Guidelines for Monaco

VMAT/IMRT Optimization.

Douglas Prah, Ph.D., DABR.,

Assistant Professor of Radiation Oncology,
Director of Medical Dosimetry
Medical College of Wisconsin.

Note these guidelines represent an initial introduction to Monaco VMAT/IMRT planning optimization.
These guidelines are not comprehensive and do not replace Elekta training and guidelines for treatment
planning. These guidelines do not replace the expertise of a qualified medical physicist. A qualified
medical physicist is responsible for understanding, reviewing, and approving radiation treatment plans.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

1 TABLE OF CONTENTS:
1 Table of Contents: ................................................................................................................................. 2
2 Overview. .............................................................................................................................................. 4
3 Calculation Properties. .......................................................................................................................... 4
3.1 Grid Settings. ................................................................................................................................. 4
3.2 Algorithm Settings......................................................................................................................... 4
3.3 Sequence Parameters for Elekta Infinity/Versa HD with Agility MLC. .......................................... 5
4 Structures. ............................................................................................................................................. 5
4.1 Forcing Electron Densities in Monaco using the CCC algorithm. .................................................. 5
4.2 Forcing Electron Densities in Monaco using the pMC algorithm.................................................. 6
4.3 Gastrointestinal Contrast. ............................................................................................................. 6
4.4 Head & Neck Structures. ............................................................................................................... 7
4.5 Thoracic Structures. ...................................................................................................................... 8
4.6 Pelvic Structures............................................................................................................................ 9
5 Beams. ................................................................................................................................................. 10
5.1 General. ....................................................................................................................................... 10
5.2 Geometry. ................................................................................................................................... 10
5.2.1 Arc Settings (Dir/Gantry Start/Arc). .................................................................................... 10
5.2.2 Increment Settings (Inc). ..................................................................................................... 10
5.2.3 Collimator............................................................................................................................ 11
5.2.4 Treatment Aids.................................................................................................................... 11
6 DVH statistics. ..................................................................................................................................... 11
7 IMRT Constraints................................................................................................................................. 11
7.1 Targets. ....................................................................................................................................... 11
7.2 External Structures...................................................................................................................... 12
7.3 OARs. ........................................................................................................................................... 12
7.3.1 Mean Dose. ......................................................................................................................... 12
7.3.2 Dose Volume Constraints.................................................................................................... 13
7.3.2.1 Single Dose Volume Constraints. .................................................................................... 13
7.3.2.2 Multiple Dose Volume Constraints. ................................................................................ 13
7.3.2.3 Maximum Dose. .............................................................................................................. 13
7.4 Constrained vs Pareto Optimization Methods............................................................................ 14

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

 7.4.1 Pareto Optimization: ........................................................................................................... 14
7.4.2 Constrained Optimization: .................................................................................................. 14
7.5 IMRT Optimization Strategy. ....................................................................................................... 15
7.5.1 Stage 1: Generate the Ideal Plan......................................................................................... 15
Step 1. Run Stage 1. ........................................................................................................................ 15
Step 2: Plan Evaluation. .................................................................................................................. 15
Step 3: Review Plan Settings. .......................................................................................................... 16
Decrease the beam increment.................................................................................................... 16
Balance Dose Gradient and Max Dose. ....................................................................................... 16
Adjust Beams. ............................................................................................................................. 16
Step 4: Adjust target and OAR constraints. .................................................................................... 16
7.5.2 Stage 2: Generate the Deliverable Plan. ............................................................................. 16
Step 1: Run Stage 2. ........................................................................................................................ 16
Step 2. Disable the 105% target quadratic overdose function. ...................................................... 16
Step 3: Plan Evaluation. .................................................................................................................. 16
Step 4. Review Plan Settings. .......................................................................................................... 17
Step 5: Adjust the optimization methods. ...................................................................................... 17

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

2 OVERVIEW.
When creating a VMAT/IMRT plan it is important that all plan settings be evaluated and adjusted prior
to the initiation of IMRT optimizations. Plan quality is directly affected by choosing the proper treatment
plan setting prior to IMRT optimization. Ideally the following list should be systematically reviewed for
each plan. Under each section site specific tips are given to help improve plan quality.

