22 chap 20 intensity modulated radiation therapy

3,448 views
3,227 views

Published on

Published in: Health & Medicine, Technology
0 Comments
7 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
3,448
On SlideShare
0
From Embeds
0
Number of Embeds
19
Actions
Shares
0
Downloads
332
Comments
0
Likes
7
Embeds 0
No embeds

No notes for slide

22 chap 20 intensity modulated radiation therapy

  1. 1. Chapter 20Intensity Modulated Radiation Therapy 1
  2. 2. Goals of Conformal Radiation Therapy • Conform dose to the tumor • Reduce normal tissue toxicity • Tumor dose escalation improve local tumor control 2
  3. 3. Conventional Conformal Therapy and IMRTConventionalConformal TherapyField shape conforms tothe outline of the target,uniform intensity acrossthe fieldIMRTNon-uniform intensityinside the field to achieveoptimum dose distribution target 3
  4. 4. Outline• Optimization (Inverse Planning)• Delivery• Quality Assurance• Clinical Applications 4
  5. 5. Inverse Planning Problem xjxj is the intensity of the j-thpencil beam.Find the optimum distribution targetof xj’s (i.e., the optimum iintensity distribution) that willgive the best target coverageand critical organ protection. organ 5
  6. 6. Optimization (Inverse Planning)Purpose: To find the ‘optimum’ intensity distribution for all beamsinvolved in a plan that will best meet the planner’s requirements.What are the requirements? Objective functions dose, dose/volume - based, biological indices - based: TCP, NTCPHow to find the optimum solution? Search algorithms deterministic methods stochastic methods(*Optimization is conceptually separated from delivery, so in thisstep we don’t need to be concerned about how it’s to be delivered.) 6
  7. 7. (Recent development)Direct Aperture Optimization:To optimize a set of apertures (shape and weight) toachieve the best dose distributions.Advantage: eliminates the step of leaf-sequencing;optimized results = deliverable results.(In fluence optimization, the optimal fluencedistribution may not be deliverable, due to limitationsof leaf transmission.) 7
  8. 8. Objective Functions based on Dose, Dose/Volume• Examples of commonly used objective functions: Etarget: (D-P)2 2critical organs: (D-Dc)2, (D-Dc) if D>Dc cdose volume conditions: no more than p% of the volume to exceed dose q.• Reasonable and mathematically convenient.• No fundamental physical basis.• Could be any other forms such as |(D-P)m|. 8
  9. 9. Objective Function Based on Dose and Dose/Volume target Serial type Parallel type Normal tissue Normal tissue 2 2wl (D-P ) l (D-Dc) No more than p V % of the volume to 2 wu(D-P ) exceed dose q. u p% q D P P l u Dc Dose constraint Dose/volume constraint 9
  10. 10. Objective Functions based on Biological Indices• Maximize TCP, minimize NTCPs.• Used in place of or in conjunction with dose-,dose/volume-based objective functions.• At present, clinical data are scarce and modelsnot well-established yet.• Not available on commercial system at present. 10
  11. 11. Examples of Optimization Methods• Deterministic: DGradient, Conjugate gradient (Eclipse,Konrad). ) Maximum likelihood (Brainlab, nuclear medicine).• Stochastic: SSimulated annealing (NOMOS). ) Genetic algorithm (Brachy therapy). 11
  12. 12. Prostate IMRT plan5-field to 81 Gy 12
  13. 13. Global vs. Local Minimum (example of local minima) Objective Function 1 Uniform target dose (100% dose): 1 solution, beams 1&2 equally weighted. target2 Critical organs: no more than 1/2 to exceed 50% dose. 1 solution, beams 1&2 equally weighted. crit. org. Critical organs: no more than 1/2 to exceed 10% dose. 2 solutions, beams 1&2 weighted either (a,b) or (b,a) ⇒ local minima !! 13
