Seminal vesicle intrafraction motion 77 of the seminal vesicles becomes smaller with increasing tumour infiltration. Based on our results and observations over all 247 fractions and 50 patients, we assume that the rigid registration method is sufficient to quantify seminal vesicle intrafraction motion. Even if the outer tip of a seminal vesicle is not completely accurately tracked, the obtained intrafraction motion may still be deemed as sufficient accurate as it is based on the main body of the seminal vesicle. The performed coverage probability analysis was based on the observed motion during the beamon period of 10 min. Motion that occurred during the preparation and (re)planning phase was not included in this analysis. This approach assumes that any intrafraction motion occurring between the daily pre-treatment scan and the beam-on period was negated. The results as presented in Table 2 show that 99% of the volume of both the left and right seminal can be covered at least 95% of the time when using 5 mm isometric volume expansion. In the case of the prostate, this threshold can be reached when using 3 mm isometric volume expansion. Especially the intrafraction rotation about the left-right axis has an effect on the coverage probability, as the rotation about this axis requires larger volumetric expansion to provide sufficient coverage. The results indicate that the seminal vesicles do not move identically. This can be observed from Figure 2, in which the right seminal vesicle has slightly larger spread in the posterior translation direction compared to the left seminal vesicle. The difference between the left and right seminal vesicle may be attributed to the asymmetrical shape and positioning of the rectum. Furthermore, it can be observed from the rotation graphs in Figure 2 and Figure 3 that the seminal vesicles are negatively correlated in the anterior-posterior and caudal-cranial rotation axis. This effect is most easily seen in Figure 3, in the graph of rotation about the anterior-posterior axis. In this graph, the seminal vesicles seem to diverge, where for the drift case the left seminal vesicle shows positive rotation, while the right seminal vesicle shows negative rotation. This effect may also be seen in the population mean lines of Figure 2, in the graphs of rotation about the anterior-posterior and caudalcranial axis. In these graphs, the population mean lines of the seminal vesicles diverge over time. This is supported by the reported Pearson’s correlation coefficients of -0.58 for the anterior-posterior and -0.47 for the caudal-cranial rotation. All these results indicate that the seminal vesicles rotate in opposite directions about the anterior-posterior and caudal-cranial axes when they are pushed by the rectum. The here presented tracking method may be used as input for a fast re-planning algorithm, which allows for intrafraction plan adaption. Examples of such re-planning algorithms are described by Kontaxis et al.27,28 Application of such methods may provide opportunities to treat high-risk prostate cancer patients with ultra-hypofractionated radiotherapy (one or two fractions). Conclusion To conclude, this is the first study to investigate six dimensions of freedom seminal vesicle intrafraction motion from 3D cine-MR imaging during actual treatments. We have shown that seminal vesicle intrafraction motion can be determined using a rigid registration method. Seminal 4
RkJQdWJsaXNoZXIy MTk4NDMw