1. Calculation Properties
2. Sequence Parameters.
3. Structure Settings.
4. Prescription.
5. Beam properties. (General/Geometry/Treatment Aids)
6. IMRT Constraints.

3 CALCULATION PROPERTIES.

3.1 GRID SETTINGS.

Acceptable Requirements:

pMC: VMAT/IMRT: 0.30 cm

pMC: SBRT: 0.20 cm
CCC: 3D Photon: 0.25 cm
eMC 3D Electron: 0.30 cm

Exception: For larger treatment volumes, Monaco may require more system memory than is available
on the workstation. In this case, increase the grid size during optimization incrementally (0.4, 0.5, 0.6
cm) until optimization occurs without exceeding the memory limitation of the workstation.

Exception: For larger SBRT lung volumes, Monaco optimization may be prohibitively long (> 5 hrs) at the
pMC: SBRT: 0.2 cm. In this case, it may be best to optimize using the pMC: VMAT/IMRT: 0.3 cm grid
spacing. However, the final dose calculation should be re-calculated at pMC: SBRT: 0.2 cm.

Note: For eMC 3D Electron plans, a smaller grid spacing (<0.3 cm) may result in a noisy dose
distribution.

3.2 ALGORITHM SETTINGS.

Acceptable Requirements:

pMC: VMAT/IMRT: Statistical Uncertainty < 1% per plan.

EMC: 3D Electron: 1,000,000 control points / cm2

Preferred: For a given plan, the overall statistical uncertainty needs to be less than 1% per plan. This can
be found in the optimization console. A plan quality of 3 to 5% per control point produces a plan that
has a smoother dose distribution, i.e. less noise in the dose distribution.
Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022
Note: For eMC 3D Electron plans, a smaller grid spacing (<0.3 cm) may result in a noisy dose
distribution.

3.3 SEQUENCE PARAMETERS FOR ELEKTA INFINITY/VERSA HD WITH AGILITY MLC.

Acceptable Settings:

Segment Shape Optimization: ON.

High Precision Leaf Position: ON.
Segment Shape Optimization Loops: Speed vs Plan Quality [Range 1-20]: 5
Maximum number of Arcs: 2
Maximum number of control points per arc [< 1024]: 180
Minimum Segment Width: 0.50
Fluence Smoothing: Medium

Preferred: When time permits set Segment Shape Optimization Loops to 20. This setting takes
significantly more time but produces higher quality VMAT/IMRT plans.

Increase State 1 plan quality: If a plan is not meeting the written planning directive in Stage 1, it will
never meet in Stage 2. Increasing the maximum number of arcs will increase the plan quality during
Stage 1.

Increase State 2 plan quality: For VMAT/IMRT plans that deteriorate from Stage 1 to Stage 2, increase
the maximum number of control points per arc; increase to as high as 512. For a higher maximum
number of control points per arc setting, be sure to decrease the increment on each beam to maximize
the gain in overall plan quality [5 to 10 degrees].

Increase State 2 plan quality: For VMAT/IMRT plans that deteriorate from Stage 1 to Stage 2,
incremental decrease the fluence smoothing until the plan quality becomes acceptable. Often for H&N
plan fluence smoothing needs to be turned off.

4 STRUCTURES.
After forcing relative electron densities, review all changes using the “synthetic CT” option available
under the “Workspace” ribbon.

4.1 FORCING ELECTRON DENSITIES IN MONACO USING THE CCC ALGORITHM.

The Monaco CCC algorithm can only convert Hounsfield Units ranging from air (rED = 0.01) to aluminum
(rED=2.456). Within Monaco, if the CT image contains HUs that exceed the calculated 2.456, the
electron density of those voxels will be force to stainless steel (rED = 6.70). Refer to TG65 “Tissue
Inhomogeneity Corrections for MV Photon Beams”.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

4.2 FORCING ELECTRON DENSITIES IN MONACO USING THE PMC ALGORITHM.
Unlike the CCC algorithm, the pMC algorithm can assign any physical rED to a structure. However, if the
material rED for a structure cannot be accurately identified, it is best to rely on the CTs estimate of the
rED of the filling.

