Introduction
Adenocarcinoma of the prostate remains the most common malignant neoplasm and the second cause of cancer-specific death among men in the United States.[1,2] For the year 2000, the number of new cases of prostate cancer in the United States is estimated to be 180,400, with 31,900 expected deaths.[1,2] Because early disease is usually asymptomatic, the need for reliable diagnostic modalities to identify patients with early-stage prostate cancer is essential if effective therapy is to be contemplated.
Determining the extent of disease in newly diagnosed patients is another critical issue. More than one-third of patients with apparent clinically localized prostate cancer already have extraprostatic disease.[3,4] Pretreatment identification of this group of high-risk patients is essential for the selection of an optimum therapy.[5] If the disease is confined to the prostate, it could be curable with definitive local therapy such as surgery or radiation. Conversely, if the disease extends beyond the confines of the gland, other treatment approaches may be more appropriate.[6]
Improvements in screening and diagnosing prostate cancer occurred with the use of prostate-specific antigen (PSA) and the development of gray-scale transrectal ultrasound (TRUS) biopsy techniques.[7-10] Years of experience have shown that TRUS-directed biopsy, while very useful, has several limitations.[11-13] Prospective TRUS imaging data have demonstrated that conventional gray-scale TRUS is slightly superior to random chance in detecting prostate cancer.[14] The present trend is to increase the number of biopsies in order to compensate for the limitations of imaging alone.
The traditional lesion-directed biopsy led to the development of the six-core or "sextant" biopsy technique. Today, the trend is to perform 8 to 10 biopsies to more adequately sample the prostate gland. These additional biopsies tend to be laterally directed.[15] This increase in the need for additional biopsies is partly due to current limitations in adequately identifying cancer with noninvasive imaging.
Imaging remains an essential part of specific management approaches to prostate cancer, such as the use of TRUS in prostate brachytherapy and cryotherapy. Imaging modalities used to diagnose and stage prostate cancer include magnetic resonance imaging (MRI) and computed tomography (CT).[5,6] Bone scans and plain x-rays are used to detect distant metastasis.[6] However, the ability of these modalities to detect and stage prostate cancer is limited.[5]
The need for a more accurate imaging test has led to the development of new imaging technologies for both the diagnosis and staging of prostate cancer. These recent advances include Doppler imaging, contrast imaging,[16-20] MRI spectroscopy, and radioimmunoscintigraphy.[21-31] This article will review and critically assess the current status of prostate cancer imaging modalities, as well as discuss the evolving role of these newer imaging technologies.
Ultrasound Imaging Modalities
Gray-Scale Transrectal Ultrasonography
The role of gray-scale TRUS as an imaging modality for the prostate was firmly established in the 1980s with the introduction of high-frequency (7-MHz) transducers, the use of real-time ultrasound imaging, the development of biplanar probes, and the clinical applicability of outpatient transrectal biopsy. Gray-scale TRUS is frequently used to guide prostate biopsies (Figure 1) and for the evaluation of a patient with a palpably abnormal digital rectal examination or abnormal laboratory tests suggestive of prostate cancer (eg, elevated PSA). In some instances, TRUS may be used to monitor responses following a prostate cancer treatment in which the prostate was left in situ.
Limitations of Gray-Scale TRUS: Although gray-scale TRUS was a major achievement in improving the diagnostic yield of prostate cancer—especially in cases of nonpalpable disease (ie, stage T1c)—experience has shown TRUS to have several limitations. The subjective nature of the exam and the expertise of the clinician may affect the interpretation of images.[13] Interpretation of real-time, gray-scale TRUS images is also restricted by limitations of human visual perception. A recent study[13] showed that an expert user was not able to discriminate among images with more than 32 gray levels out of 256 displayed on a gray-scale image. Therefore, any technical improvement of ultrasound scanners may not always be perceived by human interpretation.
Perhaps the most troubling aspect of standard gray-scale TRUS is the nonspecific, echogenic nature of the tumor itself. Very early studies[7,9] suggested that prostate cancer was associated with hyperechogenicity. Presently, prostate cancer is believed to have an echogenicity that is less than that of normal prostate gland tissue (hypoechoic), with some series reporting 60% to 97% of cases as hypoechoic.[32,33]
Hyperechoic cancers are now considered very rare, and some authors have even questioned their true existence. If present, hyperechoic prostate cancers are usually carcinomas of the comedo type or cancers that have invaded areas of calcification or corpora amylacea. Egawa et al[34] reported hyperechogenic prostate cancer to account for 1.3% of cases in a contemporary series.
