Initial clinical assessment of CT-MRI image fusion software in localization of the prostate for 3D conformal radiation therapy.

PURPOSE: To assess the utility of image fusion software and compare MRI prostate localization with CT localization in patients undergoing 3D conformal radiation therapy of prostate cancer.
MATERIALS AND METHODS: After a phantom study was performed to ensure the accuracy of image fusion procedure, 22 prostate cancer patients had CT and MRI studies before the start of radiotherapy. Immobilization casts used during radiation treatment were also used for both imaging studies. After the clinical target volume (CTV) (prostate or prostate + seminal vesicles) was defined on CT, slices from the MRI study were reconstructed to precisely match the CT slices by identifying three common bony landmarks on each study. The CTV was separately defined on the matched MRI slices. Data related to the size and location of the prostate were compared between CT and MRI. The spatial relationship between the tip of urethrogram cone on CT and prostate apex seen on MRI was also estimated.
RESULTS: The phantom study showed the registration discrepancies between CT and MRI smaller than 1.0 mm in any pair in comparison. The patient study showed a mean image registration error of 0.9 (+/- 0.6) mm. The average prostate volume was 63.0 (+/- 25.8) cm3 and 50.9 (+/- 22.9) cm3 determined by CT and MRI, respectively. The difference in prostate location with the two studies usually differed at the base and at the apex of the prostate. On the transverse MRI, the prostate apex was situated 7.1 (+/- 4.5) mm dorsal and 15.1 (+/- 4.0) mm cephalad to the tip of urethrogram cone.
CONCLUSIONS: CT-MRI image fusion study made it possible to compare the two modalities directly. MRI localization of the prostate is more accurate than CT, and indicates the distance from cone to apex is 15 mm. CT-MRI image fusion technique provides valuable supplements to CT technology for more precise targeting of the prostate cancer.