QA of Dose-Painting Plans: Risk of Overdosage in the High-Dose Regions?
K Haakansson1*, M Aznar1, L Fog1, L Specht1,2, J Rasmussen1,2, S Bentzen3, I Vogelius1, (1) Department of Radiation Oncology, Rigshospitalet, University of Copenhagen, Denmark, (2) Department of Oncology, Rigshospitalet, University of Copenhagen, Denmark, (3) Department of Human Oncology, University of Wisconsin - Madison School of Medicine and Public Health, Madison, Wisconsin.SU-E-T-167 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Deliverability of a treatment plan is generally evaluated by the gamma value (γ) and pass rate (PR) describing the fraction of measured points passing a certain γ-criteria for dose deviation (DD) and distance-to-agreement. γ does not distinguish between under/overdosage. When dose-painting to high doses in head-and-neck cancer, excessive dose to small volumes is associated with dose-limiting toxicity (Madani et al. Radiother Oncol. 2011). Overdosage of high-dose volumes is therefore a particular concern. Here we focus on the presence and detection of overdosage in high-dose regions of dose-painting plans.
40 dose-painting head-and-neck plans with five prescription levels (52.8, 60, 69.7, 73.1 and 79.7 Gy, all in 34 fractions) were measured with an EPID-based system. 14 of these plans with the poorest PR were additionally measured with a cylindrical PMMA diode phantom system. PR(γ: 3%/3mm), γ(3%/3mm) and DD were analysed. Correlations between planned dose, DD and γ were quantified by the Spearman rank correlation coefficient (ρ).
Mean PR were 98.2% (EPID, range 93.9-100) and 98.4% (PMMA, range 95.5-99.9). In the PMMA phantom, 262 diodes had a planned dose >2.2 Gy/fractions. Overdosage (between +0.004% and +6.9%) was measured in 248 of these. A high planned dose correlated with high DD (ρ=0.24) and poor γ (ρ=0.19) (p<0.001). The biological effect accentuates the problem; for example, 5% physical overdose at 80 Gy/34 fractions corresponds to +7.3% in biological effect for normal tissue (α/β=3).
The dose-painting plans were considered deliverable based on gamma PR. However, overdosage is common in the high-dose regions. This cannot be captured with the gamma PR criteria. Unfortunately, the translation of the dose to a phantom diode to a point in the patient is not easily performed. Additional QA may therefore be needed to ensure safe conduct and correct interpretation of the results of prospective dose-painting trials.