1) Image guidance is crucial for prostate radiotherapy given the need for dose escalation and tighter margins with IMRT/hypofractionation. 2) Both interfraction and intrafraction prostate motion occur randomly and can be significant without image guidance. 3) Implanted fiducial markers provide accurate localization and decrease setup errors compared to soft tissue imaging. 4) While image guidance reduces systematic errors, random errors still occur daily requiring daily imaging for optimal targeting. 5) Correcting for interfraction and intrafraction motion through real-time tracking and adaptive replanning may improve outcomes by reducing toxicity and failure rates through better dose conformity.

1. Prostate Cancer Radiotherapy:
Image Guidance
Patrick Kupelian, M.D.
Professor and Vice Chair
University of California Los Angeles
Department of Radiation Oncology
pkupelian@mednet.ucla.edu
February 2013

2. Prostate as a model for IGRT
CLINICAL NEED:
DOSE ESCALATION:
Smaller margins are a necessity (with or without IMRT)
HYPOFRACTIONATION:
Increases the importance of individual treatment accuracy
CHANGES DURING THERAPY DO OCCUR:
Motion: Significant
Deformation: Minimal for the prostate gland
SVs / LNs problematic
NATURE OF CHANGE: MOSTLY RANDOM
(Unlike anatomic changes due to tumor shrinkage, weight loss)

3. Image Guidance Methods
Different Guidance Methods Generate Different
Types of Images

4. GUIDANCE TECHNIQUES
• Soft tissue imaging: Ultrasound
CT
Large Inter / Intra user variability
• Fiducial based: X-rays
CT
Electromagnetic

5. Implantation of metallic markers
Prep: Antiobiotics day before, am and pm
Enema day of insertion
Insertion: Transrectal Ultrasound guidance
(Biopsy probe)
Needles: 18G needles
Pattern: 3 markers (R base, L base, R apex)
Time: 5 mins
Shinohara et al.
Technique for Implantation of Fiducial
Markers in the Prostate.
Urology 71, 196-200, 2008.

6. PROSTATE:
MARKER Patients Markers Intermarker distance
STABILITY n n variation
Dehnad et al. 9 19 SD: 0.5 mm
Poggi et al. 9 45 Mean: 1.2 mm
SD: 0.2 mm
Aubin et al. 7 21 Mean: –0.01 mm
SD: 0.87 mm
Litzenberg et al. 10 30 Range SD: 0.7 - 1.7 mm
Pouliot et al. 11 33 Mean SD 1.3 mm
Range SD: 0.44 - 3.04 mm
Kupelian et al. 56 168 Mean: 1.01 mm
SD: 1.03 mm
Range SD: 0.3 - 4.2 mm

7. IMPLANTED FIDUCIALS
MV X-rays kV X-rays
kV Cone Beam CT Helical MV CT MV Cone Beam
CT
VISIBLE ON X-RAY BASED IMAGING SYSTEMS

8. ARE IMPLANTED FIDUCIAL REQUIRED?
Megavoltage CT (MV CT) Cone Beam CT (CBCT)
Transverse
Transverse CB CT
MV CT
Sagittal Sagittal
MV CT CB CT

9. Prostate alignment study
MVCT scans
Radiation Therapist vs Physician
Alignment methods: Marker
Anatomy
Contours
3 patients, 112 scans
Langen et al., IJROBP, 62, pp 1517

10. Therapist vs Physician Registration
Differences ≥ 5 mm
Marker Anatomy Contour
A/P: 1 % A/P: 5 % A/P: 17 %
S/I: 2 % S/I: 10 % S/I: 31 %
Lat: 1 % Lat: 0 % Lat: 3 %
Markers decrease interuser variability
Langen et al., IJROBP, 62, 1517, 2005

11. Marker Location
Should be representative of relevant anatomy
(prostate/rectum interface)
Inadequate Placement Adequate Placement
(Anterior) (Posterior)

12. Interfraction Motion
Geometry versus Dosimetry
Dosimetric Impact of Motion

13. Interfraction Motion
Need for Daily Imaging?
74 patients, 2252 fractions, all IGRT with daily shifts.
Replay different alignment strategies;
Check residual errors vs actual daily shifts.
Significant proportions of residual errors with any scenario.
Example: Every other day imaging + apply running average;
Residual errors > 3 mm in ~40% of fractions
Residual errors > 5 mm in ~25% of fractions
Significant random component: Need daily imaging
Kupelian et al., IJROBP, 70: 1146, 2008

14. Interfraction Motion Corrections
Clinical Impact

15. Interfraction Motion / Clinical Impact
Challenges to document clinical impact:
Endpoints:
Cure; Long timeline, few events
Toxicity: Low number of significant events
Dose escalation
Implemented
Decreasing margins
simultaneously
Image Guidance
Independent effect of image guidance??

