Abstract Book

S1113

ESTRO 37

Breathe Well system for chest tracking and visual breath- hold (BH) guidance compared to the AlignRT (VisionRT, London, UK) surface imaging system for the first patient treated for left-sided breast cancer. Additionally guided vs unguided BHs are compared. Material and Methods The in-house developed Breathe Well system consists of an Intel RealSense sensor (infrared and optical), a screen and a processing unit (Figure 1a). The patient was set up on a breast board with arms placed above her head. A reference DIBH was recorded using Breathe Well during simulation by tracking the skin area around the central chest tattoo of the patient (Figure 1c). For each treatment fraction, the patient was set up following the clinical Align RT workflow. An unguided BH was recorded first with the Breathe Well system. Then the screen was turned on and the patient breathed into the BH tolerance area (reference DIBH±2mm) guided by the Breathe Well software interface, while the AlignRT system monitored the left breast area. The hypotheses were: (1) There is a correlation between the Breathe Well BH measurement and the vertical AlignRT BH measurement. (2) Breathing guidance improves BH reproducibility. Reproducibility was defined as the maximum difference between different DIBH levels [Cervino2009] and stability as the mean value of the single BH standard deviations [Stock2006].

Conclusion Breathe Well’s ability for a reproducible and stable BH guidance was verified with the AlignRT system. Visual feedback improved BH reproducibility compared to unguided BHs. EP-2038 Impact of Calypso system to treatment margin for prostate EBRT with MR-CT fusion-based workflow R. Kemppainen 1 , T. Seppälä 1 , P. Arponen-Esteves 1 , M. Myllykangas 1 , T. Kiljunen 2 , M. Tenhunen 1 1 Helsinki University Hospital, Department of Oncology, Helsinki, Finland 2 Docrates Cancer Center, Helsinki, Finland Purpose or Objective Recently introduced electromagnetic Calypso tracking devices promise to reduce geometric uncertainties of external beam radiation therapy (EBRT) of prostate cancer by allowing real-time monitoring of the treatment target. However, the implanted beacons are incompatible with magnetic resonance imaging (MRI). As MRI enables more accurate contouring of prostate compared to CT, contemporary state of the art prostate treatments are based on delineation using T2-weighted MR-images and propagation of the contours to the planning CT for treatment planning. Consequently, use of Calypso system affects total geometric uncertainty by reducing intra-fraction uncertainty and increasing registration uncertainty between CT and MRI. In this work, we evaluate the overall impact of using Calypso system for prostate EBRT with MR-based target contouring. Material and Methods The evaluation was based on published and measured estimates of spatial uncertainties in EBRT of prostate cancer. We evaluated the total geometric uncertainties for two alternative workflows at our institution by comparing the reduction of intra-fraction spatial uncertainties and increase of registration uncertainties in target definition when using Calypso system. Intra- fraction motion was assessed by using tracking recordings of 17 patients from a previously published study. CT to MR registration uncertainty was estimated from literature and by performing manual registrations for 4 patients by 3 experts. Required CTV to PTV margin was estimated for both workflows using the van Herk’s formalism and for different fractionations using Monte Carlo simulations.

Results (1) The chest motion measured with Breathe Well and the vertical AlignRT motion are correlated (p<0.001, r= 0.63) and both systems measure a similar reproducibility and stability (Table 1). Breathe Well-guided BHs stay well within the clinical AlignRT tolerance of 1cm. (2) The BH reproducibility is significantly (p<0.0001) improved with visual feedback (Table 1). The stability of the guided and unguided BHs was similar, but the unguided BHs were substantially shorter, which has potentially an impact on the results.

Results Based on the van Herk’s formalism of required CTV to PTV margins, estimated lateral vertical and longitudinal CTV to PTV margins were 7.1 mm, 7.5 mm and 8 mm with Calypso system and 5.5 mm, 6 mm and 6.5 mm when using standard implantable fiducials. According to the computer simulations, uniform margin of (w/o vs. with Calypso) 6 mm vs. 8 mm and 6 mm vs. 6 mm was needed