Dose-to-medium vs. dose-to-water: Dosimetric evaluation of dose reporting modes in Acuros XB for prostate, lung and breast cancer

Purpose: Acuros XB (AXB) dose calculation algorithm is available for external beam photon dose calculations in Eclipse treatment planning system (TPS). The AXB can report the absorbed dose in two modes: dose-to-water (Dw) and dose-to-medium (Dm). The main purpose of this study was to compare the dosimetric results of the AXB_Dm with that of AXB_Dw on real patient treatment plans. Methods: Four groups of patients (prostate cancer, stereotactic body radiation therapy (SBRT) lung cancer, left breast cancer, and right breast cancer) were selected for this study, and each group consisted of 5 cases. The treatment plans of all cases were generated in the Eclipse TPS. For each case, treatment plans were computed using AXB_Dw and AXB_Dm for identical beam arrangements. Dosimetric evaluation was done by comparing various dosimetric parameters in the AXB_Dw plans with that of AXB_Dm plans for the corresponding patient case. Results: For the prostate cancer, the mean planning target volume (PTV) dose in the AXB_Dw plans was higher by up to 1.0%, but the mean PTV dose was within ±0.3% for the SBRT lung cancer. The analysis of organs at risk (OAR) results in the prostate cancer showed that AXB_Dw plans consistently produced higher values for the bladder and femoral heads but not for the rectum. In the case of SBRT lung cancer, a clear trend was seen for the heart mean dose and spinal cord maximum dose, with AXB_Dw plans producing higher values than the AXB_Dm plans. However, the difference in the lung doses between the AXB_Dm and AXB_Dw plans did not always produce a clear trend, with difference ranged from -1.4% to 2.9%. For both the left and right breast cancer, the AXB_Dm plans produced higher maximum dose to the PTV for all cases. The evaluation of the maximum dose to the skin showed higher values in the AXB_Dm plans for all 5 left breast cancer cases, whereas only 2 cases had higher maximum dose to the skin in the AXB_Dm plans for the right breast cancer. Conclusion: The preliminary dosimetric results from our clinical study showed that the selection of either Dm or Dw in AXB is less likely to produce significant dosimetric differences in the clinical environment. However, the difference between the AXB_Dm and AXB_Dw calculations depends on the disease site, and even for the same type of disease (e.g., lung cancer), the results are patient specific.


Introduction
In conventional radiation therapy treatment planning systems (TPS), photon dose calculation algorithms typically report the absorbed dose as dose-to-water (Dw). Dose calculation algorithms employed in the TPS aim to best match the computed results with the measurements, which are performed in water phantoms. In recent years, there has been significant interest in using dose calculation algorithms that are based on Monte Carlo (MC) approach, which can report the absorbed dose in dose-to-medium (Dm) mode. In the Dm mode, the absorbed dose is computed to the medium contained in the dose voxel of the material. Siebers et al. 1 suggested that the conversion of Dm to Dw may be desirable in some of the situations when MC-based calculations are used in external beam photon radiation therapy. Currently, dosi-metric calibration protocols of external beam photon radiation therapy 2, 3 are based on the Dw mode, and the use of either Dm or Dw (after the conversion of Dm to Dw) for MC-based photon dose calculations remains a debating topic. 4 Varian's Eclipse TPS (Varian Medical Systems, Palo Alto, CA) has implemented Acuros XB (AXB) dose calculation algorithm, which can report the absorbed dose in both the Dm and Dw options. The AXB utilizes the Linear Boltzmann Transport Equation (LBTE) and solves numerically that describes the macroscopic behavior of ionizing particles as they travel through and interact with matter. 5 For the AXB_Dm, the macroscopic energy deposition cross-section and atomic density are based on the material properties of local voxel, whereas for the AXB_Dw, the energy deposition cross-sections are used for the local media. 5−7 Detail descriptions on AXB can be found in the published literature. 5−7 Several investigators 5−31 have studied the dosimetric impact of AXB using Dm reporting mode, and compared the AXB_Dm results with MC simulations, measurements, and Anisotropic Analytical Algorithm (AAA) calculations. Few studies have reported the difference between the Dm and Dw in AXB using MC simulations 23 , measurements 24,25,26 , and computed tomography (CT) data of real patients for nasopharygeal carcinoma 26 (planning technique used: intensity modulated radiation therapy (IMRT)), and soft-tissue sarcoma 27 (planning technique used: volumetric modulated arc therapy (VMAT)). Literature review shows that studies comparing AXB_Dm and AXB_Dw calculations on real clinical cases are limited. Hence, it is essential to further investigate if the selection of either Dm or Dw in AXB will make a significant dosimetric impact in the clinical environment. The major purpose of this study was to compare the dosimetric results of the AXB_Dm calculations with that of AXB_Dw calculations on real patient treatment plans for the prostate, lung, and breast cancer.

