1. The Use and Characteristics of Multileaf Collimators in
Radiation Therapy
Introduction and Purpose
According to Takahashi's article1
"Conformation radiotherapy: rotation
techniques as applied to radiography and
radiotherapy of cancer", conventional
collimator where used until the inception
of the multileaf collimator in 1965. These
conventional jaws could only shape
square and rectangular fields.
Another beam-shaping option, at the
time was Cerrobend blocks. They offer
more conformal shapes than conventional
collimator jaws. Jeraj Matjaz and Robar Vlado state that2
"typical MLSs have 40 to
120 leaves, arranged in pairs. By moving and controlling a large number of
narrow, closely abutting individual leaves, one can generate almost any desired
field shape. The advantages of MLCs are simple and less time consuming
preparation, use without needing to enter the treatment room, and simple
change or correction of field shape. The therapy expenses are lower because
individual shielding blocks are not needed, thus eliminating the need to handle
the Wood’s alloy, which is toxic. With MLC, we shorten the therapy time, and thus
also the period during which patient must remain in still position. Other
advantages are constant control and continuous adjusting of the field shape
during irradiation in advanced conformal radiotherapy. MLC has also some
disadvantages, which include a stepping edge effect, radiation leakage
between leaves, wider penumbra, and problems with generating some complex
field shapes."
Figure 1:MLCs from www.bertiehiggins.com

2. MLC Characteristics
There are three main types of MLC configurations: upper jaw replacement, lower
jaw replacement, or tertiary collimation. Upper jaw replacement is commonly
used by Elekta. Here the leaves are placed close to the source of radiation. This
can be an advantage because there is less motion of the leaves required to move
across the collimated field. This means that you can have a shorter leaf length and
thus a smaller treatment head. The disadvantage of this setup is that the leaf
width must be smaller and tolerances must be higher. Siemens uses the lower
jaw replacement setup. With this type of MLC configuration, both leaf ends and
sides match the beam divergence. The leaves can extend 10cm across the field
centerline. With third level configurations, the MLCs are just below the level of
the conventional upper jaws. Varian uses this configuration. It has the advantage
of limiting downtime in case there is an MLC malfunction becaue you can move
the leaves manually out of the field if failure occurs.2
According to Matjaz et al, "Multileaf collimators that are double focused
(Siemens design) have flat leaf ends that follow the beam divergence. The leaf
ends of Elektra and Varian MLC design are rounded. The material of choice for
leaf construction is tungsten alloy because it has one of the highest width is larger
than the penumbra generated by a focused or divergent edge. Second, the
penumbra width might change as a function of the distance of the leaf end from
the field midline. The leaf position must be detected in real-time to achieve a safe
and reliable position control. Linear encoders and video optical systems are most
commonly used for detection." Linear encoders are used for detection of leaf
positions in MLC systems where "high precision potentiometers are commonly
used. These potentiometers can detect positions of any individual leaf in the
system. For safer work two potentiometers with correlated readings are used in
this system."2

