| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Original Research |
1 Department of Radiology and Center for Imaging Science, Samsung Medical
Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-gu,
Seoul, Republic of Korea 135-710.
2 Department of Urology, Samsung Medical Center, Sungkyunkwan University School
of Medicine, Seoul, South Korea.
3 Department of Radiology, Kangwon National University School of Medicine,
Kangwon-do, South Korea.
4 Department of MRI, DIS Business Group, Philips Healthcare Korea, Seoul, South
Korea.
Received July 23, 2007;
accepted after revision November 6, 2007.
Address correspondence to C. K. Kim
(chankyokim{at}skku.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Twenty-seven patients who had increased levels of prostate-specific antigen after high-intensity focused ultrasonic ablation underwent MRI and endorectal biopsy. The MR images and biopsy results were correlated for six prostate sectors. Residual or recurrent prostate cancer after treatment was defined as local tumor progression if the biopsy results showed cancer foci. Two readers blinded to the clinical findings and biopsy results used a 5-point scale to independently assess DCE-MR images and T2-weighted and diffusion-weighted MR images. The results were compared by use of the McNemar test with Bonferroni correction, generalized estimating equations, and receiver operating characteristic analysis.
RESULTS. After high-intensity focused ultrasonic ablation, local tumor progression was pathologically detected in 54 (33%) of 162 sectors in 18 patients. The sensitivities of DCE-MRI and T2-weighted MRI with DWI were 80% and 63% for reader 1 (p = 0.004) and 87% and 70% for reader 2 (p = 0.004). The specificities of DCE-MRI and T2-weighted MRI with DWI were 68% and 78% for reader 1 (p = 0.002) and 63% and 74% for reader 2 (p < 0.001). The accuracy rates of DCE-MRI and T2-weighted MRI with DWI were 72% and 73% for reader 1 (p > 0.05) and 71% and 73% for reader 2 (p > 0.05). The areas under the receiver operating characteristic curve for DCE-MRI and T2-weighted MRI with DWI were 0.77 and 0.77 for reader 1 and 0.85 and 0.81 for reader 2.
CONCLUSION. For prediction of local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation, DCE-MRI was more sensitive than T2-weighted MRI with DWI, but T2-weighted MRI with DWI was more specific than DCE-MRI.
Keywords: diffusion-weighted MRI • dynamic contrast-enhanced MRI • high-intensity focused ultrasonic ablation • localized prostate cancer • MRI
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
High-intensity focused ultrasonic ablation causes coagulation necrosis in the targeted tissue by converting mechanical energy into heat and generating a cavitation effect [8]. The procedure is a valuable alternative in the management of well- and moderately differentiated localized prostate cancer and of local recurrence after external beam radiation therapy [9–11]. After high-intensity focused ultrasonic ablation for clinically localized prostate cancer, accurate assessment of residual or recurrent cancerous tissue is important for the planning of second-line treatment. A recent investigation [12] showed that endorectal color Doppler sonography was useful for detecting early cancer recurrence in patients who did not undergo hormonal therapy after high-intensity focused ultrasonic ablation. The coagulation necrosis and fibrosis induced by this procedure cause the tissue to become less vascular than normal prostate tissue [13, 14].
Dynamic contrast-enhanced MRI (DCE-MRI) has been reported [15] to be an effective tool for assessing the pharmacokinetics of uptake of contrast material in the prostate. Cancerous prostate tissue has enhancement patterns different from those of benign tissue (i.e., cancer usually causes nodular enhance ment before enhancement of the rest of the parenchyma and early washout of signal intensity) [16–18].
Diffusion is caused by random translational molecular motion, also known as brownian water motion [19]. Several preliminary studies have shown that prostate cancer has a lower apparent diffusion coefficient (ADC) than noncancerous prostate tissue [20–22]. Thus diffusion-weighted imaging (DWI) is showing potential for improving detection of prostate cancer [20–22].
To the best of our knowledge, no investigation of DCE-MRI or DWI for the detection of residual or recurrent prostate cancer after high-intensity focused ultrasonic ablation for localized cancer has been described in the literature. The purpose of this study was to compare the diagnostic performance of DCE-MRI with that of T2-weighted MRI with DWI for predicting local tumor progression after high-intensity focused ultrasonic ablation of localized prostate cancer.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
High-intensity focused ultrasonic ablation was performed with spinal anesthesia. A suprapubic catheter was placed preoperatively to sustain adequate urinary drainage. Transurethral re section of the prostate was undertaken to reduce prolonged urinary retention before high-intensity focused ultrasonic ablation. Immediately after transurethral resection, the targeted volume boundaries were localized, and contiguous shots of high-intensity focused ultrasound were delivered slice by slice from the apex to the bladder neck to treat the entire prostate. The suprapubic catheter was removed after the procedure if the postvoid residual urine volume was less than 50 mL.
