10
Percent Volume
Percent Volume
40
35
30
25
20
15
10
5
0
Percent Volume
Beams = 7.1 (σ = 1.1)
Rectal Dose
Prostate
Bladder Dose
R65
R40
R65 limit
R40 limit
1
2
3
4
5
6
7
8
9
Regional
Lymphatics
90
80
70
60
50
40
30
20
10
0
B65
B40
B65 limit
B40 limit
1
10
Patient Index
Proximal SVs
Percent Volume
Percent Volume
Segments = 76.1 (σ = 13.2)
Prostate
Distal SVs
40
35
30
25
20
15
10
5
0
2
3
4
5
6
7
8
9
10
Patient Index
Tx Time ~ 12.6 min
Siemens Primus, 10 MV, 10 x 10 mm2 minimum beamlet
Extended
Lymphtics
R40
R65 limit
R40 limit
1
2
3
4
5
6
7
8
9
Patient Index
Percent Volume
50
45
40
R40
R65 limit
R40 limit
3
4
5
6
7
Patient Index
Presacral
nodes
B40
B65 limit
20
10
0
B40 limit
1
R65
35
30
25
20
15
10
5
0
2
B65
2
3
4
8
9
10
5
6
7
8
9
10
Patient Index
Beams = 9.0 (σ = 0)
Rectal Dose
1
Segments = 117 (σ = 14.8)
Tx Time ~ 19.3 min
Rectum
Bladder
Bladder Dose
100
90
80
70
60
50
40
30
10
Percent Volume
Percent Volume
R65
Percent Volume
Beams = 8.7 (σ = 0.8)
Rectal Dose
40
35
30
25
20
15
10
5
0
Bladder Dose
Segments = 144.7 (σ = 11.7)
100
90
80
B65
70
60
50
40
30
20
10
0
B40
B65 limit
B40 limit
1
Tx Time ~ 23.9 min
2
3
4
5
6
7
8
9
10
Patient Index
Siemens Primus, 10 MV, 10 x 10 mm2 minimum beamlet
24
Varian 21 Ex & Siemens Primus
• 1 cm leaf width vs 5 mm leaf width
• 10 x 10 mm2 minimum beamlet vs 5 x 5 mm2
• We limit to 6-9 beam directions (primarily due to
treatment time)
• Corvus treatment planning
• Increased MU → Increase leakage → secondary
malignancies?, shielding concerns?
MSFmod = MUIMRT/MU3D CRT
Price et al. JACMP 2003
Bladder
Prostate
PTV
10 x 5 mm2 beamlets
Rectum
Bladder
Collimator 90 degrees.
Prostate
PTV
5 x 5 mm2 beamlets
Rectum
This places the short
axis of the beamlet
~perpendicular to the
prostate-rectal interface
Bladder
Prostate
PTV
10 x 5 mm2 beamlets
Collimator 0 degrees
Rectum
25
Mean values
5mm x 5mm (10.1%)
10mm x 5mm coll=0 (11.5%)
Percent of Rectum at 65 Gy
10mm x 5mm coll=90 (10.0%)
Percentage of Total Rectal Volum e
18
Original (5mm x 5 mm)
10mm x 5mm (Coll = 0)
10mm x 5mm (Coll = 90)
16
Comparisons
5mm x 5mm to 10mm x
5mm coll=0
14
p=0.004 (significant)
12
10
Comparisons
8
5mm x 5mm to 10mm x
5mm coll=90
6
p=0.85 (NOT significant)
4
2
Comparisons
0
1
2
3
4
5
6
7
8
9
10
Number of Patients
10mm x 5mm coll=0 to
10mm x 5mm coll=90
P<0.0001 (significant)
Mean values
5mm x 5mm (2055 MU)
10mm x 5mm coll=0 (1186 MU)
10mm x 5mm coll=90 (1344 MU)
4000
5mm x 5mm to 10mm x
5mm coll=0
Original 5mm x 5mm
10mm x 5mm (Coll = 0)
10mm x 5mm (Coll = 90)
3500
Monitor Units (MU)
Comparisons
Comparison of Daily Monitor Units
P<<0.001 (HIGHLY
significant)
3000
Comparisons
2500
5mm x 5mm to 10mm x
5mm coll=90
2000
P<<0.001 (HIGHLY
significant)
1500
1000
500
0
1
2
3
4
5
6
7
8
9
10
Number of Patients
26
Analysis
10 mm x 5 mm beamlets (coll 90)
5 mm x 5 mm beamlets
• Average # of segments ≈ 386
• Average # of MU ≈ 2055
• Average MSFmod ≈ 7.0
• Average # of segments ≈ 197
(~49 % reduction)
• Average # of MU ≈ 1344
(~34.6 % reduction)
• Average MSFmod ≈ 4.6
(~34.3 % reduction)
5x5 beamlet
5x10 beamlet
100%
100%
CTV
CTV
PTV
PTV
50%
50%
5x5 beamlet
PTV
100%
5x10 beamlet
100%
CTV
CTV
PTV
50%
Reductions
Segments from 141 to 81
50%
MU 1420 to 886
MSFmod 4.8 to 3.0
27
Routine QA
Prostate (Measured vs Calculated)
90%
300
80%
50%
Number of Patients
250
30%
200
150
100
50
0
[<-3.5]
[-2.5,-3.5] [-1.5,-2.5] [-0.5,-1.5] [-0.5,0.5] [0.5,1.5]
[1.5,2.5]
[2.5,3.5]
[>3.5]
Frequency Bins (% difference)
28
HooRay!!! Post Docs!!!
Weijun (Wil) Xiong
Zuoqun (Jay) Chen
Freek DuPlessis
Wei Luo
Antonio Leal Plaza
Jiajin (James) Fan
Jennifer Zhu
Xiu Xu
Sotirios Stathakis
Copernicus
Pollack-nicus
29