Feasibility of Dose Painting Using PET-Based Treatment Response
U Simoncic1,2, R Jeraj1,2, (1) Jozef Stefan Institute, Ljubljana, Slovenia, (2) University of Wisconsin, Madison, WIWE-A-217A-10 Wednesday 8:00:00 AM - 9:55:00 AM Room: 217A
The main prerequisite for biologically-conformal radiotherapy, or dose painting, is an accurate definition of the treatment targets, which is still an unresolved problem. The purpose of this study was to assess the feasibility of using early PET-based response for spatial modulation of tumor dose.
Ten patients with solid tumors were imaged with FLT PET/CT before the start of radiotherapy and during the treatment. PET-based responses were evaluated in two ways: the differences of mid-treatment and pre-treatment FLT PET/CT images, and the ratios of the images. Tumor dose was modulated using PET-based response, while integral tumor dose was kept unchanged. Non-uniformity of dose prescriptions was quantified with the relative mean difference (RMD), that quantifies the relative amplitude of modulation. The RMD is defined as the mean difference in a population, normalized to the population mean. The helical tomotherapy plans conformity was quantified with the Q95-105 factors, that represents the fraction of voxels with planned dose within 5% of the prescription.
Both PET-based responses have similar spatial distributions. Their correlation coefficients is between 0.7 and 0.9. The difference-based dose prescriptions have RMD between 3.0% and 11.8%. The ratio-based dose prescriptions are considerably more non-uniform, with RMD in range 14.0% to 32%. The conformity of the difference-based dose painting is close to the uniform dose delivery (Q95-105 in range 0.62 to 0.96), while the conformity of the ratio-based dose painting is significantly lower (Q95-105 in range 0.37 to 0.67).
These results demonstrate that dose redistribution based on the difference of two sequential PET images is feasible and could potentially be considered for implementation in a clinical setting. Dose redistribution based on ratio of two sequential PET images results dose plans with highly compromised conformity. It is also more sensitive to uncertainties in the dose prescription.