Patients
Between March 2004 and October 2006, 84 patients whose condition was
diagnosed as clinically localized prostate cancer were treated with endorectal
high-intensity focused ultrasonic ablation preceded by transurethral resection
of the prostate to reduce prolonged urinary retention. The selected population
included patients with clinical stage T1–2 disease, a prostate-specific
antigen (PSA) level less than 30 ng/mL, and no evidence of metastasis on
preoperative prostate MRI or bone scan. Radical prostatectomy was suitable for
none of the patients because of age, presence of comorbid conditions, or
choice not to undergo surgery. The PSA level was assessed 1 month post
operatively, then every 3 months in the first year after the procedure and
every 6 months there after. If an increasing PSA level equaled or exceeded 1.0
ng/mL, the PSA level was measured more frequently than the scheduled interval.
Disease-related risk was classified into three groups according to
pretreatment clinical stage, PSA level, and Gleason score.
Twenty-seven of the 84 patients (mean age, 68.5 years; range, 60–82 years) who had consecutively increasing PSA levels or a marked increase in PSA level underwent follow-up MRI for detection of residual or recurrent prostate cancer. These patients formed the study group. After MRI examinations, all 27 patients underwent prostate biopsy under transrectal sonographic (TRUS) guidance. Before the MRI examinations, the last mean level of PSA in all of the patients was 2.97 ± 2.4 (SD) ng/mL (range, 0.42–10.8 ng/mL). The mean interval between high-intensity focused ultrasonic ablation and MRI was 9.7 ± 6.1 months (range, 3–26 months). The mean PSA level before high-intensity focused ultra sonic ablation was 7.25 ± 6.0 ng/mL (range, 0.12–23.5 ng/mL). Nine of the 27 patients underwent neo adjuvant hormonal therapy before the visit to our hospital. All hormonal therapy was discontinued at the time of the high-intensity focused ultrasonic ablation. This study received the approval of the institutional review board; informed consent was waived owing to retrospective analysis of the MR images.
MRI
MRI of the prostate was performed with a 3-T system (Intera Achieva 3T,
Philips Medical Systems) with a six-channel pelvic phased-array coil. An
endorectal coil was not used. Before imaging, 20 mg of butylscopolamine
(Buscopan, Boehringer Ingelheim) was injected intramus cularly to suppress
peristalsis of the bowel. No bowel preparation was performed. After three
plain localizer images were obtained, T2-weighted turbo spin-echo images were
acquired in the three orthogonal planes (axial, sagittal, and coronal). The
parameters for T2-weighted imaging were as follows: TR/TE,
2,260–4,200/80; slice thickness, 3 mm; echo-train length, 12; interslice
gap, 0.3 mm; matrix size, 512 or 304 x 304; field of view, 18 cm; number
of signals acquired, 3; sensitivity-encoding factor, 2. The total acquisition
time of T2-weighted images in each plane was approxi mately 4 minutes 22
seconds. An axial T1-weighted turbo spin-echo sequence (4-mm slice thickness)
was performed to assess lymph nodes and pelvic bone. The total acquisition
time was approximately 3 minutes 20 seconds.
Axial DWI was performed with the single-shot echo-planar imaging technique with the following parameters: 2,300–4,400/63–65; slice thickness, 3 mm; interslice gap, 1 mm; matrix size, 112 x 112–110; field of view, 20 cm; sensitivity-encoding factor, 2; number of signals averaged, 4. Diffusion-encoding gra dients were applied as a bipolar pair at b values of 0 and 1,000 s/mm2 along the three orthogonal directions of motion-probing gradients. ADC maps were automatically constructed on a pixel-by-pixel basis. The acquisition time for DWI was approximately 1 minute 32 seconds.
Axial DCE-MRI was performed with a 3D fast field-echo sequence (7.4/3.9; flip angle, 25°; matrix size, 224 x 179; slice thickness, 5 mm; interslice gap, none; field of view, 20 cm; 11 partitions on a 3D slab). DCE-MRI was performed from the apex to the base of the prostate. The 3D volume with 11 partitions was acquired every 5 seconds, and a total of 58 images were repeated. The dynamic series consisted of one unenhanced series and the subsequent series of 57 contrast-enhanced images. A contrast-enhanced series was performed immediately after a bolus injection of gadopentetate dimeglumine (Magnevist, Bayer HealthCare) at a rate of 2 mL/s with a dose of 0.1 mmol/kg body weight through a power injector. The total acquisition time for the DCE-MR images was approximately 5 minutes 22 seconds.