  14. 14. Two-dimensional grey-scale display of the objective function Fobj as a function of beam weight 1 and beam weight 2 1.5 1.5 Fobj = 1404.5 at (0.24,1.5) 1 1 Fobj = 0 Beam 2Beam 2 at (1,1) Fobj = 745.7 at (1.5,0.35) 0.5 0.5 (a) (b) 0 0 0 0.5 1 1.5 0 0.5 1 1.5 Beam 1 Beam 1 Only dose uniformity in the Dose uniformity in the target target is required plus dose/volume constraint to the critical organ is required 14
  15. 15. Global Minimum• In Principle: I Stochastic methods can find global minimum. SDeterministic methods may get trapped in local minima.• In Practice: I It is not easily known if current solution is at global or local minimum, regardless of which methods are used. . It may not be important if the solution is clinically satisfactory. s If solution not satisfactory: ♦ for stochastic methods: change random selection parameters (e.g.annealing schedule), increase number of iterations. ♦ for deterministic methods: try different initial guess. ♦ is the new solution better? 15
  16. 16. Dose Uniformity in the TargetIMRT vs Conventional Conformal PlansThere is no Is itno issue. important ? yes IMRT should do better (at least no worse), for there are more degrees of freedom. 16
  17. 17. Dose Distributions IMRT vs Conventional Conformal Plans target 105 100 70 45°45° Wedged pair IMRT pair 17
  18. 18. Dose Volume Histograms IMRT vs Wedged Pair 100 80 IMRT Wedged PairVolume (%) 60 40 target 20 rest of volume 0 0 20 40 60 80 100 120 Dose (%) 18
  19. 19. Intensity Profiles Produced from Inverse Planning without smoothing• Local fluctuation in intensityprofile, due to numerical artifact.• Difficult to deliver, requires longerbeam-on-time.• Susceptible to treatmentuncertainties. 19
  20. 20. Smoothing of beam profiles• Smooth outside objective function (at the end of each iteration) – General method; independent of the objective function – Cannot distinguish between clinically relevant gradients and artifacts• Smooth inside objective function – Depends of particular form of objective function – Dosimetric consequences of smoothing accounted for in optimization. 20
  21. 21. Smoothing of Intensity Profiles Methods of smoothing: • Median window filter (webb, PMB 43:2785, 1998) • Savitzky-Golay (least squares fit) median window filterSavitzky-Golay(least squares fit) before smoothing after smoothing (with S-G) 21
  22. 22. Smoothing Included as Part of the Objective Function.Fobj = ∑ w (D − D ) i∈target i i p 2 For over & under dose in the targets For over dose in critical organs, may include dose-volume conditions For smoothness in intensity profile
  23. 23. Example 1: Phantom PTV OAR PTV 100 5 100Smoothing at the end of Smoothing within the each iteration objective function 23
  24. 24. PTV x Brainstem 95 50 30Smoothing at the end of Smoothing within the each iteration objective function 24
  25. 25. 100 100 80 PTV 80Volume (%) 60 60 40 Cord 40 Brainstem 20 20 0 0 0 20 40 60 80 100 120 0 10 20 30 40 50 60 70 Dose (%) Dose (%) Smoothing at the end of each iteration Smoothing within the objective function No smoothing 25
  26. 26. The ‘Skin-Flash’ Problem in Inverse PlanningConventional:Margin added to fieldedge to allow foruncertainties.IMRT: PTVIntensity remains at 0outside PTV. No skin-flash
  27. 27. Skin-Flash for Intensity-Modulated Field Simple, flat extension skin
