The BrainLAB imaging couch top is a robust light weight and low attenuating patient positioning device. It facilitates the implementation of both cone beam CT and orthogonal x-ray imaging. There is, however, potential for significant beam attenuation through this couch. Furthermore, the design of the most optimal plan for IMRT and SRS largely depends on freedom in beam incidences that can be realized by a combination of gantry and couch rotations. With these degrees of freedom, there is a possibility for the beam to pass through the couch before entering the patient resulting in unacceptable distortion of the intended dose distribution. It has been suggested that to reduce the uncertainties introduced by the couch, the attenuation effects of the couch should be modeled in the treatment planning systems such as Philips Pinnacle3[23, 25, 26], Varian Eclipse TPS
[21, 22], CMS XIO
[14, 20] and Theraplan Plus
Using CF density of 0.55 g/cm3 and foam density of 0.03 g/cm3 we obtained the best agreement between measured and Pinnacle3 TPS calculated doses. The level of agreements were mostly < 1%, but all measured doses agreed to within 2% of Pinnacle3 TPS predicted dose which is the generally accepted tolerance of 2% and 2 mm of TPS suggested by Venselaar et al.
. Our results are similar to those reported in the literature. Mihaylov et al.
 modeled BrainLAB ICT using Pinnacle3 TPS and found an agreement of 0.2% to 1.7%. Other researchers have attempted to model different imaging couches using different TPS. For example, Wagner and Vorwerk
 modeled the Varian Exac treatment couch using Eclipse TPS and found the mean agreement of 0.15% (-2.02% to 1.96%). Smith et al.
 modeled iBEAM Evo carbon fiber couch (manufactured by Medical Intelligence) using CMS Xio and Nucletron Oncentra Masterplan. Good agreement was found between measured dose and dose predicted by TPS. The study of Myint et al.
 also found that the Theraplan Plus planning system predicted the effect of the treatment Medtec (orange City, IA) carbon fiber couch on the dose distribution to better than 2%.
The CF and F densities values that resulted in the best agreement between measured and predicted dose were lower than reported by Mihaylor et al.
 et al. reported 0.7 g/cm3and 0.1 g/cm3 for CF and F respectively while we found the value to be 0.55 g/cm3 and 0.03 g/cm3 respectively. One of the main reasons of the difference is that our modeled couch had an average CF thickness of 0.61 cm instead of the 0.2 cm reported by Mihaylor et al.
. Hence it is expected that to have the expected attenuation, the density of the CF would have to be lowered to compensate for the artificially elevated CF thickness. This observation underscores the importance for an individual center to validate the couch modeling before using it for patient treatment planning. Similarly, other researchers have reported density values that were different from those reported by the manufacturer
. Elekta quoted the electron density of 1.7 ± 0.1 g/cm3 for the iBEAM carbon fiber, however Smith et al.
 measured between 0.41 – to 0.64 g/cm3. They explained the discrepancy between quoted CF and measured CF density to be due to the partial volume effect.
Pinnacle3 TPS has been used clinically for many years and its algorithms have been commissioned and validated by several authors including Bedford et al.
. Pinnacle3 has various algorithms available for dose calculation including collapse cone convolution (CCC)
[29, 30] and adaptive convolution superposition (ACS). The Pinnacle3 convolution superposition dose model is a three dimensional dose computation which intrinsically handles the effects of patient heterogeneities on both primary and secondary scattered radiation. This computation method is able to account for dose distributions in areas where the electronic equilibrium is perturbed, such as tissue-air interfaces and tissue-bone interfaces. On the other hand, an adaptive convolution superposition uses the calculation technique of CCC but with some slight modifications. The speed of the computation is increased by adaptively varying the resolution of the dose computation grid depending on the curvature of total energy released per unit mass (TERMA) and dose distribution. This technique decreases the computation time by a factor of 2-3 without compromising the accuracy of the convolution superposition calculation in the presence of heterogeneities
. In our preliminary studies, no significant difference in the predicted dose was observed between CCC and ACS. However, ACS was faster and hence all the studies reported here used ACS. Using a different TPS (XIO and Nucletron Oncentra Masterplan) Smith et al.
 found an impact of the TPS algorithm on the predicted dose. They reported that the pencil beam and convolution algorithm failed to accurately calculate the couch attenuation. However, collapsed cone and superposition algorithm calculated the attenuation within an absolute error of ±1.2% for 6 MV.
