| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Rome "La Sapienza," Via del
Policlinico, 155, 00166 Rome, Italy.
2 U.O.C. Osservazione Medica, Ospedale M. G. Vannini, Rome, Italy.
3 Istituto di Ricerche Farmacologiche "Mario Negri," Milan,
Italy.
Received August 16, 2007;
accepted after revision November 18, 2007.
Address correspondence to E. Casciani
(emanuelecasciani{at}hotmail.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. A total of 51 patients who had undergone radical prostatectomy for prostatic adenocarcinoma 10 months to 6 years before underwent a combined endorectal coil MRI and dynamic gadolinium-enhanced MRI before endorectal sonographically guided biopsy of the prostatic fossa. The MRI combined with MR dynamic imaging results were correlated with the presence of recurrence defined as a positive biopsy result or reduction in prostate-specific antigen level after radiation therapy.
RESULTS. Overall data of 46 (25 recurred, 21 nonrecurred) out of 51 evaluated patients were analyzed. All recurrences showed signal enhancement after gadolinium administration and, in particular, 22 of 24 patients (91%) showed rapid and early signal enhancement. The overall sensitivity and specificity of MR dynamic imaging was higher compared with MRI alone (88%, [95% CI] 69–98% and 100%, 84–100% compared with 48%, 28–69% and 52%, 30–74%). MRI combined with dynamic imaging allowed better identification of recurrences compared with MRI alone (McNemar test: chi-square1 = 16.67; p = < 0.0001).
CONCLUSION. MRI combined with dynamic contrast-enhanced MRI showed a higher sensitivity and specificity compared with MRI alone in detecting local recurrences after radical prostatectomy.
Keywords: contrast-enhanced MRI • MRI • prostate neoplasm • recurrence
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A higher response rate to radiation therapy has been reported in selected patients with biochemical failure and pathologically confirmed local recurrence compared with patients without biopsy-proven local recurrence [3]. The local tumor control rate with irradiation for palpable recurrence in patients who have undergone radical prostatectomy ranges from 58% to 100% [4]. Thus the recognition of a local recurrence after radical prostatectomy is a fundamental issue for therapy and follow-up of these patients.
Digital rectal examination (DRE) has been shown to be inadequate in detecting local recurrences [5]. Although endorectal sonography (transrectal ultrasonography, TRUS) is better than DRE for detecting local recurrence, it lacks specificity [6, 7]. In patients with elevated PSA levels, digitally guided biopsy of the vesicourethral anastomosis is often performed to determine if local recurrence of malignancy is present; however, a palpable abnormality is not a reliable finding because postoperative fibrosis often mimics recurrent malignancy [5]. TRUS-guided prostatic fossa biopsy currently is the most efficient and cost-effective tool for detecting local recurrence of cancer after radical prostatectomy [8]; however, the likelihood of biopsy-proven local recurrence after radical prostatectomy has been reported to vary between 38% and 54% [6, 7, 9], with nearly a third of patients requiring two or more TRUS-guided biopsy sessions to obtain a final diagnosis [10].
MRI, with its inherent superior contrast resolution, looks promising for the evaluation of local recurrence of prostatic cancer in men who have undergone radical prostatectomy [11, 12]. The aim of this study was to evaluate the sensitivity and specificity of MRI combined with dynamic contrast-enhanced MRI to detect local recurrence after radical prostatectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The mean interval from MRI to radical prostatectomy for prostatic adenocarcinoma was 32 months, ranging from 10 months to 6 years, and the mean time interval between biopsy and MRI was 2 months, ranging from 10 days to 4 months. The study protocol was approved by the local ethics committee; written informed consent was obtained before study inclusion from all patients in accordance with the national legislation and the Declaration of Helsinki. All patients underwent PSA testing and DRE before MRI. DRE was graded on the basis of likelihood of tumor presence by using the following 3-point scale: 1, normal; 2, induration; and 3, palpable nodule.
