To investigate the diagnostic accuracy of contrast enhanced transrectal ultrasound (CE-TRUS) in comparison with whole-mount radical prostatectomy specimens.

Fifty-eight subjects who underwent CE-TRUS and subsequent radical prostatectomy with whole-mount pathology were included in the study. Each patient underwent evaluation with baseline TRUS and again during CE-TRUS with intravenous infusion of perflutren lipid microsphere (Definity, Lantheus Medical Imaging, N Billerica, MA). A subjective 5 point scale was used to rate each sextant of the prostate in 3 baseline imaging modes and in 5 contrast-enhanced imaging modes. Baseline TRUS and CE-TRUS findings were compared with digitized whole-mount findings. A clustered logistic regression model was computed to compare the area under the receiver operating characteristic curve (A_z) for detection of prostate cancer by various modes of ultrasound imaging.

Among the 58 whole-mount specimens, a maximum Gleason score of 6 was identified in 29 subjects, a score of 7 was identified in 24 and a score of 8 was identified in 5. The A_z for baseline TRUS parameters was 0.55 for grayscale, 0.61 for color Doppler and 0.59 for power Doppler. CE-TRUS parameters demonstrated significant increases in A_z with the highest A_z for CE-power Doppler (0.66) and flash replenishment imaging (0.64) (P = .04 for comparison to baseline). The combination of CE-power Doppler and flash replenishment imaging resulted in improved A_z compared with baseline imaging (0.70 vs 0.59, P= .006).

Contrast-enhanced ultrasonography demonstrates greater diagnostic accuracy than baseline imaging. Diagnostic accuracy is further improved for "clinically significant" tumor volumes >1 cc.