Up to 25% of tumors are reported to be isoechoic (the same echogenicity as the normal prostate tissue).[34] Isoechoic tumors are almost impossible to detect because a clear distinction between the tumor and surrounding prostate gland tissue cannot be made on the basis of echogenicity. In such cases, the presence of secondary signs—such as glandular asymmetry, capsular bulging, and areas of attenuation—might prove helpful.
Echogenic discrepancies may again be a useless diagnostic criterion in cases where the tumor diffuses and totally replaces an entire zone or the entire gland. Overall, the presented data suggest that gray-scale TRUS alone is unreliable in diagnosing prostate cancer and must always be performed with a biopsy to evaluate for cancer.
Early studies generated enthusiasm about the role of TRUS in improving the staging of locally advanced prostate cancer.[7] However, this enthusiasm soon faded when other studies failed to demonstrate that TRUS was better than a digital rectal exam for the detection of local extension.[35] The ability of TRUS to detect a neurovascular bundle or seminal vesicle involvement is operator dependent and is associated with a high false-positive rate for seminal vesicle involvement.[5]
In general, an accurate assessment of locally advanced disease is frequently difficult; available ultrasound units do not have adequate resolution to detect the microscopic extension associated with many cases of locally advanced disease. This was reflected in staging studies that reported low predictive values of gray-scale TRUS ranging from 18% to 60%.[12] For these reasons, gray-scale TRUS is considered by most to be a nonreliable tool for staging locally advanced disease.
Doppler Imaging
The Doppler shift frequency is an effect that applies to all wave motion. Discovered by Austrian physicist Christian Doppler (1803-1853), the effect refers to a change in wave frequency caused by the motion of a wave source, receiver, or reflector. Acoustic Doppler effect is frequently experienced in our daily life, for example, with approaching and receding sirens or train whistles. Mathematically, this effect is represented by the following formula:
fD = 2f0vcosq/c
where fD is the Doppler frequency shift, f0 is the incident frequency, v is the flow velocity, c is the speed of sound in tissue, and q is the angle between the ultrasound beam and flow direction.[12] Simply put, this equation measures the difference in frequency of returning echoes and emitted frequency.
Principles of Use in Medicine: Doppler ultrasound is mainly used in medicine to detect the presence or absence of blood flow in vessels, its direction, and its characteristics. In urology, this technology has been frequently applied to detect the velocity of renal blood flow, penile vasculature, and to assess neovascularity in renal, testicular, and prostate tumors [36-39].
Flow can be detected either by pulsed-wave Doppler (which displays the frequency shift or velocity as spectral waves) or color Doppler imaging, also known as color-flow imaging.[12] Color-flow imaging provides a two-dimensional (2D), cross-sectional, real-time, color-coded Doppler shift that is superimposed on the real-time gray-scale anatomic display.[13] It displays the range of the mean frequency shift or velocities of red blood cells within flowing blood as colors of the spectrum. Flow toward the transducer is depicted in various shades of red, and flow away from the transducer is characterized by shades of blue.[40]
Normally, the prostate gland should demonstrate symmetrical, low-to-absent color-flow signal intensity, with the periurethral area exhibiting some flow and the outer gland showing minimal to no flow.[39,41,42] Several studies have demonstrated that malignant prostate tissue can sometimes be associated with abnormal vascular patterns.[43-48] Detection of these abnormal blood flow patterns within prostatic tumors is the main application of Doppler ultrasound in prostate cancer imaging.
Early Clinical Results: Early results using pulsed Doppler were disappointing,[39] and slightly better results were reported later with the use of color Doppler imaging.[42] Rifkin et al[39] reported their experience in 619 patients in whom color Doppler imaging was used as an adjunct to gray-scale TRUS. Using TRUS biopsies of foci lesions identified with gray-scale imaging or from sites showing flow abnormality on color Doppler imaging scans, they confirmed the presence of 132 cancers in 121 men. A total of 123 men (93%) had corresponding gray-scale abnormalities, and 114 men (86%) demonstrated abnormal flow on color Doppler imaging. Nine patients (7%) showed no obviously identifiable abnormality on gray-scale scanning but had distinctly abnormal flow on color Doppler scanning.