  28. 28. Detection of unexpected hotspots Beam IBeam II A PTV B PTV Cord Beam III Hotspots > 120% location > 3 cm from PTV 28
  29. 29. Detection of unexpected hotspots• IMRT may produce hotspots outside the PTV – Can be at large distances from the PTV.• Hotspots may not be appear at the standard three orthogonal planes used for plan evaluation – May be worse for non-coplanar beams• Proposed solution: – Use volume display to detect possible hotspots – “Rind” = annular region around the PTV – “Rest-of-body” = “irradiated body” - PTV - OARs – Efficiency issues: “Rest-of-body” is a large structure and fine calculation grid is required 29
  30. 30. Summary - Optimization• Objective functions may be based on dose, dose/volumeconditions, or biological indices.• Optimization methods may be stochastic or deterministic.• Local minima may exist, but there is no easy way to tellwhether a solution is at a global or local minimum, regardlessof which optimization method is used.• Optimized intensity distribution should give better dosedistribution than conventional conformal plans.• A typical case involving 103 rays, and 104 points, takes about3 to 10 minutes of computer time with the gradient method. 30
  31. 31. Summary – Optimization (practical considerations)• smoothing is necessary to reduce unnecessary fluctuations inintensity distribution.• skin flash is needed to account for treatment uncertainties(breast, head/neck).• beware of unexpected hot spots. 31
  32. 32. Delivery of IMRT• Compensator: less efficient (fabrication, re-entryinto the room between fields).• Fixed field with conventional MLC: Fcontinuous leaf motion, cstep-and-shoot.• Rotational field: Rconventional MLC, cNOMOS/MIMIC, Tomotherapy. 32
  33. 33. MIMiCMultileafIntensityModulatingCollimator 33
  34. 34. Individual leaf controls openingMIMiCMultileafIntensityModulatingCollimator 34
  35. 35. Tomotherapy 35
  36. 36. Delivery of IMRT with a Multileaf Collimator (MLC)MLC mounted on the head of a linac Close-up view of the leaves 36
  37. 37. Delivery of IMRT with a Conventional MLC sliding window method radiationbeam-ontime direction of motion Right-leaf Left-leaf P 37
  38. 38. Methods - Sliding Window Techniques Dynamic vs. Segmental (step-and-shoot) Dynamic Segmental T dBeam-on-Time d Left-leaf T c Left-leaf Right-leaf c Right-leaf b b delivered intensity delivered intensity a a ∆t 1 1 X X = ≤ v=∞ ∆x v vmax P P 38
  39. 39. The Posterior Field in a Prostate Treatment 39
  40. 40. An Intensity-modulated Field Delivered with a Conventional MLC For cordprotection 40
  41. 41. Practical Considerations• Maximum leaf speed - approx. 2.5 cm/sec on Varian MLC• Leaf transmission - approx. 1.5% on Varian MLC• Leaf edge effect - rounded leaf end on Varian MLC• Source distribution - variation of output with fieldsize & shape• Tongue & groove effect•Maximum leaf protrusion/retraction on each side -14.5 cm on Varian MLC• very low intensities not deliverable 41
  42. 42. Tongue-and-groove Effect BEAM• Varian MLC employs a tongue-and-groove groove tongue design to reduce leakage between leaves• Leaf synchronization – Van Santvoort 1996 – Webb 1997 – Longer beam-on time 50% 42
  43. 43. Splitting Large IM fields• Maximum Varian DMLC width ~ 15 cm• Larger fields (e.g. NP) split into 2 or 3 subfields• Considerations: – Smoothness of profiles; no discontinuities – Split along a straight line or a low intensity region – Use feathering to reduce effects of uncertainties 43