Preferred: Obtain a CT using a metal artifact reduction algorithm.

4.3 GASTROINTESTINAL CONTRAST.

• For CCC & pMC algorithms, the guidelines for forcing relative electron densities (rED) of GI
structures:
o Within the treatment planning system, GI contrast should be forced to a relative
electron density of water, 1.000.
o Within the treatment planning system, the GI structures SHOULD NOT be forced to an
artificial relative electron density.
o Within the treatment planning system, GI air should not be forced an artificial relative
electron density, for MR Linac only.
• During online IGRT image review immediately prior to daily treatments, the physicists should
pay close attention to large changes in the location and amount of air within the GI structures.
• During the weekly physics chart checks offline IGRT image review, the physicists should pay
close attention to large changes in the location and amount of air within the GI structures. If
large changes exist dosimetry and physics should assess the dosimetric impact of the change
and present it to the physician for review.
• During daily treatments, the therapist should pay close attention to large changes in the
location and amount of air within the GI structures. If large changes exist dosimetry and physics
should be notified to assess the dosimetric impact of the change.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

Example:

CT/SS from the initial CT simulation before electron density forcing:

CT/SS after only forcing the GI contrasts not the GI air to a relative electron:

4.4 HEAD & NECK STRUCTURES.

Tooth Fillings. In general, it is difficult to obtain accurate dental records resulting in uncertainty of
materials of the dental fillings. Given this uncertainty it is best to rely on the CTs estimate of the rED of
the filling.

• For CCC algorithm, Monaco will automatically set those voxels that exceed 2.456 to a relative
electron density of 6.70. It is recommended that only the artifacts from tooth fillings be forced.
The filling may be comprising of many different materials and should not be forced. Note that in
Monaco, the rED cannot manually be forced to 6.70. Never place a calculation point in the path
of a beam passing through the teeth. Below is a list of common dental implants materials.
• For pMC algorithm, Monaco can assign any physical rED to a structure. Ideally, obtain CT using a
metal artifact reduction algorithm. If no dental records can be obtained, it is recommended that
only the artifacts from tooth fillings be forced. The filling may be comprising of many different
materials and should not be forced, unless there are accurate and reliable dental records that
can be used to assign a rED.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

 ρ ρe ρ e(ma teri a l ) /
Ma teri a l Z eff Ma i n El ements
(g/cc) (el ectrons /cc) ρ e(wa ter)

H20 7.50 H, O 1.00 3.343E+23 1.00

Teeth 14.70 Ca, O, C, P, H, N, M, Na, F, Zn, Cl, K 2.20 6.420E+23 1.92
Ceramco 12.08 O, Si, Al, Sn, Na, K 2.60 6.61E+23 1.98
Ti 21.40 Ti 4.54 1.256E+24 3.76
Ti Alloy 21.00 Ti, Al, V 4.34 1.200E+24 3.59
Amalgam1 45.83 Ag, Cu, Sn, Zn 8.00 2.09E+24 6.25
Ni-Cr Alloy 28.14 Ni, Cr, Mo, Al, Be, Ti 7.90 2.21E+24 6.61
Amalgam2 67.40 Hg, Ag, Cu, Sn, Zn 12.00 3.039E+24 9.09

Reference: Çatli S. High-density dental implants and radiotherapy planning: evaluation of effects on dose distribution using pencil beam
convolution algorithm and Monte Carlo method. J Appl Clin Med Phys. 2015;16(5):46–52. Published 2015 Sep 8. doi:10.1120/jacmp.v16i5.5612.

4.5 THORACIC STRUCTURES.

Breast expanders.

