The utility and limitations of contrast-enhanced transrectal ultrasound scanning for the detection of prostate cancer in different area of prostate

1 Introduction

Prostate cancer is currently a common solid neoplasm in older men in western countries [1]. In recent decades in Asian countries, with the widely screening of serum prostate specific antigen (PSA), the aging of the population and the westernization of lifestyle, the incidence of prostate cancer is increasing rapidly [2]. The diagnosis of prostate cancer is confirmed by transrectal ultrasound (TRUS)-guided biopsy base on the findings of serum PSA measurement, digital rectal examination (DRE) and imaging techniques. Various imaging tools such as multi-parameter magnetic resonance imaging (mp-MRI), elastography and contrast-enhanced TRUS (CETRUS) are under investigation and proved to play increasingly important roles in the detection of prostate cancer [3–9]. CETRUS is a minimally invasive diagnostic tool, which has been used widely to enhance the visualization of the microvasculature associated with prostate cancer. Multiple studies have shown that CETRUS increase the effciency of TRUS-guided biopsies and could be used for follow-up after focal interventional therapy of prostate cancer [5–11].

Currently a major problem is that real-time fixed-plane contrast-enhanced imaging could only observe one plane at a time and is somewhat difficult to screen the entire prostate for the time and economic cost. However, the prostate cancer has a tendency to be multifocal [5, 12]. To choose one or more (e.g. three) planes with or without any suspicious lesions on the conventional TRUS to perform the real-time fixed-plane CETRUS examination is not enough, as between the selected planes, a small cancerous lesion might exist and could not be visualized. Our previous studies performed the CETRUS using the approach of real-time scanning to examine the whole gland in a short time (90 s) with a dose of 2.4 mL of Sulfurhexafluorid microbubbles contrast agent (SonoVue, Bracco SpA, Milan, Italy) [7, 8]. The results showed that this method could improve cancer detection over conventional imaging and detect higher grade prostate cancers. However, the detection of prostate cancer by contrast-enhanced imaging varied in different areas of prostate. Halpern et al. [13] demonstrated that a significant decrease in carcinoma detection by harmonic imaging targeted cores toward the apex of the prostate. We wondered whether there is the same problem with the real-time CETRUS scanning. Thus, the purpose of this study was to evaluate the ability of the real-time CETRUS scanning in the detection of prostate cancer in different areas of prostate and to discuss whether or not it has an advance over conventional TRUS.

2 Material and methods

This study was approved by the Ethics Committee of our hospital and written patient informed consent was obtained from each patient before the examination and biopsy. From July 2012 to February 2014, 228 patients (age range: 47–86 years; median: 68 years; PSA range: 4.1–20.0 ng/mL; median: 10.2 ng/mL; prostate volume range: 13.1–170.2 mL; median: 47.9 mL) underwent initial prostate biopsy were enrolled. The indication for biopsy was elevated serum PSA level of 4.0–20 ng/mL, an abnormal digital rectal examination finding, or both. Study exclusion criteria were as follow: clinical prostatitis within 1 month of biopsy, active urinary tract infection, severe cardiorespiratory disease or contraindications to the Sulfurhexafluorid microbubbles contrast agent. The preparation before biopsy is the same as the previous literature [7, 8].

An IU 22 machine (Philips Healthcare, Bothell, WA) with a C10-3v endorectal probe was used in this study. The procedure of ultrasonographic examinations was based on our previous experience [7, 8] and performed by an experienced investigator (S.W.X.), with 8 years of experience in conventional TRUS and CETRUS. Before contrast agent administration, conventional TRUS (Gray-scale and power Doppler) was performed first to scan the prostate and find the suspicious areas. Gray-scale examination followed a standard protocol of transverse imaging from base to apex, followed by sagittal imaging from right to left. The imaging settings such as gain, depth, and focus were optimized for each patient. The color window sector width of power Doppler imaging was adjusted to include the entire prostate, followed a standard protocol of transverse imaging from base to apex with the lowest possible pressure. The Doppler imaging settings were adjusted to maximize the visualization of color flow within the prostate without background noise.

