Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2363040615
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Prando, A.
Right arrow Articles by Figueiredo, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Prando, A.
Right arrow Articles by Figueiredo, E.
(Radiology 2005;236:903-910.)
© RSNA, 2005

Genitourinary Imaging

Prostatic Biopsy Directed with Endorectal MR Spectroscopic Imaging Findings in Patients with Elevated Prostate Specific Antigen Levels and Prior Negative Biopsy Findings: Early Experience1

Adilson Prando, MD, John Kurhanewicz, PhD, Alexandre P. Borges, MD, Eduardo M. Oliveira, Jr, MD and Eduardo Figueiredo

1 From the Department of Radiology, Vera Cruz Hospital, Av Andrade Neves 707, Campinas SP, 13013-161, Brazil (A.P., A.P.B., E.M.O.); Department of Radiology, University of California, San Francisco (J.K.); and MRI Advanced Application Service, GE Medical Systems, São Paulo, Brazil (E.F.). Received April 5, 2004; revision requested June 15; revision received October 7; accepted November 11. Address correspondence to A.P. (e-mail: aprando{at}mpc.com.br).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To prospectively evaluate the accuracy of transrectal ultrasonography (US)-guided biopsy directed with magnetic resonance (MR) spectroscopic imaging in patients with an elevated prostate specific antigen (PSA) level and negative findings at prior biopsy by using subsequent biopsy results as the reference standard.

MATERIALS AND METHODS: The committee on human research approved this study, and written informed consent was obtained. MR imaging and MR spectroscopic imaging were performed in 42 men (age range, 45–75 years; average age, 63.3 years; median age, 65 years) with negative findings at two or more prostatic biopsies and at digital rectal examination. MR spectroscopic data were rated on a scale of 1 (benign) to 5 (malignant) on the basis of standardized metabolic criteria. Abnormal voxels were overlaid on the corresponding transverse transrectal US images and used to perform voxel-guided biopsy of the prostate. All patients subsequently received an extended-pattern biopsy scheme.

RESULTS: Thirty-one of 42 patients demonstrated metabolic abnormalities that were suspicious for cancer (voxels with scores ≥ 4). Eleven patients with negative MR spectroscopic imaging results also had negative biopsy findings. Cancer was detected in 17 (55%) of 31 men with positive MR spectroscopic imaging findings (voxels with scores ≥ 4) with a sensitivity of 100%, specificity of 44%, positive predictive value of 55%, negative predictive value of 100%, and accuracy of 67%. In men with at least one spectroscopic voxel with a score of 5 (12 of 17 men), the sensitivity, specificity, positive and negative predictive values, and accuracy were 71%, 84%, 75%, 81%, and 79%, respectively.

CONCLUSION: Metabolic data from MR spectroscopic imaging can be transferred to transrectal US images and used to sample regions of cancer in men with rising PSA levels and negative findings at prior biopsy with good accuracy.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Many men find themselves in the clinical dilemma of having an elevated or rising prostate-specific antigen (PSA) level and at least one prostatic biopsy with negative findings (1–3). New biopsy strategies with increased numbers of systematically placed biopsy cores have been developed to decrease the false-negative rate associated with conventional sextant prostate biopsy (4–7); however, many men still find themselves in this clinical dilemma, and the best way to care for these patients remains uncertain. Helical computed tomography and conventional endorectal magnetic resonance (MR) imaging of the prostate have shown promise in the improved detection of cancer within the prostate (8,9). More recently, the addition of metabolic data provided by three-dimensional MR spectroscopic imaging to the anatomic data obtained with conventional MR imaging of the prostate has demonstrated an improvement in the localization of cancer to a sextant of the prostate (10), the estimation of extracapsular extension (11), and the assessment of the aggressiveness of prostate cancer (12). Specifically, MR spectra from regions of prostate cancer show a significant reduction or absence of citrate and polyamines, while the choline level is elevated relative to the creatine level, thus resulting in significant changes in the choline-plus-creatine–to–citrate ratio in regions of cancer (13,14).

We undertook this study to prospectively evaluate the accuracy of transrectal ultrasonography (US)-guided biopsy directed with MR spectroscopic imaging in patients with elevated PSA levels and prior negative biopsy findings by using subsequent biopsy results as the reference standard.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Patients
Between July 2002 and October 2003, 42 consecutive patients underwent a combined endorectal MR imaging and MR spectroscopic examination prior to transrectal US-guided biopsy of the prostate. The committee on human research at Vera Cruz Hospital approved all studies, and written informed consent was obtained from all patients. The patients ranged in age from 45 to 75 years (average age, 63.3 years; median age, 65 years). All patients had negative findings at digital rectal examination, and they all underwent annual repetitive PSA tests. In 28 patients, serial PSA levels were measured less than 6 months before the evaluation. No patient underwent biopsy less than 2 months before the investigation. The average PSA value was 6.8 ng/mL (range, 4.1–15.3 ng/mL).

Prostatic volume was determined by using the equation for the volume of an ellipsoid (ie, anterior-posterior diameter · width · height · 0.52). Before endorectal MR spectroscopic imaging, 10 of these men had two sextant or extended (ie, more than six cores) negative transrectal US-guided biopsy findings, 13 men had three, nine men had four, six men had five, and four men had six.

