Inhalation anesthesia and shielding devices to allow accurate preclinical irradiation of mice with clinical linac-based systems: Design and dosimetric characteristics

Highlights • Up to twelve mice can be treated simultaneously.• Inhalation anesthesia minimizes discomfort, allowing fractionated treatment protocols.• The in-house developed devices give accurate radiation doses on the intended areas.


Introduction
Radiotherapy plays an important role in many cancer treatment protocols. Although radiotherapy has improved patient survival for many tumor types, there is still a need for optimization of therapies to increase radiation efficacy, decrease radiotoxicity, and for the identification of novel radiosensitizers [1,2]. Preclinical animal studies are essential to evaluate effective drug-radiotherapy combinations, study the radiobiology of tumor tissues, and assist in the development of more effective treatments to control or cure cancer [3,4]. Due to the small anatomical size of mice, dedicated instruments have been designed. But as these instrument are costly an do not enable radiation of multiple animals at the same time [5][6][7], we have explored the use of clinical systems as alternative to irradiate mice. We developed two devices to enable the use of clinical linac systems for irradiation of mice. First, a box for irra-diation of specific targets, based on the sedation of animals by inhalation anesthesia. Second, a box for total body irradiation of free walking mice.

Anesthesia and stress
When total-body mice irradiation is desired, e.g. leukemia research, fixation of the mouse is not a prerequisite. Free moving (i.e., in 2 dimensions) of the mice inside an irradiation box is possible with a limited number of mice to avoid mutual aggression, as long as dosimetric requirements are fulfilled. However, when targeting only the area of interest whilst sparing the organs-at-risk (OAR) is desired, immobilization or anesthesia is needed during irradiation, because, amongst many other factors, stress responses influence the outcome of radiation therapy, forced immobilization of awake animals is not advisable [8,9]. Therefore, we chose to use anesthesia to immobilize the animals. Two main categories of anesthesia can be considered: injection anesthesia or inhalation anesthesia. Although injection anesthesia would be easily applica- ble because no specific equipment is needed for this, we have chosen to use inhalation anesthesia. The rationale for this is that constant inhalation anesthesia gives greater safety by limiting the risk of under or overdosing and is much more reproducible: all animals recover almost simultaneously within 5 min due to the rapid recovery time [10]. Importantly, this anesthesia is mild enough to allow repeated anesthesia of the mice, needed for repeatedfractionated-radiation regimen.

Animal irradiation and linac systems
Although several small animal irradiation systems are available on the market, most preclinical animal centers do not have access to such dedicated animal radiation systems, whilst clinical linear accelerators are often available nearby or in the same clinic. The reasons why animal radiation systems are not available range from efficacy, financial, and scientific arguments. Concerning efficacy, most dedicated animal systems can handle only one animal at a time and are therefore highly time-consuming, labor-intensive, and cost-intensive. Scientific arguments include that research irradiators often use radiation sources with lower photon energy (range of~225 kV) or cesium-137 (662 keV) as compared to clinical systems where the photon energy typically range from 6 to 15 MV. As a consequence, with the lower photon energy of the preclinical systems, skin dose is 100% while the homogeneous dose distribution over the target can only be reached if the target is irradiated from several beam directions.

Dosimetric considerations
With clinical radiation systems, the homogeneous dose distribution is guaranteed by the high photon energy. With the target at the isocenter, i.e. at 1 m from the source, single beam dose inhomogeneity (<3%) over a target volume of 1 cm 3 is not an issue. The same accounts for total body irradiation (TBI) of a mouse when radiated from opposing directions [11].
Dose buildup, as explained below, is an issue to be solved when using high energy (6-15 MV) flattened photon beams. By using a flattening filter in a linac the dose buildup will range from 14% at skin level increasing to 100% at depth d max which is 1.5 and 2.5 cm for 6 and 15 MV, respectively; that means at least half of the thickness of a normal mouse used in studies will not receive the desired radiation dose. To solve this dose buildup aspect we placed a Perspex bar with a thickness of 1.5 cm, equal to the buildup at 6 MV, in beam direction in front of the mouse [12].
If a restricted area -such as the mouse brain -is the target, OAR-irradiation by the primary beam should be avoided by extra shielding of the OAR with lead shielding blocks positioned just above the mouse. Furthermore, phantom scatter, i.e. Compton scattered photons and electrons from the mouse and surrounding materials, should be minimized as much as possible to prevent damage of the OAR. It is a tight balance between photon and electron scatter. We selected 6 MV photon beam to limit the added bar thickness to 1.5 cm polymethylmethacrylate (acrylate) and consequently limit the electron scatter. However, with this relative low photon dose, scatter to the OAR due to photon scatter cannot be completely avoided [12,13].
In the case of total-body mice irradiation, the OAR shielding is not an issue. Sufficient build-up can be created if all box walls are thick enough which is the case with 2.5 cm acrylate walls. Air gaps between the build-up material and the target area does have an influence on the administered absolute dose, therefore we measured the dose with phantoms which have a comparable size as mice.

