51. A novel algorithm (DirectDensity™) for reconstruction of simulation CT images used in radiotherapy treatment planning: dosimetric evaluation

Purpose CT imaging in radiotherapy represents the standard for treatment planning. The calculation of dose distribution requires appropriate energy-dependent calibration curves to convert Hounsfield Units (HU) into electronic densities relative to water (rEDs). To obtain optimal image quality and reduce dose, it could be useful to modulate X-ray tube potential based on patient size, but in radiotherapy treatment planning this practice requires the use of several calibration curves. A novel algorithm (DirectDensity™, DD™ recently developed by Siemens Healthcare, uses CT projection data to reconstruct images directly proportional to rEDs, independently of kVp used, thus removing the need of different curves. The purpose of the study is to evaluate this algorithm for clinical use. Methods The CIRS Density Phantom with different tissue equivalent inserts was scanned at 70, 80, 100, 120 and 140 kVp with the new Siemens Confidence CT scanner. Images were reconstructed with the DD™ algorithm to assess the correspondence of rEDs obtained from imaging with those certified. Afterwards the images of 7 patients scanned at 120 kVp were reconstructed with standard filtered backprojection (FBP) and DD™ algorithm. Treatment plans were calculated on both FBP and DD™ images by using standard HU-rEDs curve and linear CTvalue-rEDs relationship respectively (Fig. 1). Dose differences were evaluated through γ -analysis (1 mm, 1%). Results The rEDs measured from images (by reversing the linear relationship between CTvalue and rED) in regions corresponding to phantom inserts were compared to certified rEDs: differences were reported in Table 1 . Minimal dosimetric differences were found between treatment plans calculated on both FBP and DD™ images. DVHs showed differences in Dmean, D98, D2 and V95 for PTVs and OARs smaller than 1%. γ -analysis indicated negligible differences between dose distributions, with lowest passing rate at 97%. Conclusions This study proved the equivalence between DD™ and FBP, including the possibility to modulate kVp during CT acquisition, thus reducing patient dose without complicating treatment planning workflow.