Purpose: This case report discusses the safe delivery of total body
irradiation (TBI) to a patient with a left ventricular assist device
(LVAD). This treatment required radiation-dose determinations and
consequential reductions for the heart, LVAD, and an external controller
connected to the LVAD.
Methods: The patient was treated using a traditional 16MV anterior
posterior (AP)/posterior anterior (PA) technique at a source to surface
distance (SSD) of 515cm for 400cGy in two fractions. A 3cm thick
Cerrobend block was placed on the spoiler to reduce dose to the heart
and LVAD to 150cGy. The external controller was placed in a 1cm thick
plastic box to reduce neutron dose and positioned as far away as
possible, just outside the treatment fields. In vivo measurements were
made using optically stimulated luminescence dosimeters (OSLDs) placed
inside the box at distances of 2cm, 8.5cm, and 14cm from the field edge,
and on the patient along the central axis and centered behind the LVAD
block. Further ion chamber measurements were made using a solid water
phantom to more accurately estimate the dose delivered to the LVAD.
Results: The total estimated dose to the controller ranged from
135.3cGy to 91.5cGy. The LVAD block reduced the surface dose to the
patient to 271.6cGy (68.1%). The block transmission factors of the 3cm
Cerrobend block measured in the phantom were 45% at 1cm depth and
decreased asymptotically to around 30% at 3cm depth. Applying these
transmission factors to the in vivo measurements yielded a dose of
120cGy to the implanted device.
Conclusion: Physical limitations of the controller made it impossible to
completely avoid dose. Shielding is recommended. The block had limited
dose reduction to the surface, due to secondary particles, but
appropriately reduced the dose at 3cm and beyond. More research on LVADs
dose limits is required.
Introduction
There is no cure for end stage heart failure. Currently, the only viable
long-term treatment is a heart transplant. Among the absolute
contraindications for a heart transplant is the presence or history of
cancer (most types) within the previous 5 years due to concerns about
active disease. An further complicating issue is the known
cardiotoxicities of cancer therapies, including radiation, chemotherapy,
and immunobiologics, all exacerbated by pre-existing cardiac
morbidities1,2. Left ventricular assist devices
(LVADs) help to improve survival rates in cases of heart failure and can
be used as a bridge to transplant candidacy and ultimately receiving a
transplant. An LVAD is an implanted mechanical pump which is directly
attached to the left ventricle of the heart. The pump continuously
pushes blood out of the ventricle to the aorta and the rest of the body.
The LVAD consists of the mechanical pump as well as an external
controller attached by a short driveline cable, 100cm in total length.
Generally, the pump is powered by batteries located externally to the
patient. The device can also be directly plugged into an electric
outlet, if needed. As LVAD technology improves, more patients are having
them implanted and survival rates with them are increasing. This leads
to a growing number of patients concomitantly afflicted with cancer. To
bridge patients to transplant candidacy, radiation is frequently used as
part of the multidisciplinary approach to treatment.
Radiation is well known to damage electronics. While the effect of
radiation on other cardiac implantable electrical devices (CIED),
including pacemakers and defibrillators, has been well studied, there
remains limited data on radiation to LVAD3-7. Much of
the existing radiation data is in the form of in-vitro studies and case
reports8-15. None of these studies considered the
radiation sensitivity of the LVAD external controller. Here, we report
the first case of dose delivered to the controller and the first
reported case of total body irradiation (TBI) to a patient with an LVAD.
The nature of TBI treatments makes this an especially challenging case.
The goal of the TBI is to eliminate the leukemia cells within the bone
marrow, as well as suppress the immune system in order to decrease the
potential for transplanted hematopoietic stem cell rejection.
