Analysis of Prostate Deformation During Hypofractionated Radiotherapy Based on Implanted Fiducial Markers Displacement

Abstract

To report the prostate deformation during treatment based on analysis of difference in fiducial markers position in a large sample.This study included 144 patients treated with prostate SBRT (3625 cGy in 5 fractions). Every patient had implanted 4 gold fiducial markers (FMs). The same procedure to locate and mark FMs in the planning system was kept- marker in the base is FM number 1, marker in the apex is FM number 4. Treatment algorithm records the center of mass of implanted FMs for every pair of X-ray images taken during treatment and calculates the average rigid body error (RBEavg) in mm and rigid body error of all separate implanted FMs (RBEf 1-4). The RBE is defined as the distance of a fiducial during treatment from its CT position (visible on DRRs). Mean RBE, intrafraction and interfraction RBE were evaluated. The intrafraction RBE variability was defined as the standard deviation of all RBE during 1 treatment fraction. Interfraction RBE variability was defined as the standard deviation of the mean daily RBE over the whole treatment course. Regression and correlation analysis were used to evaluate correlation between FMs. Repeated measure ANOVA evaluated the change of RBE between treatment fractions.A total of 720 treatment fractions with 24453 X-ray image acquisitions were analyzed. We observed trend for higher RBE related to marker 4 (prostate base) during treatment. From 144 patients, only 1 case (0,7%) dealt with the higher RBE of FM 4 during the 1st fraction. During the 2nd fraction, FM 4 had to be disabled in 10 patients (7%), subsequently in 15 (10%), 16 (11%) and 21 (15%) patients for 3rd, 4th and 5th fraction. The mean RBEavg was 0,93 ± 0,39mm (range 0,1-11,30 mm) over the 5 fractions. The RBEavg was significantly lower for the 1st and 2nd fraction compared to others (P < 0.001). The interfraction variability of RBEavg was 0,26 ± 0,16 mm (Range 0,04 - 0,74 mm). The mean intrafraction variability of all FMs was 0,45 ± 0,25 mm. The highest Pearson correlation coefficient was observed between FM 2 and 3 implanted in the middle left and right portion of the prostate (R = 0,78; P < 0.001). Lower coefficients were observed in every combination with FM4 (range 0,66- 0,71; P < 0,001) what points to different deformation of the prostate apex.In ideal conditions prostate deformation is generally small but very sensitive to rectum and bladder filling. We observed RBE up to 11,3 mm. The overall correlation between FMs was affected by significant shifts of individual fiducials indicating that prostate is not "rigid" organ. Fiducial marker in the prostate apex was not possible to use in 28% of observations for RBE > 2.5mm. The average RBE of all FMs was significantly lower in the 1st fraction indicating prostate shrinkage during radiotherapy.L. Knybel: None. K. Resova: None. Z.Z. Cermakova: None. T. Blazek: None. T. Parackova: None. J. Cvek: None.