  44. 44. Desired and Delivered Intensity Profiles desired delivered 25 20intensity (MU) 15 10 5 0 -8 -6 -4 -2 0 2 4 6 8 x (cm) 44
  45. 45. Comparison of DMLC and SMLC Methods DMLC: • accurate delivery of the desired intensity profiles. SMLC or Step-and-shoot: • User more comfortable - resembles multi-segment conventional treatment. • Shorter beam-on-time (MU) compared to DMLC. • Longer delivery time (min.) compared to DMLC. (on some machines) • Loss of spatial & intensity resolution 45
  46. 46. Conversion fromcontinuous to discrete intensity profiles Continuous Equispaced multi-level intensity profile intensity profileResolution : x = 2 mm Resolution depending on the intensity = continuous number of levels 46
  47. 47. DVH comparison between DMLC and a 5-level SMLC delivery for a prostate case 120 DVH Comparison 100 PTV DMLC 80volume (%) 5-level SMLC 60 x-grid = 2 mm rectum 40 20 bladder femur 0 0 20 40 60 80 100 120 dose (%) 47
  48. 48. DVH comparison between DMLC and a 5-level SMLC delivery for a head/neck case DMLC 5-level SMLC x-grid = 2 mm 120 30 PTV 100 25 80 20volume (%) intensity (MU) 60 15 cord 40 10 20 5 0 0 0 20 40 60 80 100 120 -10 -8 -6 -4 -2 0 2 4 6 8 10 dose (%) x (cm) DVH comparison Delivered profile 48
  49. 49. Combined DMLC and SMLC Delivery Introduction of Beam-off Segments DMLC DMLC/beam-off segmentTo deliver very low intensity To deliver very low intensity,(e.g. cord block), opposing introduce a beam-off segmentleaves touch each other & move while leaves move from a to b.at maximum speed from a to b. T T T’ Left leaf Left leaf TimeTime Right leaf Right leaf F(x) F(x) 0 x x2 0 x x2 1 a b X 1 a b X 49
  50. 50. Leaf Pair 14 Desired intensity profile Delivered intensity profile Left leaf path Right leaf path Without beam-off segment With beam-off segment100 10090 9080 8070 7060 6050 5040 4030 3020 2010 10 0 0 -10 -5 0 5 10 -10 -5 0 5 50 10
  51. 51. Summary - Delivery• Intensity-modulated field can be delivered with aconventional MLC using the Sliding Window Technique,either in dynamic (DMLC) or segmental (SMLC or step-and-shoot) mode.• Capable of delivering (almost) any shape of intensityprofile, but limited by leaf speed, transmission, etc.• Need to account for leaf transmission, rounded leaf end,and head scatter.• Split large field into 2 or more segments, if necessary. 51
  52. 52. Summary - Delivery (cont’d)• Loss of beam-on-time efficiency relative to conventionalfields due to small field openings. ◊ prostate : 2~3 times ◊ nasopharynx: 3~4 times ◊ breast: about the same• Treatment time equal or less than conventional treatment, e.g.,5-field prostate treatment < 8 min. (excluding patient setup).• Delivery with SMLC requires less beam-on-time (MU), butlonger delivery time (min) than that with DMLC (on somemachines).• Delivery with SMLC should avoid segments of short beam-on-time for concerns of beam stability. 52
  53. 53. The Effects of Intra-Fraction Organ Motion on the Delivery of Intensity-Modulated Fields with a Multileaf Collimator• For conventional treatment, the entire treatment field is exposedduring a fraction.• For IMRT, the treatment field is divided into many sub-fields, ofwhich only one sub-field is exposed at a time. Consequently, ifthere is organ motion during treatment, portions of the treatedvolume may move in and out of the sub-field during a fraction.• Since leaf sequence is designed based on a static geometry, thepresence of organ motion will cause the actual delivered intensityprofile to be different from the planned one.