Patient selection and CT simulation
In this retrospective study, all dosimetric data were obtained from 21 st Century Oncology, Naples, Florida, USA. The patient selection was done based on the localizations of the tumor, which included prostate cancer (Group 1), stereotactic body radiation therapy (SBRT) lung cancer (Group 2), right breast cancer (Group 3), and left breast cancer (Group 4). Each group consisted five cases previously treated with external beam photon radiation therapy at 21 st Century Oncology, Naples, Florida, USA. All patients underwent standard CT simulation on a Phillips Brilliance CT Scanner. The CT scans were acquired using 512 × 512 pixels and 2.5 mm slice thickness. After the CT simulation, DICOM CT images were transferred to the Eclipse TPS for contouring and planning purpose.

Contouring
All four groups of patients included the planning target volume (PTV), which was expanded from the clinical target volume (CTV) drawn by the physician. Additionally, organs at risk (OARs) were delineated based on the axial CT images. The OARs of each group of patients are provided in Table 1.

Treatment planning
Treatment plans of all four groups of patients were generated in the Eclipse TPS (version 11.2) using RapidArc, IMRT, and 3D conformal radiation therapy (3DCRT) (Energy: 6 MV X-ray; Machine: TrueBeam (Varian Medical Systems, Palo Alto, CA, USA)). For prostate cases, the total dose prescribed to the PTV was 81 Gy with a daily dose of 1.8 Gy in 45 fractions. The prostate cancer treatment plans were generated using a single-full-arc. For the SBRT lung cases, the total dose prescribed to the PTV was 50 -60 Gy with a daily dose of 10 or 12 Gy in 5 fractions. The SBRT lung cancer treatment plans were generated using 5-field IMRT (n = 4) and 2-partial-arc RapidArc (n = 1) techniques.
For the right breast cancer cases, the total dose prescribed to the PTV was 45 Gy (n = 4) and 46.8 Gy (n = 1) with a daily dose of 1.8 Gy. The treatment plans were generated using 2-tangent fields IMRT (n = 4) and 2-tangent fields 3DCRT (n = 1) techniques. For all 5 left breast cancer cases, the total dose prescribed to the PTV was 45 Gy with a daily dose of 1.8 Gy in 25 fractions. The treatment plans were generated using 2-tangent fields IMRT (n = 3) and 2-partial-arc Rapi-dArc (n = 2) techniques. The isocenter in all treatment plans was placed at the center of the PTV.
The treatment plans of all four groups of patients were optimized using the Varian Eclipse Progressive Resolution Optimizer (version 11.2) with an objective of meeting the planning objectives of institutional guidelines. After the optimization process, each treatment plan was calculated using AXB_Dm (version 11.2) and AXB_Dw (version 11.2). The dose calculation grid size was set to 2 mm for all the cases. The calculated treatment plans were normalized such that the prescribed dose covered at least 95% of the PTV.

Plan evaluation
The dose-volume histograms (DVHs) of both the AXB_Dm and AXB_Dw were generated in the Eclipse TPS. Treatment plans were evaluated for various dosimetric parameters. For a comparative purpose, the relative difference (∆) in the corresponding dosimetric parameter (for example, mean dose, maximum dose, etc.) between the AXB_Dm and AXB_Dw plans of the same case was calculated using Equation 1.
where, x is the corresponding dosimetric parameter in the AXB_Dw and AXB_Dm plans of the same case.

Results
The dosimetric results in the AXB_Dm and AXB_Dw plans of all four groups of patients are presented in Figures 1-4.

Prostate cancer
For the prostate cancer, in comparison to the AXB_Dm plans, the AXB_Dw plans produced a higher mean dose to the PTV (∆avg=0.8%), rectum (∆avg=0.4%), bladder (∆avg=1.4%), and femoral heads (∆avg=2.5%). The relative rectal volume exposed to radiation was higher in the AXB_Dw plans with the average relative differences of 0.6% for V30, 0.9% for V50, and 1.3% for V70. Similarly, the relative bladder volume exposed to radiation was higher in the AXB_Dw plans with the average relative differences of 1.4% for V30, 2.3% for V50, and 4.5% for V70.