16. Image Guidance: Avoiding Systematic Errors
IMPACT ON CLINICAL OUTCOMES: TUMOR CONTROL
Impact of Rectal Distention at the Time of Initial Planning
MDACC Dutch Trial
de Crevoisier et al, IJROBP, 62, 965-973, 2004 Heemsbergen et al., IJROBP, 67, 1418–1424, 2007

17. NO IMPACT OF RECTAL DISTENTION ON
RELAPSE FREE SURVIVAL IN PATIENTS TREATED WITH IMAGE GUIDANCE
Cleveland Clinic Rectal volume at simulation
Biochemical Relapse Free Survival
N=488 1 <50 cc
BAT
IMRT .8
> 50<100 cc
> 100 cc
Rectal volume .6
on Planning CT
.4
Med FU: 60 mos
.2
p=0.18
0
0 12 24 36 48 60 72 84 96
Time (months)
Kupelian, IJROBP (70), 1146, 2008

18. TREATMENT MARGINS TOO SMALL:
INCREASED FAILURES ??
Engels, IJROBP, 74: 388-391, 2009

19. TREATMENT MARGINS TOO SMALL??
N=213 6 mm lateral, 10 mm otherwise No guidance
N=25 3 mm lateral, 5 mm otherwise Fiducial/Guidance
bNED at 5 years (median follow-up 53 months):
No guidance (large margins): 91%
Guidance (small margins): 58% p=0.02
On multivariate analysis, biochemical failure predictors were;
High Risk Group
Low RT Dose
Rectal Distention
Guidance (small margins)
Engels, IJROBP, 74: 388-391, 2009

20. CHHiP Trial
Localized Prostate Cancer; N=3026
1st randomization: Dose:
74 Gy at 2 Gy per fx vs 60 Gy at 3 Gy per fx
2nd randomization: Image Guidance:
Image Guidance vs no Image Guidance
3rd randomization: Margins

21. INTRAFRACTION MOTION

22. Prostate Real Time Motion Studies
Adapted from Ghilezan, IJROBP, 62, 406–417, 2005
Author (year) Obs Method Sampling Motion (mm)
Kupelian (2005) 1157 Calypso 9-11 min >3 mm motion for >30 secs in 41%
Continuous >5 mm motion for >30 secs in 15%
Ghilezan (2005) 18 Cine MRI 1hr q 6 sec Range S.D.: 0.7-1.7
Mah (2002) 42 Cine MRI 9 min q 20 s Range S.D.: 1.5-3.4
Padhani (1999) 55 Cine MRI 7 min >5 mm motion in 29%
Khoo (2002) 10 MRI 6 min q 10 s Range S.D.: 0.9 -1.7
Kitamura (2002) 50 Fluoro 2 min Range S.D.: 0.1-0.5
Dawson (2000) 4 Fluoro 10–30 sec Range S.D.: 0.9–5.3
Malone (2000) 40 Fluoro q 20 s >4-mm motion AP 8% SI 23%
Nederveen (2002) 251 Fluoro 2–3 min Range S.D.: 0.3-0.7
Shimizu (2000) 72 Fluoro 9 min Median: AP 0.7, SI 0.9
Vigneault (1997) 223 EPID -- No displacement
Huang (2002) 20 US (2) 15–20 min apart Range S.D.: 0.4 – 1.3

23. Intrafraction Motion
Electromagnetic Tracking (Prostate)
Calypso Micropos
Wireless Wire
Permanent Removable

24. Kupelian, IJROBP, 67: 1088-1098, 2007
High
Continuous Transient Persistent Frequency
Drift Excursion Stable
Excursion Excursion Erratic

25. MONITORING INTRAFRACTION MOTION IN THE CLINIC
Calypso, Electromagnetic tracking
Clinical protocol: 3 mm threshold
Realignment only between beams
Patients N=29 Fractions: N=963, mean= 33 /patient
Events: Mean Frequency Range
(indiv pt)
No motion >3 mm, no intervention 59% 10-100%
Motion >3 mm, transient, no intervention 14% 0-42%
Motion >3 mm, realignment between beams 25% 0-85%
MD disagree with therapist intervention 1% 0-8%
(interuser variability)
Kupelian 2009