BrainLAB iPlan RT Dose is based on the well established pencil beam superposition algorithm. Both the imaging couch top and the frameless extension have been modeled in iPlan RT Dose. However, to the best of our knowledge, no independent validation of the modeling accuracy has been reported in the literature. Table
6 shows an excellent agreement between iPlan predicted dose and measured dose. Examination of the modeled couch in iPlan RT Dose revealed that the modeled electron density were 1.70 g/cm3 (HU = 1240) and 0.11 g/cm3 (HU = -890) for the CF and foam respectively. The density numbers are similar to those quoted for iBEAM by Elekta (1.7 ± 0.1 g/cm3 for CF and 0.075 ± 0.0005 g/cm3 for foam)
 but are higher than those we found to produce high agreement with measured data in pinnacle3 TPS. As previously explained one of the possible reasons for the difference is that our modeled couch had an average CF thickness of 0.61 cm instead of the true value of 0.2 cm. So by simple ratio, our modeled CF is 3 times thicker and therefore the appropriate modeled density should be 3 times lower to predict the true attenuation (0.55x3 = 1.65 ~ 1.7).
Some of the other issues that have to be considered when including the couch in the treatment planning is that the patient must be positioned reproducibly on the treatment couch as compared to the imaging couch (CT couch). This is because as demonstrated in this study and other studies
[13, 23], the left to right shifts in patient position will result in beam path length in the couch being different and results in different degree of attenuation. So, it is important that some form of indexing be implemented. The “in/out” or longitudinal position of the patient is not as critical as the left to right position because there is no variation in path length.
One of the advantages of megavoltage over kilovoltage radiotherapy is skin sparing due to the buildup region. However, any material in contact with the patient’s skin during radiotherapy may cause a bolus effect and therefore the introduction of carbon fiber couch has the potential of causing a loss of skin sparing
. It is important to assess this loss in skin sparing because lack thereof could result in side effects such as induced erythema, moist desquamation and permanent hair loss
. The loss of skin sparing from megavoltage photons when using CF has been documented in the literature for different couch designs
[32–35]. The early work by Meara and Langmack
 noted an increase in build up with respect to Dmax of 47-56% through different carbon fiber combinations compared to 17.8% with no CF for 6 MV photon. Butson et al.
 reported that for the Varian ExactTM tabletop, the maximum skin dose (defined at a depth of 0.15 mm), for a 10 cm × 10 cm 6MV photon field was 55% compared to 19% for open field.
We did not actually measure the skin dose. However, we used the Pinnacle3 TPS to predict the skin dose. The ability of the Pinnacle3 TPS to accurately model dose at the buildup region has been reported by other researchers including Spezi et al.
, who found that Pinnacle3 TPS underestimated the buildup dose of 5% at the depth of 3 mm and to 2% for the depth between 5 and 10 mm. However, calculations and experiment agree very well below the extended buildup region
. So, one can assume that our results in Table
7 will agree to measured data to within 5%. To limit the loss of skin sparing due of CF, Mihaylov
[32, 33] has suggested using mixed beams that is using higher photon energies for the beam traversing the CFC. They reported that substantial skin sparing ranging from 5% to more than 49% can be achieved for individual cases
. They however noted that one caveat to this proposal is the well debated issue of neutron production with the use of high photon energy in IMRT.
The significance of modeling the couch is in the impact on the dose delivery to the target. It is well document that a 5% variation in dose delivered to the tumor can affect the therapeutic ratio
. It is therefore imperative to limit all sources of variation to a minimum. Preceding sections and previous studies
[36, 37] have demonstrated that ICT has an impact on the dose delivered to the target when an incident beam traverses the couch. The magnitude of its impact will however depend on the type of treatment, photon energy used, field size and total number of beams intersecting the couch. There are various clinical situations where the beams will traverse the couch or the head rest. For instance, four field box for the pelvic irradiation, AP/PA for the lungs, head and neck IMRT, prostate IMRT and brain IMRT. In our previous study
 we demonstrated that if the couch is not accounted for, up to 3% and 1.6% decrement in tumor dose for prostate and head and neck patients respectively can occur.