MR Imaging
Conventional MR imaging was performed with a 1.5-T MR imager (Signa; GE Medical Systems, Milwaukee, Wis). Patients were examined by using the body coil for signal acquisition and a combination of a pelvic phased-array coil (Torso PA; GE Medical Systems) with a commercially available balloon-covered endorectal coil (Endo ATD; Medrad, Pittsburgh, Pa) for signal reception. The balloon-covered endorectal coil was inflated with 90 mL of air for the first 20 patients.

This amount of air was well tolerated by the majority of patients. To reduce the high magnetic field susceptibility at the air-tissue interface and improve the quality of MR spectroscopic imaging data, we replaced the air with the same amount of liquid perfluorocarbon in the remaining 22 patients. On MR images, the prostate was evaluated with transverse spin-echo T1-weighted MR images by using the following parameters: repetition time msec/echo time msec, 575/minimum; section thickness, 3 mm; matrix, 256 x 224; two signals acquired; field of view, 13 cm; intersection gap, 0 mm; bandwidth, 20.83 kHz. Transverse and transverse-oblique T2-weighted images were obtained with the following parameters: 3500/130; three signals acquired; section thickness, 3 mm; intersection gap, 0 mm; field of view, 13 cm; matrix, 256 x 224; bandwidth, 20.83 kHz. For the transverse images, phase encoding was in the anterior-to-posterior direction. T2-weighted sagittal MR images were obtained with the following parameters: 4000/150; two signals acquired; section thickness, 5 mm; intersection gap, 2 mm; field of view, 15 cm; matrix, 256 x 192; bandwidth, 41.67 kHz. For the T2-weighted sagittal MR images, phase encoding was in the superior-to-inferior direction.

MR Spectroscopic Imaging
After review of the transverse T2-weighted images, a spectroscopic imaging volume was selected to maximize coverage of the prostate while minimizing the inclusion of periprostatic fat and rectal air. Three-dimensional MR spectroscopic data were acquired by using water and a lipid-suppressed double-spin-echo point-resolved spectroscopy sequence, which was optimized for quantitative detection of both choline and citrate. Water and lipid suppression was achieved by using the spectral-spatial pulses capable of both volume selection and frequency selection (15).

Outer voxel saturation pulses were also used to eliminate signals from adjacent tissues, especially periprostatic lipids and rectal wall tissue (16).

Data sets were acquired as 16 x 8 x 8 phase-encoded spectral arrays (1024 voxels; nominal spatial resolution, 0.34 cm3; 1000/130; acquisition time, 17 minutes. The total examination time was 1 hour, including coil placement and patient positioning.

Three-dimensional MR spectroscopic imaging data were processed, aligned with the corresponding MR imaging data, displayed, and analyzed by using Functool software (GE Medical Systems). The raw spectral data were apodized with a 1-Hz Gaussian function, and Fourier transformation was performed in the time domain and three spatial domains. The estimation of choline, creatine, and citrate peak parameters (ie, peak area, peak height, peak location, and line width) was accomplished by using an iterative procedure that allowed initial identification of statistically significant peaks (signal-to-noise ratio > 5:1) in the magnitude spectrum and enabled estimation of a frequency shift that best aligned these peaks with the expected locations of choline, creatine, citrate, and residual water. Subsequently, the spectra were phased by using the phase of the residual water resonance and baseline corrected by using a local nonlinear fit to the nonpeak regions of the spectra. Choline, creatine, and citrate peak areas were obtained with numerical integration over a frequency range determined with the metabolite peak locations and widths. In regions of healthy tissue, the polyamine peak that resonates between choline and creatine could not be sufficiently resolved, and it was incorporated in the area of the peak choline-plus-creatine level. In the final 22 patients, a software upgrade allowed us to use the voxel shifting capability to better align the spectral data with the anatomy.

MR Data Analysis
MR spectroscopic imaging data were overlaid on the corresponding transverse T2-weighted MR images and evaluated in consensus by three radiologists (A.P., A.P.B., and E.M.O., each with 2 years of experience in the interpretation of MR spectroscopic images of the prostate) to determine which voxels were suitable for analysis. Images were considered suitable if they were not contaminated by insufficiently suppressed water or lipids (17) (Fig 1). Suitable spectroscopic voxels were rated as optimal, fair, or poor on the basis of spectral quality, and they were subsequently scored according to a recently developed standardized five-point scale (18). Specifically, a study was considered to be of optimal spectral quality if the signal-to-noise ratio of all metabolites was greater than 10, all metabolic resonances were well resolved, and there were no baseline distortions due to residual water or lipids. A study was considered to be of fair spectral quality if the signal-to-noise ratio of all metabolites was between eight and 10, all metabolic resonances were reasonably well resolved, or there were minimal baseline distortions due to residual water or lipids. Studies with lower signal-to-noise ratios and substantial lipid contamination were considered to be of poor spectral quality. The quality of spectral data was rated as optimal in 23 patients, fair in 10, and poor in nine.



View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Graph shows proton spectrum taken from a three-dimensional MR spectroscopic imaging data set of a patient with normal prostatic gland (score, 1). Note the choline (Co) peak at 3.2 ppm, polyamine (Pa) peak at 3.1 ppm, creatine (Cr) peak at 3.0 ppm, and citrate (Ci) peak at 2.6 ppm.