Kelly and associates[49] reported a sensitivity of 96% for TRUS alone and 87% for color Doppler imaging. The addition of color Doppler imaging increased the positive-predictive value (PPV) from 0.53 using TRUS alone to 0.77—but at a cost of reduced sensitivity. In only 1 case out of 158, color Doppler imaging suggested the diagnosis of malignancy independently of TRUS. The authors concluded that color Doppler imaging improves the PPV of TRUS, but has no demonstrable value over TRUS alone in the diagnosis of prostate cancer.
Correlation With Biopsy Findings: Newman and coworkers[50] correlated color Doppler imaging results with the histologic findings from site-specific transrectal core biopsies. In this study, color Doppler imaging had a sensitivity of 49%, specificity of 93%, and PPV of 62%, and was able to detect at least one focus of carcinoma in seven patients with no gray-scale abnormalities. The findings indicated that focal peripheral zone hypervascularity on color Doppler imaging is associated with an increased likelihood of prostate cancer or inflammation on biopsy—often without a focal gray-scale abnormality. Although the authors suggested that color Doppler imaging may help identify an appropriate site for biopsy, they stated that a negative color Doppler imaging scan should not preclude biopsy, since it has a limited sensitivity in the detection of all sites of cancer.
Patel and Rickards[51] investigated the discriminative value of the amount of color flow on Doppler ultrasound within the peripheral zone of the prostate. The histologic outcome of 274 guided biopsies was correlated with the grade of color flow on ultrasound hard-copy images. They noted that normal color flow was seen with both normal and abnormal prostate biopsies. Results of the study demonstrated that with the greatest color flow, specificity for the diagnosis of an abnormal prostate (either cancer or prostatitis) is very high, and that with prostatitis, a markedly increased color flow reflects the severity of inflammatory cellular reaction. Nevertheless, the authors concluded that grading the amount of color flow with Doppler ultrasound is of limited diagnostic usefulness.
Predictive Value of Tumor Vascularization: The prognostic significance of detecting an increased flow within a given tumor has also been investigated. Findings of several studies have indicated that tumor vascularization may correlate with its potential for rapid growth and distant metastasis.[44-47] Brawer et al[46] demonstrated that the histologic determination of microvessel density in prostatic carcinoma is an independent predictor of pathologic stage, and, hence, malignant potential.
Thus, quantification of tumor angiogenesis may guide prostate cancer treatment strategies. This issue was partly addressed in a study by Littrup et al,[52] in which cancers with a high color flow had significantly higher mean Gleason scores than cancers without demonstrable color flow, particularly in African-American men.
Our group previously demonstrated that color-coded Doppler flow within the tumor correlates with both tumor grade and stage, and that increased flow is associated with a higher Gleason score and higher incidence of seminal vesicle invasion. Our study findings also indicated that increased flow noted on color Doppler imaging is independently predictive of the likelihood of biochemical relapse following radical prostatectomy.[53]
Drawbacks of Color Doppler Imaging: Although studies suggest that color Doppler imaging has potential prognostic significance, it still has two major pitfalls: overlap with prostatitis and low sensitivity in the detection of tumor blood flow within a prostatic carcinoma. Its failure to detect tumor blood flow may be partly attributed to the fact that the presence or absence of tumor flow may be influenced by tumor size or volume, with tumors smaller than 2 mm in diameter being avascular and those at least 1 cm3 possessing vascularity.[54] Other technical factors may also contribute to the failure of tumor blood flow detection: (1) The limited spatial resolution on color Doppler makes it difficult to detect blood flow in very small vessels, and (2) low-volume flow results in frequency shifts below the noise level, and consequently, cannot be detected.[13]
Power Doppler: More recently, the amplitude or power of the Doppler shift (known as power Doppler) has been encoded into color Doppler and used to detect flow. The main advantage of power Doppler is its ability to detect slower flow with less reliance on the Doppler angle.[55]
Our group recently compared the accuracy of gray-scale TRUS, color Doppler imaging, and power Doppler imaging in the detection of prostate cancer, and then assessed the influence of operator experience on test results.[14] Our study indicated that gray-scale TRUS and Doppler imaging are minimally effective in prostate cancer detection, that there is no benefit of power Doppler over color Doppler imaging, and that there is no apparent benefit from increased operator experience with Doppler imaging. We did, however, observe that with gray-scale TRUS or Doppler, foci abnormalities were 2.5 times more likely to contain cancer than adjacent tissues with normal ultrasound findings. While Doppler flow studies may provide some prognostic data, the routine use of this technique for the diagnosis of prostate cancer does not appear to be warranted at the present time.
