3. LIMITATIONS OF CONVENTIONAL RT
• True spatial extent of disease
• Shape & location of normal structures
• Efficient planning & delivery
• Optimal dose distributions
• Resulting in:
• large safety margins
• Tumor dose often compromised to prevent normal tissue complications - higher
probability of local failure
6. RATIONALE OF IMRT
• More conformal and homogenous dose distribution
• Sharper falloff dose at PTV boundary – lesser dose to normal tissues
• Tumor dose escalation (TCP), reduction of normal tissue dose (NTCP – less morbidity)
• SIB - radiobiologic advantage
• Lower cost and burden to the patient
• Adaptive therapy - revision of treatment plan according to imaging of tumor reduction and
organ movement during the course of RT
7.
8. IMRT
• IMRT is a form of 3DCRT in which a computer-aided iterative optimization
process is used to determine customized nonuniform fluence
distributions to attain certain specified dosimetric and clinical objectives.
9. INVERSE PLANNING
• The word “inverse” is used in reference to the established body of
mathematical inverse problem-solving techniques, which start at the final
or desired result and works backwards to establish the best way to
achieve it.
• So-called inverse treatment planning starts by describing a goal, i.e., a
series of descriptors characterizing the desired absorbed dose distribution
within the tumor, with additional descriptors designed to spare normal
tissues.
11. PATIENT & TUMOR CHARACTERISTICS
• Cooperative – prolonged treatment sessions
• Irregularly shaped tumor
• Near-critical structures
• Few or no organ motion concerns
• Higher than conventional doses indicated
12. PATIENT POSITIONING AND IMMOBILISATION
• Comfortable
• Reproducible
• Suitable for beam entry with minimum accessories in beam path)
• IMRT is highly conformal – techniques to reduce internal organ motion
(simulation protocols)
16. • Radio-opaque fiducial – presumed isocentre
• Site dependant protocols
• Scout – topogram area
• Slice thickness
• Data transfer to TPS
VOLUMETRIC DATA ACQUISITION
17. TREATMENT PLANNING SYSTEM
• Image registration, segmentation or contouring
• Planning and Dose calculations
• Plan Evaluation
• Data Storage and transmission to console
• Treatment verification
18.
19. IMAGE REGISTRATION
• Aligns different image data sets into a common coordinate system
• Point to point fitting
• Line or curve matching
• Surface matching
• Volume matching
20. APPLICATIONS OF IMAGE REGISTRATION
• Fusion of preop scan to postop treatment planning scan to define the target volume
• Mapping CNS structures (more clear on MRI) to CT image for planning-fusion
• For organ motion studies
• Image guidance
• 4D CT
• Image registration allows computation of cumulative doses from multiple plans done on
different image sets for same patient
21. • Automated
• Transform (mapping between spaces)
• Linear (rigid)
• Non-linear
• Cost (score defining relative quality of
a match)
• Deformable
• Inter-fractional scans (organ shift and
deformation)
• Intra-fractional registration (respiratory
time scale)
• Multimodality image registration
IMAGE REGISTRATION
22. IMAGE SEGMENTATION
• Slice by slice delineation of anatomic region of interest –
manual/automatic - time consuming and labour intensive
• Qualitative and quantitative evaluation of treatment plan
• Reconstructed sagittal & coronal images provide
additional orientation cues & are useful in defining
spatially consistent volumes of interest
• Edge detection and edge tracking – high contrast
structures
23. • Volume definition is prerequisite for
3-D treatment planning
• To aid in the treatment planning
process & provide a basis for
comparison of treatment outcomes
VOLUME DEFINITION
24. • Prior to treatment planning: (anatomical)
• Gross tumor volume (GTV)
• Clinical target volume (CTV)
• Organs at risk (OAR)
• During treatment planning: (geometrical)
• Planning target volume (PTV)
• Planning organ at risk volume (PRV)
• Remaining volume at risk (RVR)
• Internal target volume (ITV)
• At the end of treatment planning:
• Treated volume (TV)
• Irradiated volume (IRV)
VOLUME DEFINITION
25.
26.
