Intra-fraction motion of the prostate is not increased by patient couch shifts

Fifteen patients with histologically confirmed adenocarcinoma of the prostate were included in this analysis. All patients received norm-fractionated IMRT with 6 MV photons in our institution with a cumulative dose ranging from 70.0 to 76.0 Gy, depending on the tumor stage. Average age of patients was 72.0 ± 9.0 years (median 76.4 years, range 53.2 to 85.9 years).

Intra-fraction motion of the prostate was tracked during 359 fractions by four-dimensional perineal ultrasound. The Clarity ultrasound system (see Fig. 1) was used in combination with an auto-scanning perineal ultrasound probe [10] which provided estimates of the prostate position at a rate of about one to two Hertz. During this time, a total of 22 couch shifts of up to 31.5 mm along different axes were recorded.

For each couch shift, the (absolute) position of the couch and the position of the prostate (relative to the couch) were plotted for 3 min, centered on the shift. The plots were evaluated for any visible impact of couch movements on prostate motion relative to the couch.

Prostate motility was defined as the standard deviation of prostate positions. It was calculated for each 1 min before, during, and after each couch shift.

The prostate position closely followed along the couch position in every case. Consider the example in Fig. 2. It was selected because of all recordings it features the largest amplitude of couch shifts. At 00:00, the prostate was off by 6.5 mm in absolute terms (relative to the beam isocenter), and the couch position was at zero (relative to the beam isocenter). Between 01:00 and 02:00 several couch shifts were performed, with a net offset of the couch of −6.5 mm (relative to the beam isocenter). The prostate closely followed all of these shifts (relative to the beam isocenter) such that there was no remaining residual error after the shifts, with the prostate at 0.0 mm after 02:00. During all of these shifts, the position of the prostate relative to the table continued its intra-fraction motion, however at much smaller amplitude and not visibly affected by even the largest shifts at a speed of 9.7 mm per second. Similar patterns of motion were observed during the 22 recorded couch shifts, all of which were plotted in Fig. 3. In no case there was any visible influence of the couch shifts (relative to the beam isocenter, red lines in Fig. 3) on intra-fraction motion of the prostate (relative to the couch, blue lines in Fig. 3).

The standard deviation of prostate positions (relative to the couch) as a measure of prostate motility was calculated for each of the 3 min shown in the plots. Average motility before the shifts was 0.26 mm ± 0.14 mm (mean ± SD, median 0.22 mm, range 0.08 to 0.61 mm); average motility during the shifts was 0.27 mm ± 0.16 mm (mean ± SD, median 0.22 mm, range 0.05 to 0.64 mm); average motility after the shifts was 0.17 mm ± 0.09 mm (mean ± SD, median 0.15 mm, range 0.06 to 0.51 mm); see Fig. 4.

Thus, motility did not significantly increase during couch shifts (by 3 % on average, p = 0.88) and even slightly decreased after a couch shift (by 37 % on average; p = 0.02). The decrease in motility after a couch shift could be a statistical artifact. Tentatively, it could be because patients tautened when experiencing couch shifts. In any case, this effect needs further measurements for confirmation and explanation.

Couch shifts for real time correction of intra-fractional setup errors did not result in any relevant secondary motion of the prostate. In particular, motility did not significantly increase during couch shifts.

In conclusion, the strategy of real time correction of intra-fractional setup errors seems feasible as far as our results with regard to possible secondary motion of the prostate are concerned.

Written informed consent was obtained from the patient for the publication of the accompanying image.