Estimation of the prostatic volume by transrectal ultrasonography

Biplane planimetry as a new method

Kimura A,Kurooka Y,Kitamura T et al:Biplane planimetry as a new method for prostatic volume calculation in transrectal ultrasonography.Int J Urol 1997;4:152-156.

Abstract
Background: A new prostatic volume calculation method ; biplane planimetry utilizing prostatic contours of both cross and sagital sections is proposed.
Methods: Based on the cross and sagital contours, a model composed of the sequentially arranged copies of the cross section which are reduced so that the anteroposterior diameters of the copies fit the contour of the sagital section. The formula of biplane planimetry is the sum of the square of the reduced rates of the height multiplied by the area of the maximum cross section and by slice thickness. Tomograms of 150 patients who underwent ultrasonography with a transrectal chair-type probe were used for the coparison of this method with ellipsoid volume calculation or prolate ellipse volume calculation.
Results: The biplane planimetric volume calculation was the most accurate among the three methods.
Conclusion: The formula of our biplane planimetry is very simple, so it is easy to be incorporated in ultrasonic console having the funtion of distance and area measurement.

Introduction

Estimation of the prostatic volume by transrectal ultrasonography is now widely used for the evaluation of conservative treatment of prostatic cancer or hypertrophy and to improve the specificity of prostatic specific antigen. Though the most accurate method for sonometrics is step-section planimetric volume calculation, easier methods such as ellipsoid volume calculation or prolate ellipse volume calculation are more frequentry used. This is partly because a ratcheted stepping device is required to take tomograms at 5-mm intervals.
The ellipsoid volume calculation compared with step-section planimetry has been reported as having the tendency to underestimate the volume by 20%. Compared with ellipsoid volume calculation, prolate ellipse volume calculation seems to be more accurate, being calculated from both transverse and sagital sections. However, only three diameters are measured in this method.It means only six points of the prostatic contours are used for calculation.
In order to utilize full information from biplane section, we propose a new calculation method ; biplane planimetry. In this new method, prostatic contours of both cross and sagital sections are traced. Based on the cross and sagital contours, a non-ellipsoidal model is created. The model is composed of the sequentially arranged copies of the cross section which are reduced so that the anteroposterior diameters (height ; H) of the copies fit the contour of the sagital section. Because the areas of the copies are reduced in proportion with the square of the reduced rates of the height (H), the formula of biplane planimetry is
l x Amax x Σ(Hi/Hmax)2
in which l is a stepped interval of the arrangement of copies, Amax is the area of the maximum cross section, Hmax is the height of the maximum cross section, and Hi are heights measured in certain intervals in the sagital section where the reduced copies should be arranged.
The accuracy of this method was compared with ellipsoid volume calculation and prolate ellipse volume calculations.
Materials and Methods
Tomograms of 150 patients who underwent transrectal ultrasonotomography using an Aloka SSD-60 or a Toshiba SSL-51C with a transrectal chair-type probe were used for the analysis. Serial tomograms were taken at 5-mm intervals. Fifty cases had prostatic cancer (two having stage A disease, 21 stage B, seven stage C, 20 stage D). Fifty cases with benign prostatic hypertrophy underwent prostatectomy, and histologital confirmation was obtained. The weight of resected adenomas ranged from 5g to 140g. Remaining 50 cases are those admitted to our hospital with mild symptoms such as a feeling of incomplete emptying of the bladder or those who were suspected of having abnormalities of the prostate by digital rectal examination during screening. Urinalysis, examination of prostatic secretion, biochemical tests, and ultrasonography showed no abnormalities. These were categorized as normal groups.
Serial tomograms taken at 5-mm intervals were input into a personal computer via videosignal. The prostatic contour of each tomogram was traced manually, and area (Ai) and height (Hi) were measured. In the widest cross section, height (Hmax), transverse diameter (width ; W), and area (Amax) were measured. The number of slices multiplied by 0.5 cm was defined as cephalocaodal diameter (length ; L).
Because real prostatic volume was not obtained in the 150 cases, true prostatic volume is defined as equal to that calculated by step-section planimetric volume calculation ( V0 ).
The formula for step-section planimetric volume calculation is
V0=0.5cm x ΣAi
The formula for biplane planimetri volume calculation is
V1=0.5cm x Amax x Σ(Hi/Hmax)2
since the area of reduced copy is Amax x (Hi/Hmax)2.
The formula for ellipsoid volume calculation is
V2=8Amax2/3pW
which assumes that the gland is ellipsoidal in shape, where width (W) is defined as a hypothetical axis of rotation of area (A).
The formula for prolate ellipse volume calculation is
V3=p HmaxWL/6
in which the shape of the prostate is considered cuboidal.
The rates of the error for V1, V2, and V3 were calculated by dividing the difference of the value from V0 by V0.
Statistical analyses were done with Student's t-test.

