Clinical implementation of intensity-modulated arc therapy (IMAT) for rectal cancer

In rectal cancer, combined radiotherapy and chemotherapy, either pre- or postoperatively, is an accepted treatment. Late small bowel (SB) toxicity is a feared side effect and limits radiation-dose escalation in a volume-dependent way. A planning strategy for intensity- modulated arc therapy (IMAT) was developed, and IMAT was clinically implemented with the aim to reduce the volume of SB irradiated at high doses and thus reduce SB toxicity. We report on the treatment plans of the first 7 patients, on the comparison of IMAT with conventional 3D planning (3D), and on the feasibility of IMAT delivery.

Seven patients, who were referred to our department for preoperative (n = 4) or postoperative (n = 3) radiotherapy for rectal cancer, gave written consent for IMAT treatment. All patients had a planning CT in prone position. The delineation of the clinical target volume was done after fusion of CT and MRI, with the help of a radiologist. For the IMAT plan, arcs were generated using an anatomy-based segmentation tool. The optimization of the arcs was done by weight optimization (WO) and leaf position optimization (LPO), both of which were adapted for IMAT purposes. The 3D plans used one posterior and two lateral wedged beams, of which the outlines were shaped to the beam's-eye view projection of the planning target volume (PTV). Beam WO was done by constrained matrix inversion. For dose-volume histogram analysis, all plans were normalized to 45 Gy as median PTV dose. Polymer gel dosimetry (PGD) on a humanoid phantom was used for the validation of the total chain (planning to delivery). IMAT treatments were delivered by an Elekta SliPlus linear accelerator using prototype software with the same interlock class as in clinical mode.

The IMAT plan resulted in 3 to 6 arcs, with a mean delivery time of 6.3 min and a mean of 456 monitor units (MU) for a 180 cGy fraction. The minimal dose in the PTV was not significantly different between 3D and IMAT plans. Inhomogeneity was highest for the IMAT plans (14.1%) and lowest for the 3D plans (9.9%). Mean dose to the SB was significantly lower for the IMAT plans (12.4 Gy) than for the 3D plans (17.0 Gy). The volume of SB receiving less than any dose level was lower for the IMAT plans than for 3D plans. Integral dose was lower in the IMAT plans than for the 3D plans (respectively 244 J and 262 J to deliver 45 Gy). Differences between the PGD measured dose and the calculated dose were as small for IMAT as for 3D treatments.

IMAT plans are deliverable within a 5-10-minute time slot, and result in a lower dose to the SB than 3D plans, without creating significant underdosages in the PTV. PGD showed that IMAT delivery is as accurate as 3D delivery.