The target box
The design of the radiation device for the mice-brain setting is shown in detail in Fig. 1A-D. Fig. 1A shows a box with 4 compartments in which inlays can be placed to separate the mice. Each compartment contains 3 anesthesia outlets (referred to as masks) where the mice can be positioned with their teeth fixed to an outlet with constant flow of anesthesia gas allowing simultaneously mice-brain irradiations. In Fig. 1B, lead blocks are positioned shielding the OAR-areas whilst the target areas remain uncovered. Fig. 1C, D shows mice fixed in their masks sedated by isoflurane anesthesia. Fig. 1D shows the additional shielding of those mice with a 7 cm extra lead shielding block just above each mouse. Fig. 1F shows the box and anesthesia device setup on the accelerator. Mice can be anesthetized groupwise on-site just before radiation in a separate anesthesia box, called the induction chamber after which the mice can be transferred to the target box. Anesthesia can be administered by an isoflurane nebulizer (XGI-8 Gas Anesthesia System, Xenogen) both to the induction chamber and the target box. The whole system can be transported easily on a trolley. Except for positioning the mice in their masks, the anesthesia circuit is a closed loop where the excess of anesthesia gas is removed through carbon filters (Fig. 1F). After induction of anesthesia mice are transferred to the irradiation box and positioned in their individual masks (Fig. 1B). Anesthesia masks are available in different sizes to enable accurate positioning of each mouse with respect to the primary beam. The target box is closed during irradiation for optimal control of the anesthesia flow inside the box and prevents contamination of the mice (Fig. 1E).
For a subcutaneous tumor or a tumor in one of the extremities of the animal, an alternative setup of the target box can be configured where the so-called ''flank-only inserts" are used. In this configuration the mouse masks are rotated over 90°and the lead blocks are positioned along the body line such that the OARs are protected (Fig. 1G). In this configuration, 4 mice can be treated simultaneously. Different inserts are available to enable left or right-sided irradiation. The further process of anesthesia and positioning of the mice and lead blocks remains the same.

The total body irradiation box
For total body irradiation (TBI) of mice, a box inside a box concept is used. The inner box contains the mice (mouse-box) with an inside measure of 30 * 30 * 3.2 cm 3 (length, width, height), each wall of 0.3 cm thickness. This inside height of this box does not allow the animals to climb on top of each other thus preventing overlapping each other. The lid contains breathing openings with air-filter; the box can also be used for short-time mice transport, allowing sterile transport from the animal facility towards the linac. In case of irradiation, the mice-box is positioned inside the outer box (irradiation-box) exact fitting to the mice-box and with thicknesses of about 2.5 cm for all walls. The breathing openings are covered by the upper (removable) wall of the irradiation box during the beam-on time ( Fig. 2A, B).

Control dose measurements
To quantify administered dose in irradiated mice, we made near-cylindrical mouse-like phantoms, sized 25 Â 25 Â 50 mm, in which radiochromic EBT-3 film fits in a slit in the middle (Fig. 2C). These phantoms were positioned at different locations in the mouse-box simulating the brain-irradiation setting, as well as in the total-body irradiation box using varying settings of those   (Fig. 2B). The film was oriented perpendicular to the primary beam; in addition, the film was oriented parallel to the beam too in the TBI box dosimetry to obtain dose profile in both directions (Fig. 2F). The X-and Y-jaws were set at the desired irradiation geometry [14]. Dosimetric details are given in section C and Appendix.