 
Primary scores ranged from 1 to 5, and they were assigned on the basis of the mean healthy ratio of the choline-plus-creatine–to–citrate ratio (18). The polyamine peak occurs between the creatine and choline peaks, and currently, it cannot be entirely resolved from these peaks; therefore, it is incorporated into the choline and creatine peak areas. The mean choline-plus-creatine–to–citrate ratio was defined as 0.22 ± 0.013 on the basis of a previously published study that used the same MR spectroscopic data acquisition and processing used in this study (18). A score of 1 was assigned to voxels with a choline-plus-creatine–to–citrate ratio greater than or equal to 1 standard deviation of the mean healthy value. A score of 2 was assigned to voxels with a choline-plus-creatine–to–citrate ratio of more than 1 and less than or equal to 2 standard deviations above the mean healthy value. A score of 3 was assigned to voxels with a choline-plus-creatine–to–citrate ratio of more than 2 and less than or equal to 3 standard deviations above the mean healthy value. A score of 4 was assigned to voxels with a choline-plus-creatine–to–citrate ratio of more than 3 and less than or equal to 4 standard deviations above the mean healthy value. A score of 5 was assigned to voxels with a choline-plus-creatine–to–citrate ratio of more than 4 standard deviations above the mean healthy value (Fig 2). Further adjustments were made to the primary score on the basis of the elevation of choline-to-creatine ratio, the loss of polyamines on the basis of increased resolvability of choline and creatine, and spectral signal-to-noise ratio, as previously described.



View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Graph shows proton spectrum taken from a three-dimensional MR spectroscopic imaging data set of a voxel containing prostate cancer (score, 5). Note the marked elevation of the choline (Co) peak and the reduction of the polyamine (Pa) and citrate (Ci) peaks. The choline-plus-creatine–to–citrate ratio is more than 4 standard deviations greater than the mean healthy value.

 
When the choline-to-creatine ratio was greater than or equal to 2 with a primary score of 2 or 3, the overall score was increased to 4. When the choline-to-creatine ratio was less than 2 or there was no reduction in polyamines with a primary score of 4 or 5, the overall score was decreased by one (ie, 3 or 4, respectively). We considered poor signal-to-noise ratio when it was less than 8 for voxels with a score of 3 to 5 and less than 5 for voxels with a score of 1 or 2. In the presence of a poor signal-to-noise ratio, a score of 1 became a score of 3, a score of 2 or 4 became a score of 3, and a score of 5 became a score of 4 (18). This scoring scale was developed for the metabolic identification of peripheral zone cancers. In the central gland, it is more difficult to identify prostate cancer because of the overlap of the MR imaging and MR spectroscopic findings for prostate cancer and predominantly stromal benign prostatic hyperplasia (19); central gland tumors were not included in this investigation.

For all patients who underwent MR spectroscopic imaging–guided biopsy, the metabolic abnormalities consisted of multiple contiguous MR spectroscopic voxels with scores of 4 or 5 (ie, clusters of abnormal voxels).

Special attention was given during MR spectroscopic imaging to spectral analysis of the median portion of the peripheral zone. Specifically, voxels that included the ejaculatory ducts and urethra were disregarded because these voxels often have high choline levels that are similar to those in patients with cancer because of the glycerophosphocholine in the fluids that are contained within these structures.

Biopsy
To evaluate the accuracy of transrectal US-guided biopsy directed with MR spectroscopic imaging, all patients underwent extended-pattern biopsy (12 cores total, two cores of each sextant). Patients with abnormal MR spectroscopic imaging voxels (scores ≥ 4) also underwent biopsy with MR spectroscopic guidance. All the cores were labeled according to their sextant location and corresponding abnormal voxels.

Focal hypointense areas in the peripheral zone on T2-weighted MR images were not subjected to guided biopsy unless they were associated with a region classified as abnormal at MR spectroscopic imaging (score ≥ 4). Among the 42 patients studied, 31 had MR spectroscopic imaging abnormalities (score of 4, 5, or both) and 11 patients did not have MR spectroscopic imaging abnormalities (score of 3, 2, or 1).

After patient preparation (which included antibiotic therapy and a cleansing enema) and administration of local anesthetic, a complete transrectal US evaluation of the prostate was performed prior to biopsy. The prostate gland was evaluated with a 7-MHz endorectal transducer with the patient in the left lateral decubitus position.

US transverse scans were obtained to reproduce the same gland morphologic findings obtained with T2-weighted transverse MR images to better localize the suspicious MR spectroscopic findings. These scans were obtained by using internal and external anatomic landmarks (ie, external sphincter, veromontanum, surgical and anatomic capsule, urethra, neurovascular bundle, hypertrophic central gland nodules, and seminal vesicles). Additional useful topographic information was obtained with the correlation of the transverse images with the midsagittal T2-weighted scout view (Fig 3). By using these criteria, the suspicious MR spectroscopic areas were projected as accurately as possible on US scans, and transrectal US–guided biopsy was performed. To facilitate the correlation of both methods, similar image zoom factors were used to obtain a similar peripheral zone thickness. Direct voxel-guided biopsy was then performed, with removal of two or three cores from the areas classified as abnormal with MR spectroscopic imaging. All biopsies were performed by two radiologists (A.P. and A.P.B., with 9 and 7 years of experience in transrectal US–guided biopsy, respectively) who were also involved in the acquisition and interpretation of the MR images and MR spectroscopic images. The biopsy cores were labeled to reflect the location of the biopsy and posterior end of the biopsy core. An extended-pattern transrectal US biopsy scheme sampling the base, midgland, and apex was also performed in these patients.