27. REMAINING VOLUME AT RISK (RVR)
Difference between volume enclosed by external contour of patient and that of CTVs and
OARs
Definition of an RVR and its inclusion in the treatment plan (at least in the form of dose
constraints) is essential in IMRT.
Without such limits, the optimization software could craft excellent dose distributions for the
CTV and OAR but cause toxic irradiation levels in otherwise uncontoured tissues.
28. BEAM’S EYE VIEW (BEV)
• Demonstrates geometric coverage of target
volume by the beam
• Shielding and MLC’s are designed on BEV
• Best gantry, collimator and couch angles can be
identified
29. ROOM’S EYE VIEW (REV)
• Viewing point simulating any arbitrary location
within the treatment room
• Overall treatment technique geometry and
placement of isocenter
30. MLC – MILLENIUM 120
• 60 leaves on each bank
• Central 40 leaves on each bank - 5mm wide at the isocentre plane
• Peripheral 20 leaves on each bank (10 each side) – 1cm wide
• Max velocity – 3cm/s
• Minimum gap width – 5mm
32. STATIC IMRT
• Conceptually simple
• Multiple subfields with uniform
intensities
• Beam off in between treatment –
lower MUs
• Step and shoot / segmental MLC
• Prolonged treatment time for
complex plans
DYNAMIC IMRT
• Complex
• Sweeping leaves
• Fields with varying intensities
• Smoothly varying intensity profiles
• Beam is on throughout leaf motion-
higher MUs
• Sliding window IMRT
33.
34. OPTIMIZATION
• Find the best physical and technically possible treatment plan
• Fulfill the specified physical and clinical criteria
• Mathematical technique aims to minimize or maximize a score under
certain constraints
• Beamlet optimization and aperture based optimization
35.
36.
37. BEAM ANGLE/ENERGY SELECTION
• Common – 5/7/9 equally spaced (360/n)
• non-opposing coplanar beams of 6MV
• Larger number of beams
• better conformality
• more treatment time and complexity
• >10MV – high neutron dose – increases probability of 2nd malignancy
38. BEAM ANGLE/ENERGY SELECTION
• Shortest path to irradiate targets and avoid OARs
• Two parallel opposed beams not used
• less beam variables for optimization
• Non-coplanar beams - better dose distribution
39. BEAMLETS (BIXELS / PENCIL BEAMS)
• Small photon intensity element - subdivide an intensity-modulated beam for intensity
distribution optimization or dose calculation.
• Each beamlet - fixed number of intensity levels.
• Beamlet width is limited to MLC leaf width.
40. BEAMLETS (BIXELS / PENCIL BEAMS)
• Beamlet length (step size of MLC leaf movement) defined as the smallest step in the leaf
travel direction is specified by the user.
• Smaller beamlet size / larger number of intensity levels - better spatial or intensity
resolution but requires more MLC seg.
43. OPTIMIZATION
• Physical criteria: dose coverage
• Biological criteria: TCP and NTCP calculation
• Total objective function (score) derived from these criteria
• Priorities defined to tell the algorithm the relative importance of different objectives (penalties)
• Algorithm maximize the score based on criteria and penalties
44. PLANNING OBJECTIVE
PTV:
• Lower objective:
• 100% volume = 100% prescription dose
• Upper objective:
• None of PTV volume receive > 100% dose
OAR:
None of the OAR volume receive any dose
Non realistic: Never practically achievable
45. ANALYTICAL METHODS
• Mathematical techniques in which desired dose distribution is inverted by using a back
projection algorithm.
• Reverse of a CT reconstruction algorithm – 2-D images reconstructed from one-
dimensional intensity functions
• Assume dose distribution is result of convolutions of a point-dose kernel and kernel
density
46. ANALYTICAL METHODS
• Then reverse is also possible - deconvolving a dose kernel from desired dose distribution
gives kernel density or fluence distribution
• Fluences then projected onto beam geometry to create incident beam intensity profiles.
