Softwarové nástroje pro evaluaci kvality sCT na základě 3D gamma analýzy pro analýzu karcinomu prostaty

Abstract

The use of magnetic resonance imaging (MR) in radiation oncology has been increasing in recent years. The main advantage over computed tomography (CT) is better soft tissue contrast, a feature that contributes to more accurate delineation of targets and organs at risk. The dose distribution calculation algorithm needs electron density data that can be obtained by CT. CT and MR image registration therefore appears to be an ideal combination. This approach is often used in prostate cancer treatment planning. A disadvantage of this approach is the different filling of the bladder and rectum during time-lapse data acquisition. A solution to this problem should be synthetic CT (sCT) generated from MR using one of several commercially available algorithms. One of them is the MRCAT Pelvis (MR for Calculating Attenuation) available on the Philips Ingenia MR scanner. The next advantage of MR - only radiotherapy is the reduction of radiation exposure by omitting a CT scan. Although the quality of the MRCAT Pelvis algorithm has already been verified by several studies, there is a lack of a fast and simple tool for MR - only radiotherapy commissioning. The MRGamma software tool was created as part of this thesis. This software can evaluate the quality of sCT obtained by any algorithm using 2D or 3D gamma analysis by comparing sCT with data from a classic CT scanner. Furthermore, the created MRGamma application was applied to a cohort of 26 patients with prostate cancer who were treated at the Department of Oncology and Radiotherapy at the Hradec Králové University Hospital between 2015 and 2018, and whose sCT were created using the MRCAT algorithm. Two sets of criteria were used to calculate the gamma index – hard (HU_dif = 70 HU, DTA = 3 mm) and soft (HU_dif = 100 HU and DTA = 3 mm). For all patients, the area that met the criteria that ≤ 1 was evaluated using both 2D and 3D gamma analysis. For the 3D gamma analysis with soft parameters, the average percentage of areas that met the criteria ≤ 1 was (95.90 ± 1, 83) %. For the 3D gamma analysis with hard parameters, the average percentage of areas that met the criteria ≤ 1 was (94.22 ± 2.34) %. For the 2D gamma analysis with soft parameters, the average percentage of areas that met the criteria ≤ 1 was (94.87 ± 2.09) %. Finally, for the 2D gamma analysis with hard parameters, the average percentage of areas that met the criteria ≤ 1 was (92.76 ± 2.50) %. The results showed that the MRCAT algorithm provides high-quality data that closely matches standard CT and provides HU values with sufficient accuracy for dose distribution calculation. The developed MRGamma application was tested on a cohort of patients and could therefore become a fast tool for MR - only radiotherapy commissioning in radiotherapy departments.