3070 - Dosimetric Impact of Inter-Lesion Position Variation and Organ Deformation for Multiple Liver Tumors Treated with DIBH SBRT

Purpose/Objective(s):

Liver-directed RT plans often target multiple lesions. Single isocenter (ISO) technique is fast to plan and deliver but may introduce greater setup uncertainty. The goal was to estimate the dosimetric impact of inter-lesion position variation (ILPV) and organ deformation on SBRT plan with multiple targets treated with single ISO.

Materials/Methods:

Sixty pairs of simulated tumors were placed at most left/right, anterior/posterior, and superior/inferior locations along the central axes of liver on the planning CT (pCT) for 20 patients. An SBRT plan with a single ISO and 5 mm GTV-to-PTV margin was constructed for each pair. The ILPV was calculated as the difference between a pair of peri-lesional liver anatomy-based rigid registrations of the pCT with the pre-treatment CBCT. The dosimetric impact of ILPV was estimated by shifting the ISO by ILPV/2 and recalculating dose.

To assess the dosimetric impact of organ deformation: (1) the CBCT was deformably stitched with the pCT to generate an extended CBCT; (2) A second deformable image registration (DIR) was performed to deform the pCT and propagate contours onto the extended CBCT; (3) the propagated contours were reviewed and manually adjusted if needed; (4) the dose was recalculated on the deformed pCT, which represented the anatomy-of-the-day.

The ILPV-shifted plan and the DIR-recalculated plan were compared against the original plan, in terms of dose-volume-histogram (DVH) parameters (V100% for tumor coverage, Dmax and D2cc for luminal OARs).

Results:

The median inter-lesion distance was 10 (range, 5 - 17) cm. The median ILPV was 3 (0 - 15) mm. ILPV was not correlated with the inter-lesion distance. ILPV exceeded 5 mm for 17 of the 60 (28%) lesion pairs.

In the ILPV-shifted plans, the median PTV coverage was reduced by 3% (-25% - 2%). The reduction in PTV coverage was strongly correlated with ILPV. For ILPV > 5 mm, 97% of plans had a reduction > 5%, whereas for ILPV <= 5 mm, only 14% had a reduction > 5%. Though there was no difference in the median of any OAR DVH parameters, luminal OAR constraint was violated in 7 of 60 (12%) plans. In all 7 plans, the distance between OAR and GTV was <= 5 mm.

In the DIR-recalculated plans, the median PTV and GTV coverage was reduced by 14% (-47% - 10%) and 0% (-28% - 11%), respectively. The reduction in PTV coverage was moderately correlated with ILPV. Though there was no difference in the median of any OAR DVH parameters, luminal OAR constraint was violated in 13 of 60 (22%) plans. In 11 of the 13 plans, the distance between OAR and GTV was <= 5 mm.

Conclusion:

In this worst-case scenario dosimetric study, organ deformation added an additional 11% reduction in PTV coverage to the 3% by ILPV. It also increased cases violating OAR constraint by 10% over the 12% by ILPV. When ILPV is large and no lesions is close to luminal OARs, larger PTV margins may improve target coverage. When any lesion is close to luminal OARs, two-isocenter or online adaptive planning (adapt-to-shape) techniques may help to improve target coverage and ensure OAR sparing.