Correlation between prostate volume measured by transrectal and suprapubic ultrasound: can we do without TRUS?

Introduction & Objectives : Transrectal ultrasound (TRUS) is established as the gold standard prostate imaging tool in daily clinical practice, mostly to determine prostate volume (Pvol). Pvol is important for decision-making in prostate cancer to determine prostate-specific antigen density (PSAD, with a suspicious threshold of >15%) as well as in benign prostate hyperplasia (BPH, with a threshold of >80 mL not to perform transurethral resection of the prostate, TURP). TRUS however is not as widely available as suprapubic ultrasound (SPUS), more expensive and rather unpleasant for the patient. We investigate whether Pvol measurement with TRUS or SPUS differ significantly. Materials and Methods : We included 55 patients who underwent TRUS at Ghent University Hospital from March until May 2012. All SPUS measurements were performed by a urology resident, who also performed 18 (33%) of TRUS measurements. In 37 cases (67%) TRUS was performed by another urologist blinded to results from SPUS. In both examinations Pvol was calculated with the ellipse formula after measurement of all three prostate diameters (anteroposterior, transversal and craniocaudal). Correlation between Pvol by TRUS and SPUS was determined with Pearson correlation coefficient and differences between correlation coefficients with Fisher Z transformation. Univariate and multivariate linear regression was performed to account for influence of body mass index (BMI), bladder volume and blinding on the ratio Pvol(TRUS)/Pvol(SPUS). Results : Median Pvol was 32,1 mL (range 8,4-154,0) with SPUS and 35,5 mL (range 10,1-117,0) with TRUS. Overall Pearson correlation coefficient was 0,919 (P30% between SPUS and TRUS Pvol (range 0,5%-58,1%). SPUS and TRUS were inconclusive in 5 patients (9%) for a PSAD threshold of 15% and in 2 patients (4%) for a TURP threshold of 80mL. Both on uni- and multivariate linear regression lower BMI (univariate P=0.006) and blinding between observers (univariate P=0.008) predicted for lower ratio Pvol(TRUS)/Pvol(SPUS). Conclusion : Although there is a strong correlation between Pvol measurement by SPUS and TRUS, in almost one in three patients both measurements will differ >30%. Lower BMI predicted for more accurate SPUS measurement