  54. 54. Organ Motion Relative to Leaf Motion Parallel Perpendicular T d d’Beam-on-Time Left-leaf c’ Right-leaf c b’ b Desired intensity a a’ X Y P P
  55. 55. Intra-fraction Breathing Motion (Breast Treatment) A = ±3.5 mm, τ = 4 sec. organ motion  leaf motion Leaf pair #14: average gap = 4.51 cm. Single fraction Averaged over 30 fraction planned profile ******* delivered profile, motion averaged over 30 fx
  56. 56. Intra-fraction Breathing Motion (Breast Treatment) A = ±3.5 mm, τ = 4 sec. organ motion  leaf motion 105 103 100 50 IMRT static IMRT motion averaged over 30 fx
  57. 57. Intra-fraction Breathing Motion (Lung Treatment) A = ±7.0 mm, τ = 4 sec. organ motion ⊥ leaf motion planning images acquired1 cm planned profile ******* delivered profile, motion averaged over 30 fx
  58. 58. Intra-fraction Breathing Motion (Lung Treatment) A = ±7.0 mm, τ = 4 sec. organ motion ⊥ leaf motion planning images acquired 105 100 90 50 20 IMRT static IMRT motion averaged over 30 fx
  59. 59. Intra-fraction Breathing Motion (Lung Treatment) A = ±7.0 mm, τ = 4 sec. organ motion ⊥ leaf motion planning images acquired 120 100 GTVvolume (%) 80 static PTV 60 motion lungs 40 20 0 0 20 40 60 80 100 120 Dose (%)
  60. 60. Summary of Effects of Intra-fraction Organ Motion • Effects of intra-fraction organ motion can be calculated, provided the pattern of motion is known. • Effects can be calculated for single or multiple fractions. • For breast treatment, if the amplitude < 3 ~ 5 mm, effects appear to be insignificant over multiple fractions. • Penumbra broadening at field edge is common to both conventional and intensity-modulated fields. • If in-field effects significant, alternative means will be needed, e.g., compensator, breath hold, gating.
  61. 61. Quality Assurance• Machine Performance (DMLC): : Film test FGap test• Machine Performance (Step-and-Shoot): : Dose/MU vs. MU D Flatness, symmetry vs. MU• Patient Dosimetry: : Record & verify, file check-sums (each fraction) ( Independent MU check, portal image (each patient) ( Log file analysis, chamber measurement, film dosimetry (periodically or new software, technique, or site) 61
  62. 62. Film test1 mm bands errors introduced ← - 0.5 mm ← - 0.2 mm ← + 0.2 mm ← + 0.5 mm 62
  63. 63. Gap error → Dose error 20.0 Range of gap width 15.0% Dose error Gap error (mm) 10.0 2.0 1.0 5.0 0.5 0.2 0.0 0 1 2 3 4 5 Nominal gap (cm) 63
  64. 64. Gap Test 100 MU Gap error → Dose error 5mm gap 20.0 Range of gap width 15.0 chamber% Dose error Gap error (mm) 10.0 2.0 1.0 5.0 0.5 0.2 0.0 0 1 2 3 4 5 64 Nominal gap (cm)
  65. 65. DMLC Output vs Gantry / Collimator Angles 1.03 1.03 1.02 1.02Relative output 1.01 0.2 mm 1.01 1.00 1.00 0.99 0.99 0.98 0.98 245 0-0 445 0.97 90-0 0.97 8 8 270-0 7 7 8 8 8 8 8 7 7 8 8 t-9 ov-9 n-9 pr-9 n-9 ug-9 ct-9 8 180-0 t-9 ov-9 an-9 pr-9 n-9 ug-9 ct-9 Oc N Ja A Ju A O Oc N J A Ju A O 90-90 Date 270-90 Date 65