All dynamic data sets were transferred to an independent workstation (ViewForum, Philips Medical Systems) for measurement of wash-in rate, wash-out rate, maximal enhancement, maximal relative enhancement, and time-to-peak and for generation of parametric images. Time–signal intensity curves were first fit to a general exponential signal-enhancement model. Dynamic images were converted to a five-parameter model: wash-in rate, wash-out rate, maximal enhancement, maximal relative enhancement, and time-to-peak. The wash-in rate is defined as the maximum slope between the time of onset of contrast inflow and the time of peak enhancement on the time–signal intensity curve. The wash-out rate is defined as the negative slope of the late part of the exponential curve. Maximal enhancement is the concentration at which the exponential curve becomes level and the exponential constant defines the time-to-peak.
The postprocessing procedure (
6 minutes per patient) was performed by
a radiologist, who extracted a threshold value of five pa ram eters optimal
for differentiating can cerous from normal tissue. The radiologist generated
parametric images using the same threshold for all patients. At parametric
imaging, pixels with a wash-in rate, wash-out rate, maximal enhancement,
maximal relative enhancement, and time-to-peak greater than the threshold
values were color-coded over the enhanced fast-field echo images.
Endorectal Prostate Biopsy
TRUS-guided biopsy was performed on all 27 patients with an 18-gauge needle
mounted on a spring-loaded commercial biopsy device (Biopty Gun, Bard). After
induction of local periprostatic anesthesia, sextant biopsy cores (upper,
middle, and lower on each lobe) were each obtained from all of the patients by
two genitourinary radiologists who performed TRUS-guided biopsy on 17 and 10
patients, respectively. At TRUS biopsy, the same two radiologists were aware
of the MRI results and were asked to direct the biopsy carefully toward the
part of the sextant they believed contained a suspicious lesion on MRI. All
biopsy cores were labeled to identify the biopsy location, and all specimens
were evaluated by one experienced uropathologist (4 years of experience in
prostate biopsy), who was blinded to the imaging results. The mean interval
between MRI and biopsy was 22.6 ± 17.9 days (range, 1–75 days).
The mean volume of residual prostate tissue was 12 ± 4.3 mL (range,
6–22 mL).
MRI Analysis
In the follow-up period after high-intensity focused ultrasonic ablation, a
positive result of a prostate biopsy was the finding of residual or newly
developed recurrent cancer in the remnant prostatic tissue. It was almost
impossible, however, to determine whether the finding was incompletely treated
viable tumor that continued to grow or was new tumor. Therefore, in this
study, residual or newly developed recurrent prostate cancer was considered
local tumor progression if the prostate biopsy result showed any cancer foci
[23].
Two readers were asked to identify the presence of cancer in six prostate sectors (upper, middle, and lower on each lobe). Because high-intensity focused ultrasonic ablation results in morphologic deformity of the prostate and indistinct zonal anatomy, we did not subdivide the prostate into peripheral zone and central gland. For the prediction of local tumor progression of prostate cancer, two readers used a 5-point scale to report their confidence that tumor was present in each of the six prostate sectors: 1, definitely no tumor present; 2, probably no tumor present; 3, tumor possibly present; 4, tumor probably present; 5, tumor definitely present.
Two readers independently reviewed the MR images on a PACS workstation
(PathSpeed, GE Healthcare). Reader 1 had 6 years of experience ( 700
studies), and reader 2 had 4 years of experience ( 450 studies). The
readers knew only that the PSA level was increasing; each was unaware of the
other's findings and of the patient data. They first independently assessed
T2-weighted and diffusion-weighted images without know ledge of the results of
DCE-MRI. More than 2 weeks after the first assessment, the two readers
independently evaluated DCE-MR images.