Traditional TBI techniques use opposing fields, AP/PA or laterals, to
deliver a therapeutic dose. This does not allow for the same level of
precise dose delivery as more advanced treatment techniques for
localized cancers. This creates a tradeoff between limiting dose to the
LVAD while ensuring sufficient dose to the surrounding bone marrow. In
addition, these TBI treatments utilize large open radiation fields with
sufficient flash to cover the entire patient. This is quite different
from the majority of case reports in localized radiotherapy where the
LVAD controller is usually far away from the radiation field. Even
outside of the direct radiation field, there is out-of-field dose to
consider from scatter dose and linear accelerator leakage. As a result,
the external controller cannot be completely removed from all radiation.
This report details the treatment approach used to manage this complex
patient.
Furthermore, accurately assessing the dose delivered to the LVAD during
a TBI poses unique challenges. In-scatter, back-scatter, and electron
contamination all contribute to uncertainty in dose
calculations16,17. In addition, there is a lack of
information on out-of-field doses in treatments with extended SSDs such
as TBI treatments18. Therefore, direct measurements
are necessary to estimate dose to the LVAD and the external controller.
Methods and Materials
Patient Background
The patient in this case study is a 36-year-old male, with a history of
acute myeloblastic leukemia. Chemotherapy is believed to have caused
cardiomyopathy leading to cardiogenic shock with an ejection fraction of
18% and the subsequent need for a left ventricular assist device
(LVAD). To bridge the gap to transplant candidacy, the patient was
implanted with an Abbot HeartMate 3 LVAD approximately 5 months prior to
radiation treatment19,20.
TBI Simulation
The patient was simulated and treated using a traditional 16 MV anterior
posterior (AP)/posterior anterior (PA) technique. High energy x-rays
(>10MV) produce less dose variation form the central
axis17. However, known neutron contamination in high
energy x-rays is detrimental to electronics21,22. At
the time of treatment, alternative TBI treatments using 6MV were not
commissioned in our clinic. As such, the patient was classified as
high-risk and associated recommendations made in the report of AAPM Task
Group 203 were followed, including having members of the cardiology team
present for every treatment4.
During simulation, the patient laid on their left side facing the
treatment machine. A 2.54cm thick acrylic beam spoiler was placed in
front of the patient near the skin surface to increase the surface dose
in the build-up region at a source-to-surface distance of 500 cm,
simulating the AP treatment. The patient was then positioned such that
the distance from the spoiler to their umbilicus was 15cm. Measurements
of the patient’s thickness were acquired at the umbilicus, head, neck,
shoulder, mediastinum, hip, mid-thigh, knee, and ankle. The midline
separation was measured to be 26cm. Off-axis distances and the
spoiler-to-patient separation were also measured for each anatomical
landmark. Planar imaging was done with 16MV beams and film in this
position to ascertain the position of the heart and LVAD within the
patient. This entire process was repeated with the patient on their
right side facing away from the machine to simulate the PA treatment.
The PA simulation setup of the patient is shown in Figure 1A
The patient was prescribed 400cGy to the midline in two 200cGy daily
fractions, with the AP and PA fields equally weighted.
After careful discussion about the patient’s safety, the area
encompassing the LVAD and heart was prescribed a reduced total dose of
150cGy (37.5% prescription), which was the minimum that would be
achievable due to scatter dose from behind the Cerrobend block. The size
and shape of the Cerrobend was determined by the physician based on the
planar imaging acquired during simulation. An in-house TBI calculator
was used to calculate the necessary thickness of lead compensators to
optimize dose homogeneity at each anatomical site measured and the
required thickness of Cerrobend to block the heart and LVAD (LVAD
block). From this, the LVAD block was calculated to be 3 cm thick. Lead
compensators were fabricated out of 1.69mm thick lead sheets to be
attached to the gantry mount.