  66. 66. Beam Stability: Dose per MU• Iso-centric setup, farmer- type ion chamber, in solid water phantom• Checked both short and long term stability.• For > 2MU, dose per MU is within +/- 2%. For > 2MU, dose per MU is within +/- 2%. 66
  67. 67. Beam Stability: Flatness, Symmetry 0.1 0 GUN-TARGET DIRECTIONS ymmetry CROSS-PLANE DIRECTION 0.05 0.00 -0.05 0.1 0Flatnes s 0.05 0.00 -0.05 1 10 1 00 Total MUs Delivered 67
  68. 68. Measured vs Calculated Sum of 5 IMRT fields 1.03 1.02Dmeas / Dcalc 1.01 1.00 Linac-1 0.99 Linac-2 0.98 0.97 0.96 mean = 0.993, s = 0.008 0.95 0 100 200 300 400 Patient # 68
  69. 69. Port Film for Intensity-Modulated FieldUse left leaves of first segment and right leaves of lastsegment as the port film field. IMRT field Port film 69
  70. 70. Dose Reconstruction from Log & DVA Files Nasopharynx - Rt Lateral Overlay Dose differences (log-dva) dva dva log filelog file 70
  71. 71. Log File Analysis 1 Dose 0.2 0.9 max = 32 cGy 0.15 0.8 B 0.1 Dose Difference (cGy) 0.7 A 0.05 0.6MU fraction 0.5 0 0.4 -0.05 0.3 -0.1 0.2 0.1 -0.15 0 -0.2 -8 -6 -4 -2 0 2 4 6 8 Distance from axis (cm) 71
  72. 72. Independent Verification of IMRTA good QA practice. In the US, it may also be a regulatoryrequirement. For IMRT, hand calculation is impractical, anindependent verification program is needed. Leaf sequence file, beam-on-time, input jaw settings, machine data table. Including effects of scattered source, rounded leaf end, Calculate delivered mid-leaf & between-leaf transmission,intensity distribution tongue-and-groove effect, and scatter from the leaves. Calculate dose Calculate dose (cGy) . 72
  73. 73. Calculation of Delivered Intensity DistributionLeaf sequencing file: describes leaf positions as a function ofbeam-on time (MU).Step-and-shoot time T1 T2 T3 T4 Leaf positions + + + + dynamic time 0 T Leaf positions + ••• + ••• + •••+ 73
  74. 74. Between leaf transmission Tongue-and/or-groove effect BEAM BEAMgroove tongue 74
  75. 75. Ring-shaped Field first segment second segment 75
  76. 76. Varian Clinac-2100C MLC tongue & groove 76
  77. 77. Intensity distribution in the penumbra region as a functionof the distance from the rounded leaf end. beam mid-leaf Rounded leaf-end tongue or groove 1 direct exposure under mid-leaf 0.8 under tongue or groove intensity 0.6 0.4 0.2 under the leaf 0 ε -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 77 distance from the leaf-end (cm)
  78. 78. Variation of Output with Field Shape/Size Backprojection to the Source Plane Source plane MLC opening MLC plane Isocenter plane (x,y)  1 if ( x, y ) inside opening φi = ∫∫ Oi ( x , y ) S ( x − x , y − y )dx dy Oi ( x, y ) =  source  0 if ( x, y ) outside opening φ = ∫∫ φ ( x , y ) S ( x − x , y − y )dx dy φ = ∫∫ source Oi ( x , y ) S ( x , y )dx dy i source =φ ⊗S 78
  79. 79. Scattered Intensity from the MLC – For large fields can contribute up to 5% intensity Pencil beam 1.5 to 2% transmission leaf Scatter fromScatter from the leafthe leaf 10 5 0 5 10 transmission Distance from pencil beam 79