View larger version (127K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Midsagital T2-weighted (4000/150) MR scout view shows how transverse T2-weighted MR images were obtained.

 
Several patients had hypointense areas on T2-weighted MR images in the peripheral zone that were evaluated in consensus by the three radiologists (A.P., A.P.B., and E.M.O.). These hypointense areas did not undergo directed biopsy unless they were associated with voxels with scores of 4 or 5 because we considered these areas to have been sampled with extended-pattern biopsy. Since transition-zone biopsies were previously performed in the majority of patients studied and because of the complications in the identification of cancer within the central gland on MR spectroscopic images, we did not include central gland cancers in this analysis.

In the 31 patients who underwent MR spectroscopic imaging targeted biopsies and an extended-pattern biopsy, 19 procedures were performed by one radiologist (A.P.), and 12 were performed by another radiologist (A.P.B.). In the remaining 11 patients who underwent only an extended-pattern biopsy, six procedures were performed by one radiologist (A.P.), and five were performed by another radiologist (A.P.B.).

Histopathologic Analysis
Histopathologic analysis was performed for all biopsy samples and used as the standard of reference for the MR spectroscopic imaging findings. Histopathologic analysis was performed by a pathologist with 24 years of experience in prostate pathology. The histopathologic analysis included determination of the (a) Gleason score, (b) number of positive cores, (c) percentage of involvement of cancer in each core, and (d) presence of atypical small acinar proliferation, high-grade prostatic intraepithelial neoplasia, and prostatic atrophy.

Statistical Analysis
Descriptive statistical data (ie, sensitivity, specificity, and positive and negative predictive value) were determined by using statistical software (Stata, release 8.0; Stata, College Station, Tex) with two different dichotomized ratings (ie, voxels having scores of ≥4 and having at least 1 voxel with a score of 5); biopsy results served as the reference standard. A {chi}2 test was used to determine whether there was a significant difference in sensitivity and specificity between the MR spectroscopic imaging–guided biopsy and the extended-pattern biopsy at a 95% confidence level. A P value of less than .05 indicated a statistically significant difference.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Better spectral resolution was accomplished in the final 22 patients of our study after endorectal coil air was replaced with perfluorocarbon. This product allowed us to reduce magnetic field susceptibility problems with the rectal air–tissue interface and improved spectra (Fig 4 ).



View larger version (30K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a. Three-dimensional spectra obtained in a patient by using (a) air and (b) perfluorocarbon within the endorectal coil. Note the improvement of the spectra, with reduced noise in the baseline and better metabolite discrimination. Polyamines (Pa) peak resolved from choline (Co) and creatine (Cr) and the citrate (Ci) doublet.

 


View larger version (28K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b. Three-dimensional spectra obtained in a patient by using (a) air and (b) perfluorocarbon within the endorectal coil. Note the improvement of the spectra, with reduced noise in the baseline and better metabolite discrimination. Polyamines (Pa) peak resolved from choline (Co) and creatine (Cr) and the citrate (Ci) doublet.

 
Direct voxel-guided biopsies were performed in the areas that appeared to be abnormal at MR spectroscopic imaging, as illustrated in Figure 5 . An extended-pattern transrectal US biopsy scheme was performed in all patients in our study (Fig 6).



View larger version (141K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a. Images were obtained in a 65-year-old man with a PSA level of 7.8 ng/mL, negative findings at digital rectal examination, and three previous negative biopsies. (a) Transverse T2-weighted fast spin-echo (3500/130) image. Note the areas of low signal intensity in the right (*) and left (arrows) portions of the basal peripheral zone. Note a large hyperplastic nodule (HN) in the left transition zone and the surgical capsule (C), which will be used as an internal landmark for adequate transrectal US-guided biopsy. (b) Spectroscopic imaging data overlaid on this MR image shows a grade 5 voxel (voxel 15) (c) with a highly elevated choline (Co) level, an unchanged creatine (Cr) level, a loss of polyamine, and a reduced citrate doublet (Co + Cr/Ci = 1.92), which is more than 4 standard deviations greater than the mean healthy value. (d) The abnormal voxel was projected on the corresponding transverse US image, directed voxel-guided biopsy was performed, and prostate cancer (Gleason score of 3 + 3) was diagnosed. The internal prostate landmarks used in this case were the surgical capsule (C) and a large hyperplastic nodule (HN) on the left transition zone, which was previously identified on the T2-weighted transverse image.