3. Acceptance Testing
According to the AAPM Task Group 50 report3
, "MLCs should function according
to manufacturer specifications. Acceptance testing provides the opportunity for
the user to become familiar with the MLC and to confirm that it does in fact meet
the stated criteria for acceptance. These tests do not guarantee long-term
accuracy and reliability. As with other equipment, frequent QA testing should be
performed initially, and as confidence builds, the frequency may be relaxed to
balance effort with anticipated need." One acceptance test to perform, is
checking the mechanical axes alignment. Axes that should be included are the
following: gantry, collimator, couch rotation, and jaw and leaf symmetry with the
collimator axis. Optical axes alignment should also be checked by doing light and
radiation coincidence tests. This compares fields with opposing collimator angles.
" This test will also detect flat collimator faces that are out of focus with the
source. Collimator and gantry spoke shots are also useful and should be
registered to the mechanical isocenter. Any misalignment is generally more
serious for collimators which are closer to the source due to geometric
magnification. Therefore, focused MLCs that replace the conventional jaws
require the most careful alignments, while MLCs with rounded leaf faces which
are located below the jaws are usually within tolerances met by the jaws.
Accordingly, these parameters should be tested for the following situations: (1)
jaws or backup diaphragms alone and (2) selected leaf ends and sides from
selected locations within the leaf banks, across the full range of motion, at 0°, 90°,
180°, and 270° gantry angles."3
During acceptance testing MLC performance will also need to be checked. Boyer
et al states in the TG 50 report that "The width of the x-ray attenuation of a leaf at
isocenter is sensitive to the source-to-MLC distance, and it should be verified
during acceptance. The errors in leaf position can be compensated for using
software corrections at the time the apertures are configured. However, for the
sake of uniformity among machines, this condition should be corrected during
installation of the MLC." Leaf position should also be calibrated according to the
vendor technique. Varian uses an optical beam projected over the leaf ends and
then extends them one at a time until the beam is completely blocked. Shaft
encoders help to determine the leaf positions.3

4. Figure 2: Shows Alignment of 5cm Strips Formed By MLCs (Image from TG-50 Report)
Acceptance tests for MLCs also include: leaf travel, leaf speed, transmission, and
leakage between leaf faces in the closed position. The leaf speed test is to verify
the maximum speed of the leaves as well as a smooth motion. The leaf travel test
checks to see if the leaves can reach their maximum range. The leaf transmission
test verifies the interleaf transmission and transmission beneath the leaves and
jaws combined. Leakage between the leaf faces in the closed position also needs
to be checked, especially when rounded leaves are used. The last check in
acceptance testing of MLCs involves using field-shaping software to make an
irregularly shaped field to test be for using the MLCs clinically.3
Commissioning of MLCs
During commissioning, interleaf transmissions should be less than 2%. Although,
average transmissions can be used instead, if needed. Central axis profiles should
also be obtained as TPRs or TMRs and PDDs. Penumbra data should be added to
the treatment planning system (TPS) if it has that capability. Flat and rounded
leaf edges have different penumbra to account for. Profiles for symmetric and
asymmetric fields should be acquired to check off-axis ratio.
MLC Quality Assurance
According to Klein et al and the task group 142 report4
, a matched segment test
should be performed weekly. It is sometimes called the picket fence test. Some of
the leaf parameters that affect dose delivery for IMRT include leaf positional
accuracy and transmission values. Simple tests, such as the picket fence test "can
assess positional accuracy qualitatively by the matching of sequential segments
and leaf transmission, particularly interleaf. We recommend the picket fence test
be performed weekly with a careful examination of the image acquired by static
film or on-line portal image."4

5. Monthly test include: the travel speed test, backup diaphragm settings, setting
vs radiation field for 2 patterns, and leaf position accuracy. The travel speed test
is to account for gantry rotation which may affect leaf motion due to gravitational
effects imposed on the leaf carriage system. Loss of travel speed can result in
increased beam holds or gap width errors. MLC travel speed is evaluated with
vendor software or by MLC log file evaluation.4
The backup diaphragm setting test
is only done for Elekta machines and must be within 2mm. The setting versus
radiation field test is done on non-IMRT machines and has a tolerance of 2mm.
Leaf position accuracy is a test done on IMRT machines and has a tolerance of
1mm for leaf positions of an IMRT field for four cardinal gantry angles.4
Figure 3: From TG 142
Annual tests include the following: MLC transmission test, leaf position
repeatability test, MLC spoke shot, coincidence of light field and x-ray field test,
segmental IMRT test, and moving window IMRT test. The tolerance for these tests
are listed in figure 3, from the table in the task group 142 report.