• For CCC algorithm, it is recommended that only the artifacts from breast expanders ports be
forced. The expander is typically comprised of Niobium metal (RED = 6.813), the largest
component, and a Titanium alloy (rED = 3.60) and should not be forced. Monaco will
automatically set those voxels that exceed 2.456 to a relative electron density of 6.70 (I.e.
stainless steel rED). Note that in Monaco, the rED cannot manually be forced to 6.70. Never
place a calculation point in the path of a beam passing through the breast expander port. Below
is a list of common dental implants materials.
• For pMC algorithm, the rED should be forces to the corresponding value; any physical rED can be
assigned to a structure. If no implant records can be obtained, it is recommended that only the
artifacts be forced and CT be obtained using a metal artifact reduction algorithm. Unless there
are accurate and reliable records that can be used to assign a rED, the breast expander may be
comprising of many different materials and should not be forced.

ρ ρe ρ e(ma teri a l ) /
Ma teri a l Z eff Ma i n El ements
(g/cc) (el ectrons /cc) ρ e(wa ter)

H20 7.50 H, O 1.00 3.343E+23 1.00

Plastic Expander Empirically Derived 1.60
Ti Alloy 21.40 Ti, Al, V, Fe, C, O, N, H 4.54 1.205E+24 3.60
Niobium 41.00 Nb 8.74 2.28E+24 6.81

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

4.6 PELVIC STRUCTURES.
High-Z implanted hip prostheses.

Caution: When possible, treatment beams should avoid the hip implant, exit dose is acceptable.

• For CCC algorithm, it is recommended that only the artifacts from hip implant be forced. The hip
implant is typically comprised of CoCrMo alloy (RED =6.74), Stainless Steel alloy (8.83) or
titanium alloy (RED = 3.6 – 3.73) and should not be forced. When possible, treatment beams
should avoid the hip implant, exit dose is acceptable. Monaco will automatically set those voxels
that exceed 2.456 to a relative electron density of 6.70. Note that in Monaco, the rED cannot
manually be forced to 6.70. Never place a calculation point in the path of a beam passing
through the hip implant. Below is a list of common dental implants materials.
• For pMC algorithm, the rED should be forces to the corresponding value. If the material cannot
be identified with record use the CT to estimate the material and force the entire rED of the
structure to the corresponding material tyle. Note that due to the volume of the implant the
artifact will be significant.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

 ρ ρe ρ e(ma teri a l ) /
Ma teri a l Z eff Ma i n El ements
(g/cc) (el ectrons /cc) ρ e(wa ter)

H20 7.50 H, O 1.00 3.343E+23 1.00

Stainless Steel 26.70 Fe, Cr, Ni, Mo, Mn, Si, C, P, S 8.10 2.300E+24 6.88
Co-Cr-Mo 27.60 Co, Cr, Mo, Ni, Mn, Si, Fe, C 7.90 2.255E+24 6.75
Ti Alloy 21.40 Ti, Al, V, Fe, C, O, N, H 4.54 1.205E+24 3.60
Reference: Mark B. Hazuka, Geoffrey S. Ibbott, Jeannie J. Kinzie, Hip prostheses during pelvic irradiation: Effects and corrections, International
Journal of Radiation Oncology*Biology*Physics, Volume 14, Issue 6, 1988, Pages 1311-1317, ISSN 0360-3016, https://doi.org/10.1016/0360-
3016(88)90412-9.

5 BEAMS.

5.1 GENERAL.
Most treatment plans will achieve the written planning directive using two beams (collimate 20 and 340
degrees) using two arcs per beam.

5.2 GEOMETRY.
5.2.1 Arc Settings (Dir/Gantry Start/Arc).
Unless the contralateral structures need to be completely avoided, it is best to use complete arcs (360
degrees) and allow the optimizer to minimize dose on the contralateral side of the patient.

Preferred: Due to the end point of the kVCB, it is best to start the collimator at CCW. This is more
efficient use of the treatment unit’s time.