26. Intrafraction Motion: Dosimetric Consequence
FRACTIONATION, Cumulative doses, D95%
MDACCO: N=16 patients, full course
Prostate - No Delay
Pat. 1
5
Pat. 2
a n g e fr o m p D95
Pat. 3
%P DosechChange, la n
Pat. 4
0
Pat. 5
Pat. 6
-5 Pat. 7
Pat. 8
Pat. 9
-10
e r ce n t
Pat. 10
Pat. 11
Pat. 12
-15
Pat. 13
Pat. 14
-20 Pat. 15
0 5 10 15 20 25 30 35 Pat. 16
Fract ion number
Langen, IJROBP, 2009

27. Intrafraction Motion: Dosimetric Consequence
FRACTIONATION, Cumulative doses, D95%
MDACCO: WORST CASE (WORST INTRAFRACTION MOTION)
P rostate - Delay (Pat. 8)
5 Fractions
10
P erc en t c h an g e from p la n
% Dose Change, D95
5
0
-5
- 10
- 15
Cumulative daily
- 20
- 25
0 5 10 15 20 25 30 35
Fraction number
Langen, IJROBP, 2009

28. INTRAFRACTION MOTION
CLINICAL IMPACT

29. CORRECTION OF
INTERFRACTION AND INTRAFRACTION
MOTION
CLINICAL IMPACT
“AIM” STUDY

30. Intrafraction Monitoring Correction – Clinical Impact?
Assessing the Impact of Margin (AIM) Reduction Study
Sandler, et al. Urology:75(5):1004-8, 2010
Acute toxicity comparison (PreRT and End of treatment QOL scores):
Group 1 Group 2 (Historical control)
Sanda et al., NEJM 2008 358(12) 1250-61
EM tracking (Calypso) Guidance method, if any, not reported
(2 mm threshold: “Institutional norms”
realigned in 57% of fx)
3 mm post margins Varying margins “5-10 mm”
81 Gy ~ 75-80 Gy
N=64 patients N=153 patients
IMRT, no hormones IMRT, no hormones
2008-2009 2003-2006

31. Intrafraction Monitoring Correction – Clinical Impact?
PreRT PostRT Difference (Post – Pre)
Benefit from smaller margins oron
EPIC
Domain
Study
Mean
EPIC
Mean
EPIC
95% CI
Difference
motion management? Difference
Score Score
AIM 91.8 89.8 -1.9 [-9.0, 5.1]
Bowel /
rectal NEJM
94.4 78.5 -16.0 [-19.4,-12.5]
Control
Urinary AIM 84.5 80.6 -4.0 [-10.0, 2.1]
irritation /
obstruction NEJM
86.6 70.1 -16.5 [-19.8, -13.3]
Control
Assessing the Impact of Margin (AIM) Reduction Study
Sandler, et al. Urology:75(5):1004-8, 2010

32. Reducing anatomic variations: Intrarectal balloon
Immobilization? Unclear
van Lin et al, IJROBP, 61, 278, 2005
Court et al, RO, 81, 184, 2006
Wang et al, RO, 84, 177, 2007
Heijmink et al, IJROBP, 73, 1446, 2009
Improve rectal dosimetry /
Decrease late rectal bleeding
Patel et al, RO, 67, 285, 2003
Teh et al, Med Dosim, 30, 25, 2005
van Lin et al, IJROBP, 67, 799, 2007

33. DAILY POSITIONAL VARIATION
BALLOON VERSUS PROSTATE GLAND
Day 5
3
2
1
4

34. BALLOON: INTRAFRACTION MOTION
DURING PROSTATE SBRT
Time = 0 mins
+38
+28
+10 mins
+18
+5

35. ENDORECTAL BALLOON:
DECREASE IN INTRAFRACTION MOTION?
Nijmegen + MDACC Orlando: Smeenk et al. ASTRO 2010.
Electromagnetic tracks in 30 patients
1143 tracks available for analysis
15 patients without balloon
15 patients with balloon
% 3D motion > 3 mm
20% No ERB Balloon decreases
15% but does not
10% eliminate
ERB
5% intrafraction motion
0%
200 500 1000
Time (s)