  80. 80. Scattered Intensity from the MLC Source plane intensity Scattered intensity from the leaf Smlc MLC opening 10 5 0 5 10 MLC plane Distance from pencil beam φmlc ( x, y ) =Isocenter plane ∫∫ B( x , y )S isocenter mlc ( x − x , y − y )dx dy plane  0 if ( x, y ) inside opening φmlc = B ⊗ S mlcBi ( x, y ) =   1 if ( x, y ) outside openingB ( x, y ) = ∑ Bi ( x, y )∆ti 80 i
  81. 81. Dose Calculation Algorithm for IM Fields Pencil Beam Convolution pencil beam Intensity-modulated field d1 d2 pencil beam kernel dose distribution 81
  82. 82. Comparison Between Calculation and Measurement Film Calc. Tongue-and-groove Between leaf transmission 82
  83. 83. A Field used in a Nasopharynx Treatment Sup 6 MVxAnt Post Film Plan Inf 83
  84. 84. 7-field Nasopharynx - cylindrical phantom/coronal 84
  85. 85. Dosimetry (Lung - PA field) 2 cmPlanFilm Film - Plan Overlay Difference 85
  86. 86. Lung PA-field: Dose difference (film – calc.) MLC scatterT&G Source 86
  87. 87. Milestone of IMRT at MSKCCOct. 95: 1st IMRT (boost) at MSKCCSpring 96: Prostate 81 Gy, all IMRT (5 fields)*96-97: Prostate 86 Gy, all IMRT (5 fields)98: Head / Neck99: BreastTo date: over 2,000 prostate, H/N, breast, and other IMRT patients *Has become Standard Treatment 87
  88. 88. IMRT for Prostate History: Prostate Dose Escalation Study 7020 - 7560 - conventional8100 - conventional and IMRT (with rectal sparing) ~ 500 patients with IMRT 8640 - IMRT (with rectal sparing) 40 patients 9180 - IMRT (with rectal and urethra sparing) 88
  89. 89. 81-Gy Prostate Plans – 3D Conformal Plan 6-field to 72 Gy, 6 cone-down fields to 81 Gy Rectal wall target bladder wall 89
  90. 90. 3D Conformal Plan - Prostate 81 Gy 100 D=77Gy V=90% 75Volume (%) D=75Gy 50 V=30% target bladder 25 rectum femurs 0 0 25 50 75 100 Dose (Gy) D=72Gy 90
  91. 91. 81 Gy - Transverse8640 8100 7560 5000 2500Conventional Plan IMRT Plan 91
  92. 92. Prostate 81-Gy Plans 100 100 standard 100 optimizedVolume (%) 75 75 75 50 50 50 25 target 25 25 rectum 0 70 80 90 0 0 0 25 50 75 100 0 25 50 75 100 100 100Volume (%) 75 75 50 50 25 bladder 25 femur 0 0 0 25 50 75 100 0 25 50 75 100 Dose (Gy) Dose (Gy) 92
  93. 93. PSA Relapse-free Survival in Unfavorable Patients by Dose 100 81 Gy (65) 75.6 Gy (193) Percent PSA Relapse-free Survival 64.8-70.2 (134) 80 66 % 60 p=0.05 43% 40 p=0.006 21% 20 Unfavorable T1-3 PSA >10; Gleason >7 0 0 12 24 36 48 60 72 84 96 108 Months 93
  94. 94. Grade 2 Rectal Toxicity by Dose 80 20 75.6 Gy (446)Percent Grade 2 Rectal Toxicity 81 Gy (61) 15 85 10 90 p< 0.001 64.8-70.2 Gy (364) 595 81 Gy IMRT (171) 0 24 48 72 96 120 Months 94
  95. 95. Nasopharynx Traditional Treatment e e50 Gy70 Gy77 Gy PTV Cord, Brainstem 95
  96. 96. MSKCC Conformal Treatment50 Gy70 Gy77 Gy PTV Cord, Brainstem 96
  97. 97. IMRT Treatment50 Gy70 Gy77 Gy PTV Cord, Brainstem 97
  98. 98. Comparison of nasopharynx treatment techniques 50 Gy 70 Gy 77 Gy Traditional Conformal IMRT 98
  99. 99. Posterior Field in a Nasopharynx Treatment 99
  100. 100. Nasopharynx IMRT 100