MRI Technique
Eligible patients underwent a combined endorectal coil MRI and dynamic
contrast-enhanced MRI before TRUS-guided biopsy of the prostatic fossa. The MR
images were obtained with a 1.5-T system (Signa, GE Healthcare) using a
combination of an endorectal coil and a pelvic phased-array coil. The
endorectal coil was positioned with the patient in a lateral recumbent
position, and 100 mL of room air was insufflated after IV administration of 1
mg of glucagon (Glucagen, Nordisk) to reduce intrinsic rectal movements and
discomfort during exami nation. Transverse T1-weighted spin-echo seq uences
from the aortic bifurcation to the symphysis pubis to evaluate pelvic lymph
node and bone status (TR range/TE, 500–700/12; section thickness, 5 mm;
1-mm gap; two signals acquired; field of view, 26 cm; matrix, 512 x 224;
no phase wrap) were performed. T2-weighted fast spin-echo sequences with the
driver equili brium fast-recovery technique were performed in the transverse,
sagittal, and coronal planes of the pelvis using the following parameters:
4,000–5,000/102; echo-train length, 15–17; sec tion thickness, 3
mm without interval; field of view, 14 cm; three signals acquired; matrix, 320
x 192. A multisection T1-weighted 3D spoiled gradient-echo sequence (TE,
1.8; flip angle, 12°; sections, 15; section thickness, 4 mm; no
intersection gap; slab, 26–28; field of view, 260 mm; matrix, 160
x 256 [reconstruction, 512]; 21 seconds) in the axial plane was
performed before and during IV bolus injection of a paramagnetic gadolinium
chelate (0.1 mmol per kilogram of body weight of gadopentetate dimeglumine
[Magnevist; Bayer HealthCare]) by means of a power injector with an injection
rate of 4 mL/s followed by a 15-mL saline flush
[13]. Overall dynamic
contrast-enhanced MRI time was 5–6 minutes. All dynamic data sets were
transferred to an independent workstation (Leon ardo VD30B, Siemens Medical
Solutions) used to perform dia gram analysis of time–signal intensity
curves.
|
The initial signal increase is the maximum signal intensity within the
first 90 seconds after contrast administration compared with the signal
intensity of the unenhanced image. It is calculated using the following
formula: Initial signal increase
.
The start of enhancement was calibrated by using the external iliac artery.
The postinitial signal behavior describes the course of the signal curve
between 90 and 325 seconds after contrast administration. The signal intensity
value after 325 seconds is expressed in relation to the maximum value in the
initial phase as follows: Postinitial signal behavior
(Fig. 1).
According to previous reports [11, 12], MR images were classified on the basis of recurrence signal intensity, as follows: 1, normal (no soft-tissue nodule); 2, dubious (soft-tissue nodule hypointense or hyperintense to muscle on T1-weighted sequences and markedly hyperintense to muscle on T2-weighted sequences); and 3, positive (soft-tissue nodule hypointense on T1-weighted sequence and mildly hyperintense on T2-weighted sequences). It is known from previous literature that dynamic gadolinium-enhanced MRI perform ed before radical prostatectomy in patients with known carcinoma shows the imaging features of carcinoma [13–16]. These studies have shown that prostate carcinoma is associated with tumoral angiogenesis leading to abnormal contrast en hancement patterns.
In prostate recurrences, dynamic-enhanced MRI grading is not already established. For this reason, we decided to use the following grading (Fig. 1): 1, normal (initial signal increase absent or < 50% for any postinitial signal course); 2, dubious (initial signal increase between 50% and 100%, postinitial signal course as steady increase or plateau); and 3, positive (initial signal increase between 50% and 100% and postinitial signal course as continuous decrease, or initial signal increase > 100% for any postinitial signal course). We defined as steady increase an increase up to 10% and a steady decrease or wash-out a decrease up to 10% in the postinitial curve.