5.2.2 Increment Settings (Inc).

During Stage 1 of optimization, Monaco generates ideal fluence maps using the increment specified. The
smaller the increment the more the ideal fluence maps. If a plan is struggling during Stage 1, consider
decreasing the increment. During Stage 2 of optimization the optimizer creates MLC shapes to
reproduce the ideal fluence obtained in Stage 1.

General Recommendations:

H&N: Increment = 5 to 10

Chest: Increment = 5 to 20

Pelvis: Increment = 15 to 30

Preferred: Use different increments for each beam, this gives the optimizer more ideal fluence angles to
achieve the desired dose distribution.

Increase State 1 plan quality: If a plan is not meeting the written planning directive in Stage 1, it will
never meet in Stage 2. Decreasing the increment will increase the plan quality during Stage 1.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

5.2.3 Collimator.
Preferred: For VMAT plans, to reduce intra-leaf leakage high dose bands: it is preferred to start with at
least two treatment beams, using two different collimator settings: 20 and 340 degrees.

Note: If each beam is a complete arc, then each additional beam needs a different collimator angle.

5.2.4 Treatment Aids.

Treatment aids including couch and bolus should be assigned.

6 DVH STATISTICS.
The written planning directive should always be entered into Monaco. This makes evaluation of the plan
in real-time possible.

Preferred: Instead of using a “Maximum Dose” criterion, use a “Maximum Absolute Volume that
Receives Dose” using 0.03 cc at the maximum dose. Since pMC is statistical in nature, recalculating at a
lower statistical uncertainty (even less than 1% per plan, e.g. 1% per control point) reduces doses to
these structures. Most clinical trials evaluate maximum point dose at volumes > 0.03 cc.

7 IMRT CONSTRAINTS.

7.1 TARGETS.
Preferred: Use two target penalties for each target. If the structure overlaps with another target, use a
0.0 cm shrink margin to increase target coverage.

Note: The first target penalty should always be the 95% coverage of the lowest coverage constraint.
Note that, if a 98% coverage constraint is used in the first target penalty constraint, the system will
struggle to provide target coverage for more complex plans.

For Stage 1 or Stage2: Start by prioritizing the highest dose targets first. However, if you are struggling
with coverage on some of the other targets, and the highest dose level target is relatively small, it is best
practice to place the largest structure as the first structures in the list. After the optimizer makes
changes to the MLC shapes it will rescale the dose to prioritize coverage for the first target.

Surface Margin: Surface margin decreases the target volume off the skin surface by a user defined
amount, typically 3mm. Even if PTV_eval structures are used, enabling surface margin will improve the
plan quality. This is defined for each target penalty.

Auto-Flash: Auto-flash will extend MLC beyond the skin surface by the user defined amount. If patient
swelling occurs this likely will reduce the need to create a new VMAT to accommodate the anatomical
changes. This is typically used for breast VMAT plans. This is defined for each structure. It is located
under properties.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

Recommended Constraints for each Target Structure:

• Target penalty: Minimum Volume: 95%; Prescription: 100% RadRx

• Target Penalty: Minimum Volume: 98%; Prescription: 98% RadRx
• Quadradic Overdose: 105% RadRx; RMS Dose Excess: 2 cGy. (Typically, only used during Stage 1)
• Quadradic Overdose: 110% RadRx; RMS Dose Excess: 2 cGy.

7.2 EXTERNAL STRUCTURES.

Instead of using rings, Monaco allows the control of the dose gradient by using multiple quadratic
overdose constraints at different doses and shrink margins. Ideally the external structure should be
located directly below the targets.