36. Deformation / Rotations
Dosimetric Impact

37. Interfraction Anatomic Variations: Daily MVCTs
Alignment on markers Alignment on mid gland prostate
Prostate Prostate
Seminal Vesicles Seminal Vesicles
Courtesy: Chester Ramsey

38. DOSIMETRIC IMPACT OF DEFORMATION
DOSE RECALCULATION
Plan Treatment CHECK ISODOSE LINES
RECALC
RECONTOUR:
MANUAL OR DEFORMABLE REGISTRATION CHECK DVHs
DVH

39. DOSIMETRIC IMPACT OF INTERFRACTION DEFORMATION
Legend Legend
10 patients Rectum
Prostate D95
Rectal volume @ 2 Gy (cc)
70 Plan Plan
2.3 Mean Mean
60 SD SD
Margins:
Fraction D95 (Gy)
Max
2.2
50 Min
Max
Min
6 mm 2.1
40
4 mm Posterior
30
2.0
20
1.9
400 daily scans
10
Alignment on marker 1.8
0
1.7 1 2 3 4 5 6 7 8 9 10
Recontours 1 2 3 4 5 6 7 8 9 10
Patient
Patient
Dose recalculation
Prostate: Adequate coverage
Rectum: Large variations in some patients
Kupelian, IJROBP, 66, 876-82, 2006

40. DEFORMATION
Challenge: Independent movement of prostate vs nodes
J. Pouliot, From Dose to Image to Dose: IGRT to DGRT, Panel on On-Board Imaging : Challenges and Future Directions,
ASTRO 49th Annual Meeting in Los Angeles, Ca, Oct. 29, 2007.
Xia et al., Comparison of three strategies in management of independent
movement of the prostate and pelvic lymph nodes. Med Phys. 2010
Sep;37(9):5006-13.

41. Independent Motion of Prostate vs Nodes
Solution: Don’t Treat Pelvic Lymph Nodes!
RTOG 9413
RTOG 94-13. N=1323. 4 group randomization.
No difference between Pelvis vs Prostate only RT
Biochemical failure free survival (Phoenix definition)
Whole Pelvis
Prostate Only
Lawton et al , IJROBP, 69, 646-655, 2007

42. GETUG Trial: Prostate only versus Pelvic RT
Pommier et al, JCO, 2007
Low Risk
High Risk
Progression Free Survival

43. FUTURE:
REAL TIME RADIOTHERAPY
Adjustments at the time of delivery
(similar to surgery)

44. Ultimate solution for deformations;
Real-Time Radiotherapy
Assessment and adjustments:
• Daily (all fractions)
• On-line
• Intra-fraction variations included
• Deformable registration software
• Dose accumulation (inter/intrafraction)
• Real-time adaptation
X-rays / CT / PET / MRI

45. Real-Time Radiotherapy:
Continuous soft-tissue imaging with
automated plannning / delivery
In-room MRI / MRI guidance
ViewRay system
0.35 T on-board MRI

46. Pilot (Navigation) Scans
20 sec Pilot Scan

47. Automatically Identify & Locate Tissues

48. Predict Dose On Demand

49. Optimize On Demand

50. Track Tissues & Control Therapy

51. Physician’s Workstation: Overall review and supervision

52. FUTURE:
… HYPOFRACTIONATION / SBRT …
BOOSTING OF INTRAPROSTATIC LESIONS
INCORPORTATING SBRT IN HIGHER RISK CASES
NORMAL TISSUE AVOIDANCE
URETHRA
RECTUM SUBSECTIONS / ANUS
ERECTILE TISSUES
SPHINCTERS

53. TECHNIQUE
SELECTIVE INTRAPROSTATIC BOOST
VS
FOCAL THERAPY

56. CONCLUSIONS
Image Guidance is necessary with dose escalation.
Assessing the dosimetric impact of radiation delivery
techniques is important, particularly with hypofractionation.
Systematic errors during delivery (e.g. rectal distention at
simulation) will introduce the most consequential errors.
The dosimetric impact of motion will differ in various moving
and deforming areas: e.g. tumor vs prostate vs SVs vs LNs.
In the future, real time adjustments might be necessary for
tight margin / large fraction treatments.

57. Prostate Cancer Radiotherapy:
Image Guidance
Patrick Kupelian, M.D.
Professor and Vice Chair
University of California Los Angeles
Department of Radiation Oncology
pkupelian@mednet.ucla.edu
February 2013