The diagnostic criteria for local tumor pro gression on T2-weighted MRI with DWI were considered when high focal signal intensities at a b value of 1,000 s/mm2 on DWI were found as lesions of low focal signal intensity with a low ADC relative to the surrounding benign tissue on ADC maps and the lesions had a length of more than 5 mm in the transverse diameter. In addition, because of low signal-to-noise ratios and imaging distortion of the ADC maps, T2-weighted MRI was used to understand the entire anatomy of the residual prostate tissue. At DCE-MRI, the presence of asymmetric wash-in rate, wash-out rate, maximal enhancement, maximal relative enhancement, and increased time-to-peak are highly indicative of prostate cancer [17, 24, 25]. On MRI the areas of noncancerous tissue in the prostate after high-intensity focused ultrasonic ablation were considered to have the following features: signal intensity similar to that of normal nonirradiated prostate, low focal signal intensity (length < 5 mm in transverse diameter) relative to the surrounding prostatic tissue on the ADC map, and lack of enhancement or ill-defined delayed gradual enhancement on DCE-MRI. Quantitative analysis for T2-weighted MRI with DWI and for DCE-MRI was not performed.
Statistical Analysis
We dichotomized the 5-point scoring system to evaluate the sensitivity,
specificity, positive predictive value, negative predictive value, and overall
accuracy for the prediction of local tumor progression of prostate cancer
after high-intensity focused ultrasonic ablation. Scores of 3–5
indicated presence, that is, local tumor progression. For prediction of local
tumor progression of prostate cancer, the sensitivity and specificity of
DCE-MRI and T2-weighted MRI with DWI were compared by use of the McNemar test.
Bonferroni correction was used to adjust for multiple comparisons. Overall, in
162 prostate sectors of 27 patients, the accuracy rates of DCE-MRI and
T2-weighted MRI with DWI were analyzed by use of generalized estimating
equations to account for clustering effects from multiple measurements in the
same patient [26].
Interobserver agreement for prediction of local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation was calcu lated with nonweighted kappa statistics. The following qualitative terms were used to describe the strength of various kappa values: 0–0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agree ment; and 0.81–1.00, near-perfect agreement [27].
The area under the receiver operating characteristic (ROC) curve was calculated with the scores from the prediction of local tumor progression of prostate cancer. Statistical analysis was performed with Rockit 0.9B (Department of Radiology, University of Chicago) and the SAS software package (version 8, SAS Institute). Two-tailed tests were used, and values of p < 0.05 were con sidered significant. For Bonferroni correction of the McNemar test result, p < 0.0083 was considered to indicate a statistically significant difference.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Table 1 shows the diagnostic performance of DCE-MRI and T2-weighted MRI with DWI in prediction of local tumor progression of prostate cancer in 27 patients who underwent high-intensity focused ultrasonic ablation (Figs. 1A, 1B, 1C, 1D, 1E and 2A, 2B, 2C, 2D, 2E, 2F). The sensitivities of DCE-MRI and T2-weighted MRI with DWI were 80% and 63% for reader 1 (p = 0.004) and 87% and 70% for reader 2 (p = 0.004). The specificities of DCE-MRI and T2-weighted MRI with DWI were 68% and 78% for reader 1 (p = 0.002) and 63% and 74% for reader 2 (p < 0.001). The accuracy rates of DCE-MRI and T2-weighted MRI with DWI were 72% and 73% for reader 1 (p > 0.05) and 71% and 73% for reader 2 (p > 0.05). For interobserver agreement, the kappa values of DCE-MRI and T2-weighted MRI with DWI were 0.59 and 0.46, respectively, which showed moderate agreement. Findings of DCE-MRI were false-positive for 34 (32%) of the prostate sectors for reader 1 and 40 (37%) of the prostate sectors for reader 2 (Fig. 3A, 3B, 3C, 3D, 3E). Findings of T2-weighted MRI with DWI were false-positive for 24 (22%) prostate sectors for reader 1 and 28 (26%) of the prostate sectors for reader 2.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
For prediction of local tumor progression of prostate cancer, the areas under the ROC curves for DCE-MRI and T2-weighted MRI with DWI were 0.77 (95% CI, 0.67–0.85) and 0.77 (0.69–0.85) for reader 1 and 0.85 (0.77–0.92) and 0.81 (0.74–0.89) for reader 2 (Fig. 4). No statistical difference between the use of DCE-MRI and the use of T2-weighted MRI with DWI was found for either reader (p > 0.05).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After high-intensity focused ultrasonic ablation, complete loss of prostate zonal anatomy occurs, and the central gland cannot be differentiated from the peripheral zone. Furthermore, prostate tissue can have a diffuse or multifocal area of low signal intensity on T2-weighted MRI [13]. These findings suggest difficulty in differentiating benign tissue and local tumor progression in remnant prostate tissue. After high-intensity focused ultrasonic ablation of localized prostate cancer, early detection of local tumor progression is important because it can affect decisions about second-line treatment.