Recommended Starting Constraints for each Target Structure:

• Quadratic Overdose: 110% RadRx; RMS Dose Excess: 2 cGy. All Shrink Margins off. (This controls
the overall maximum dose in the entire plan).
• Quadratic Overdose: 100% RadRx for each target; RMS Dose Excess: 2 cGy. Use a shrink margin
of 0.0 cm the target of interest. For targets plan in the presence of higher dose targets, add an
appropriate shrink margin.
• Quadratic Overdose: 75% RadRx for each target; RMS Dose Excess: 10 to 20 cGy. Use a shrink
margin of 1.0 cm the target of interest. For targets plan in the presence of higher dose targets,
add an appropriate shrink margin. (If the optimizer struggles it might be necessary to increase
the RMS).
• Quadratic Overdose: 50% RadRx for each target; RMS Dose Excess: 50 to 100 cGy. Use a shrink
margin of 2.0 cm the target of interest. For targets plan in the presence of higher dose targets,
add an appropriate shrink margin. (If the optimizer struggles it might be necessary to increase
the RMS).

Exception: For breast VMAT plans, only use a quadratic overdose constraint to control the overall global
max point dose. Do not constrain the dose gradient with additional quadratic overdose constraints as
this will increase the dose to the lungs and heart. Dose spill is not important for breast plans, refer to 3D
breast plans.

7.3 OARS.
7.3.1 Mean Dose.
The best constraint to control mean dose for an OAR is parallel. If there is not much overlap with the
target no shrink margins are needed. If a shrink margin is used the mean organ damage needs to be
reduced to achieve the mean dose constraint.

Recommended Starting Constraints for each OAR Mean Dose Constraints:

• Parallel: Reference Dose = Mean Dose; Mean Organ Damage = 50%; Power Law Exponent= 4.

Note: A power law exponent of 4 is easier to achieve than a lower value. If a plan is easily achieved it is
best practice to rerun with a lower power law, e.g. Power Law = 1. A power law exponent of 4 impacts

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

the dose distribution centered at the reference dose. As the power law exponent decreases broadens
effect broadens.

7.3.2 Dose Volume Constraints.

7.3.2.1 Single Dose Volume Constraints.

Instead of using a “DVH Overdose” constraints use a parallel constraint. When a parallel constraint is
used in combination with multi-criterial optimization dose superior organ sparing will result.

Recommended Starting Constraints: Single Parallel Constraints

• Parallel: Reference Dose = Dose_1; Mean Organ Damage = Volume_1; Power Law Exponent= 4.

Note: A power law exponent of 4 is easier to achieve than a lower value. If a plan is easily achieved it is
best practice to rerun with a lower power law, e.g. Power Law = 1. A power law exponent of 4 impacts
the dose distribution centered at the reference dose. As the power law exponent decreases broadens
effect broadens.

7.3.2.2 Multiple Dose Volume Constraints.

There are two strategies using the biological cost functions when multiple DVH constraints are used.
Serial constraints work well on the following structures: rectum, bladder, and mandible.

Recommended Starting Constraints: Strategy 1: Serial Constraints.

• Serial: Reference Dose = Dose_1; Power Law Exponent= 1-to 20.

Note: Choose a starting reference dose that corresponds to the dose you want at the 50% coverage
point. The slope of the DVH is controlled by the power law exponent. A k=1 corresponds to a mean
organ dose while a k=20 corresponds to a maximum dose. A good starting point is 5. Once the optimizer
starts running this can be changed to reflect the circumstances.

Recommended Starting Constraints: Strategy 2: Multiple Parallel Constraints

• Parallel: Reference Dose = Dose_1; Mean Organ Damage = Volume_1; Power Law Exponent= 4.
• Parallel: Reference Dose = Dose_2; Mean Organ Damage = Volume_2; Power Law Exponent= 4.
• Parallel: Reference Dose = Dose_3; Mean Organ Damage = Volume_3; Power Law Exponent= 4.

Note: A power law exponent of 4 is easier to achieve than a lower value. If a plan is easily achieved it is
best practice to rerun with a lower power law, e.g. Power Law = 1. A power law exponent of 4 impacts
the dose distribution centered at the reference dose. As the power law exponent decreases broadens
effect broadens.

