CEUS in the differentiation between low and high-grade bladder carcinoma

Abstract

Introduction

Bladder cancer ranks 4th overall in the number of newly diagnosed cancers and 10th in causes of cancer deaths. More than 90% of all cases of bladder cancer are transitional cell carcinoma (TCC). The goal of this study is to confirm the usefulness of low mechanical index contrast-enhanced ultrasonography (CEUS), also in association with time–intensity curves, in the differentiation between high- and low-grade bladder malignant lesions.

Materials and methods

From February 2006 to February 2012 we recruited 144 patients. All patients underwent grayscale ultrasonography (US), color-Doppler ultrasonography (CDUS) and contrast-enhanced ultrasonography (CEUS). Subsequently all patients underwent cystoscopy and TURB.

Results

Histological diagnoses were: 88 high-grade carcinomas (61.1%), and 56 low-grade carcinomas (38.9%). Sensitivity and specificity of CDUS were 87.5% (126/144) and 60%, respectively. Sensitivity and specificity of CEUS were 90.9% and 85.7%, respectively. Sensitivity and specificity of TIC were 91.6% (132/144) and 85.7%, respectively.

Discussion and conclusions

CEUS is a reliable noninvasive method for differentiating low- and high-grade bladder carcinomas since it provides typical enhancement patterns as well as specific contrast-sonographic perfusion curves.

Introduction

Bladder cancer ranks 4th overall in the number of newly diagnosed cancers and 10th in causes of cancer deaths. More than 90% of all cases of bladder cancer are transitional cell carcinoma (TCC), with most of the remainder squamous cell carcinoma (5%), adenocarcinoma (2%), or rhabdomyosarcoma (1%). Thirty percent of patients with bladder tumors initially present with muscle-invasive or metastatic disease. Of these, 50% will die 2–3 years after the diagnosis despite aggressive local therapy. On the other hand, 70% of TCC are referred to as non-muscle-invasive or superficial bladder cancer at the initial presentation. These are tumors confined to the mucosa (70%) or lamina propria (30%). Approximately 50%–70% of these tumors will recur, with 10%–30% showing grade and stage progression [1-6].

Transurethral resection of bladder tumor (TURB) is the only procedure which allows the execution of a correct local staging. The goal of TURB is to enable a correct diagnosis by the pathologist, which means including bladder muscle in the adequately sized resection biopsies. The strategy of resection depends on the size of the lesion. Small tumors (less than 1 cm) can be resected en bloc, where the specimen contains the complete tumor plus a part of the underlying bladder wall including bladder muscle. Larger tumors have to be resected separately in fractions, which include the exophytic part of the tumor, the underlying bladder wall with the detrusor muscle and the edges of the resection area. At least the deeper part of the resection specimen must be referred to the pathologist in a separate labeled container to enable him to make a correct diagnosis. Furthermore the resection should include the removal of the bladder wall surrounding the lesion for at least 1 cm of diameter, to allow an assessment of eventual alterations of the mucosa [7–11].

Over recent years ultrasonographic contrast media have been used in order to assess the neoangiogenesis of these neoplasms; the goal of this study is to understand if it is possible to distinguish high-grade from low-grade lesions using specific contrastographic parameters, in order to provide better information for the execution of TURB [12].

Materials and methods

From February 2006 to February 2012 we recruited 144 patients, age range 51–76 (mean age 68), 48 (33.3%) women and 96 (66.6%) men. Urinary cytology was positive in 96 (66.6%) of cases.

All patients underwent grayscale ultrasonography (US), color-Doppler ultrasonography (CDUS) and contrast-enhanced ultrasonography (CEUS). Subsequently all patients underwent cystoscopy and TURB. The time between CEUS and TURB was less than 15 days. Sonographic examination was performed on Technos Mpx, Mylab 70 Vx Gold (Esaote, Genova, Italy), and Toshiba Aplio VX (Osaka, Japan) using a 2–5 MHz multi-frequency broadband convex transducer. Examination was performed when the urinary bladder was adequately full (neither empty nor overfilled, but sufficiently full in order not to overestimate the lesions). For each session sonographic examination consisted of conventional grayscale imaging using tissue harmonic imaging (THI).

The location and size of the lesions detected during grayscale examination were recorded. CDUS was performed using scanning parameters set for maximum sensitivity to slow flow while the power output was increased to maximum. The color gain was increased until just prior to the appearance of random noise. The pulse repetition frequency was set at the lowest possible level.

