Sixteen patients aged between 51 and 73 (mean, 62) years who underwent preoperative TRUS and radical prostatectomy for prostate cancer were included in this study. Using a HDI 3000 scanner (Advanced Technology Laboratories, Bothell, Wash) with using a 10MHz transverse sector probes, and calculating prolate ellipse volume by means of the (height × length × width × π/6) formula, prostate volume was measured and then compared with real specimen volume (Fig. 1).

The volume of gross pathologic specimens (prostate only) was calculated using the formula for prolate ellipse volume calculation, measuring three unequal axes of the prostatectomy specimen within 1hr of excision. Triaxial measurements of the prostate were thus obtained. The transverse diameter was recorded at the point of maximal transverse diameter perpendicular to the anteroposterior plane of the prostate. The anterior-posterior diameter was recorded in the transverse plane at a point perpendicular to the greatest transverse diameter, while the longitudinal diameter was recorded as the distance between the junction of the bladder neck and the prostate, and the prostatic apex at the genitourinary diaphragm. Prior to measurement, all specimens were immersed in normal saline to minimize tissue dessication.

Statistical analysis using a paired t test was used to determine the difference between the volume measured by TRUS and real specimen volume. A paired sample test was used for correlation of these two volume measurements.

Real specimen volume was 16.5-80.6 (average, 31.7±16.1) cc. Prostate volume measured by TRUS using height determined from transaxial scans was 22.0-78.3 (average 29.8±14.8) cc, and using height determined midsagitally was 21.4-72.4 (average 32.7±13.6) cc. Prostate volume measured either transaxially or midsagittally correlated closely with real specimen volume. When height was measured in one of these planes, a paired t test revealed no significant difference between prostate volume according to TRUS and real specimen volume (p = .411 and p = .740, respectively), nor were there significant differences between the findings of transaxial and midsagittal scanning (p = .570). A paired sample test, however, showed that TRUS prostate volumes determined transaxially showed a higher correlation coefficient (0.833) and lower standard deviation (9.04) than those determined midsagittally (0.714 and 11.48, respectively) (Table 1) (Fig. 2).

Prostate volume was measured by means of the prolate ellipse volume calculation (height × length × width × π/6), a formula used for ellipsoids with three unequal axes. In the early stages of prostate volume measurement by TRUS, maximal height and width measurements were obtained by axial scanning at the mid-gland level which appeared largest. Length was defined as the distance from the proximal external sphincter to the bladder base, as seen on midsagittal scans (11).

To avoid the salami effect, however, height measurement was corrected by sagittal projection in a plane perpendicular to length measurement: salami can be sliced in many different ways; to obtain larger slices, some prefer to cut it obliquely (13). It has been found that height measured transaxially was accurate, and high correlation with real volume was found (11), but as far as we know, no previous study has investigated whether prostate volume measured by TRUS is more accurately calculated using the height measurement obtained by transaxial scanning or by midsaggital. Thus, we determined whether, when measuring prostate volume by TRUS, height is more accurately determined by transaxial or midsagittal scanning, comparing the volume finding thus obtained with real specimen volume.

Using the height measurement obtained in our study by transaxial or by midsaggital scanning, a paired t test demonstrated excellent correlation between TRUS prostate volume and real specimen volume (p = .411 and p = .740, respectively).

Nor were there significant differences between the findings of transaxial and midsagittal scanning (p = .570). A paired sample test, however, showed that TRUS prostate volume determined transaxially showed a higher correlation coefficient (0.833) and lower standard deviation (9.04) than that determined midsagittally (0.714 and 11.48, respectively). We therefore suggest that when determining prostate volume by TRUS, volumes calculated using the height measurement obtained transaxially are more accurate.

Terris et al. (14) have claimed that since the point of juncture between the prostate apex and distal urethra is frequently poorly visualized, prolate ellipse volume calculation may be technically difficult. Likewise, definition between the base of the prostate and the seminal vesicle and bladder neck is often unclear. Accurate height, they believe, is more difficult to determine by midsagittal scanning than by transaxial. Using experimental models, Kim et al. (15) investigated the accuracy of prostate volume calculated by TRUS using height measurements determined transaxially and midsagitally. They concluded that the transaxial mode was more accurate.

Our results are supported by the two studies above. Because we compared the prostate volume obtained by TRUS with real specimen volume, we believe that our study was more physiologic than that of kim et al. (15). In addition, prostate weight or volume of water displaced may provide a more precise measure of prostate size; manipulation of the pathologic specimen (removal of seminal vesicles; dissection of periprostatic fat, connective tissue, and residual bladder neck tissue) prior to sectioning may, however influence the accuracy of the findings. Because we calculated specimen volume using the formula for prolate ellipsoids with three unequal axes (as described for TRUS prostate volume), our result may be more accurate.

During prostate volume measurement by TRUS and processing of the prostate specimen, our study suffered from several limitations. First, since ultrasound is a dynamic modality, operator-dependent factors may contribute to the differences observed between TRUS-derived prostate volume estimates and the volume of the corresponding specimen. Second, after surgical removal of the prostate, varying amounts of residual tissue are left behind, depending on the surgeon and the surgical method; a comparison of the volume of surgically removed tissue with ultrasonically estimated volume may, therefore, be inaccurate. Third, prostate tissue looses between 15 and 42% of its weight after tissue cauterization and due to the effect of irrigating solutions (16, 17).

Henneberry et al. (8), however, who estimated prostate volume pre-and post-operatively in 29 patients (18 of whom had undergone transurethral resection, and 11 open prostectomy), found good correlation in the 11 open prostatectomy cases, but that almost all of the 18 resected specimens weighed less than estimated preoperatively (8). In our study, prostatectomy was performed by open surgery, and we therefore expected less prostate tissue shrinkage.

The value of our study is that we measured prostate volume by TRUS using the height determined by transaxial and midsagittal scanning, and then comparing our finding with a real prostatic specimen. Thus, we investigated which method was most accurate and therefore most useful for clinical evaluation and determination.

In conclusion, prostate volume measured by TRUS closely correlates with real prostate volume. Furthermore, we suggest that when measuring prostate volume in this way, height is more accurately determined by transaxial than by midsagittal scanning.