In an evaluation [13] of the MRI appearance of the prostate after high-intensity focused ultrasonic ablation of localized cancer, the ablation–induced abnormalities (on fat-saturated contrast-enhanced T1-weighted images, the lesion appeared as an unenhanced hypointense zone surrounded by a 3- to 8-mm-thick peripheral rim of enhancement) seemed to disappear within 3–5 months. According to the results of that study, MRI findings were not predictive of the histologic results of treatment, especially the presence of residual cancer foci after high-intensity focused ultrasonic ablation. A limitation of the study, however, was that functional methods, such as DCE-MRI, DWI, and spectroscopic imaging, were not used to predict the presence of residual viable tumor.
The availability of 3-T MRI units with increased signal-to-noise ratios has enabled the use of parallel imaging technique and a phased-array coil to obtain good-quality images of the prostate [20, 21]. Owing to increased signal-to-noise ratios at 3 T, DWI has received attention as a promising method of imaging prostate cancer. After high-intensity focused ultrasonic ablation, however, the prostate zonal anatomy can be lost or become indistinct, causing difficulty in differentiating the central gland and the peripheral zone in residual prostate tissue. In our study, T2-weighted MRI was used because with only DWI and an ADC map it might have been difficult to visualize the complete anatomy of the residual prostate tissue.
To the best of our knowledge, no investigation of the use of DCE-MRI or DWI findings for predicting local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation has been reported in the literature. We therefore performed DCE-MRI and DWI for this purpose in patients with elevated PSA levels. The sensitivities of DCE-MRI and T2-weighted MRI with DWI were 80% and 63% for reader 1 and 87% and 70% for reader 2, and a statistically significant difference (p = 0.004) between DCE-MRI and T2-weighted MRI with DWI was found for the two readers. The specificities of DCE-MRI and T2-weighted MRI with DWI were 68% and 78% for reader 1% and 63% and 74% for reader 2, and a statistically significant difference (p < 0.003) between DCE-MRI and T2-weighted imaging with DWI was found for the two readers. However, the accuracy rates of DCE-MRI and T2-weighted MRI with DWI were similar. For interobserver agreement, DCE-MRI and T2-weighted MRI with DWI had moderate agreement. In prediction of local tumor progression of prostate cancer, ROC analysis showed that the areas under the ROC curve of DCE-MRI and of T2-weighted MRI with DWI were 0.77 and 0.77 for reader 1 and 0.85 and 0.81 for reader 2. No statistical difference between DCE-MRI and T2-weighted MRI with DWI was found for either reader (p > 0.05).
For predicting local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation in this study, the rates of false-positive findings on DCE-MRI were 32% for reader 1 and 37% for reader 2. However, the rates of false-positive findings on T2-weighted MRI with DWI were 22% for reader 1 and 26% for reader 2. An explanation of the false-positive findings on DCE-MRI may be that we did not exclude the central gland in the MRI analysis. We made the exclusion because the indistinct zonal anatomy after high-intensity focused ultrasonic ablation can cause enhancement of benign prostatic hypertrophic nodules in the residual prostate tissues to mimic local tumor progression of prostate cancer. In terms of the false-positive findings on DWI, fibrosis due to high-intensity focused ultrasonic ablation or benign prostatic hypertrophic nodules in residual prostate tissue can decrease ADC, a finding that can mimic local tumor progression of prostate cancer.
This study had limitations. First, the study population was small. For higher statistical power, larger enrollment numbers would have strengthened +the results. Second, the study lacked a histologic reference standard. The MRI findings were correlated only with the results of TRUS-guided prostate biopsy. A suspected cancerous lesion in prostate tissue on MRI might not be accurately targeted at TRUS-guided prostate biopsy. Conversely, both MRI and random biopsy might have missed residual or recurrent cancer foci. However, a whole-mount pathologic specimen of the prostate cannot be obtained from patients treated with high-intensity focused ultrasonic ablation. Therefore, site-by-site comparison of biopsy and MRI findings is a reasonable way of estimating the clinical value of DCE-MRI and DWI. Third, quantitative assessment in residual prostate tissue, including cancerous and noncancerous tissue, after high-intensity focused ultrasonic ablation was not performed for DCE-MRI and DWI. Further studies are needed to assess the effect of quantitative analysis of DWI and DCE-MRI on diagnostic performance in predicting local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation. Finally, recall bias might have been present in this study because the population was small and the time between MRI readings was relatively short.
In conclusion, for prediction of local tumor progression of prostate cancer after high-intensity focused ultrasonic ablation, DCE-MRI was more sensitive than T2-weighted MRI with DWI, and T2-weighted MRI with DWI was more specific than DCE-MRI.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