After conventional TRUS examinations were completed, the real-time CETRUS scanning was carried out. A dual-scan mode was used so that CETRUS and low mechanical index (MI) gray-scale TRUS images were shown on the screen simultaneously to assure good overlap of these 2 modalities. The settings of the machine parameters were as follows: MI, 0.06; gain, 60%; focus, one-point focus, located in the middle-anterior part of the gland. Sulfurhexafluorid microbubbles contrast agent was used as the ultrasound contrast agent. The microbubbles was prepared in a standard fashion and infused intravenously with a dosage of 2.4 mL, followed by a 0.9% normal saline solution flush (5 mL). Scanning was done with the lowest possible pressure on the prostate to minimize the effects from probe pressure on prostatic vascularity as far as possible. The timer was activated at the beginning of microbubbles administration. Scanning was started 2-3 seconds after microbubbles reach to the prostate capsule, followed a standard sequence of transverse imaging from base to apex (about 4-5 seconds), then from apex to base, and repeated. Scanning was lasted continuously for about 90 seconds. The entire examination lasted about 10–15 minutes. Conventional TRUS and the whole real-time CETRUS scanning process were stored on the hard disk incorporated in the scanner.

The images of conventional TRUS were analyzed by two experienced investigators (B.J.D. and S.J.Z., with 5 and 9 years of experience in conventional TRUS, respectively), who were blinded to the clinical and CETRUS information. If they did not agree on the evaluation results, the images were evaluated by another experienced investigator (J.G.X, with 13 years of extensive experience in conventional TRUS) until a consensus was reached. The images of CETRUS were analyzed by two experienced investigators (H.L.L. and J.D., with 9 and 10 years of experience in CETRUS, respectively), who were blinded to the clinical and conventional TRUS information. If they did not agree on the evaluation results, the images were evaluated by another experienced investigator (F.H.L., with 11 years of extensive experience in CETRUS) until a consensus was reached.

As transition zone (TZ)-tumors represent a special diagnostic challenge to contrast-enhanced imaging; we mainly focused on the peripheral zone (PZ), and used the following criteria on the basis of our clinical experience [7, 8]. Conventional TRUS: 1) focal hypoechoic lesion; 2) isochoric mass evidenced by focal contour distortion in the PZ; 3) ill-defined of PZ and TZ; 4) focal asymmetric/increased flow. CETRUS: 1) rapid contrast enhancement (compared with TZ, contralateral half of the PZ or ipsilateral PZ tissue) (Fig. 1); 2)increased contrast enhancement (Fig. 1); 3) no/low enhancement; 4) asymmetric appearance of intraprostatic vessels; 5) diffused appearance(included: ill-defined of PZ and TZ; diffused rapid and increased contrast enhancement of the PZ, nearly as same as the TZ; distorted intraprostatic vessels; with or without no/low enhancement; etc.). Marks (urethra, cyst, calcification, nodule, area of heterogeneous echotexture, contour deformity in low MI gray-scale TRUS images) were recorded to assure good overlap of CETRUS and biopsy US.

OPEN IN VIEWER

Fig.1

A 70 years old man with PSA level of 6.48 ng/ml, prostate volume of 35.5 ml. A, Grey-scale image demonstrates no echotexture abnormality, contour deformity or ill-defined of peripheral zone and transition zone. B,Power Doppler image demonstrates no significant increased flow in the prostate. C,Contrast-enhanced transrectal ultrasound scanning image demonstrates a rapidly enhancing lesion with greater contrast enhancement in the right base, mid-gland and apex (C1, C2, C3, respectively) (arrows). Biopsies targeted to these suspicious areas revealed Gleason 7 prostate cancer.

TRUS–guided biopsy was performed by 2 investigators (S.W.X. and W.X.). When there was no abnormal finding, a 10-site biopsy was taken randomly, five from each prostate side (two sites from the base, two sites from the mid-gland, one site from the apex). And then if any suspicious lesion indicated by conventional TRUS or CETRUS, targeted biopsy samples were taken from the corresponding site. Biopsy specimens were labeled according to the gland location and fixed with a 10% formaldehyde solution in separate bottles, then reviewed by an experienced pathologist. Each biopsy specimen was reported as prostate cancer with an assigned Gleason score, or as prostatic intraepithelial neoplasia, prostatitis, or benign prostatic tissue. All no cancerous findings were grouped together.

All statistical calculations were performed using SPSS 20.0 software (SPSS, Chicago, IL) with P < 0.05 was considered as statistically significant. Significant prostate cancer was defined as Gleason score ≥ 7(3+4). The McNemar test was used to evaluate the differences in detection of prostate cancer (overall, significant, and insignificant) between conventional TRUS and CETRUS. The McNemar test and chi-square test were used to compare the differences between conventional TRUS and CETRUS for the detection of prostate cancer in different areas (right, left; base, mid-gland, apex) on a per-site basis. The receiver operating characteristic (ROC) curve was plotted to evaluate the diagnostic performance of CETRUS for prostate cancer detection in different areas. The diagnostic performance was expressed as the area under the ROC curve (AUC).

