ESTRO 37 Abstract book

S183

ESTRO 37

Purpose or Objective High-risk cervical cancer patients are treated on the pelvic as well as the higher para-aortic region (PAO) using a 0.5cm CTV-PTV margin, according to the EMBRACE ll protocol. The margins from the previous clinical protocol were reduced from 0.8cm to 0.5cm. Therefore, we decided to reduce the default restriction of 4 degrees rotation in the setup of the pelvic to 2 degrees to ensure the coverage of the PAO region. If this restriction was exceeded the patient was repositioned. The first aim of this study was to assess whether 2 degrees restriction ensures geometrical PAO region coverage. The second aim was to investigate the relation between the error in position of the PAO region and the rotation of the pelvic area. Material and Methods For this study we used 159 CBCT scans of 12 high-risk cervical patients who were recently treated in our clinic. As part of their treatment patients received daily online CBCT imaging. We also included scans that were not acceptable for treatment (>2 degrees rotation). To assess whether the PAO region is sufficiently covered, we first defined two regions of interest (ROI). The first ROI was defined in the pelvic area. The second ROI was defined from L5 to the upper border of the PTV area (figure 1a). The pelvic ROI was registered using an automatic bony anatomy registration (translations and rotations). This registration was converted into table correction values (translation only; excluding rotations). The PAO ROI was registered using a grey value algorithm with translations only. To assess the coverage of the PAO region, we calculated the difference between the registrations of the two ROI’s, based on translations only; the residual error. Sufficient geometrical coverage was assumed in case this difference is less than 0.5cm. Furthermore we investigated whether the residual error correlates with the X-rotation found in the pelvic area. Finally, a prediction of the residual error was made based on rotations found in the pelvic area assuming a rigid spinal column, and was compared with the residual errors found in the clinic.

X-rotations found in the pelvic area had a moderately strong correlation (R 2 =0.36) with the residual error in the Z-direction (figure 1b). The predicted residual error was generally larger than the errors found in the clinic.

Conclusion Our study showed that a restriction of 4 degree s rotation in set up of the patient would ensure a geometrical coverage in the PAO region. The rotations in the pelvic area only moderately influence the position of the PAO region. This might be explained by deformation of the lumbar and sacral spine. Reducing the threshold on rotations from 4° to 2° is not necessary in clinical practice when margins of 0.5cm are being used. PV-0365 CASPIR Trial: Calculi as an alternative to fiducial markers for IGRT in localised prostate cancer A. O'Neill 1 , R. King 2 , K. Crowther 3 , S. Osman 1 , S. Jain 1 , A.R. Hounsell 1 , J.M. O'Sullivan 1 1 Queens University Belfast, Centre for Cancer Research & Cell Biology, Belfast, United Kingdom 2 Belfast Health and Social Care Trust, Medical Physics Department, Belfast, United Kingdom 3 Belfast Health and Social Care Trust, Radiotherapy Department, Belfast, United Kingdom Purpose or Objective Image guided Radiotherapy (IGRT) for prostate cancer (PCa) frequently employs surgically implanted fiducial markers (FMs). However, it is estimated that up to 35% of prostate radiotherapy patients have prostatic calculi (PC) visible on treatment cone beam CTs (CBCT). The purpose of this clinical trial is to directly compare FMs with PCs as We report data from a single institution prospective clinical trial investigating the feasibility of using prostate calcifications as natural FMs for IGRT. Patients planned for standard prostate radical EBRT +/- brachytherapy were eligible. Prior to CT planning, 3 gold fiducial markers were inserted into the prostate by the trans- perineal route under TRUS guidance. PCs visible within the PTV were contoured. Participants were aligned to FMs for EBRT using daily CBCT image guidance on a Varian TrueBeam linac. Off-line, a single experienced user analysed CBCTs using Image Registration in Eclipse (version 13.6). Random and systematic set-up errors were determined based on FMs, PCs (where present), prostate gland (PG) and bony landmarks (BL), enabling derivation of CTV-PTV margins for each data set. Results 30 participants with PC have been recruited. Data from the first 20 patients (385 fractions) resulted in 2695 individual image registrations (8085 individual data points). The calculated PTV margins based on each reference structure are summarized in Table 1. The maximum difference between the CTV-PTV (PC) margin and CTV-PTV (FM) margin is 2.2mm in the X or L/R direction. Margins required for FM, PG and PC in the Y and Z dimensions are comparable, with a maximum difference of 0.5mm between CTV-PTV (PC) and CTV-PTV (PG) . an aid to prostate IGRT. Material and Methods

Results The mean difference of the residual error of the POA region and their standard deviations is shown in table 1 Of the 159 CBCT’s only 3 times the residual error exceeded 0.5cm even though 58 CBCT-scans showed X- rotations above 2 degrees.