 


View larger version (157K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b. Images were obtained in a 65-year-old man with a PSA level of 7.8 ng/mL, negative findings at digital rectal examination, and three previous negative biopsies. (a) Transverse T2-weighted fast spin-echo (3500/130) image. Note the areas of low signal intensity in the right (*) and left (arrows) portions of the basal peripheral zone. Note a large hyperplastic nodule (HN) in the left transition zone and the surgical capsule (C), which will be used as an internal landmark for adequate transrectal US-guided biopsy. (b) Spectroscopic imaging data overlaid on this MR image shows a grade 5 voxel (voxel 15) (c) with a highly elevated choline (Co) level, an unchanged creatine (Cr) level, a loss of polyamine, and a reduced citrate doublet (Co + Cr/Ci = 1.92), which is more than 4 standard deviations greater than the mean healthy value. (d) The abnormal voxel was projected on the corresponding transverse US image, directed voxel-guided biopsy was performed, and prostate cancer (Gleason score of 3 + 3) was diagnosed. The internal prostate landmarks used in this case were the surgical capsule (C) and a large hyperplastic nodule (HN) on the left transition zone, which was previously identified on the T2-weighted transverse image.

 


View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5c. Images were obtained in a 65-year-old man with a PSA level of 7.8 ng/mL, negative findings at digital rectal examination, and three previous negative biopsies. (a) Transverse T2-weighted fast spin-echo (3500/130) image. Note the areas of low signal intensity in the right (*) and left (arrows) portions of the basal peripheral zone. Note a large hyperplastic nodule (HN) in the left transition zone and the surgical capsule (C), which will be used as an internal landmark for adequate transrectal US-guided biopsy. (b) Spectroscopic imaging data overlaid on this MR image shows a grade 5 voxel (voxel 15) (c) with a highly elevated choline (Co) level, an unchanged creatine (Cr) level, a loss of polyamine, and a reduced citrate doublet (Co + Cr/Ci = 1.92), which is more than 4 standard deviations greater than the mean healthy value. (d) The abnormal voxel was projected on the corresponding transverse US image, directed voxel-guided biopsy was performed, and prostate cancer (Gleason score of 3 + 3) was diagnosed. The internal prostate landmarks used in this case were the surgical capsule (C) and a large hyperplastic nodule (HN) on the left transition zone, which was previously identified on the T2-weighted transverse image.

 


View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5d. Images were obtained in a 65-year-old man with a PSA level of 7.8 ng/mL, negative findings at digital rectal examination, and three previous negative biopsies. (a) Transverse T2-weighted fast spin-echo (3500/130) image. Note the areas of low signal intensity in the right (*) and left (arrows) portions of the basal peripheral zone. Note a large hyperplastic nodule (HN) in the left transition zone and the surgical capsule (C), which will be used as an internal landmark for adequate transrectal US-guided biopsy. (b) Spectroscopic imaging data overlaid on this MR image shows a grade 5 voxel (voxel 15) (c) with a highly elevated choline (Co) level, an unchanged creatine (Cr) level, a loss of polyamine, and a reduced citrate doublet (Co + Cr/Ci = 1.92), which is more than 4 standard deviations greater than the mean healthy value. (d) The abnormal voxel was projected on the corresponding transverse US image, directed voxel-guided biopsy was performed, and prostate cancer (Gleason score of 3 + 3) was diagnosed. The internal prostate landmarks used in this case were the surgical capsule (C) and a large hyperplastic nodule (HN) on the left transition zone, which was previously identified on the T2-weighted transverse image.

 


View larger version (59K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6. Schematic diagram of an extended-pattern biopsy (base view). Far lateral sites (regions 7–12) are added to the sites that are usually sampled with a conventional sextant biopsy strategy (regions 1–6). A region containing an abnormal spectroscopic voxel (V) with a score of 5 (eg, definitely cancer) and sampled with two cores is illustrated on the midline of the peripheral zone.

 
In the 31 patients who underwent MR imaging and MR spectroscopic imaging and had abnormal spectroscopic voxels (score of 4, 5, or both) that were considered positive for metabolic abnormalities, cancer was detected in 17 (55%) patients with a sensitivity of 100%, specificity of 44%, and positive predictive value of 55%. In all 11 patients with negative MR spectroscopic imaging findings (score of 1, 2 or 3), there was no evidence of cancer at pathologic analysis; the negative predictive value was 100% and accuracy was 67% (Table 1). The other 14 of 31 patients with positive scores of 4 or 5 showed no evidence of cancer, but four (29%) of these patients had focal prostatic atrophy at histopathologic analysis (Fig 7).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Biopsy Findings in Patients with Voxel Score of 4 and/or 5

 


View larger version (37K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7. Graph shows a proton spectrum taken from a three-dimensional MR spectroscopic imaging data set of a voxel in a prostate with diffuse focal prostatic atrophy that was histopathologically observed in all cores obtained from the peripheral zone. Note the moderate elevation of the choline (Co) level, the reduction of the citrate (Ci) level, and the preservation of the polyamine (Pa) peak. Cr = creatine.

 
In Table 2, the analysis is restricted to men with at least one spectroscopic voxel with a score of 5. In this group of patients, carcinoma was detected in 12 (75%) of 16 patients with positive MR spectroscopic imaging findings, yielding a sensitivity of 71%, specificity of 84%, positive predictive value of 75%, negative predictive value of 81%, and accuracy of 79%.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Biopsy Findings in Patients with at Least One Voxel Score of 5

 
In Table 3, the extended-pattern biopsy and the voxel-guided biopsy findings were compared. In the extended-pattern biopsy group, a total of 252 sextants underwent biopsy, and findings in 10 (4%) were positive for cancer.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Findings in Sextants that Underwent Extended Pattern Biopsies and MR Spectroscopy–guided Biopsy

 
In the MR spectroscopic imaging biopsy group, 42 (44%) of 96 abnormal sextants (containing clusters of voxels with scores of 4 and/or 5) were positive (P < .01).