All patients subsequently underwent real-time continuous examination using a low mechanical index (MI) (range: 0.04–0.1) after administration of SonoVue (Bracco, Milan, Italy). SonoVue was administered as an intravenous hand-injected fast bolus of 2.4 ml, followed by 5 ml of a saline flush. Images in the ideal scanning plane were displayed in real-time by slightly changing the scanning plane to portray the whole area of the lesion and the surrounding bladder wall. The true subtraction effect can be obtained only by using a low MI. This technique is based on signal amplitude subtraction (not only frequency subtraction) and is possible thanks to the combination of two different components: the harmonic signal coming from SonoVueTM and the dynamic threshold of low amplitude signals which suppress the low amplitude signals returning to the transducer.

The entire examination was digitally recorded with still images recorded on the optic disk. These images were subsequently reviewed on a frame-by-frame basis. The number, size, location and morphological features of the lesions were studied.

Quantitative analysis was subsequently performed using perfusion software.

TURB, executed by one operator, and the successive histologic exam lead to the finding of 8 high-grade Ta (5.6%), 56 low-grade Ta (38.8%), 72 high-grade T1 (50%) e 8 high-grade T2 (5.6%).

Imaging analysis

Grayscale conventional US was used to characterize the lesions of the bladder wall as sessile or polypoid masses. Lesions with a broader base compared to their height were classified as sessile masses, while lesions which were taller than they were wide at their base were classified as polypoid masses. Using CDUS, the patterns of the vascular structures were assigned a score and classified as follows: absent (score 0), mild (one vascular pole; score 1) or strong (more than one vascular pole; score 2).

Images were analyzed off-site at consensus conferences to determine the size of the lesion, baseline and CDUS appearance as well as signal intensity of the enhancement after Sonovue administration. The enhancement pattern was assigned a score: 1 (mild enhancement) or 2 (strong enhancement) as compared to the surrounding urinary bladder wall.

The images were reviewed by two radiologists well skilled in urological sonography and with more than 5 years of experience in CEUS examinations. Perfusion software was then used to estimate the perfusion pattern (Qontrast v3.0, Bracco, Milan, Italy). This quantitative software analyzes the temporal sequence of the images. The signal intensity is calculated for every pixel for every second, thus generating maps of perfusion parameters. The most important parameters are: maximum intensity value (SI max); enhancement time (time needed to reach 50% of value) and slope (the rapidity of the signal intensity growth rate).

Time-intensity curves (TIC) were extracted from regions of interest in the lesion and the closest bladder wall, and TIC qualitative analysis was performed. Three TIC shapes were identified: shape A = wash-in phase followed by a plateau (normal bladder wall); shape B = wash-in phase followed by a slow wash-out phase (low-grade carcinoma); shape C = wash-in phase followed by a rapidly descending curve (high-grade carcinoma).

All patients underwent TURB. Histological examination classified the lesions according to the 2004 WHO classification. In histopathological staging, grade I tumors were classified as “low-grade” and grades II and III tumors as “high-grade”.

Results

No clinically significant contrast agent related side effects were experienced by the patients, who well tolerated the sonographic examination.

Histological diagnoses were: 88 high-grade carcinomas (61.1%), and 56 low-grade carcinomas (38.9%).

There were 80 polypoid lesions (55.6%) and 64 sessile lesions (44.4%). Among the 80 polypoid lesions, 32 were high-grade (40%) and 48 were low-grade (60%).

Among the 64 sessile lesions, 56 were high-grade (87.5%) and 8 were low-grade (12.5%).

Mean size of the low-grade lesions (1.75 ± 0.62 cm) was significantly smaller than the mean size of the high-grade lesions (2.74 ± 0.81 cm) (p < 0.001; Student's t-test).

At CDUS 18 lesions showed no flow signal (score 0), 12/18 were low-grade bladder cancer and 6/18 were high-grade; 73 presented mild vascularization (score 1), 30/73 were low-grade, while 43/73 were high-grade; 53 were characterized by an intense vascularization and assigned score 2, 39/53 were high-grade and 14/53 were low-grade. Sensitivity and specificity of CDUS were 87.5% (126/144) and 60%, respectively.

At CEUS Intensity score analysis, 80/88 high-grade carcinomas were score 2, and 8/88 were score 1; 48/56 low-grade carcinomas were score 1, and 8/56 score 2. Sensitivity and specificity of CEUS were 90.9% and 85.7%, respectively.

