Construction of a plan library for high risk prostate cancer patients in proton therapy
Abstract
High-risk prostate cancer patients can receive treatment of both the pelvic lymph nodes and seminal vesicles, in addition to the prostate based on an estimated risk of lymph node spread. Due to the large volume being irradiated, these patients could be an attractive sub-group for proton therapy. The primary advantages of proton therapy is that it causes less damage to healthy surrounding tissues than photon therapy dose and improves the conformity of the dose to the target. In regard to normal tissue, proton therapy therefore provides great potential in sparing normal tissue and reduce unwanted side-effects, but if the reduced dose leads to a reduction in toxicity has yet to be documented. The involvement of three target structures and the independent motion of these regions make delivery of the treatment with protons challenging. Adapting the treatment to the target motion of the patients during the course of treatment could be a solution [Yan D. et al., 2010]. A resource effective method since online re-optimisation is not yet feasible when delivering the treatment, can be to construct a plan library with different positions of the prostate in regards to the lymph nodes, which previously has been applied for radiotherapy [Xia P. et al., 2010]. This Master project will investigate if this strategy can be extended to proton therapy. The plan library was generated based on an initial analysis of prostate motion relative to bony anatomy in 18 patients (Study A). The output from study A was then used as an input for study B. In study B, we created a pool of three plans in addition to a standard plan for each patient basing on a planning computed tomography (pCT) and optimized each plan to accommodate a presumed prostate position. The three plans were for the prostate in three different locations with respect to the pelvic lymph nodes, including the original prostate contour and two contours shifted +/-5mm in an anteriorposterior (AP) direction and +/-4.6mm in cranial-caudal (CC) direction simultaneously. Prostate only (CTV67.5), prostate together with seminal vesicles (CTV60) and lymph nodes (CTV50) were the clinical target volumes (CTVs); and planning target volumes (PTVs) were created around the CTVs with an isotropic margin of 5mm. We compared this strategy to the standard treatment strategy with respect to the doses on the targets, rectum and bladder. The standard plan had a larger margin compared to the plan library with 5mm margin for PTV50 around CTV50, 3mm margin in the left-right and 10mm margins in both the cranial-caudal and anterior-posterior directions for PTV67.5 around CTV67.5; and 5mm margin in left-right and 10mm in both cranial-caudal and anterior-posterior directions for PTV60 around CTV60. The prostate margin was determined uisng the Van Herk's formula as part of the motion analysis in study A. The results showed similar doses to the targets as in the standard treatment approach but with reduced gEUD to the rectum (k=12) and bladder (k=8). Similar doses were also obtained for both small (k=4) and large (k=4) bowels in the two strategies. The plan library approach for treatment of high-risk prostate cancer patients therefore proved feasible.