• Disadvantage: negative beam weights
47. ITERATIVE METHOD
• Beamlet weights iteratively adjusted to minimize cost function (quantitative representation of
deviation from the desired goal)
• Cn is the cost at the nth iteration
• 𝐷0(r) is the desired dose at some point r
• 𝐷𝑛(r) is the computed dose at the same point
• W (r) is the weight factor in terms of contribution to the cost from different structures
• Sum is taken over a large N number of dose points
48. ITERATIVE METHOD
• For targets:
• cost is root mean squared difference between the desired (prescribed) dose and the realized dose.
• For the designated critical normal structures:
• cost is the root mean squared difference between zero dose (or an acceptable low dose value) and the
realized dose
• Overall cost is sum of costs for targets and normal structures, based on their respective weights
• Optimization algorithm attempts to minimize overall cost at each iteration until desired goal achieved
49. OPTIMIZATION ALGORITHMS
• Deterministic method (Gradient Search)
• Same setup and initial conditions gives same solution always
• Fast in reaching optimum
• Stochastic method (Computer Simulated Annealing)
• Slow in reaching optimum but can provide the best optimum solution
50.
51.
52. LEAF SEQUENCING
• A set of leaf position and corresponding monitor units to be delivered by the treatment machine
• Initial step involves generation of ideal fluence pattern that satisfy the optimum solution for the
objective function
• Then in second step; that is converted to deliverable leaf sequences
• Degradation in the quality of plan may occur due to head scatter and leakage through MLC leaves
• Discrepancy between absorbed-dose distribution obtained after deliverable leaf segmentation and
that suggested by the optimized plan is called convergence error.
53. • Correction based:
• Pencil beam
• Model based:
• Convolution superimposition
• Monte Carlo
DOSE CALCULATION ALGORITHMS
54.
55. CORRECTION BASED ALGORITHMS
• Based on measured data obtained in cubic water phantom.
• Corrections applied
• Attenuation correction for contour irregularity
• Scatter correction
• Geometric correction (inverse square law)
• Attenuation corrections for beam intensity modifiers
• Attenuation corrections for tissue heterogeneities
56. PENCIL BEAM ALGORITHM
• The dose at a point resulting from each individual beamlet is calculated using the product of
• MUs delivered
• Inverse square law
• Tissue-Maximum Ratio (TMR)
• Output factor
• Transmission factor
• Off-axis ratio
• Total dose at a point is summation of doses contributed by individual pencil beams
• Limited for 3-D (lung-tissue) interfaces where electronic equilibrium is not fully established
57. MODEL BASED ALGORITHMS
• Computes dose distribution with a physical model that simulates the actual radiation transport
• Ability to model primary photon energy fluence incident at a point and the distribution of energy
subsequent to primary photon interaction
• Able to simulate the transport of scattered photons and electrons away from the interaction site
58. CONVOLUTION SUPERIMPOSITION
• Separately considers transport of primary photons and that of scatter photon and electron emerging
from primary photon interaction
59. MONTE CARLO
• Computer program (MC code) that simulates the transport of millions of photons and particles
through the matter.
• It uses fundamental laws of physics to determine the probability of distribution of interactions.
• Larger the number of particles to be simulated (Histories) greater is the accuracy of prediction.
• But this increases computational time. Hence only a small randomly selected sample is simulated.
• This predicts the average behavior of all particles in the beam
60. PLAN EVALUATION
• Isodose lines
• Color wash
• Dose-volume histograms (DVH)
• Cumulative DVH
• Differential DVH
• Dose distribution statistics
61. CONFORMITY INDEX
• Degree to which high-dose region conforms to target volume, usually the
PTV
• CI = TV/PTV
62. HOMOGENEITY INDEX
• Uniformity of the absorbed-dose distribution within target volume
• Perfectly homogeneous dose to PTV:
• spike (a delta function) in the differential DVH or
• vertical drop of the cumulative DVH line for the PTV at that absorbed dose
• Differential DVH for a PTV for reasonable treatment plan has a near Gaussian shape
tightly distributed around the mean absorbed dose
63.