7.3.2.3 Maximum Dose.

To control the maximum point dose for most structures, a quadratic overdose function is the best
function to use. Try to limit the use of the “Max Dose” constraint. The only time it is suggested to use a
maximum dose constraint is when the statistical uncertainty is set to “per control point,” for example
neural structures such as the spinal cord or cranial nerves. However, in most cases quadratic overdose
dose is adequate.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

Recommended Starting Constraints for Maximum OAR Dose:

• Quadradic Overdose: Dose: Max Dose; RMS Dose Excess: 2 cGy.

7.4 CONSTRAINED VS PARETO OPTIMIZATION METHODS.

Constrained and Pareto optimization techniques are how the optimizer automatically assigns weights to
each cost function dynamically during the optimization process. As the optimizer runs the weights will
change in real-time. They also provide insight into which constraints the system is struggling to achieve.
This gives clues on which constraints might not be achievable. The same optimization settings are not
needed for Stage 1 and Stage 2. MCO can also be turned on and off at any time. See the definitions
below for Constrained and Pareto Optimization Methods.

Constrained Optimization will benefit from the use of multi-criterial optimization (MCO). The optimizer
will look for global minimum in the weighting function instead of searching for local minimum. This
increases the total optimization time but makes significant improvements to OARs.

Helpful Hint: Sometimes MCO might push an OAR too hard in Stage 2. In this case it might be beneficial
to manually set a weight to something small, 1 or less, especially if the constraints are easily achieved.

Preferred: The preferred technique is constrained optimization with MCO utilization. If a plan cannot be
met, progressively move down the list until you can achieve a suitable plan. Note OARs dose will get
progressively worse in exchange for target coverage as you move down the list.

1. Run Stage 1 and Stage 2 using Constrained Optimization and with MCO enabled for all
structures.
2. Run Stage 1 using Constrained Optimization and Stage 2 using Constrained Optimization with
MCO disabled for all structures.
3. Run Stage 1 using Constrained Optimization and Stage 2 using Pareti Optimization with MCO
disabled for all structures.
4. Run Stage 1 & 2 using Pareto Optimization and with MCO disabled for all structures.

7.4.1 Pareto Optimization:

Pareto Optimization Overview:

• Dose limiting function will not be spared. Target coverage will be prioritized.
• Targets Penalties are set to 1; OARs weight will be limited to 10.
• Dose not work well with MCO.

7.4.2 Constrained Optimization:

Constrained Optimization Overview:

• Dose limiting cost functions will always be achieved (OARs will always be met).
• Only when these are met, the dose to the targets will be prioritized.
• Use this to your advantage and follow the clues.
• The weight, iso-effect, and relative impact on the IMRT constraints is your key to identifying
conflicts and achieving your planning objectives.
• Targets Penalties are set to 1; OARs weight will NOT be limited.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

 • Works well with MCO.

Constrained Optimization Prioritization:

• 1st Order Constraints

o Goals will always be met.
o Serial, Parallel, Quadratic Overdose, Max Dose
• 2nd Order Constraints
o Goals will be met UNLESS there is a 1st Order constraint.
o Quadratic Under Dose, Under Dose DVH
• 1st Order Objective
o Goal will be met unless a 1st or 2nd Order Constraints prevents this.
o Target EUD, Target Penalty
• 2nd Order Objective
o Goal will be met or succeeded unless Constraints prevent it and UNLESS 1st order
objectives are not met.
o Cost functions that have “Multi Criterial” option

Tip: During optimization it is difficult to determine plan progress, especially during the longer Stage 2. It
is beneficial to open the optimization console located on the “Workspace” >> “Controls” Menu. Filter for
the string “Shapes change”. This will list which SSO loops have completed.

7.5 IMRT OPTIMIZATION STRATEGY.

For most cases, minor changes are necessary during optimization. The more frequently the end user
generates a plan for a specific disease site and dose fractionation regime, the more heuristic technique
will develop. Once consistent high quality treatment plans are being generated using similar plan
settings and optimization constraints, a plan template should be saved and used for future cases. Plans
templates will save a significant amount of time during optimization. In most cases, only minor changes
to constraints need to be made to maximize plan quality.