Combination of the extended-pattern biopsy and MR spectroscopic imaging–guided biopsy results (Table 4) yielded a sensitivity of 85%, specificity of 89%, positive predictive value of 58%, negative predictive value of 97%, and accuracy of 89% (P < .05).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Findings Combining Extended Pattern Biopsy and MR Spectroscopy–guided Biopsy

 
The biopsy findings in patients with sextants containing areas that appeared hypointense on T2-weighted MR images showed a sensitivity of 88%, specificity of 28%, positive predictive value of 45%, negative predictive value of 78%, and accuracy of 52% (Table 5).


View this table:
[in this window]
[in a new window]

  		TABLE 5. Biopsy Findings in Patients with Sextants Showing Hypointense Areas on T2-weighted Images

 
In our study, the average prostatic volume was 82 g (mean prostate volume, 62 g; range, 35–164 g). The average prostate volume in patients with cancer was higher than that in patients without cancer (87 g vs 58 g, respectively). Five of 13 patients with positive biopsy findings had very large prostates (>75 g).

In the group of patients with abnormal MR spectroscopic imaging findings, subsequent extended-pattern biopsy revealed additional regions of cancer in eight of 31 patients (only one additional core of cancer per patient).

In the 17 patients in whom cancer was detected with MR spectroscopic imaging and confirmed at biopsy, 10 (59%) had at least one site of cancer located toward the midline of the peripheral zone.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
A recent report (20) has demonstrated that MR imaging and MR spectroscopic imaging have the potential to enable the identification of cancer foci and direct transrectal US-guided biopsy in patients with negative findings at previous transrectal US biopsy. Similarly, our results show that radiologists who perform MR imaging and MR spectroscopic imaging examinations and transrectal US-guided biopsy can transfer metabolic data from MR spectroscopic imaging to transrectal US images and effective use this data to sample regions of cancer in men with rising PSA levels and prior negative findings at biopsy. The sensitivity, specificity, and overall accuracy of MR spectroscopic imaging targeted transrectal US-guided biopsy was good and could be changed on the basis of metabolic criteria used to define the region of metabolic abnormality. Specifically, when MR spectroscopic imaging scores of 4 and 5 were used to target regions of abnormality, a high sensitivity (100%) and negative predictive value (100%) were obtained; however, a much lower specificity (44%) and positive predictive value (55%) were also obtained. The use of only spectroscopic voxels that demonstrate a more dramatic metabolic abnormality (ie, MR spectroscopic imaging score of 5) to target regions of cancer yielded a higher specificity (84%) and positive predictive value (75%) and a lower sensitivity (71%) and negative predictive value (81%). The overall accuracy of detection was greater when MR spectroscopic imaging scores of 5 (79%) were used than when MR spectroscopic imaging scores of 4 and 5 (67%) were used for targeting regions of abnormality for transrectal US-guided biopsy.

An MR spectroscopic imaging score of 5 is more accurate than an MR spectroscopic imaging score of 4 and 5; however, from the clinical standpoint, it is adequate to include scores of 4 and 5 to avoid missing any patient with prostate cancer. Comparison of the extended-pattern biopsy results when the voxel-guided biopsy results showed a significant increase in cancer detection from 4% to 44%, respectively (P < .01). By using the combined approach (ie, extended-pattern and voxel-guided biopsy), there was an overall gain in accuracy (89%) when compared with voxel-guided biopsy only (67%, score of 4 and 5). Regarding the fact that the extended-pattern biopsy protocol showed cancer in eight of 17 patients, it should be emphasized that regions that had suspicious MR spectroscopic imaging findings were not avoided; therefore, they might have been resampled during the extended-pattern biopsy, particularly in patients with small prostates (30–60 g).

We must keep in mind that our accuracy with MR spectroscopic imaging reflects a prediction of biopsy results. In comparison, conventional MR imaging alone showed a positive predictive value of 40.0%, a negative predictive value of 94.4%, and an overall accuracy of 69.7% in targeting cancer for biopsy; this was in a patient population at higher risk for having cancer than the patients in this study (9). Specifically, of 33 patients in the MR-based biopsy study, nine had negative findings at one biopsy, and 77% (24 patients) had positive findings at digital rectal examination. Also, the serum PSA level was higher (4.8–36.0 ng/mL; median PSA level, 11.7 ng/mL) in these patients than in the patients in our study (9). In this study, all patients had negative findings at digital rectal examination and negative findings at two or more biopsies, and the average PSA level was only 7.6 ng/mL (range, 4.1–15.3 ng/mL). The earlier stage of cancer demonstrated in this study better reflects the type of cancer that would be present in patients who would be clinically referred for prostate MR imaging and MR spectroscopic imaging. The presence of several hypointense areas on T2-weighted MR images obtained with normal MR spectroscopic imaging may be a result of postbiopsy changes, since all patients had at least two negative biopsies. These findings could explain the lower accuracy in the biopsy results of sextants containing only hypointense areas on T2-weighted images.