  101. 101. 3D dose display of the 65 Gy isodose surface IMRT 3D Conformal Traditional Spinal cord/Brain stem Eye PTV 101
  102. 102. 100 100 CORD 80 PTV 80%VOLUME 60 60 40 40 Conf . Conf . 20 Trad . 20 Trad . IMRT IMRT 0 0 0 2000 4000 6000 8000 0 2000 4000 6000 8000 DOSE (cGy) DOSE (cGy) 100 100 MANDIBLE 80 80 PAROTID%VOLUME 60 60 40 40 Conf. Conf . Trad. 20 Trad . 20 IMRT IMRT 0 0 0 2000 4000 6000 8000 0 2000 4000 6000 8000 102 DOSE (cGy) DOSE (cGy)
  103. 103. Plan ComparisonPlan PTV Cord Mandible Time Mean V95 Max. Max V95 Plan TreatTrad. 68 Gy 79% 48.5 Gy 77 Gy 27% 1h 0.4 hConf. 73 Gy 97% 47.5 Gy 82.5 Gy 18% 12 h 0.75 hIMRT 76 Gy 99% 39.0 Gy 77 Gy 9% 8h 0.5 hAvg. for 3 patients 103
  104. 104. Fusion of CT, MRI, and PET Scans for IMRT Planning 104
  105. 105. Parotid Sparing in 100 N0 disease 80 %Volume 60 PTV54 40 No PS 20 PS 0 0 20 40 60 80 Dose (Gy) 100 80 PAROTID %Volume 60 No PS PS 40 2030 36 42 48 54 Gy 0 0 20 40 60 80 PTV 54 Parotid Dose (Gy) 105
  106. 106. Optimization Parameters and Target -Normal Tissue Proximity 100 PTV70 80 % Volume 60 40 Lt. Cochlea 20 040 47 55 63 70 Gy 0 20 40 60 80 Dose (Gy) PTV 70 PTV 54 Cochlea 106
  107. 107. IMRT for Nasopharynx – Hong Kong 107Kam et al, IJROBP 60, 1440 (2004)
  108. 108. IMRT for Head/Neck - Finland Stimulated secretion 108Saarilahti et al, Radiother Oncol 74, 251 (2005)
  109. 109. Breast Plans - IMRT vs. Wedged Pair 108 104 100 90 50 10Wedged pairIMRT 109
  110. 110. DVH: Coronary Artery Region 100 80 Standard Intensity ModulatedVolume (%) 60 40 20 0 0 20 40 60 80 100 120 Dose (%) 110
  111. 111. DVH: Contralateral Breast Standard IMRT Dose (%) 111
  112. 112. Breast Plan Comparison PLAN PTV Ipsilateral Lung Arteries Max. Dose V50 V100 Mean Dose IMRT 112% 12.5% 6.6% 21.3% Conv. 120% 14% 10.2% 32%Avg . for 10 patients 112
  113. 113. CTI/Siemens PET/CTCT PET PET/CT 113
  114. 114. RPM Respiratory Gating System Infrared camera Reflective respiration monitorFirst RG-IMRTprotocol patienttreated in On/off beam gateOctober, 2001 114 Control workstation Treatment machine, CT, or PET
  115. 115. IMRT 3DCRT 115
  116. 116. IMRT vs. 3DCRT: DVH for lungs and PTV IMRT 3DCRT 116
  117. 117. Whole Abdomen Treatment (planning study only) Target: whole abdomen. Critical Organs: bones, kidneys, liver. conventional IMRTAP-PA Standard distanceExtended distance Two isocenters Nine fields 117
  118. 118. Whole Abdomen - Isodose Distributions108 100 90 60 50 coronal sagittal 118
  119. 119. IMRT AP-PAPTV KidneysLiver Bones 119
  120. 120. Summary• Optimized IMRT plans can produce better dose distributionsthan conventional conformal plans, in terms of both targetcoverage and normal organ sparing.• Delivery of IMRT with a conventional MLC is practical,either in continuous or in step-and-shoot mode.• Comprehensive QA is needed, for machine performance andpatient-specific dosimetry.• IMRT has been implemented for prostate, head/neck, breastand other sites. Early experience on prostate and nasopharynxthus far show encouraging results. 120

×