Using multi-criterial optimization (MCO) in combination with biologic cost functions ensures that the
lowest achievable OARs doses are achieved during optimization. Plans will typically far overshoot their
goals.

7.5.1 Stage 1: Generate the Ideal Plan.

Step 1. Run Stage 1. It is important to let the optimizer run through Stage 1 completely the first time.

Step 2: Plan Evaluation. After Stage 1 has completed, evaluate the dose distribution and DVH statistics
for the plan. If certain criteria are not meeting, check the relative impact and the optimizer assigned
weight. Determine which cost functions are having a difficult time achieving. Typically, if the weight
exceeds 1 at the end of Stage 1, the optimizer was struggling achieve the plan goals. It is best practice at
this point to review constraints and make sure that no errors were made setting up the plan.
Incorporate changes and repeat Step 1.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

Step 3: Review Plan Settings. If no errors were made creating the cost function and the plan quality is
not acceptable, it is best to evaluate the plan settings again. Consider the following changes:

Decrease the beam increment. Remember often reducing the increment requires an increase in the
number of control points per arc. Increasing the number of control points per arc will not improve Stage
1 plan quality.

Balance Dose Gradient and Max Dose. Consider relaxing the dose gradient by increasing the shrink
margins on the quadratic overdose function or relaxing the RMS on low dose cost functions. Relax the
105% quadratic overdose function RMS, may be increase to 10 to 50.

Adjust Beams.

• Increase the number of arcs per beam. The default starting point is typically 2, however
increasing this number will give the optimizer different idea fluence maps to achieve the desired
plan quality. This is often required for complex hypo-fractionated plans.
• Add additional beams; remember that each common coplanar beam will need a different beam
increment (inc).
• Add couch kicks to each beam. For example, for a brain plan consider adding a vertex field, for a
lung plan consider adding a couch kick of 10 and 350 degrees to each beam (this can help with
the V20).

Incorporate changes and repeat Step 1.

Step 4: Adjust target and OAR constraints. After Step 3 has been exhausted, it might be necessary to
discuss the trade-offs of the treatment plan with the physician, the plan is over constrained. You might
be required to give up a little on certain OARs to maintain target coverage. The physician will need to
make a decision.

Incorporate changes and repeat Step 1.

7.5.2 Stage 2: Generate the Deliverable Plan.

Step 1: Run Stage 2. Run Stage 2 using Constrained Optimization and with multi-criterial optimization
enabled for all structures. Be patient sometime the plan will take the entire time to come together.

Step 2. Disable the 105% target quadratic overdose function. If the optimizer is struggling with
105% target quadratic overdose function, then disable it and repeat Stage 2: Step 1.

Step 3: Plan Evaluation. After Stage 2 has completed, evaluate the dose distribution and DVH statistics
for the plan. If certain criteria are not meeting, check the relative impact and the optimizer assigned
weights. Determine which cost functions are having a challenging time achieving. Typically, if the weight
exceeds 1 at the end of Stage 2, the optimizer was struggling achieve the plan goals. Sometime MCO
pushes certain constraints excessively hard. Consider disabling MCO for the function or entering a
manual weight of 1 or less on that cost function if the constraint is already met. Try backing off on
structures that are struggling. This may help the overall plan quality. Incorporate changes and repeat
Step 1.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022

Step 4. Review Plan Settings. If the plan cannot maintain the same plan quality that was achieved
during Stage 1, plan setting may need to be reviewed again. Go to Stage 1: Step 3.

Step 5: Adjust the optimization methods. Try progressively moving through the list of alternatives
optimization methods:

1. Run Stage 1 using Constrained Optimization and Stage 2 using Constrained Optimization with
MCO disabled for all structures.
2. Run Stage 1 using Constrained Optimization and Stage 2 using Pareto Optimization with MCO
disabled for all structures.
3. Run Stage 1 and Stage 2 using Pareto Optimization and with MCO disabled for all structures.

Incorporate changes and repeat the appropriate Stage & Step.

Prepared for Monaco version 5.51.10. Last Edited on: 8/17/2022