Treatment
Both treatment fields were 40x12cm2 with the
collimator set to 90 degrees. The LVAD was switched from battery power
to external power to minimize the chance of power disruption during the
treatment. As shown in Figure 1B, the external controller for the LVAD
was placed inside a 1cm thick plastic box to reduce neutron dose and
positioned outside the treatment field above the patient such that the
field edge was approximately coincident with bottom of the box. The
controller could not be moved any further out of the field due to the
finite length of the driveline cable connecting it to the rest of the
LVAD. High-Z shielding was not used due to the concern of creating
secondary particles. The appropriate LVAD block was placed on the beam
spoiler for each treatment field. Figure 1B shows the placement of the
AP block. Planar film imaging was used to confirm the position of the
block. If necessary, the position of the block was adjusted, and imaging
was repeated. For the first fraction, in vivo measurements were obtained
by placing two Landauer nanoDot optically stimulated luminescence
dosimeters (OSLDs) on the patient for each treatment field to measure
the entrance dose along the central axis and directly behind the LVAD
block. In addition, OSLDs were placed on the inside the box with the
controller at distances of 2cm, 8.5cm, and 14cm from the field edge, as
shown in Figure 1C. In the figure, the OSLDs are placed on the outside
of the box for ease of visualization. The OSLDs were not placed directly
on the controller to minimize the handling of the controller and to help
expedite the treatment process. Before, during, and after each
treatment, the LVAD was interrogated by the cardiology team to monitor
the operational parameters of the LVAD. Before, during, and after each
treatment, the LVAD was interrogated by the cardiology team to monitor
the operational parameters of the LVAD.
Dose Estimation
In addition to the in vivo measurements taken during treatment, a
series of phantom-based measurements were acquired to estimate the dose
more accurately to the LVAD at depth. First, ion chamber measurements in
solid water were made to estimate the depth dose behind the LVAD block.
An Exradin A12 ion chamber (Standard Imaging, Middleton, WI), with a
radius of 0.61 was used for all measurements. The overall size of the
solid water phantom was kept constant at 30x30cm2 and
a 24cm thickness. The thickness of the phantom was chosen to match the
patient thickness, while the height and width were limited by the size
of the solid water available. However, the size was deemed sufficiently
large to account for scatter. The phantom was placed 15cm from the
spoiler, centered behind the LVAD block location. To acquire a depth
dose curve, ion chamber measurements were taken at depths of 1, 2, 3, 4,
6, 9, 12, and 15cm. The effective point of measurement for each of these
points was shifted upstream by 0.6 multiplied by the chamber radius
upstream, as recommended for cylindrical ion chamber
dosimetry23,24. These measurements were first taken
with the LVAD block placed on the spoiler identical to the treatment
setup and the process was repeated with no LVAD block in place. The
ratios of these two measurements determined a depth-dependent
transmission factor of the LVAD block. To estimate the transmission
factor at the surface, an OSLD was placed on the surface of the phantom
to measure the dose with and without the LVAD block present. The final
dose estimate to the LVAD motor was based on CT imaging taken
subsequently to the TBI treatment. This imaging was used to ascertain
the depth of the LVAD motor from the AP and PA directions. From this,
the dose to the motor was estimated.
Results
The imaging from the AP and PA simulation and treatment of the patient’s
chest can be seen in Figure 2. The treatment images include the LVAD
blocks. Daily interrogation of the LVAD showed no transient effects
during or immediately after radiation. A 15-month follow up reported no
ill adverse effects from his TBI conditioning regimen from a cardiac
perspective. In fact, the ejection fraction had recovered to 65%, and
LVAD team is considering removal and disconnection of LVAD. It appears
the risk of damage or injury to the LVAD device, and heart from the
neutron contamination or radiation exposure were well reduced. As shown
in Table 1, the in vivo OSLD measurements showed good agreement
between the expected dose of 100cGy and measured dose to the central
axis. The AP and PA doses were 103.45 and 102.20cGy, respectively. The
OSLDs behind the LVAD block measured 66.49 and 73.61cGy for the same
treatment fields. This equates to an average surface dose of 70.05cGy
behind the block and a relative transmission factor of 68.1%.