Prostate volume has been shown to be an important factor in men with rising PSA levels and negative findings at biopsy, since there is an inverse relationship between prostate size and the likelihood of detecting prostate cancer with transrectal US-guided biopsy (21). Thus, one would have expected more cancer to be found in smaller prostates in this study. In fact, the average prostate volume was higher in patients with cancer than in patients without cancer (87 g vs 58 g, respectively). Moreover, five of 13 patients with positive biopsy findings had very large prostates (>75 g). This finding is consistent with the fact that more cancer is missed with transrectal US-guided biopsy in men with larger prostates, since all of the men examined had negative findings at transrectal US-guided biopsy. Additionally, this finding suggests that in this population of men, MR imaging and MR spectroscopic imaging can help target cancers for transrectal US-guided biopsy.

There was also an interesting trend in the location of cancer that was missed at the original transrectal US-guided biopsy but was detected with MR spectroscopic imaging–targeted transrectal US–guided biopsy. Ten (59%) of the 17 patients with cancer had at least one positive core located toward the midline of the peripheral zone. This could be explained by the fact that in most systematic transrectal US-guided prostate biopsy schemes, the midline area does not routinely undergo biopsy. This result suggests that the addition of a midline biopsy to a systematic biopsy scheme should improve the detection of cancer.

Of the 14 patients with positive MR spectroscopic imaging findings (score of 4, 5, or both), four (29%) had results that were negative for cancer but showed focal proliferative atrophy in the area with abnormal voxels. We might speculate about the reason why focal prostatic atrophy (22) manifests as a false-positive MR spectroscopic imaging finding. This is probably because of the proliferative variation (eg, increased cellular membrane), which could explain elevation of the choline level, without modification of the polyamine peak.

There are several limitations associated with our study. First, MR spectroscopic imaging is a new technology, and the quality of the MR spectroscopic imaging data will improve with maturation of the technology and increased operator experience. Specifically, all MR spectroscopic imaging studies were evaluated in this study, regardless of spectral quality that changed during the course of the study. The introduction of a perfluorocarbon in the endorectal coil instead of air dramatically reduced magnetic field susceptibility problems due to the rectal air-tissue interface, and it resulted in dramatically improved spectra in the final 22 patients examined. Additionally, data analyses improved during the course of the study because of the introduction of voxel shifting capabilities, which allowed better alignment of the spectral data with anatomy in the later patients. Finally, the transfer of spectral abnormalities onto the transrectal US images used for prostate biopsies is currently a manual process that is susceptible to localization errors. To further investigate this source of error, all patients studied will return in the near future to undergo repeat biopsy.

Additional limitations, such as the lack of MR spectroscopic imaging evaluation of the transition zone, must be considered; however, a recent study has shown that the broad range of metabolite ratios of the transition zone cancer precludes its differentiation from benign prostatic hyperplasia (19). As we know, up to 30% of prostate cancers arise in this area (23). Another limitation is the need for a very high level of expertise in MR spectroscopic imaging and prostatic biopsy targeting.

Our findings suggest that MR spectroscopic imaging targeting of transrectal US–guided biopsies could improve the determination of whether men with rising PSA levels and negative findings at prior biopsy have cancer. To validate this hypothesis, however, a larger number of patients must be studied with standardized MR spectroscopic techniques.