Standard of Reference
A patient was considered to be clinically free of local recurrence if
biopsy was negative and the PSA level did not increase after surgery for at
least 1 year after the MR evaluation for local relapse or if biopsy was
negative and a pelvis MRI or bone scan or 18F-FDG PET scan yielded
positive results without evidence of tumor in the postprostatectomy fossa.
The prostatic bed biopsy had to be performed between 1 and 3 weeks after MRI. After patient preparation (which included antibiotic therapy and a cleansing enema) and administration of a local anesthetic, a complete TRUS evaluation of the prostatic fossa was performed before biopsy. The prostatic fossa was evaluated with a 7-MHz endorectal transducer with the patient in the left lateral decubitus position. Sonographic transverse, coronal, and sagittal scans were obtained to reproduce the same morphologic findings obtained with T2-weighted transverse, coronal, and sagittal MR images to better localize the suspicious MRI findings. These scans were obtained using internal and external anatomic landmarks (i.e., external sphincter, vesicourethral anastomosis, and posterior wall of the bladder).
By using these criteria, the suspicious MR areas were projected as accurately as possible on sonograms, and TRUS–guided biopsy was performed. If MRI showed an abnormal tissue mass, biopsies of this mass were taken; if no abnormalities were detected, random biopsies of the anastomotic site were performed. Random guided biopsy of the prostatic fossa usually included removal of two cores from either side of the anastomosis, one toward the bladder neck and one toward the external urethral sphincter. All biopsies were performed by radiologists who were also involved in the acquisition and interpretation of the MR images and dynamic contrast-enhanced MRI.
Statistical Analysis
Results of DRE, MRI, and dynamic contrast-enhanced MRI were analyzed
considering grade 1 as negative (i.e., no recurrence), whereas grades 2 and 3
were considered positive (i.e., recurrence). Once the concordance of MRI and
MRI combined with dynamic contrast-enhanced MRI results with those of the
reference standard was assessed, the following statistics were calculated:
sensitivity, specificity, and diagnostic accuracy. The McNemar test was used
to test for differences between experimental procedures in the concordance
rate with the standard. The relationship between risk of recurrence and PSA
level was assessed by means of a logistic regression model, including tumor
stage as a covariate. In this model, PSA level was factored as a dichotomous
variable using a PSA level of 0.4 ng/dL to stratify patients into low PSA or
high PSA groups [17]. Results
are expressed as point estimates and their 95% CIs are derived by binomial
distribution. All p values are two-tailed. Analyses were performed
using SAS System version 8.20 (SAS).
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Overall, MRI correctly classified 23 of 46 patients with a diagnostic accuracy of 48% (95% CI), 33–63%. Eleven of 21 patients without prostate cancer recurrence and 12 of 25 patients with prostate cancer recurrence were correctly classified with a specificity of 52%, 30–74% and a sensitivity of 48%, 28–69%. MRI alone was not superior to DRE in concordance rate with the reference standard (McNemar test: chi-square1 = 1.47; p = 0.22).
On dynamic contrast-enhanced MRI of the 25 patients with prostate recurrence, one was classified as grade 1, two as grade 2, and 22 as grade 3 (Fig. 2A, 2B, 2C). All 21 patients without prostate recurrence were classified as grade 1 (Fig. 3A, 3B, 3C). Only one recurrence was not detected by MRI before or after contrast administration or by biopsy but was considered present for PSA reduction after radiation therapy. MRI combined with dynamic contrast-enhanced MRI showed a diagnostic accuracy of 94%, (95% CI) 82–99%, sensitivity of 88% (69–98%), and specificity of 100% (84–100%).
|
|
|
|
The mean diameter of soft-tissue nodules revealed at MRI was 1.5 cm (range, 0.4–4.0 cm). Local recurrence revealed at MRI was most common at the perianastomotic site, occurring in 52% of cases. Patients with a PSA value higher than 0.4 ng/dL have a three-times higher probability of recurrence than patients with a value below that cutoff value, even if the result does not reach statistical significance (odds ratio [OR], 2.9; [95% CI], 0.63–13.4; p = 0.773), whereas the effect of stage was statistically significant for patients with stage T3 relative to patients with stages T2–T1 (OR, 4.8; 1.3–17.3; p = 0.023).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patients with high PSA levels (> 2 ng/dL) and negative TRUS findings do
not appear to benefit from biopsy of the vesicourethral anastomosis because
they do not have a local recurrence but rather a distant metastasis
[19]. In this study, mean PSA
value before MRI was 1.084 ng/dL (range, 0.1–6 ng/dL) in patients
affected by recurrence. Using 0.4 ng/dL as a cutoff value, the risk of
developing a recurrence was three times higher in patients with PSA > 0.4
ng/dL compared with patients with PSA 
0.4 ng/dL, but the difference was
not statistically significant (OR, 2.93; 0.63–13.59; p =
0.773).
In cases of normal TRUS findings of the prostatic fossa, only 20% of patients had a positive biopsy, whereas 62% of those with a suspected lesion had proven local recurrence [8]. Saleem et al. [7] also reported a statistically significant higher risk of cancer detection in cases of positive TRUS findings. Indeed, although the overall positive biopsy rate was 50%, in cases of TRUS–suspected lesions [19], it increased to 67%.
Our study shows the prevalence of recurrence to be 54.4%, according to the literature [10], and the rate of TRUS–guided biopsy recurrence identification to be 88%. This high accuracy was due to the useful information obtained from MRI performed before biopsy and used to obtain similar coronal and sagittal scans with sonography. In fact, MRI has the potential to direct endorectal biopsy to these sites and may thus lead to a better diagnostic yield. Endorectal surface coil MRI is a reasonable alternative technique for use in the detection of local recurrence.
Until now in the literature, we know of only two studies evaluating the ability of endorectal MRI to detect local recurrence after radical prostatectomy, and neither of them used dynamic imaging. In the first (prospective) study, Silverman and Krebs [11] achieved a high sensitivity and specificity (100%) evaluating 41 patients with T1- and T2-weighted sequences and T1-weighted images with fat-suppression technique after gadolinium administration. In this study, all 31 recurrences were isointense to muscle in T1 whereas they appeared unequivocally hyperintense in 18 (58%) and slightly hyperintense in 13 (42%) T2-weighted sequences. These 13 slightly hyperintense softtissue nodules were not considered unequivocally suspicious for recurrent disease using only the unenhanced sequences. However, all nodules showed signal enhancement after gadolinium administration, strengthening the suspicion that they were recurrences.
In the second (retrospective) study of 48 patients, Sella et al. [12] achieved a high sensitivity (95%) and specificity (100%) using T1- and T2-weighted sequences, without sequences obtained after gadolinium administration. All 42 local recurrences seen on MR images were isointense on T1-weighted sequences and slightly hyperintense to muscle on T2-weighted sequences (in the article, tumor signal intensity was defined on T2-weighted sequences as similar to that of the pelvic muscle, slightly higher than that of muscle, or much higher than that of muscle).
There is some discordance between these two studies regarding signal intensity in T2-weighted sequences and the utility of contrast medium. In the first study [11], recurrence is definite when signal is unequivocally hyperintense on T2-weighted sequences and contrast medium was useful to confirm the diagnosis. In the second study [12], all recurrences appeared slightly hyperintense on T2-weighted sequences without gadolinium administration.
In our study, MRI alone showed a poorer accuracy in detecting recurrences, probably due to the smaller size (between 0.4 and 3.0 cm) of the recurrences compared with those in the study of Silverman and Krebs [11] and the study of Sella et al. [12], 0.7–3.8 cm and 0.8–4.5 cm, respectively.
Recurrences were, in most cases (12/14 patients, 85%) slightly hyperintense to internal obturator muscle on T2-weighted sequences as found by Sella et al. [12] and in fewer cases (11/19 patients, 58%) markedly hyperintense on T2-weighted sequences. In fact, nodules that appear slightly hyperintense or markedly hyperintense on T2-weighted sequences may represent not only recurrences but also prostatic or seminal vesicle residues with different amounts of fibrosis.
Eight markedly hyperintense nodules (redefined as normal) were scars (four cases), fibrotic residues of seminal vesicles (three cases), and hematoma (one case) and two slightly hyperintense nodules were fibrotic residues of seminal vesicles. As opposed to residual prostatic tissue, retained seminal vesicles do not secrete PSA [20, 21]. They may, however, pose a diagnostic challenge for postsurgical follow-up of patients [22].
After dynamic contrast-enhanced MRI in our study, all benign nodules showed signal enhancement of less than 50% in the early phase, whereas all recurrences showed fast signal enhancement in the early phase followed by plateau or washout. Not performing dynamic contrast-enhanced MRI on the eight nodules showing signal enhancement of less than 50% would have led us to consider them as recurrences. In contrast to what was reported by Silverman and Krebs [11], in our experience there are benign nodules that show signal enhancement after gadolinium administration.
Our study showed that MRI combined with dynamic contrast-enhanced MRI considerably improved diagnostic accuracy. This is particularly notable because although MR spectroscopy is useful for detecting a recurrent cancer after radiation therapy [23, 24] or hormone deprivation therapy [25], it is not yet sufficiently robust to evaluate small local recurrence because MR spectra are of poor quality due to lipid contamination. Currently, there appears to be no evidence in the literature of the diagnostic value of MR spectroscopy for identification of local recurrence of prostate cancer after radical prostatectomy.
Previous studies performed before radical prostatectomy in patients with known carcinoma have described the imaging features of carcinoma at dynamic gadolinium-enhanced MRI [13–16, 26–28]. These studies have shown that, similar to other tumors, including breast carcinoma and carcinoma of the uterine cervix, prostate carcinoma is associated with tumoral angiogenesis leading to abnormal contrast enhancement patterns. Tumor enhancement in the early arterial phase is more intense and more rapid. In more-delayed acquisitions, 3 to 8 minutes after injection, tumor washout is more rapid and more pronounced. It is a reasonable assumption that dynamic contrast enhancement in prostatic local recurrence tissue might be different from that in adjacent tissue because of the increased microvessel density of prostatic recurrence tissue [22–24]. In our study, these differences were quantified and used to discriminate prostatic local recurrence tissue from postsurgical fibrosis, residual prostatic tissue, or residual seminal vesicle. We believe that in prostatic local recurrence tissue, the relative peak enhancement was the optimal parameter. MRI and dynamic contrast-enhanced MRI performed from the vesicoureteral junction to the bladder dome were able to identify perianastomotic recurrences (52%) and recurrences localized in less common sites: retrovesical space (20%), bladder neck (16%), and circumferential (12%). Prior studies using MRI [12] or TRUS [8] confirmed that the most common site for recurrence was the vesicourethral anastomosis.
This study has some limitations. First, the number of patients studied is relatively small to form a definitive conclusion. Second, we applied dynamic contrast-enhanced MRI parameters used for prostate and breast cancer because data were not available on prostate cancer recurrence.
In conclusion, we think that MRI alone, without contrast medium, is actually insufficient for detecting prostate cancer recurrence. Contrast medium should be used not only for qualitative but also for quantitative study, with diagram analysis of time–signal intensity curves (dynamic contrast-enhanced MRI) because many confounding factors may be present after radical prostatectomy.
Acknowledgments
The authors thank Stefano Caprasecca, Dino D'Amico, Simona Tuzi, Nicola
Pellicciari, Fabio Fiocco, and Dario Di Luzio.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
