Purpose: Post-prostatectomy radiotherapy (PPRT) is delivered to the prostate and seminal vesicle bed, with the aim of eradicating local microscopic disease. PPRT is employed in both the adjuvant and salvage settings with curative intent. The vesico-urethral anastomosis (VUA) is the most common site of local recurrence following radical prostatectomy, and accurate identification of this site is pivotal in defining the inferior extent of the Clinical Target Volume (CTV) in most radiotherapy protocols. These protocols often rely on computed tomography (CT) scans for VUA delineation. CT has limited soft tissue resolution in comparison to Magnetic Resonance Imaging (MRI), which may more accurately identify the VUA. Accurate CTV delineation is essential to minimise both the risk of geographic miss and unnecessary normal tissue irradiation. The aim of this study is to identify systematic differences between the identified anastomotic sites on CT versus MRI -

Methods and Materials: Thirty consecutive patients treated with post-prostatectomy radiation at Austin Radiation Oncology Centre were identified. Planning CT scans were fused to T2 weighted MRIs by a bony then soft tissue match. The VUA, defined as the axial slice immediately inferior to the last slice where urine is visualised, was contoured by a single observer blinded to MRI findings. The VUA was contoured on MRI by the same observer, blinded to CT findings. Images were reviewed by a diagnostic radiologist. Positional differences of the VUA were recorded for each patient and analysed using a one sample t-test.

Results: Eight patients received adjuvant treatment and the 22 received salvage radiotherapy. Nineteen patients had pT3a disease and a further 2 patients had pT3b disease. The remainder had pT2a (n=2), pT2b (n=3) and pT2c (n=4) disease. Nineteen patients had positive surgical margins, with 11 being apical. The location of the VUA was more inferior on the MRI scan in all but three patients. The difference in location ranged from 0.00 – 1.00cm. The mean distance between the VUA on the CT and MRI scan was 0.45cm (95% CI 0.35 – 0.54, p<0.0001).

Conclusion: The use of CT alone in PPRT planning results in estimations of VUA position that are more superior than seen on MRI, in the majority of patients. This has implications for inadequate coverage inferiorly, which is of particular concern in patients with positive apical margins. Planning MRI scans enable more accurate identification of the inferior extent of the CTV and may prevent geographic miss.