The overall per-patient cancer detection rate with biopsy was 39.0% (89 of 228). 48 (53.9%) of these were significant cancer. The overall cancer detection rate for CETRUS-targeted biopsies and conventional TRUS-targeted biopsies was 32.9% (75/228) versus 27.6% (63/228) (P = 0.008; McNemar test). The significant cancer detection rate for CETRUS-targeted biopsies and conventional TRUS-targeted biopsies was 21.1% (48/228) versus 18.4% (42/228) (P = 0.031; McNemar test) and the insignificant cancer detection rate for CETRUS-targeted biopsies and conventional TRUS-targeted biopsies was 11.8% (27/228) versus 9.2% (21/228) (P = 0.146; McNemar test), respectively. 11 (4.8%) patients were only detected for nontargeted biopsies, and all these targeted-missed cases were Gleason 3 + 3 = 6 cancer (Table 1).

On a per-biopsy basis, of all 2280 specimens, prostate cancer was detected in 376 sites (16.5%) from 89 patients. Findings in 273 (60.4%) of 452 CETRUS scanning-targeted biopsies were positive for prostate cancer, while 203 (47.3%) of 429 baseline TRUS-targeted biopsies had findings positive for prostate cancer. 183 (48.7%) of 376 sites diagnosed with cancer were detected with both CETRUS scanning and baseline TRUS-targeted biopsies. In 83 (22.1%) sites of prostate cancer, no suspicious findings were found in both CETRUS scanning and baseline TRUS. The average number of targeted biopsies per patient was 4.0 cores (616 cores/153 cases). 180 (47.9%) and 196 (52.1%) sites were diagnosed with cancer in right and left lobe, respectively. For different planes, cancer was detected with the highest frequency in apex (89/456, 19.5%), followed by mid-gland (162/912 17.8%) and base (125/912, 13.7%). The outcome of histology and ultraonography by biopsy site according to the area of prostate are shown in Table 2.

CETRUS scanning had higher sensitivity, specificity (except right lobe), accuracy, positive predictive value (PPV) and negative predictive value (NPV) than baseline TRUS in total, right and left lobe (all P-values<0.05; McNemar test; chi-square test). The sensitivity were statistically significant greater for CETRUS scanning than for baseline TRUS in all areas except left base and right apex (all P-values<0.05; McNemar test). The accuracy were statistically significant greater for CETRUS scanning than for baseline TRUS in all areas except left mid-gland and right apex (all P-values<0.05; McNemar test). The specificity was statistically significant greater for CETRUS scanning only in left apex (P = 0.007; McNemar test). The PPV was statistically significant greater for CETRUS scanning than for baseline TRUS in right base and left apex (P = 0.011 for right base; P = 0.005 for left apex; chi-square test). The NPV was statistically significant greater for CETRUS scanning only in right base (P = 0.04; chi-square test) (Table 3).

ROC analyses for detection of prostate cancer with CETRUS scanning according to the area of prostate are shown in Table 4. ROC analysis showed the CETRUS Scanning totally got the AUC of 0.816. The AUC of CETRUS was higher in left lobe than right lobe (0.837 vs. 0.793). It was most accurate at the base (0.833) of the prostate, then mid-gland (0.826), and lowest in apex (0.772).

At present, patients with elevated serum PSA level are often subjected to multiple sites prostate biopsy guided by TRUS. However, conventional gray scale or Doppler imaging TRUS had some limitations for the detection of prostate cancer based on their unsatisfactory sensitivity and accuracy [14–16]. As the tissue of prostate cancer is associated with increased microvascular density [17, 18], microbubbles which could pass through the pulmonary circulation and reach the microvasculature of organs had been used to detect the malignant prostate nodules. The microbubbles were first used to enhance the Doppler techniques, supposedly increasing the sensitivity from 65 to 93% [19–22]. And many reports showed that compared with systematic biopsy, enhanced Doppler imaging targeted biopsy improved the detection of prostate cancer with fewer biopsy cores and could detect higher Gleason scores cancers [23–25]. However, Taverna et al. raised that prostate cancer detection rate does not significantly improve with the use of colour Doppler ultrasonography with or without contrast agent [26]. The acoustic power with Doppler imaging may destroy a large proportion of microbubbles before they reach the microvascular. Therefore, diagnostic techniques need to be improved.