6. Component Replacement
After an MLC is replaced it is important to ensure accurate leaf positioning by
the calibration of leaf positions. According to Klein et al,4
"through the calibration,
the measured signals, such as voltages from the potentiometers or pixel numbers
from a solid-state camera, and the actual leaf positions establish a one-to-one
relationship. Periodic checking and recalibration are also needed to ensure the
integrity of the controlling system. The Varian MLC calibrates the leaf positions
using narrow infrared beams built into the collimator assembly that transect the
paths of the leaves.
The calibration procedure is carried out automatically each time the MLC
operating system is initialized. Each leaf is driven through its range of travel. As a
given leaf intersects the infrared beam, the values returned by its position
encoders are acquired. These values are used along with equation to calibrate the
leaf position. The calibration values are saved in a table for use by the control
system. In the Philips MLC system, which uses a video optical controlling
mechanism, four reference reflectors are fixed on the head structure. The
positions of the four reflectors establish a fixed frame of reference, which
requires film exposures of regular fields with different field sizes set by a set of
default calibration values. The actual field sizes measured from the films set the
final calibration. During operation, the positions of the four reference reflectors
are acquired and checked every 0.1 sec. The installation of an MLC on existing
equipment should be accompanied by measurements sufficient to realign the
original equipment, if necessary."
Radiation Safety Concerns
Boyer et al3
states that the " assessment of safety with accelerators and
associated devices is tested only minimally in a manufacturer’s acceptance
procedures. Additional safety tests are warranted because of the increased
complexity of an MLC. The use of multiple, conformed MLC fields in either static
or dynamic modes will render the conventional use of visual inspection as a daily
verification of field shapes impractical or impossible. Active interlock checks
should be carried out for leaf and jaw positional tolerances. These measurements
should include assessment of software interlocks, hardware interlocks, and other
possible independent systems. Non-active interlocks designed to prevent

7. unauthorized motions should be tested. These would include procedures such as
dynamic imaging of field shape, motion enable power line interrupt, etc.
Communication link interlocks are provided to ensure that the heavy data traffic
that flows between the control computers and the accelerator hardware is not
corrupted. Means of intentionally corrupting the data should be carefully
discussed with the manufacturer. Tests should be devised to demonstrate that
the interlocks are functioning to detect true positive data errors. Interlock checks
to ensure the software will not allow a trailing edge of a leaf to be unshielded by
the jaws must be performed."
When it come to radiation safety, it is important to remember events like the
one that took place in March of 2005, at St. Vincent hospital in New York, in which
the MLCs were not even in place during the treatment of Scott Jerome-Parks. This
overtreatment left him, according to Walt Bogdanich of the New York Times,5
"deaf, struggling to see, unable to swallow, burned, with his teeth falling out, with
ulcers in his mouth and throat, nauseated, in severe pain and finally unable to
breathe".
Conclusion
Multileaf collimators are a great leap forward in radiation oncology from the
limitations of jaws and Cerrobend blocks. With MLCs, we have to ability to further
increase the dose to cancer while limiting the dose to normal structures through
IMRT. In the future, I see greater numbers of smaller MLC leaves. This will help
shape the dose further and offer more conformity to the tumor.

8. References
1. Takahashi, S., "Conformation radiotherapy: rotation techniques as applied
to radiography and radiotherapy of cancer", Acta Radiol. Suppl. 242 (1965)
1–142
2. Matjaz J, Vlado R. Multileaf Collimator in Radiotherapy. Radiol Oncol 2004;
38(3): 235-40.
3. Boyer A, Biggs P, Galvin J, et al. Basic Applications of Multileaf Collimators:
Report of Task Group No. 50 Radiation Therapy Committee. American
Association of Physicists in Medicine 2001.
4. Klein E, Hanley J, Bayouth J, et al. Quality Assurance of Medical
Accelerators: Task Group 142 Report. Med. Phys. 36(9), 2009.
5. Bogdanich W, Radiation Offers New Cures, and Ways to Do Harm. New York
Times. January 23, 2010.