gms | German Medical Science

Backgrounds: Low-Z inhomogeneities such as lung and air column create dose perturbation in proton beam therapy (PBT) for lung and liver treatments. Without proper inhomogeneity correction, dose perturbation causes inaccurate patient treatment. The impact of low-Z material on proton dose distributions is studied and compared with those obtained from the treatment planning system.

Material and methods: Urie et al [1] in 1984 showed that low-Z bulk inhomogeneity can be corrected by scaling the range shift, DR, by equivalent density, Ρ_eq for a thickness, t, as:

ΔR = (1-t×Ρ_eq/Ρ_w). This concept is revisited by measurement at our center for 208 MeV beam with a maximum range of 28 cm. Measurements were made in a solid water-cork (Ρ=0.25g/cm³) slab phantom mimicking tissue-lung with parallel plate ion chamber embedded in the phantom. Measurements were made in reference condition (10 cm aperture, 16 cm range, 10 cm SOBP). The slab phantom was scanned and data transferred to the CMS treatment planning system (TPS) for dose calculation with proper electron density correction. The inhomogeneity correction factor, CF(E, d) is defined as ratio of dose at the same depth in low-Z medium to the dose in water: D(E,d)_m/D(E,d)_w. The CF were measured and compared with the results from TPS.

Results: The depth dose for various SOBP measured in cork has identical pattern. Figure 1 [Fig. 1] shows the partial depth doses in cork for different SOBP by matching the distal portion of the depth doses. The magnitude of ΔR is approximately 3 cm in cork per cm of water which does not agree with Urie'e density scaling. Note that distal end dose is reduced linearly with depth. The slope of SOBP and the distal fall off is nearly constant for all SOBP, and are found to be -0.6%/cm and -22.7%/cm, respectively. There is a significant range broadening in cork. The longitudinal penumbra (20–80%) increases linearly in cork as shown in Figure 2 [Fig. 2]. The calculated CF from TPS were compared for same geometry to that of the measurements and showed a fairly good agreement within 1.02±0.04 for all SOBP.

Conclusions: Low-Z inhomogeneity causes linear dose reduction in SOBP with a slope of (-0.6%/cm) and the measured ΔR is not equated by density scaling. The distal edge of the SOBP has a slope of -22.7%/cm which is much wider than that in water. Lateral distribution also increases linearly. The measured CF agreed with that from TPS to ±4%.