TL;DR
This study reveals that the choice of prescription isodose line significantly impacts the difference in target dose calculations between Monte Carlo (MC) and pencil beam (PB) algorithms in lung SBRT. Higher prescription isodose lines lead to greater discrepancies, with PB overestimating doses more substantially, particularly at the target's periphery.
What Problem Does This Paper Address?
Accurate dose calculation is critical for the efficacy and safety of lung stereotactic body radiotherapy (SBRT). A known challenge is the discrepancy between dose calculations from the more accurate Monte Carlo (MC) method and the faster, but less precise, pencil beam (PB) algorithm. PB algorithms tend to overestimate target doses, especially in the heterogeneous environment of the lung. While factors like tumor size and location are known to influence these differences, the impact of the normalized prescription isodose line—a fundamental parameter in treatment planning—remained largely uninvestigated. This study aims to quantify this effect and understand its underlying causes.
Methods
The researchers employed a two-pronged approach:
- Patient Study: They retrospectively analyzed plans from 12 lung SBRT patients. For each patient, four treatment plans were generated using the PB algorithm, each prescribed to a different isodose line (60%, 70%, 80%, and 90%). These plans were then recalculated using the MC algorithm. The differences in target dose coverage (specifically PTV D95, Dmean, and D5) and plan heterogeneity were compared across these different prescription isodose lines for each patient.
- Virtual Phantom Study: To isolate the cause of the observed isodose line effect, a simulated phantom study was conducted. Spherical virtual tumors were created, and plans were generated with uniform beam margins. This allowed for a controlled investigation into how varying isodose lines, tumor sizes, and beam energies influenced the MC vs. PB dose differences.
Datasets
- Patient Data: 48 treatment plans from 12 lung SBRT patients were used. These plans were originally treated at the University of Nebraska Medical Center.
- Virtual Phantom Data: 20 virtual phantom plans were created. These included variations in spherical tumor diameters (15 mm, 30 mm, 50 mm) and beam energies (6 MV, 10 MV, 15 MV), in addition to the four prescription isodose lines.
Key Insights and Ideas
- Isodose Line Matters: The choice of prescription isodose line is not merely an aesthetic or convention-based decision; it demonstrably influences the magnitude of dose calculation discrepancies between MC and PB algorithms.
- Peripheral Dose Errors: The primary impact of the isodose line is on the peripheral target dose, specifically affecting dose coverage metrics like PTV D95. Central dose metrics (Dmean, D5) are less affected.
- Higher Isodose = Higher Error: The discrepancy between MC and PB calculations increases as the prescription isodose line increases from 60% to 90%. This means PB overestimates the dose more significantly at higher prescription levels.
- Beam Margin is the Culprit: The phantom study revealed that the isodose line effect is driven by the effective beam margin. Higher prescription isodose lines necessitate larger beam margins to achieve coverage, leading to more low-density lung tissue being included in the beam path. PB algorithms struggle more with these heterogeneous interfaces, leading to greater errors.
- Tumor Size and Beam Energy Exacerbate the Effect: The isodose line dependence is more pronounced for smaller tumors and higher beam energies, suggesting these factors amplify the challenges posed by heterogeneous dose calculations.
Key Results
- PTV D95 Loss: For patient plans, the MC vs. PB reduction in PTV D95 (a measure of minimum dose coverage) monotonically increased with higher prescription isodose lines (60% to 90%).
- No Trend for Mean/High Dose: No significant isodose level-dependent trend was observed for PTV Dmean or D5.
- Phantom Study Confirmation: The virtual phantom study quantitatively confirmed that higher prescription isodose lines led to larger MC vs. PB coverage losses (measured by PTV V100 ratios < 1, indicating PB overestimation).
- Magnitude of Effect: The difference in PTV D95 loss between the 90% and 60% prescription plans for individual patients ranged from 4% to 12%.
- Tumor Size/Energy Impact: For a 30mm GTV, the MC-to-PB V100 ratio decreased (indicating larger PB overestimation) with increasing beam energy and increasing prescription isodose line.
Limitations and Future Work
- PTV as Surrogate: The study used the Planning Target Volume (PTV) as a surrogate for the target. The authors acknowledge that for MC calculations, especially with accurate heterogeneity correction, the Gross Tumor Volume (GTV) might be a more appropriate endpoint, and the isodose line effect could differ if GTV-based endpoints were used.
- Historical Data Interpretation: The findings highlight the importance of considering the prescription isodose line when comparing dose calculations from different algorithms or when interpreting historical treatment data. Future work could involve re-evaluating historical cases using this understanding.
- Clinical Implications: The authors suggest that this newly identified dependence should be considered when re-evaluating historical cases calculated by PB-type algorithms, especially those with varying prescription isodose lines, to better correlate clinical outcomes with realistic target doses calculated by more accurate MC algorithms. They also propose that fluence optimization and GTV-based prescription might be more suitable for MC-based SBRT.
Summary updated