Sun Ho Kim, MD

Although 3D ultrasound is now widely available, the number of nonobstetric applications for which it is used remains limited. When the modality is used for nonobstetric purposes, it is generally for gynecological applications, vascular or superficial soft-tissue imaging, or hepatobiliary imaging.

The prostate is an ideal organ for 3D ultrasound because it is motionless and small enough to be covered by a single acquisition step. But the modality has rarely been used for prostate imaging, owing to problems with image quality and an anticipated minimal clinical impact.

Advances in ultrasound technology now make it possible to acquire volume data for larger objects more rapidly. The high-quality multiplanar images produced are comparable to those obtained with 2D ultrasound. The role of 3D ultrasound in clinical prostate imaging should consequently be reevaluated.

Three-D ultrasound images of the prostate should ideally be acquired using an endocavitary probe with a mechanized drive. The transducer elements in such probes automatically sweep through the region of interest (volume box) while the probe is held stationary. The freehand method, in which the probe is moved manually, is inconvenient and not as accurate.

Automated endocavitary 3D probes should be in close contact with the anterior rectum wall and kept stationary during data acquisition. The volume box should be adjusted to cover the maximum axial length of the prostate compactly, thereby increasing speed without losing image quality. Most endocavitary volume probes can sweep as wide as 90º, though some newer models can cover up to 120º. The 90º sweep may not be sufficient to cover a hypertrophied prostate. The angle can be wi¬dened by drawing back the probe, though this may increase the artifact and make probe fixation more difficult.

Digitally stored volume data can be manipulated and presented in various ways. Multiplanar displays can show three perpendicular planes (axial, sagittal, and coronal) simultaneously. The volume can also be explored by scrolling through any plane or by rotating to obtain an optimal view. The highest resolution will be in the acquisition plane.

Three-D rendering provides realistic, lifelike images that can be comprehended easily by patients and clinicians alike. The rendering algorithms used most often are maximum intensity projections (MIP) and transparent and surface-rendering modes (Figure 1). These algorithms rely on differences in acoustic impedance at tissue interfaces (e.g., fluid/soft tissue) and can be combined to yield an optimal display.

The multiplanar display is the one most commonly used for prostate imaging. Three-D rendering of this organ is relatively difficult because the prostate is surrounded by fatty tissue, not fluid, and does not itselfcontain fluid. The difference in acoustic impedance at tissue interfaces in and around the prostate is consequently smaller than for many other organs.

Transrectal ultrasound should be used to make accurate volume measurements of the whole prostate and the transitional zone.

 Although ellipsoid volume calculation is widely used in 2D transrectal ultrasound, this method has errors of about 20%. These errors may not be acceptable if data will be used to calculate prostate-specific antigen densities or inform the management or follow-up of prostate cancer patients. This method also relies heavily on individual practitioners' skill and experience, which can vary widely.

Three-D ultrasound enables actual volumes to be calculated even in irregularly shaped structures. The most well-known method involves tracing the perimeter of a structure in multiple parallel planes. The narrower the increments between planes, the more accurate the volume calculations will be. An alternative rotational tracing method uses software to divide the 360º circumference of a structure into precise increments. This technique, known as virtual organ computer-aided analysis (VOCAL), is reportedly reproducible.

Gray-scale 3D transrectal ultrasound can increase both the sensitivity and specificity of prostate cancer detection. This is due to its improved visualization of the gland and focal lesions, especially on coronal images. An early study showed 3D transrectal ultrasound to be superior to 2D transrectal ultrasound in depicting a tumor's presence and extent.

 Static volume contrast imaging can improve the delineation of focal lesions on multiplanar displays. This technique uses thin volumes to produce 2D slice images, similar to a projection of 3D images. Images have enhanced contrast and suppressed speckle when compared with 2D images (Figure 2).

Researchers conducting a pilot study found 3D trans¬rectal ultrasound to have an overall accuracy of 94% for staging prostate cancer.

 This finding compares favorably with the staging accuracy of 2D transrectal ultrasound (72%). The 3D approach proved to be 80% sensitive and 96% specific for local staging, with a positive predictive value (PPV) of 90% and a negative predictive value (NPV) of 96% for the detection of extracapsular extension (Figure 3).

This improvement in cancer detection and staging is solely dependent on lesions being visualized. Isoechoic cancers are difficult or impossible to see on 2D and 3D ultrasound. Switching to 3D transrectal ultrasound may increase sensitivity, but on its own, this change will not improve the detection and staging of prostate cancer significantly.

Color Doppler and power Doppler imaging can aid the detection and characterization of prostate cancers seen on 2D and 3D ultrasound. More time is needed for data acquisition when using Doppler imaging with 3D ultrasound, though gray-scale and power Doppler images can be acquired in a single step. Power Doppler signals can be removed or viewed separately from gray-scale images during postprocessing (Figure 4).The use of 3D Doppler ultrasound for cancer detection can reportedly increase the likelihood of a positive biopsy from 75% to 85%.

 A separate study has found 3D power Doppler ultrasound to have 92.4% overall sensitivityfor initial diagnosis and 72% specificity (PPV 80.6%,NPV 88.2%).

Signs of extracapsular extension can be seen more clearly on 3D power Doppler ultrasound. Prostate cancers can appear as regular avascular capsules or irregular avascular capsules. They may also have blood vessels crossing the capsule. The latter type is believed to be highly significant for extracapsular involvement.

 Three-D ultrasound is superior to 2D ultrasound for the demonstration of such findings on multiplanar views.Contrast enhancement has been used in many studies to increase the diagnostic accuracy of 2D ultrasound. One such investigation reported that adding contrast in¬creased the sensitivity of prostate cancer detection from 38% to 65% without significant loss of specificity (from 83% to 80%).