Contrast-specific harmonic imaging techniques use a lower MI energy of insonation to evaluate microbubbles without destroying them. Therefore microbubbles could reach the microvascular within cancerous lesions and could increase the visualization of microvasculature. As many studies showed, Contrast-specific harmonic imaging was often performed using the approach of fixed-plane which at a time just one plane can observed [5, 12]. It required about 2 min to finish the examination of each plane and 3–5 min between each injection to allow sufficient microbubbles breakdown. And a dose of 2.4 mL of Sulfurhexafluorid microbubbles contrast agent is needed each time. So it is somewhat difficult to screen every plane in the prostate for the time consumption and economic cost. However, the datas indicated that specimens obtained by radical prostatectomy showed the majority cases had multifocal tumors [5, 12]. The insufficient fixed-plane CETRUS examination might miss the small cancerous lesions.

We tried to use the real-time CETRUS scanning to examine the whole gland in a short time (90 s) with a dose of 2.4 mL of Sulfurhexafluorid microbubbles contrast agent. The results of the previous studies revealed significant improvement of the cancer detection with CETRUS scanning over conventional TRUS and higher grade prostate cancers were detected by CETRUS scanning [8]. And CETRUS scanning performed better in patients with lower PSA level and modest prostate volume [7]. In the present study, we evaluated the diagnostic capabilities of real-time CETRUS scanning and conventional TRUS for the detection of prostate cancer in different areas of prostate, and compared with pathological findings of biopsy. The results of the study revealed that compared with conventional TRUS, CETRUS could increase the detection rates of overall and significant cancer; CETRUS had a significant advantage for detecting prostate cancer in different areas. These results suggested that CETRUS scanning could be used to guide biopsy and improve the detection of prostate cancer but not obviously increased the cost and mean time of the examination.

Halpern et al. [27] evaluated the effect of patient position on the observed flow pattern with high-frequency Doppler TRUS and reported that flow asymmetry in patients who underwent biopsy may have been related to patient position. Likewise, in the present study, ROC analysis showed that the AUC of CETRUS was higher in left lobe than right lobe. It seemed that the patient position might also impact the enhancement of CETRUS. More investigations are needed.

The results revealed that CETRUS scanning performed best in base, then mid-gland, apex lastly. Our finding was in agreement with that of Halpern et al. [13], who demonstrated that the majority of positive sites missed by targeted biopsy were found to be at the apex of the prostate by harmonic imaging. These findings suggest that we must focus more on the apex. The reasons for the unsatisfactory efficiency of CETRUS scanning in apex may be as follows: First, the component of apex is mainly the PZ, lacked the comparison of the TZ, which may adverse to the detection of prostate cancer; Second, we performed the scanning from base to apex in a short time, then from apex to base, and repeated. We may miss some small and extremely rapid enhancement lesions in apex. And repeat intravenous injections of contrast agent or adding other imaging modalities may improve the visualization and detection of cancer in the prostate in future.

In previous reports, CETRUS findings suggestive of cancer have been defined as rapid contrast enhancement, increased contrast enhancement, no/low enhancement and and asymmetric appearance of intraprostatic vessels [5, 8]. However, as our experience, we found a portion of prostates had multiple suspicious findings, included the following: ill-defined of PZ and TZ; diffused rapid and increased contrast enhancement of the PZ, nearly as same as the TZ; distorted intraprostatic vessels; with or without no/low enhancement. These prostates had great possibilities to be multiple-site (more than 6 sites) and more aggressive (Gleason score ≥7) prostate cancer. So we added this diffused appearance we called as one of the suspicious findings. We considered that when this appearance existed, fewer targeted biopsy cores could be performed which may reduce the biopsy related risks and patient complaints.

The present study is limited by the fact that we just compared ultrasonographic findings with the histopathlogic outcome of 10-core biopsy; in addition, lesions that were not visible on TRUS imaging were not biopsied and may have not been detected by systematic biopsy. This might led to false evaluation of the diagnostic accuracy of CETRUS. Further study correlated with thin-section, whole-mount prostatectctomy is needed. The second limitation is that the targeted and nontargeted biopsies were done by the same operators, which could cause a bias towards nontargeted biopsies.

In conclusion, the CETRUS scanning had a significant advantage over conventional TRUS for the detection of prostate cancer in different areas. On the basis of that CETRUS scanning much more easily missed the prostate cancer in apex; we must focus more on the apex and may add other imaging modalities to improve the visualization and detection of cancer in the prostate.