Three TIC shapes were identified: shape A = wash-in phase followed by a plateau (normal bladder wall); shape B = wash-in phase followed by a slow wash-out phase (low-grade carcinoma); shape C = wash-in phase followed by a rapidly descending curve (high-grade carcinoma). All low-grade carcinomas except 8 followed shape B, while all high-grade carcinomas except 4 followed shape C. Therefore, the sensitivity and specificity of TIC were 91.6% (132/144) and 85.7%, respectively.

There were no significant differences between sensitivity of CDUS and CEUS, between CDUS and TIC's shape, and between CEUS and TIC's shape (p > 0.05; McNemar's test). Specificity of CEUS and TIC's shape was significantly higher than that of CDUS (p < 0.05; McNemar's test).

There was no significant difference between the specificity of CEUS and TIC (p > 0.05; McNemar's test).

Discussion

We used a second generation contrast medium, Sonovue, and a quantitative tool, Qontrast, which provides automatic signal–intensity curves, previously used for differentiating breast lesions.

Unlike CT and MRI contrast agents, which diffuse into the extravascular space to enter the equilibrium phase, SonoVue is a blood pool agent and does therefore not have a true equilibrium phase. In the present study, CEUS with SonoVue was a useful tool in the differentiation between low-and high-grade bladder cancer lesions, since it significantly increased the sensitivity of CDUS from 87.5% to 90.9% using only CEUS enhancement intensity score and to 91.6% if also TICs were considered. CEUS led to a significant improvement of the specificity of CDUS from 60% to 85.7% (p < 0.05). However, there were no significant differences in the sensitivity of CDUS versus CEUS, CDUS versus TIC, and CEUS versus TIC (p > 0.05; McNemar test).

It should be underlined that the new quantitative software seems to be a reliable tool which increases the usefulness of contrast-enhanced US. It analyses temporal sequences of images, thus calculating the signal intensity for every pixel for every second resulting in maps of perfusion parameters.

Although both low- and high-grade malignant lesions are hypervascular, some characteristic features are helpful in differentiating them, such as enhancement score and particularly TIC shapes. In fact, 80/88 cases of high-grade carcinoma were score 2, and 8/88 score 1; 48/56 cases of low-grade carcinoma were score 1, and 8/56 score 2.

However, more interestingly, significantly higher sensitivity and specificity were achieved by using TIC. Three TIC shapes were identified; all of them had a wash-in phase followed by shape A, a plateau (normal bladder wall), shape B, a slowly declining curve indicating low-grade carcinoma; shape C, a rapidly declining curve indicating high-grade carcinoma. All low-grade carcinomas except 8 showed shape B, while all the high-grade carcinomas except 4 showed shape C. High-grade bladder cancers are characterized by shape C TIC corresponding to late enhancement peak (>28 s), higher SI (SI max >50%), and by a fast wash-out (Fig. 1); whereas low-grade carcinomas present faster wash-in (<26 s), a lower maximum SI <45% and a slow wash-out (Fig. 2).

Figure 1.

High grade bladder tumor. (a) Grayscale US. (b) Color-Doppler shows strong flow signal inside the lesion (score 2). (c) CEUS shows mild enhancement of the lesion (score 1). (d) Time-intensity curve shows a shape C represented by a slow wash-in, with high SI and fast wash-out.

Figure 2.

Low-grade bladder tumor. (a) Grayscale US. (b) Color-Doppler shows one pole of flow signal inside the lesion (score 1). (c) CEUS shows mild enhancement of the lesion (score 1). (d) Time-intensity curve shows a shape B revealing fast wash-in, but lower level of SI, with slow wash-out.

However, CEUS has some drawbacks. One is the limited ability to depict flat lesions, such as carcinoma in situ; however, this issue was not addressed in our paper. However, the combination of contrast-enhancement pattern and the TIC shapes could indicate the histopathology of bladder lesions, thus increasing the specificity of Color-Doppler US without a statistically significant reduction of sensitivity.

Conclusions

CEUS is a reliable noninvasive method for differentiating low- and high-grade bladder carcinomas since it provides typical enhancement patterns as well as specific contrast-sonographic perfusion curves. It should be underlined that the new quantitative software seems to be a reliable tool which increases the usefulness of contrast-enhanced US.

In conclusion, the combination of contrast-enhancement pattern and TIC shapes may indicate the histopathology of bladder lesions, thus increasing the specificity of color-Doppler US without a statistically significant reduction in sensitivity.

Conflict of interest statement

The authors have no conflict of interest to declare with respect to this article.