Contrast-enhanced ultrasound does have limitations, though. It is highly dependent on the practitioner's experience and expertise, and objective and reproducible data are difficult or impossible to obtain. Methods of data storage are also limited. The use of a videotape recorder to store data has been tried, but comparison of contrast enhancement at the same level is difficult with this method. Data postprocessing is also impossible.

Contrast-enhanced 3D ultrasound can overcome these limitations. The method is not examiner-dependent if performed according to a fixed protocol. Objective information on enhancement is saved in the volume data. This can be evaluated on a separate workstation or PC at any time after completion of the examination. 

Postprocessing can enhance image quality and allow identification of subtle enhancements. Unlike the 2D approach, evaluation is not restricted to the axial plane. Three-D and multiplanar evaluation can additionally increase sensitivity and specificity for the detection of abnormally enhancing lesions.

Two options exist for performing contrast-enhanced ultrasound. One approach involves using the contrast agent to increase the Doppler shift signal. This then increases the sensitivity of power Doppler measurements. Published data indicate that contrast-enhanced power Doppler 3D ultrasound could increase the diagnostic accuracy of prostate cancer imaging from 61% to 86%.

 Another option is to use contrast-enhanced harmonic imaging. This method uses rich harmonic signals generated by the destruction of contrast microbubbles. In¬termittent scanning ensures that the bubbles are not destroyed too quickly.

MRI has shown that cancers and normal prostate tissue have different vascular characteristics. An early wash-in rate has been shown to be a good predictor of cancer.

 This type of "parametric imaging" is time-consuming and complex, though. Researchers investigating dynamic contrast-enhanced power Dop¬pler 2D ultrasound, using images recorded on videotape, measured the number of pixels in power Doppler signals in different regions of the prostate. They reported minimal time-to-peak results to be the best predictor of cancer. Diagnostic accuracy was 78%.

Researchers at our institution have performed dynamic contrast-enhanced 3D ultrasound in 33 patients. The results of this unpublished study are promising. We found that the technique detected 21% more cancerous regions than conventional 2D transrectal ultrasound could (Figure 5).

The results of our investigation into dynamic contrast enhanced 3D ultrasound are preliminary. The study also has some limitations; selection bias (all patients were suspected for cancer clinically), retrospective analysis, and types of cancer (more diffuse types than focal) might affect the results. Nonetheless, we  believe that this approach to prostate imaging is feasible. An objective study is now required to confirm this.Three-D transrectal ultrasound may, in the future, be used in a number of other clinical applications. It could replace CT for the localization of brachytherapy seeds and help guide the implantation of probes during the cryoablation of localized prostate cancer. The modality could also be used to assess the prostate before and after treatment.

Three-D ultrasound has the potential to be superior to 2D ultrasound in prostate imaging. Storing objective data for offscreen viewing allows consensus decisions to be made and reduces interobserver variability. Gray-scale 3D transrectal ultrasound appears to be more effective than 2D transrectal ultrasound at detecting and evaluating focal lesions.

Local staging of cancers is potentially as accurate as MRI-based staging. Power Doppler and contrast enhancement can be used to increase diagnostic accuracy when lesions are isoechoic. Future research should identify additional clinical applications for this modality and refine 3D ultrasound techniques.

References
1. Mehta SS, Azzouzi AR, Hamdy FC. Three dimensional ultrasound and prostate cancer. World J Urol 2004;22(5):339-345.
2. Bega G, Lev-Toaff AS, O'Kane P, et al. Three-dimensional ultrasonography in gynecology: technical aspects and clinical applications. J Ultrasound Med 2003;22(11):1249-1269.
3. Raine-Fenning N, Campbell B, Collier J, et al. The reproducibility of endometrial volume acquisition and measurement with the VOCAL imaging program. Ultrasound Obstet Gynecol 2002;19:69-75.
4. Hamper UM, Trapanotto V, DeJong MR, et al. Three-dimensional US of the prostate: early experience. Radiology 1999;212(3):719-723.
5. Kim SH, Lee JM, Lee KH, et al. Four-dimensional volume contrast ultrasound imaging of the gallbladder compared with tissue harmonic imaging: preliminary experience. Eur Radiol 2004;14(9):1657-1664.
6. Garg S, Fortling B, Chadwick D, et al. Staging of prostate cancer using 3-dimensional transrectal ultrasound images: a pilot study. J Urology 1999;162(4):1318-1321.
7. Moskalik A, Carson PL, Rubin JM, et al. Analysis of three-dimensional ultrasound Doppler for the detection of prostate cancer. Urology 2001;57(6):1128-1132.
8. Sauvain JL, Palascak P, Bourscheid D, et al. Value of power Doppler and 3D vascular sonography as a method for diagnosis and staging of prostate cancer. Eur Urol 2003;44(1):21-31.
9. Halpern EJ, Rosenberg M, Gomella LG. Prostate cancer: contrast-enhanced us for detection. Radiology 2001;219(1):219-225.
10. Sedelaar JP, van Leenders GJ, Goossen TE, et al. Value of contrast ultrasonography in the detection of significant prostate cancer: correlation with radical prostatectomy specimens. Prostate 2002;53(3):246-253.
11. Buckley DL, Roberts C, Parker GJ, et al. Prostate cancer: evaluation of vascular characteristics with dynamic contrast-enhanced T1-weighted MR imaging-initial experience. Radiology 2004;233(3):709-715.
12. Goossen TE, de la Rosette JJ, Hulsbergen-van de Kaa CA, et al. The value of dynamic contrast enhanced power Doppler ultrasound imaging in the localization of prostate cancer. Eur Urol 2003;43(2):124-131.

Sun Ho Kim, MD

Articles

Three-D adds accuracy to prostate evaluation