SBRT lung cancer
For SBRT lung cancer, the mean dose in the AXB_Dw plans was consistently higher for the heart (∆avg=1.3%) and esophagus (∆avg=0.7%) but almost identical for the PTV (∆avg=-0.1%). In evaluating the mean dose to the ipsi-lung and contra-lung, no clear trend was observed, with AXB_Dw plans producing higher values in 2 cases for the ipsi-lung and in 3 cases for the contra-lung. For the ipsi-lung, the V5, V10, and V20 were similar in the AXB_Dw and AXB_Dm, with an average relative difference within ±0.4%. For the V5 of the contra-lung, the averaged difference is -0.2% (range, -1.3% to 0.1%) with higher values in the AXB_Dw plans in 3 cases. The maximum dose to the spinal cord was consistently higher (∆avg=2.1%) in the AXB_Dw plans, with a relative difference ranged from 1.4% to 4.4%.

Right breast cancer
For the right breast cancer, the difference in the PTV maximum dose between the AXB_Dw and AXB_Dm plans ranged from -0.3% to -1.5%. The mean dose in the AXB_Dw plans was higher for the ipsi-lung (∆avg=1.2%) and almost identical for the majority of the cases for the contra-lung and heart (except for case #2 and #5). On average, the AXB_Dm plans had higher values for the maximum dose to the skin (∆avg=-0.3%) and contra-breast (∆avg=-0.9%). In the case of ipsi-lung, on average, the AXB_Dw plans had higher values for the V5 (∆avg=1.2%) and V20 (∆avg=0.8%) when compared to the AXB_Dm plans.

Discussion
The dosimetric impact of Dm and Dw reporting mode in AXB was investigated for the prostate, SBRT lung, and breast cancer. The AXB_Dm and AXB_Dw plans were evaluated based on the results derived from the DVH in the Eclipse TPS. The preliminary results from the clinical cases in this study showed that the differences in dosimetric results between the AXB_Dm and AXB_Dw plans depend on the tumor type. For instance, the mean PTV dose in the AXB_Dw plans was slightly higher (relative difference up to 1.0%) for the prostate cancer when compared to the mean PTV dose in the AXB_Dw plans for the SBRT lung cancer (relative difference within ±0.3%). Kan et al. 26 and Fogliata et al. 27 also reported higher values calculated by AXB_Dw when compared to the ones calculated by AXB_Dm.
The analysis of the OAR results for the prostate cancer (Figure 1) showed that the AXB_Dw plans consistently produced higher values for the bladder and femoral heads but not for the rectum. In the case of SBRT lung (Figure 2), a clear trend was seen for the heart mean dose and spinal cord maximum dose, with AXB_Dw plans producing higher values when compared to the AXB_Dm plans. However, the difference in the lung doses between the AXB_Dm and AXB_Dw plans did not always produce a clear trend, with the relative difference ranged from -1.4% to 2.9%. For both the left and right breast cancer, we observed higher maximum dose to the PTV as well as higher mean dose to the heart in the AXB_Dm plans, and this was true for all cases. The evaluation of the maximum dose to the skin showed higher values in the AXB_Dm plans for all 5 left breast cancer cases, whereas only 2 cases had higher maximum dose to the skin in the AXB_Dm plans for the right breast cancer cases.
The results presented in this study demonstrated that the difference between the AXB_Dm and AXB_Dw calculations is dependent on the tumor type, and even for the same type of tumor, the results are patient specific. The use of Dm vs. Dw for external beam photon radiation therapy is an interesting debating topic for the medical physics community. 4 On one hand, the supporters of Dm reason that (i) conversion from Dm to Dw adds uncertainty in dose calculations due to uncertainties in computed stopping power ratios; (ii) changing to Dm will have minimal impact on the treatment protocols; (iii) Dm is more likely to provide a better measure of biological response; and (iv) conversion of the Dm to Dw defeats a potential advantage of using MC-based dose calculation algorithms. On the other hand, the supporters of Dw argue that (i) commissioning beam data are always measured in water; (ii) clinical experience in terms of tumor/tissue response is based on Dw; (iii) dosimetry calibration protocols are based in water; and (iv) conversion from CT numbers to media results uncertainty in the medium type and composition.

Conclusion
The preliminary dosimetric results from our clinical study showed that the selection of either Dm or Dw in AXB is less likely to produce significant dosimetric differences in the clinical environment. However, the difference between the AXB_Dm and AXB_Dw calculations depends on the disease site, and even for the same type of disease (e.g., lung cancer), the results are patient specific. Future studies need to include large cohort of clinical cases with different types of cancer. Also, it is recommended to investigate the dosimetric impact of the treatment technique (e.g., 3DCRT, IMRT, and VMAT) on the AXB_Dm and AXB_Dw calculations for different types of cancer.