Use of Synthetic CT for Magnetic Resonance–Only Based External Beam Pelvic Radiation Therapy of Cervical Cancer

Abstract

The superiority of T2 MRI over CT for cervical cancer tumor delineation is established by GEC-ESTRO. With the emergence of integrated MRI/LINAC technology, MR-only pelvic radiation needs to be studied in the definitive treatment of cervical cancer. In this study, we evaluate the impact of MR on pelvic targets and organs at risk (OARs) demarcation and dose distribution. Eight patients treated for cervical cancer staged FIGO IIA2 to IIIB for which a supine CT simulation and supine MRI of pelvis was performed were reviewed. All patient data was de-identified in accordance with institutional IRB requirements. Diagnostic MR Images were acquired on a 1.5T scanner. The MR images were manually co-registered with CT using Velocity software. CT body was segmented into 3 tissue types known as Dixon map: bone, fat, muscle/organs, by thresholding of CT number, and assigned HU numbers of 300, -100, and 0, respectively, to generate synCT. Treatment planning from CT was recalculated on the synCTs. Contouring of the CTV was performed by a radiologist on CT and T2MR to ensure accuracy. OARs were contoured on both CT and registered T2-weighted MR images. Four-field box conformal 3D-CRT planning to 45 Gy was used for plan generation. Treatment plans from CT was recalculated on the synCTs using same beam arrangements and monitor units. Dosimetric statistical analysis of the targets and OARs was performed using t-test with an alpha of 0.05. The volumetric difference between CT-CTV (mean 292.70 cm3, range 68.40-1139.90 cm3) and MR-CTV (mean 88.96 cm3, range 30.30-331.70 cm3) was not statistically significant (SS) (p=0.06). The mean D99 for CT and MR uterocervix was 45.32 and 44.99 Gy, respectively, and V100 for CT and MR uterocervix was 226.00 and 176.96, respectively, with no statistical difference between D99 or V100 for CT and MRI uterocervix (p=0.26 and 0.24, respectively). All dosimetric data can be found in Table 1 demonstrating non-significant difference on max doses for all OARs.Abstract 3614; Table 1Mean dosimetry metrics and statistical analysis using student's T-test comparing CT to MRICT CTVMRCTVCT UtcxMRUtcxCT BladMR BladCT RectMR RectCT L FHMRL FHCT R FHMR R FHMean Vol cm3292.790.0292.7238.3Mean Max dose Gy47.747.747.947.341.642.340.242.2t-Test P-Value0.060.240.360.060.750.26 Open table in a new tab The use of MRI for external beam planning provides optimal tumor visualization, providing potential clinical impact by improving target coverage. Our findings suggest that defining the CTV on MRI may be more accurate given the inherent MR advantage for pelvic organ visualization. The average CTV volume was smaller on MRI, which could ultimately lead to improved OAR sparing and reduced toxicities. We feel that this small pilot study warrants further investigation in order to fully evaluate the advantage of MR only based external beam planning for cervical cancer.