3-T MRI with phased-array coil in local staging of prostatic cancer

To evaluate the diagnostic accuracy of a 3-T magnetic resonance imaging (MRI) system with a phased-array coil (3T MRI) in the local staging of prostatic cancer.

Between July 2004 and September 2007, 59 patients (mean age 66 years) with a histologic diagnosis of prostatic cancer underwent 3-T MRI with a phased-array coil. A total of 42/59 patients underwent a radical prostatectomy within 3 weeks of the MRI examination. Two radiologists with differing experience in the interpretation of prostatic imaging used a 1-5 scale score to assess extracapsular spread, seminal vesicle and neurovascular bundle infiltration, and prostatic apex involvement. The anatomopathologic examination conducted on histologic macrosections was the reference test used to evaluate the results of 3-T MRI. Interobserver reliability was assessed using the k value.

The sensitivity, specificity, and accuracy values obtained by the expert radiologist were 68%, 92%, and 83%, respectively, compared to 50%, 85%, and 71% for the identification of extracapsular spread and 81%, 62%, and 84% compared to 63%, 50%, and 55% for apex involvement. Interobserver reliability was good (k=0.71). Seminal vesicle infiltration was correctly identified in four of five cases and neurovascular bundle infiltration was identified in four of four cases.

Despite presenting diagnostic accuracy values lower than those reported in literature using 1.5-T endorectal coil MRI, the use of 3-T MRI with a phased-array coil could constitute a valid alternative to MRI techniques using endorectal coils in selected patients. Direct comparative studies between the two methods on large caseloads are required to confirm this hypothesis.