64. DOSE STATISTICS
• 𝐷𝑚𝑒𝑎𝑛, 𝐷2% (near-maximum), 𝐷98% (near-minimum) for PTV and CTV
• 𝐷𝑚𝑎𝑥 for serial OARs
• 𝐷𝑚𝑒𝑎𝑛, 𝑉𝐷 for parallel OARs
• 𝐷𝑚𝑎𝑥, 𝐷𝑚𝑒𝑎𝑛, 𝑉𝐷 for serial-parallel organs
65. DVH - QUANTITATIVE
• Summarizes entire 3D dose distribution into a graphical 2D format for each anatomic structure of interest
• Provides quantitative information (how much dose is absorbed in how much volume)
• Volume under consideration is divided into a 3-D grid of volume elements (voxels)
• Volume's dose distribution is divided into dose bins, and voxels are grouped according to dose bin without
regard to anatomic location.
66.
67. Cumulative / integral DVH Differential / Direct DVH
Plot of volume of a given structure receiving a certain
dose or higher as a function of dose.
is a plot of volume of a given structure receiving a dose
within a specified dose interval (or dose bin) as a
function of dose.
Any point on cumulative DVH curve shows the volume
that receives the indicated dose or higher.
The differential form of DVH shows extent of dose
variation within a given structure.
For instance, 71 voxels (or 350 cm3 or 71% of organ)
received 2 Gy or more.
For instance, 12 voxels (or 60 cm
3
or 12% of organ)
received 2 Gy or more but less than 3 Gy.
Each bin represents volume, or percentage of volume (y
axis), that receives a dose equal to or greater than an
indicated dose (x axis).
A plot of number of voxels in each bin (x axis) vs bin
dose range (y axis) is a differential DVH.
size of dose bin determines the height of each bin of
differential DVH.
70. QUALITY ASSURANCE
• IMRT can produce steep gradients between the target and OARs
• Goal is to ensure that the system computes absorbed-dose distributions in patients as
accurately as possible.
• IMRT results in a 2- to 10-fold increase in number of delivered monitor units compared
with conventional therapy
• Leakage through leaf ends and between leaves is extremely important
73. PATIENT SPECIFIC QA
• If the intensity modulated field boundary matches the planning boundary
• Machine instructions driving the leaves produce the planned absorbed-dose distribution
• Compare absorbed-dose distribution in phantom with the treatment planning computer for
the same irradiation condition
• Compare planned leaf motions with that recorded on MLC log files
• Confirm initial and final positions of MLC for each field by a record-and-verify system
• In vivo absorbed-dose measurements.
74. TREATMENT DELIVERY
• Patient alignment verification using portal imaging is necessary
• More important as margins used are smaller
• Shape of each aperture is defined by terminal positions of the leading leaf
tips and the starting positions of the trailing leaf tips
• Treatments may be delivered remotely or automatically under computer
control
75. GAMMA INDEX
• Dose Difference
• directly compares measured dose at each point to corresponding calculated dose at
that point
• works very well in areas of uniform dose but can produce high failure rates within
dose gradients due to errors of alignment between the measured and calculated
dose.
• Distance-to-Agreement
• measure distance between a point on measured dose distribution and nearest point
on calculated distribution with equal dose
• work well in evaluating regions of high dose gradient but becomes over-sensitive in
regions homogeneous dose.
76. GAMMA INDEX
• Combine dose difference and DTA into a single metric that is useable in both areas of dose gradient
and homogeneous dose
• Defines a 2-dimensional space with one axis being dose difference and the other being distance-to-
agreement
• Γ is computed as the Euclidean distance, normalized to passing criteria, for each point
• The minimum value of Γ for a point is taken to be γ and is used for evaluation.
78. GAMMA INDEX
• Dominated by dose difference in homogeneous regions
• Dominated by distance-to-agreement in heterogeneous regions
• Commonly evaluated using:
• Distance criteria of 2-3mm
• Dose difference criteria of 2-5% (3% is most common)
• Pass rate criteria of 90% or 95%
• Threshold dose of 10%
• Tolerance Limits: γ pass rates should be ≥95% using 3%/2mm tolerance and 10% threshold.
• Action Limits: γ pass rates should be ≥90% using 3%/2mm tolerance and 10% threshold.