Hypothesis / aims of study
The measurement of prostate volume informs the clinical diagnosis and treatment of benign prostatic hyperplasia (BPH), including the decision to treat with a 5-alpha-reductase inhibitor or surgery (1). The most widely used method for measuring prostate volume is the ellipsoid method (EM; V = L × H × W × 0.523 cm3) on transrectal ultrasound (TRUS) (2). However, some patients are unable or reluctant to undergo TRUS. In these cases, computed tomography (CT) is often performed. Here, we compared prostate volumes measured on TRUS and CT images in the axial, coronal, and sagittal views.
Study design, materials and methods
This was a single-center retrospective study. Among the males who underwent prostate ultrasonography at our hospital from January to December 2017, those who subsequently underwent multidimensional enhanced abdominal-pelvic CT within 6 months were included in our study. Patients who were diagnosed with acute prostatitis or underwent prostate surgery were excluded. Prostate volume was derived from the ellipsoid volume measured on ultrasound and axial, coronal and sagittal CT scans (Fig. 1). The volumes were compared using Pearson’s correlation tests and linear regression tests. Agreement between modalities was assessed using Bland-Altman analysis, which is widely applied to evaluate the concordance of measurements of the same variable obtained by two different methods. It is useful for identifying the degree of discrepancy (bias) between observations, presence of outliers, and data trends. If there is a difference between two volume measurements within the limits of agreement (standard deviation × 1.96), they can be considered interchangeable. Here, the clinical limit was designated as the mean value ± 10 cm3.
Results
TRUS scans were performed in 1,844 patients for 1 year, and multidimension CT scans were performed in 38 patients within 6 months of the date of ultrasonography. The median patient age was 69 (range: 48–85) years, and the median PSA was 2.59 (range: 0.21–45.73) ng/ml. The width and height correlation coefficients obtained for TRUS and axial CT were 0.868 and 0.844 (p<0.001 and p<0.001), respectively, and the length correlation coefficients were 0.689, 0.704, and 0.750 (p<0.001, p<0.001, and p<0.001) for the axial, coronal, and sagittal views, respectively. All volumes from CT scans showed a correlation coefficient of ≥ 0.9 (0.921, 0.957, and 0.970; p<0.001, p<0.001, and p<0.001, respectively). The mean differences between the VolTRUS and VolAx/VolCor/VolSag were -2.3 ± 6.4, -6.8 ± 4.6, -8.9 ± 3.9, respectively. The Bland-Altman plot, used to compare the TRUS volume with each calculated volume showed that axial measures for three patients (7.9%) and coronal measures for two patients (5.3%) were outside the clinically acceptable range. The sagittal volume was not outside the clinically acceptable range in any case (Fig. 2).
Interpretation of results
When the clinical limit of the estimated volume difference was limited to a mean value 10 cm3, only the sagittal volume showed agreement with TRUS. Therefore, in cases with sagittal view CT scans, VolTRUS can be calculated by subtracting the mean value of 8.9 3.9 cm3 from VolSag. Furthermore, since all of the VolSag values were greater than the VolTRUS values in all patients, physicians could reconsider BPH diagnosis in patients with a small VolSag and limit 5-alpha-reductase treatment if their VoSag is < 30 cm3.