Figure 3 shows the amount of radiation to the controller box and the
percentage of the prescribed dose to the central axis, excluding imaging
dose. In this region, the out-of-field dose dropped perfectly linearly
(R2 = 1.000), ranging from 33.83% (135.30cGy) to
22.87% (91.46cGy) at 2cm and 14.5cm, respectively. Extrapolating the
results, the dose at the field edge is estimated to be 35.6% of the
prescription (142.5cGy). The dose decreased away from the field edge at
0.91% (3.65cGy) per centimeter. Based on the location of the
controller, average dose to the controller is estimated to be 27.5%
(110cGy) for the course of the entire treatment.
The result of the transmission factor measurements can be seen in Figure
4. The OSLD based transmission factor at the surface is 67.9%. This
agrees very well with the in vivo measurements taken. Below the
surface, the relative transmission drops until leveling out around 30%
at a depth around three centimeters. Starting at a depth of eight
centimeters, the dose ratio slowly increased, reaching 31% at a depth
of 15cm. The higher values in the region upstream of 3cm can be
attributed to a couple factors: in-scatter from the beam spoiler and
secondary particles from the LVAD block. The slight increase at the
distal edge of the measurements is likely a result of backscatter from
the wall beyond the phantom setup and in-scatter from other surfaces in
the treatment room. Beam hardening beyond the block may also have
contributed to this increase.
The chest CT of the treated patient was used to localize the LVAD motor
within the patient. Based on this imaging, the motor, including the
housing, was estimated to have a width of 5.7 cm. From the AP direction
the motor had a depth ranging from 6.3 to 12.0 cm. From the PA
direction, the depth was 7.1 to 12.8 cm. These values fall within the
flat region of Figure 4, where the LVAD block transmission was measured
around 30%. Based on the estimated total of 200cGy delivered by each
field, the final estimated dose to the LVAD motor is 120cGy. However,
this estimate has some caveats, was noted in the Discussion section.
Discussion
Based on the measurements acquired, much of the LVAD pump within the
patient is estimated to have received around 120cGy. However, a couple
factors complicate this estimate. The high-Z titanium shell of the LVAD
pump causes attenuation and backscatter. The amount of dose penetrating
the titanium to the distal end of the pump itself is likely less than
the estimated 30%. However, the dose deposition to the proximal region
would have been slightly elevated. These elements would slightly offset,
the extent of which is challenging to determine.
Multiple case studies have reported on patients with LVADs receiving
external beam radiation6-15. All these studies looked
at directed radiation, with none involving a patient receiving TBI or
LVAD controller dose. These studies looked at multiple generations of
devices as new models are commonly being
released19,24. The case studies suggest a dose
tolerance of as much as 7500 cGy11. However, one case
study identified a patient who received stereotactic body radiation
therapy to the lung who later had very significant ventricular
tachycardia burden following treatment which was not accurately recorded
by their device12. This failure may be a result of
radiation damage. Data around other implanted cardiac devices,
defibrillators and pacemakers, has shown a large degree of variability
in dose sensitivity in these devices26. This is likely
the case with LVADs, as well.
Most of the studies on LVAD radiation sensitivity have been limited to
treatments using low energy photons (<10 MV). X-ray treatments
produce neutrons above when photon energies exceed 10 MV, which is
associated with malfunctions of contemporary implantable cardiac
devices23. It is reasonable to assume a similar
correlation with LVADs. Gossman et al found that LVADs (n = 2) did not
have any changes in pump operation during radiation with 18 MV X-rays
dosed 64-75 Gy14. More studies are required to fully
determine the effect of neutrons on these devices.
We can be optimistic for this patient’s future, although the
cardiovascular disease in stem cell transplant survivors remains a
concern27.
Conclusion
This case documents the total body irradiation administered to a patient
with a HeartMate 3 LVAD and concomitant cancer. Despite direct radiation
to the patient’s LVAD motor and scatter radiation to the controller, no
inappropriate device function was found during device interrogations
performed throughout treatment or within 6 months following treatment.
The current literature on the effects of radiation on LVADs is limited.
As an increasing number of cardiac device-dependent patients will need
cancer treatment, it is imperative to understand the best treatment
approaches that can be safely offered to this unique population. Further
research to address the safety of radiation therapy in patients with
LVADs devices is needed.