   FOOTNOTES
 

Abbreviations: PSA = prostate specific antigen

Author contributions: Guarantors of integrity of entire study, A.P., A.P.B., E.M.O.; study concepts, A.P., J.K.; study design, A.P.; literature research, A.P., J.K., A.P.B., E.M.O.; clinical studies, A.P., A.P.B., E.M.O.; experimental studies, E.F.; data acquisition, A.P.B., E.M.O.; data analysis/interpretation, A.P., J.K.; statistical analysis, A.P., A.P.B., E.F.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, A.P., J.K.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Ellis WJ, Brawer MK. Repeat prostate needle biopsy: who needs it. J Urol 1995;153:1496–1498.[CrossRef][Medline]
  2. Keetch DW, Catalona WJ, Smith DS. Serial prostatic biopsies in men with persistently elevated serum PSA levels. J Urol 1994;151:1571–1574.[Medline]
  3. Djavan B, Ravery V, Zlotta A, et al. Prospective evaluation of prostate cancer detected on biopsies 1, 2, 3 and 4: when should we stop? J Urol 2001;166:1679–1683.
  4. Levine MA, Ittman M, Melamed J, Lepper H. Two consecutive sets of TRUS-guided sextant biopsies of the prostate for the detection of prostate cancer. J Urol 1998;159:471–475.[CrossRef][Medline]
  5. Babaian RJ, Toi A, Kamoi K, Troncoso P, Sweet J, Evans R. A comparative analysis of sextant and an extended 11-core multisite directed biopsy strategy. J Urol 2000;163:152–157.[CrossRef][Medline]
  6. Stewart CS, Leibovich BC, Weaver AL, Lieber MM. Prostate cancer diagnosis using a saturation needle biopsy technique after previous negative sextant biopsies. J Urol 2001;166:86–92.[CrossRef][Medline]
  7. Matlaga BR, Eskew A, McCullough DL. Prostate biopsy: indications and technique. J Urol 2003;169:12–19.[CrossRef][Medline]
  8. Prando A, Wallace S. Helical CT of prostate cancer: early clinical experience. AJR Am J Roentgenol 2000;175:343–346.[Abstract/Free Full Text]
  9. Perrotti M, Huan KR, Epstein RE, et al. Prospective evaluation of endorectal MRI to detect tumor foci in men with prior negative prostatic biopsy: a pilot study. J Urol 1999;162:1314–1320.[CrossRef][Medline]
  10. Scheidler J, Hricak H, Vigneron DB, et al. Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging—clinicopathologic study. Radiology 1999;213:473–480.[Abstract/Free Full Text]
  11. Yu KK, Scheidler J, Hricak H, et al. Prostate cancer: prediction of extracapsular extension with endorectal MR imaging and three-dimensional proton MR spectroscopic imaging. Radiology 1999;213:481–488.[Abstract/Free Full Text]
  12. Kurhanewicz J, Vigneron DB, Nelson SJ. Three-dimensional magnetic resonance spectroscopic imaging of brain and prostate cancer. Neoplasia 2000;2:166–189.[CrossRef][Medline]
  13. Kurhanewicz J, Vigneron DB, Hricak H, Carroll P, Narayan P, Nelson SJ. Three-dimensional H-1 MR spectroscopic imaging of the in situ human prostate with high (0.24–0.7-cm3) spatial resolution. Radiology 1996;198:795–805.[Abstract/Free Full Text]
  14. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer. J Magn Reson Imaging 2002;16:451–463.[CrossRef][Medline]
  15. Star-Lack J, Pauly J, Vigneron DB, Kurhanewicz J, Nelson SJ. Improved solvent suppression and increased spatial excitation bandwidths for 3-D PRESS CSI using phase-compensating spectral/spatial spin-echo pulses. J Magn Reson Imaging 1997;7:745–757.[Medline]
  16. Schricker AA, Pauly JM, Kurhanewicz J, Swanson MG, Vigneron DB. Dualband spectral-spatial RF pulses for prostate MR spectroscopic imaging. Magn Reson Med 2001;46:1079–1087.[CrossRef][Medline]
  17. Males RG, Vigneron DB, Star-Lack J, et al. Clinical applications of BASING and spectral/spatial water and lipid suppression pulses for prostate cancer staging and localization by in vivo 3D 1H magnetic resonance spectroscopic imaging. Magn Reson Med 2000;43:17–22.[CrossRef][Medline]
  18. Jung JA, Coakley FV, Vigneron DB, et al. Endorectal MR spectroscopic imaging of the prostate: investigation of a standardized evaluation system. Radiology 2004;233:701–708.[Abstract/Free Full Text]
  19. Zakian KL, Eberhardt S, Hricak H, et al. Transition zone prostate cancer: metabolic characteristics at 1H MR spectroscopic imaging—initial results. Radiology 2003;229:241–247.[Abstract/Free Full Text]
  20. Yuen JS, Thng CH, Tan PH, et al. Endorectal magnetic resonance imaging and spectroscopy for the detection of tumor foci in men with prior negative transrectal ultrasound prostate biopsy. J Urol 2004;171:1482–1486.[CrossRef][Medline]
  21. Vashi AR, Wojno KJ, Gillespie B, Oesterling JE. A model for the number of cores per prostate biopsy based on patient age and prostate gland volume. J Urol 1998;159:920–929.[CrossRef][Medline]
  22. De Marzo AM, Marchi VL, Epstein JI, Nelson WG. Proliferative inflammatory atrophy of the prostate: implications for prostatic carcinogenesis. Am J Pathol 1999; 155:1985–1992.[Abstract/Free Full Text]
  23. McNeal JE, Redwine EA, Freiha FS, Stamey TA. Zonal distribution of prostatic adenocarcinoma: correlation with histological pattern and direction of spread. Am J Surg Pathol 1988;12:897–906.[Medline]



This article has been cited by other articles:


Home page
 Am. J. Roentgenol.Home page
H. Choi and J. Ma
Use of Perfluorocarbon Compound in the Endorectal Coil to Improve MR Spectroscopy of the Prostate
Am. J. Roentgenol., April 1, 2008; 190(4): 1055 - 1059.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
A. Shukla-Dave, H. Hricak, C. Moskowitz, N. Ishill, O. Akin, K. Kuroiwa, J. Spector, M. Kumar, V. E. Reuter, J. A. Koutcher, et al.
Detection of Prostate Cancer with MR Spectroscopic Imaging: An Expanded Paradigm Incorporating Polyamines
Radiology, November 1, 2007; 245(2): 499 - 506.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
N. G. Costouros, F. V. Coakley, A. C. Westphalen, A. Qayyum, B. M. Yeh, B. N. Joe, and J. Kurhanewicz
Diagnosis of Prostate Cancer in Patients with an Elevated Prostate-Specific Antigen Level: Role of Endorectal MRI and MR Spectroscopic Imaging
Am. J. Roentgenol., March 1, 2007; 188(3): 812 - 816.
[Abstract] [Full Text] [PDF]

Home page
 RadioGraphicsHome page
Y. J. Choi, J. K. Kim, N. Kim, K. W. Kim, E. K. Choi, and K.-S. Cho
Functional MR Imaging of Prostate Cancer
RadioGraphics, January 1, 2007; 27(1): 63 - 75.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Prando, A.
Right arrow Articles by Figueiredo, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Prando, A.
Right arrow Articles by Figueiredo, E.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE