| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genitourinary Imaging |
1 From the Departments of Radiology (T.S., L.H.S., C.N.O., H.H.), Urology (P.W.S., P.T.S.), and Medicine (H.I.S.), Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021. From the 2002 RSNA scientific assembly. Received January 3, 2003; revision requested March 10; final revision received September 12; accepted October 8. Supported by National Institutes of Health grant R01 CA76423. Address correspondence to H.H. (e-mail: hhricak@rsna.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Endorectal MR images were reviewed retrospectively in 82 patients who underwent prostatectomy. The interval between prostatectomy and MR imaging ranged between 0.5 and 13.0 years (mean, 3.25 years). Local recurrence was considered present if there was no evidence of distant metastases and there was a positive biopsy result, subsequent reduction in prostate-specific antigen (PSA) level after radiation therapy of the pelvis, or serial MR imaging findings of increased tumor size. Local recurrence on MR images was assessed for location, size, signal intensity, and invasion of adjacent structures. All images were reviewed independently by two readers who were blinded to clinical information.
RESULTS: Thirty-four of 82 patients did not meet inclusion criteria. Forty-one of 48 remaining patients had clinically documented local recurrence, which MR imaging depicted in 39 of 41 (95%) patients. Seven of 48 patients had no evidence of local or distant metastases, and none had positive MR imaging findings. Sensitivity of MR imaging was 95%, and specificity was 100%. Local recurrences were perianastomotic in 12 (29%) patients and retrovesical in 17 (40%), within retained seminal vesicles in nine (22%), and at anterior or lateral surgical margins in four (9%). All local recurrences were hyperintense to adjacent pelvic muscles on T2-weighted MR images. The mean diameter of tumors was 1.4 cm (range, 0.8–4.5 cm). PSA levels at MR imaging in patients with clinically proved recurrences ranged from undetectable to 10 ng/mL (mean, 2.18 ng/mL).
CONCLUSION: MR imaging depicts a high proportion of local recurrence after prostatectomy. The authors propose incorporation of endorectal coil MR imaging in the diagnostic paradigm of patients who have undergone prostatectomy and are suspected of having local recurrence.
© RSNA, 2004
Index terms: Prostate neoplasms, MR, 844.121411 • Prostate neoplasms, surgery, 844.39, 844.451
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
When a patient’s PSA level begins to increase or a nodule is palpable at digital rectal examination, recurrence is suspected, and a diagnostic work-up is generated. Clinically, there are four main categories of recurrence: PSA-only relapse, local recurrence in the prostatectomy bed, distant metastases (most commonly nodal or osseous), and a combined local and distant recurrence. It is critical to distinguish among these four types of recurrence for the delivery of appropriate treatment. The type of recurrence is difficult to determine clinically, since an increasing PSA level is rarely associated with symptoms or findings at physical examination (2,10,11).
A number of nomograms have been developed to help estimate whether a recurrence is local or distant. These nomograms are based on clinical parameters, such as pathologic stage and tumor grade at the time of prostatectomy, PSA doubling time, time interval between surgery and PSA relapse, and others (9,12,13).
Imaging plays a central role in the detection of metastatic disease. Computed tomography (CT) and technetium 99m bone scanning are commonly used for the detection of nodal or osseous metastases in a patient who has undergone prostatectomy and is suspected of having recurrence (14–16). However, imaging is not commonly used in the detection of local recurrence. Transrectal ultrasonography (US) and CT are neither sensitive nor specific in the identification of local relapse in the postprostatectomy fossa or in the differentiation of a recurrence from postsurgical scarring (17–20).
Magnetic resonance (MR) imaging, with its inherent superior contrast resolution, shows promise in the evaluation of the postsurgical pelvis (21,22). The purpose of the present study was to evaluate endorectal coil MR imaging in the depiction of local recurrence after radical prostatectomy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patient age at the time of diagnosis ranged between 42 and 72 years (median age, 59 years). All patients underwent radical retropubic prostatectomy (23), with the interval between surgery and MR imaging ranging between 0.5 and 13 years (mean, 3.25 years; median, 2.25 years). The clinical information gathered for each patient included surgical histopathologic findings and tumor stage; PSA levels and dynamics at the time of diagnosis, before and after surgery, and at the time of MR imaging; time interval from surgery to PSA relapse and from surgery to MR imaging; and any pre- or postsurgical treatment the patient received (eg, radiation therapy, hormonal therapy, and prostate fossa biopsy). In patients whose MR imaging findings were validated by using clinical parameters, the minimum clinical follow-up was 1 year after MR imaging.
MR Imaging Technique
All MR images were obtained with a 1.5-T whole-body MR imager (Signa; GE Medical Systems, Milwaukee, Wis). Patients were examined in the supine position. A body coil was used for excitation, and a pelvic phased-array coil (GE Medical Systems) was used in combination with a commercially available balloon-covered expandable endorectal coil (Medrad, Pittsburgh, Pa) for signal reception. Transverse T1-weighted spin-echo MR images (repetition time msec/echo time msec, 700/8; section thickness, 5 mm; intersection gap, 1 mm; field of view, 24 cm; matrix, 256 x 192; transverse frequency direction to prevent obscuration of the pelvic nodes by endorectal coil motion artifact; and one signal acquired) were obtained from the aortic bifurcation to the symphysis pubis. Transverse thin-section high-spatial-resolution and coronal T2-weighted fast spin-echo MR images (5,000/96 [effective]; echo train length, 16; section thickness, 3 mm; intersection gap, 0 mm; field of view, 14 cm; matrix, 256 x 192; anteroposterior frequency direction to prevent obscuration of the ureterovesicle anastomosis by endorectal coil motion artifact; and three signals acquired) of the fossa were obtained after prostatectomy.
MR Image Analysis
A retrospective analysis of MR images was performed. Images were assessed for the presence of local recurrence, pelvic lymphadenopathy, or pelvic bone metastases. If a patient underwent more than one MR examination, the authors evaluated the images obtained at the time recurrence was clinically suspected, with any prior or later studies assessed at a different time for validation purposes only.
Local recurrences seen on MR images were assessed for location, size, signal intensity, and invasion of adjacent structures. In view of the sonographic and urologic literature, location was recorded as one of the following: perianastomotic (around the urinary bladder and/or membranous urethra), retrovesical (between the urinary bladder and rectum), within retained seminal vesicles, or at the anterior or lateral surgical margins of the prostatectomy bed (eg, abutting the levator ani muscles). Tumor size was assessed in the transverse plane with the maximal diameter and the greatest perpendicular measurement recorded. Signal intensity was defined relative to the adjacent pelvic muscles. On T1-weighted MR images, tumor signal intensity was scored as similar to, lower than, or higher than that of muscle. On T2-weighted MR images, signal intensity was defined as similar to that of the pelvic muscles, slightly higher than that of muscle, or much higher than that of muscle if it approached the signal intensity of fluid.
All images were reviewed independently by two readers (H.H., L.H.S.) with more than 10 years of experience in reading prostate MR images. At the time of interpretation, both readers were blinded to the indication for imaging, as well as any clinical information regarding disease status at the time of MR imaging (eg, PSA level). Findings were recorded by an independent third author (T.S.).
Standard of Reference
Prior to initiation of the study, a system for clinical validation of local recurrence was established by means of consensus between two urologists (P.T.S., P.W.S.) and a medical oncologist (H.I.S.). Validation was achieved by means of one of the following methods: a positive biopsy result from the postprostatectomy fossa, reduction in PSA level following radiation therapy, or serial MR images that demonstrated at least a 20% increase in the size of a suspicious pelvic soft-tissue mass. A patient was considered to be clinically free of local recurrence if (a) the PSA level never increased following surgery and remained undetectable (<0.1 ng/mL) for at least a year after the MR evaluation for local relapse or (b) a bone scan or fluorine 18 (18F) fluorodeoxyglucose positron emission tomographic (PET) scan yielded positive results outside of the pelvis without evidence of tumor in the postprostatectomy fossa (ie, administration of pelvic radiation did not result in a reduction of PSA level). Pelvic bone or nodal disease was considered proved if there was a positive biopsy result or if a bone scan or PET scan had abnormal findings concordant with those of MR imaging.
Statistical Analysis
Descriptive statistics were used to calculate the sensitivity and specificity for detection of local recurrence with MR imaging. The standards of reference for validating these results were those described above; therefore, only patients who had adequate information available were included in this analysis. Descriptive statistics were also used for the purpose of characterizing MR features of local relapses. The Spearman correlation coefficient was used to evaluate the association between PSA level and size of depicted local recurrence, and 95% CIs were calculated.
Interreader agreement of MR image interpretation regarding local relapse was assessed by using 
statistics, with a value of less than 0.20 indicating poor agreement; 0.20–0.39, fair agreement; 0.40–0.59, moderate agreement; 0.60–0.79, substantial agreement; and 0.80 and higher, excellent agreement. If there was discordance between readers, the senior reader’s MR imaging observations were used for all statistical analyses except the calculations.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Findings
Seven of the 48 (14%) patients showed no MR imaging or clinical evidence of pelvic local recurrence (Table 1). As defined by our standards of reference, local recurrence was identified clinically in 41 patients. In 39 of these 41 patients, MR imaging demonstrated at least one soft-tissue mass in the postprostatectomy fossa. All sites visible with MR imaging were validated clinically as a local recurrence of prostate cancer. In this group of patients, 15 of 39 (38%) had a biopsy-proved recurrence (Figs 1, 2), 17 of 39 (44%) showed reduction of PSA level following radiation therapy (Fig 3), and seven of 39 (18%) had follow-up MR images that showed an increase in the size of the prostatectomy fossa mass (Fig 4). Two of 48 (4%) patients had indeterminate MR imaging findings. In these two patients, MR images demonstrated abundant postsurgical fibrosis, and the differentiation between postsurgical changes and tumor recurrence could not be made. In both patients, local recurrence was documented clinically, thus representing false-negative MR imaging results. There were no false-positive MR findings. These data resulted in a sensitivity of 95% (95% CI: 0.83, 0.99) and a specificity of 100% (95% CI: 0.59, 1.00) for the demonstration of local relapse at MR imaging (Table 2). Interobserver agreement was high regarding the presence of local recurrence at endorectal coil MR imaging, with a value of 0.91. The overall discordance rate between the two readers was 4%. Findings are summarized in Tables 1 and 2.
|
|
|
|
|
|
|
Local Recurrence
A total of 42 sites of local recurrences were demonstrated in 39 patients. The location within the postprostatectomy fossa was perianastomotic in 12 (29%) lesions (Fig 5), retrovesical in 17 (40%) (Fig 6), within a retained seminal vesicle in nine (22%) (Fig 7), and at the anterior or lateral surgical margin in four (9%) (Fig 8). The maximal transverse dimension of a mass representing the local recurrence averaged 1.4 cm (range, 0.8–4.5 cm). PSA level at the time of MR imaging in patients with a local recurrence on MR images ranged from undetectable to 10 ng/mL (mean, 2.18 ng/mL; median, 1.02 ng/mL). We found no association between PSA level at the time of MR imaging and size of the depicted local recurrence with a calculated Spearman correlation coefficient of r = 0.29 (P = .07; 95% CI: 0.03, 0.55). These data are presented in Figure 9.
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Imaging plays a central role in the evaluation of metastatic disease. CT, monoclonal antibody scanning, and 18F fluorodeoxyglucose PET have been shown to be beneficial in locating pelvic lymphadenopathy in patients who have undergone prostatectomy (15). Technetium 99m bone scanning is considered the standard of reference for the detection of osseous metastases and has a pivotal role in the follow-up of patients with this disease. However, imaging for local recurrence has not been used widely.
All three cross-sectional imaging modalities (CT, MR imaging, and US) have previously been evaluated in the detection of local recurrence following prostatectomy. In clinical practice, the most commonly used modality is transrectal US. Leventis et al (18) evaluated the accuracy of transrectal US and showed that it was more sensitive than digital rectal examination (76% vs 44%, respectively), albeit less specific (67% vs 91%, respectively). The overall transrectal US–guided biopsy detection rate in their patient cohort was 41%. The greatest chance of achieving a diagnosis of a local recurrence with transrectal US–guided biopsy was if the patient had a PSA level higher than 4 ng/mL with positive digital rectal examination and transrectal US findings (18).
In a CT study on the detection of local recurrence in 22 patients with biopsy-proved recurrences, only 36% of recurrences were detected, all of them larger than 2 cm (17). Silverman and Krebs (21) have shown the value of endorectal coil MR imaging and have reported it to be sensitive (100%) and specific (100%) in 31 men with biopsy-proved local recurrence following prostatectomy. Our results confirm their findings; our level of sensitivity for detecting a recurrence was 95%, with a specificity of 100%. Our study provides further data on MR localization of recurrence sites and correlates MR imaging findings with clinical parameters that define a local recurrence. These issues are relevant to emerging treatment regimens.
All recurrences in the study of Silverman and Krebs (21) were in the perianastomotic region. In our study, only 29% of the recurrences were perianastomotic. An additional 40% were retrovesical. These two sites are generally well identified at transrectal US and carry the highest detection rate for transrectal biopsy in attempts to confirm a tumor recurrence (18). Thirty percent of local recurrences in our study occurred elsewhere in the pelvis, however, at sites that are often missed with transrectal US and biopsy (within retained seminal vesicles or at the lateral or anterior surgical margins). MR imaging has the potential to direct transrectal biopsy to these sites and may thus lead to a better diagnostic yield. An additional benefit of MR imaging with use of both an endorectal and a pelvic phased-array coil, as in our study, is the ability to also evaluate pelvic lymph node and bone status, thus allowing detection of all sites of pelvic relapse in a single examination.
Unlike the limited interobserver agreement for presurgical staging of prostate cancer (24), agreement between readers for detection of local recurrence in our study was excellent ( = 0.91). This was also the case in the study of Silverman and Krebs (21), who had 100% reader agreement.
In our cohort, we found no association between PSA level and the depicted size of a recurrence. Fifteen (38%) patients with local recurrence had a PSA level lower than 0.4 ng/mL at the time of MR imaging, and in two patients, the PSA level was not measurable. This may be partially due to the multidisciplinary approach to the treatment of prostate cancer with a combination of hormonal, immunologic, nutritional, and other forms of therapy. Twenty-five (64%) patients had a PSA level below 1.5 ng/mL at the time of MR imaging, which is currently the PSA level under which the American Society of Therapeutic Radiation Oncology recommends to begin salvage radiation therapy (25), thus further reinforcing the potential value of MR imaging.
It is challenging to prove that a patient who has undergone prostatectomy has a local recurrence, since pathologic confirmation is often not available. We based our validation criteria on a constellation of pathologic, laboratory, and imaging findings, similar to those used by urologists and oncologists in their daily practice. This challenge is likely unrelated to the retrospective nature of our study but rather to the profile of disease and disease management of prostate cancer. We recognize the potential for bias in excluding from analysis the patients who could not have their disease status validated, which is a setback of a retrospective study. Patient selection for referral for MR imaging is another potential source of bias, resulting from this being a retrospective study. It is possible that our sample group is not adequate to achieve statistical significance; this is in part due to the lack of guidelines for patient referral, resulting in a relatively small number of patients who have undergone both prostatectomy and endorectal MR imaging. Additionally, the difficulty of clinical verification of a local recurrence greatly limits the size of our study group. Because of the size of our study group, 95% CIs for our specificity and correlation coefficient are wide. Although our results may lack statistical stability, our patient cohort reflects the population being referred for pelvic MR imaging. We therefore believe that the trends shown in our study are valid. The lack of clear guidelines as to when to refer a patient for MR imaging after prostatectomy raises the need for a large prospective study aimed at defining such guidelines.
Currently, treatment of a local recurrence following prostatectomy is most commonly delivered by means of radiation therapy. When relying on clinical factors alone, radiation oncologists often face a dilemma regarding when to begin salvage radiation therapy. This is particularly true when a patient has a profile of a rapidly increasing PSA level (doubling time of 
6 months), which may indicate the existence of lymph node or osseous metastases. We believe that the ability to visualize a local recurrence may provide a guide for initiating therapy in such cases. Additionally, radiation therapy is usually not a treatment option for patients with metastatic prostate cancer; however, sometimes local and metastatic relapses occur simultaneously (26,27). Potentially, a patient with a large pelvic mass, even if he has metastatic disease, may benefit from pelvic radiation for palliation, similar to patients with advanced primary tumors of the prostate (28). MR imaging may play a role in the detection of such local pelvic masses, as well as demonstrating the effects of a pelvic mass on the surrounding anatomy and possible invasion into adjacent structures. The ability to demonstrate measurable disease on MR images may also provide a means for following the response to therapy, be it radiation therapy, hormonal therapy, chemotherapy, or others. With the increasing availability of modern radiation techniques, such as three-dimensional conformal radiation therapy and intensity-modulated radiation therapy, MR imaging may, in the near future, also assist in the planning of radiation fields. The clinical utility of depicting a local recurrence with MR imaging clearly goes beyond detection alone.
MR imaging may demonstrate a high proportion of local recurrence following prostatectomy; the ability to depict and localize small recurrences in patients with minimally increased PSA levels makes it a valuable tool with clinical relevance. We propose incorporating endorectal coil MR imaging in the evaluation of patients who have undergone prostatectomy and are suspected of having local recurrence.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, H.H.; study concepts, T.S., H.H., P.T.S., H.I.S., P.W.S.; study design, T.S., H.H., P.W.S.; literature research, C.N.O., T.S., P.W.S.; clinical studies, T.S., H.I.S., P.T.S., P.W.S.; data acquisition, H.H., L.H.S., T.S., C.N.O.; data analysis/interpretation, H.H., L.H.S., T.S.; statistical analysis, T.S., C.N.O., H.H.; manuscript preparation, T.S., C.N.O.; manuscript definition of intellectual content and editing, H.H.; manuscript revision/review, P.T.S., H.I.S., P.W.S., L.H.S.; manuscript final version approval, H.H.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Casciani, E. Polettini, E. Carmenini, I. Floriani, G. Masselli, L. Bertini, and G. F. Gualdi Endorectal and Dynamic Contrast-Enhanced MRI for Detection of Local Recurrence After Radical Prostatectomy Am. J. Roentgenol., May 1, 2008; 190(5): 1187 - 1192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Stephenson, P. T. Scardino, M. W. Kattan, T. M. Pisansky, K. M. Slawin, E. A. Klein, M. S. Anscher, J. M. Michalski, H. M. Sandler, D. W. Lin, et al. Predicting the Outcome of Salvage Radiation Therapy for Recurrent Prostate Cancer After Radical Prostatectomy J. Clin. Oncol., May 20, 2007; 25(15): 2035 - 2041. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Hricak, P. L. Choyke, S. C. Eberhardt, S. A. Leibel, and P. T. Scardino Imaging Prostate Cancer: A Multidisciplinary Perspective Radiology, April 1, 2007; 243(1): 28 - 53. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Sella, L. H. Schwartz, and H. Hricak Retained Seminal Vesicles After Radical Prostatectomy: Frequency, MRI Characteristics, and Clinical Relevance Am. J. Roentgenol., February 1, 2006; 186(2): 539 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Beyersdorff, K. Taymoorian, T. Knosel, D. Schnorr, R. Felix, B. Hamm, and H. Bruhn MRI of Prostate Cancer at 1.5 and 3.0 T: Comparison of Image Quality in Tumor Detection and Staging Am. J. Roentgenol., November 1, 2005; 185(5): 1214 - 1220. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. I. Scher, M. J. Morris, W. K. Kelly, L. H. Schwartz, and G. Heller Prostate Cancer Clinical Trial End Points: "RECIST"ing a Step Backwards Clin. Cancer Res., July 15, 2005; 11(14): 5223 - 5232. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Schoder, K. Herrmann, M. Gonen, H. Hricak, S. Eberhard, P. Scardino, H. I. Scher, and S. M. Larson 2-[18F]Fluoro-2-Deoxyglucose Positron Emission Tomography for the Detection of Disease in Patients with Prostate-Specific Antigen Relapse after Radical Prostatectomy Clin. Cancer Res., July 1, 2005; 11(13): 4761 - 4769. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. J. Kelloff, J. M. Hoffman, B. Johnson, H. I. Scher, B. A. Siegel, E. Y. Cheng, B. D. Cheson, J. O'Shaughnessy, K. Z. Guyton, D. A. Mankoff, et al. Progress and Promise of FDG-PET Imaging for Cancer Patient Management and Oncologic Drug Development Clin. Cancer Res., April 15, 2005; 11(8): 2785 - 2808. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