Anesthesia
Athymic nude Foxn1 nu mice can be anesthetized by 2-3% isoflurane inhalation at a flow of 1 l/min. Mice are first anesthetized group-wise in the induction chamber after which the individual mice are carefully positioned in the radiation device. At this time, anesthesia can be lowered to 1.5-2.5% isoflurane. Anesthesia induction and positioning of the 12 mice can be done within 5 min. After radiation, mice can be returned to their home cages, and typically they will recover from anesthesia within a couple of minutes.

Target box
For dosimetry purposes, the exposed films were readout with our standard film dosimetry facility [15,16]. Dose profiles and absolute dose were read out for six mouse positions (Fig. 2D). Fig. 2E shows a representative dose profile. The observed mean absolute dose is 290 cGy ± 1.7% (1 SD), deduced from the flat profile part. If the tumor is situated at 5 mm from the actual field edge, i.e. set by the additional shielding blocks and defined by the 50% points, it will receive 95% (±1.0% 1 SD) of that absolute dose. Combining the above, a tumor will receive 276 cGy (±2.0% 1 SD) during 300 monitor units (MU) under the condition of the 5 mm distance between target and actual field edge. In practice, that value will be slightly higher (approximately 2%) as targets are positioned in general on top of the mice's skin, i.e. just below the bar, while in our measurement films are positioned more centrally in each mouse phantom. Additional information is given in the Appendix.

TBI box
With respect to dose calculation: Fig. 2G illustrates that radiotherapy dose calculation engines like Eclipse-AAA (Varian, Palo Alto, USA) are not capable to predict the measured outcomes for the moving phantoms. Therefore, in the TBI box administered doses were measured with six mouse phantoms located in the device, in various random settings at distances of 0-11 cm from each other (Fig. 2B). For 6 mouse phantoms present in the box, the absolute dose in the geometrical center of each mice phantom was constant at about 336 cGy ± 1.1% (1 SD) applying an AP-PA irradiation technique. As the absolute dose value might be influenced by the mutual distance of the mice phantoms, this effect was investigated for the AP-irradiation technique by considering the absolute dose with two phantoms: a stationary phantom at the geometrical center of the TBI box and a moving phantom at several distances of that stationary phantom. In the stationary phantom the absolute dose does not change irrespective the position of the other phantom; in the moving phantom an increase up to 5% is observed. (Fig. 2G). The reason is the more near position of that phantom to the acrylate wall of the box with increasing distance from the stationary phantom. Considering this result, a total of six phantoms were positioned randomly and dose measured after applying the AP-PA irradiation technique: a dose increase of 4% was observed in all phantoms, caused by a phantom scatter balance between mutual radiation and shielding of neighboring phan-toms. Therefore, in experimental settings the standard number of mice is 6 per box; allowed variation is limited between 4 and 8 mice per box. Experience over many years learns that mice move freely in the box during irradiation, i.e. our experiments seem to reflect real situation.

Conclusions
Many preclinical radiation experiments are performed with mice. For these experiments, it is important that the dose can be delivered accurately and that the discomfort of the animals is minimized. To enable this on clinically available linac systems, we developed a target box and a TBI box. We quantified that both devices can be used to deliver absolute dosages within an error range of circa 1% (1 SD); and a dose variation of up to 1.5% (1.1% SD) over each mouse. Moreover, the target box setup allowed us to anesthetize with isoflurane, position, and radiate up to 12 mice within 5 min. The target box has already proven its value in the identification of radiosensitizers for the treatment of pediatric brain tumors and glioblastoma [17,18], whilst the TBI box has been used in several studies, such as those for head-and-neck cancer [19].
In conclusion, the target box allows easy immobilization and positioning of the mice, the mice remain well controlled under anesthesia during the radiations, and both devices support accurate administration of the radiation dose.

Studies in animals
All animal experiments were done after authorization of the protocols by the local (VUmc animal welfare board) and national (CCD, central committee for animal experiments) authorization boards and all the guidelines have been followed.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. and below the head of the mice no additional material is present to limit phantom scatter as much as possible.
Anterior-to-Posterior and Posterior-to-Anterior (AP-PA) irradiation of mice in the TBI box yields over each mouse in all directions a dose variation of 1.5% at most (averaged over 6 mouse phantoms; ±1.1% 1 SD) for the dose at the geometrical center of each mouse phantom. The used field size is 34 * 34 cm 2 , i.e. 2.5 cm irradiation of the sidewalls of the outer box. Used photon beam energy is 15 MV.