Background:
Proper staging and localization of prostate cancer is essential in directing the final treatment for a patient with high risk disease. Staging is determined by the digital rectal exam, prostatic specific antigen, transrectal ultrasound imaging and biopsy findings. Magnetic resonance imaging (MRI) of the prostate has been used as an adjunct to the above studies. We have previously reviewed our endorectal coil MRI studies and found the staging accuracy, sensitivity, and specificity to be 80%, 92%, and 69% respectively. The accuracy in detecting seminal vesicle (SV) involvement was 84%. In March 2001 we reverted to performing pelvic phased array coil MR images. We reviewed the accuracy of these studies for the detection and localization of extracapsular extension of disease (ECE).
Methods:
All patients who had pelvic phased array coil MRI of the prostate prior to a radical prostatectomy from March 2001 through February 2005 were included in this review. MRI studies were reread by 2 radiologists and pathology specimens were reexamined by a pathologist and the location of ECE was noted by both. The sensitivity, specificity, positive, and negative predictive value of the MRI in detecting ECE was determined. The location of the ECE noted on the MRI was correlated with the actual location of the disease. Chart review was performed.
Results:
The mean time between biopsy and the MRI was 59 days (earliest 8 days). 24% of the patients had T3 disease. The pelvic phased array coil MRI detected ECE with a sensitivity, specificity, positive, and negative predictive value of 62.5%, 88-92%, 62.5-71%, and 88%. The length of ECE did not differ between the false negatives and the true positives. SV involvement was incorrect 100% of the time. In both patients with a false positive reading the MRI was performed 20 days after the biopsy. Therefore if the MRI was performed more than 20 days after the biopsy the specificity increases to 100% without changing the sensitivity. The location of the ECE detected on MRI correlated with the final pathology in location 33.3%- 40% of the time.
Conclusions:
The sensitivity of the pelvic phased array coil is lower than it was with the endorectal coil MRI studies while the specificity is higher. The ability to accurately detect SV involvement is worse.

Back to Final Program