Results
The rates of the error for biplane planimetric volume calculation(V1), ellipsoid volume calculation(V2), and prolate ellipse volume calculation(V3) in 50 normal cases were shown in fig.3.
V2 showed a tendency to underestimate the volume. V3 was more accurate than V2. V1 was the most accurate among the three methods. The mean and standard deviation of the error rate of these methods in the 50 normal cases were shown in table 1. The diferencies between V1 and V2 and V1 and V3 were statistically significant. In table 1, those in 50 cases with hypertrophy and those in 50 cases with cancer were also shown. V1 was the most accurate also in cases with hypertrophy. The rate of errors of V1 was smaller than V2 in cases with cancer, while the deference between V1 and V3 was not significant.

Discussion

As a new method to calculate prostatic volume from biplane sonograms, we proposed biplane planimetry. It was demonstrated that biplane planimetry was more accurate than prolate ellipse volume calculation or ellipsoid volume calculation. Littrup et al6) reported that prolate ellipse volume calculation was more accurate than ellipsoid volume calculation based on the prostatic volume measurement in 100 patients. Terris et al stated that pHW2/6 was more accurate than pHWL/6, because L correrated poorly with real length. Our resulls were in accord with Litrup et al.Since Terris's formula ; pHW2/6 uses only diameters obtained from cross section, the calculated volume seems to fluctuate according to the tilt of the scanner. As salami slices are of different heights depending on the inclination of the knife, the shape of the cross section changes as the scanning angle changes. By cutting 20 prostatic three-dimensional models created inside a computer, we previously demonstated that ellipsoid volume calculation fluctuated markedly in accordance with changes of the sccaning angle of the transverse section.
Compared with prolate ellipse volume calculation, biplane planimetry uses full information obtained from biplane sections. Prolate ellipse volume calculation uses only six points. Terris et al stated that measurement of cephalocaodal diameter (length ; L) is technically difficult, since the point of juncture between the prostatic apex and distal urethra frequently is poorly visualized. This is why pHW2/6 was more accurate than pHWL/6 in their study.
However, poor visulization of juncture between the prostatic apex and distal urethra does not yield significant error in biplane planimetry. Because H near the margin is small, the square of the H (biplane planimetry is the function of the sum of H2) becomes too small to affect affect the value of biplane planimetry.
In this study, only cross sections taken at 5-mm intervals were available in the 150 cases. Using chair-type radial scanner, sagital sections were not taken. In biplane planimetry, anteroposterior diameters ( height ; H) are designed to be measured at certain intervals in sagital section. However, in this study, height measured in step-section was used in stead of that measured in sagital section. Similary, prolate ellipse volume was calculated using length defined as the number of slices multiplied by 0.5 cm in stead of length measured in sagital section.
Therefore, there is a possibility that the evaluation of biplane planimetry using cross and sagital section yields different results from our study. However, Terris et al demonstrated that L determined from stepped axial section did not nessesarily inaccurate compared with L determined from sagital plane. This indicates that when step sections are available, sagital section is not essential.
Aarnink et al have been successful in outlining the prostate in the ultrasonic images automatically. Manual tracing of the prostatic contours has been inevitable not only in the step-section planimetric volume calculation, but also in the ellipsoid volume calculation. In the prolate ellipse volume calculation, an operator has had to input six points to measure the three diameter. Among these volume calculation techniques, the step-section planimetry is most accurate one but extremely time-consuming. Therefore, automated volume determination using edge detection technique and step-section planimetry will become a routine. However, as they stated, the automated method can not determine the prostate contours correctly when image quality is poor. Besides, it is inevitable that the level of gain and dinamic range affects edge detection, which varries markedly depending hospitals, machines, and operators. It takes sufficient times for automated determination to be a world wide routine prosedure.
The formula of our biplane planimetry is very simple, so it is easy to be incorporated in ultrasonic console having the funtion of distance and area measurement. The interval of the copy arrangement can be arbitralily determined from 5mm to 1mm or less according to the capacity of the console.

References

1. Benson MC, Whang IS, Olsson CA, McMahon DJ, Cooner WH. The use of prostate specific antigen density to enhance the predictive value of intermediate levels of serum prostate specific antigen.J Urol, 821:817-821,1992.
2.Watanabe H, Igari D, Tanahashi Y, Harada K, Saitoh M: Measurements of size and weight of prostate by means of transrectal ultrasonotomography. Tohoku J Exp Med,114:277,1974.
3.Terris MK, Stamy TA:Determination of prostate volume by transrectal ultrasound.J Uro,145:984-987,1991.
4. Littrup PJ, Williams CR, Egglin TK, Kane RA. Determination of prostate volume with transrectal US for cancer screening. Part 2. Accuracy of in vitro and in vivo techniques. Radiology 1991;179:49-53
5. Kimura A, Hirasawa K, Aso Y. Accuracy of ellipsoid method for estimatio n of prostatic weight.Jpn J Med Ultrasonics 1991;18:620-625.
6. Dahnert WF. Determination of prostate volume with transrectal US for cancer screening. Radiology 1992;183:625-627
7.Kimura A, Kurooka Y, Hirasawa K, Kitamura T, Kawabe K:Accuracy of prostatic volume calculation in transrectal ultrasonography.Int J Urol 1995;2:252-256.
8.Aarnink R.G., Huynen A.L., Giesen R.J.B., Rosette J.J.M.C.H., Debruyne F.M.J. and Wijkstra H.:Automated prostate volume determination with ultrasonographic imaging.J.Urol.,153:1549-1554,1995