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


     

DOI: 10.1148/radiol.2281011808
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 Swindle, P.
Right arrow Articles by Mountford, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Swindle, P.
Right arrow Articles by Mountford, C.
(Radiology 2003;228:144-151.)
© RSNA, 2003

Genitourinary Imaging

Pathologic Characterization of Human Prostate Tissue with Proton MR Spectroscopy1

Peter Swindle, MD, FRACS, MS, Simon McCredie, MD, FRACS, Peter Russell, MD, FRCPA, Uwe Himmelreich, PhD, Mohammed Khadra, FRACS, PhD, Cynthia Lean, PhD and Carolyn Mountford, DPhil

1 From the Institute for Magnetic Resonance Research and Department of Magnetic Resonance in Medicine, University of Sydney, Block 3 Level 3, Royal North Shore Hospital, St Leonard’s, New South Wales, 2065, Australia. Received November 9, 2001; revision requested January 23, 2002; final revision received November 1; accepted January 2, 2003. Supported by Australian NH&MRC grant no. 991337. Address correspondence to C.M. (e-mail: caro@imrr.usyd.edu.au).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
PURPOSE: To assess the accuracy of magnetic resonance (MR) spectroscopy in documenting the chemical features of human prostate tissue and to ascertain if there are chemical criteria of diagnostic importance.

MATERIALS AND METHODS: Seventy-seven prostate tissue specimens (peripheral zone, n = 61; transitional zone, n = 16) from 43 patients were analyzed with MR spectroscopy. Histologic features were compared with MR spectroscopic data. Statistical analysis was undertaken with analysis of variance and computer software.

RESULTS: Histologically identified carcinomas were determined by using MR spectroscopy with a sensitivity of 100% and a specificity of 94%. Histologically benign tissue from patients without carcinoma of the prostate was distinguished from malignant tissue with a sensitivity of 100% and a specificity of 94%. When benign specimens from patients with cancer elsewhere in the prostate were included in the database, MR spectroscopy helped distinguish benign prostatic hyperplasia from adenocarcinoma with a sensitivity of 97% and specificity of 88%. Depleted citrate and elevated choline levels alone were not accurate markers of malignancy, since citrate levels remain high when a small amount of malignant disease is present. Carcinomas missed at routine histologic examination were identified with MR spectroscopy and confirmed with specialized, nonstandard histologic examination.

CONCLUSION: By comparing the intensity of resonances assigned to choline, creatine, lipid, and lysine, MR spectroscopy can depict prostate carcinoma with a high degree of sensitivity and specificity. Citrate and choline resonances alone are not sufficiently accurate markers for distinguishing between various patterns of prostatic disease.

© RSNA, 2003

Index terms: Magnetic resonance (MR), spectroscopy • Prostate, biopsy, 844.1261 • Prostate, MR, 844.12145 • Prostate neoplasms, MR, 844.316, 844.324


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Prostate carcinoma is the most common cancer in men in Australia (1,2) and the United States (3). Histologic examination is the standard in the diagnosis and classification of prostate neoplasms (4–6). A major limitation in the management of prostate cancer is an inability to distinguish early those cancers that will progress and become life threatening (7). Histopathologic examination plays a central role in determining patient prognosis and care (8–10); however, even when the stage and grade have been assigned with histologic assessment after radical prostatectomy, there is considerable variation in the outcome of patients (11).

The prevalence of prostate cancer is high, occurring in 42% of men aged 50 years or older, but the risk of dying of this disease is only 3.6% (12). A method is needed that will aid in distinguishing patients who have a clinically indolent prostate cancer and will survive from patients who have an aggressive cancer and will die from their disease within a few years (12). The solution to this dilemma is an accurate and noninvasive way of determining the location, extent, and biologic potential of prostate cancer. By improving the ability to define subsets of patients in whom cure is possible and necessary, the effectiveness of current and future treatments will be enhanced.

Magnetic resonance (MR) spectroscopy has the potential to provide such diagnostic and prognostic information. The documentation of the MR chemical fingerprint from samples of cervix (13), brain (14), thyroid (15), colon (16,17), ovary (18), breast (19), esophagus (20), and liver (21) tissue has proven that MR spectroscopy can assist physicians in detecting and diagnosing disease in biopsy specimens with a sensitivity and specificity of 95%–100%. MR spectroscopy combined with a statistical classification strategy (22) applied to the MR and clinical records can also provide prognostic information. MR spectroscopic analysis of an aspirate from a breast lesion (19) aids in the identification of nodal status from chemical information in the primary tumor (23).

Others have reported MR spectroscopic analysis of the prostate both in vivo (24,25) and ex vivo (26–29). In vivo MR spectroscopy of the prostate can be performed with a high level of technical success (25) with use of decreased citrate levels and increased choline levels as diagnostic markers of carcinoma in the prostate (25,27–29). Spermine, spermidine, and other polyamines (30,31) are also listed as diagnostic. Citrate and choline alone do not, however, allow localization of prostate cancer (32,33) and prediction of extracapsular extension (34) with the sensitivity and specificity required for clinical use.

The National Research Council of Canada undertook an MR spectroscopic study of prostate biopsies and developed a statistical classification strategy with an accuracy of 96.6% (35). The statistical classification strategy is dependent on the correct histologic data if the computer-based strategy is to generate an accurate classifier (22). We have demonstrated that a routine hospital histologic examination is often inadequate and that benign lesions with invasive cancer elsewhere in the organ can generate a chemical fingerprint different from that of histologically benign tissue from an otherwise healthy organ (36), which is typical of a "field change" (37).

The purpose of our study was to assess the efficacy of MR spectroscopy in documenting the chemical features of human prostate tissue and to ascertain if there are chemical criteria of diagnostic importance.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Patients provided their written consent after the nature of the experimental procedure was explained, and the relevant ethics review boards approved our study.

Surgeons from the Sydney University, Australia, teaching hospitals took responsibility for patient selection, follow-up, and specimen handling. MR spectroscopic data were collected at the Institute for Magnetic Resonance Research in Sydney, Australia. Histologic data were obtained from collaborating hospitals, and assessment of the MR spectroscopic sample was performed by the director of pathology at the Institute for Magnetic Resonance Research (P.R.).

Patient Selection
Patients were included in this study if they had prostate cancer and were undergoing radical retropubic prostatectomy or transurethral resection of the prostate. A number of patients who did not have prostate cancer but were undergoing radical cystectomy, open prostatectomy, or transurethral resection of the prostate were also included.

Patients who underwent prostate surgery at Royal Prince Alfred, Concord Repatriation General, and Strathfield Private hospitals—all located in Sydney, Australia—were included in this study. Patients ranged in age from 45 to 80 years. Specimens were collected from patients who underwent open or radical retropubic prostatectomy, radical cystoprostatectomy, and transurethral resection of the prostate. Consecutive patients who underwent radical retropubic prostatectomy from January to October 1997 at Royal Prince Alfred and Concord Reparation General hospitals were included. Not all patients who underwent transurethral resection of the prostate were included. We used a randomized selection process, since there were too many patients at this clinic to be included in our study. As a consequence, every fourth patient who underwent a transurethral resection of the prostate was included. Patient demographics and prostate specific antigen levels are representative of published variables for patients who underwent transurethral resection of the prostate.

Indications for radical retropubic prostatectomy were clinically localized carcinoma of the prostate (clinical stage T1c–T2) in patients with a life expectancy of more than 10–12 years. Stage of cancer was determined on the basis of the findings of digital rectal examination, abdominopelvic computed tomographic examination, whole-body bone examination, and the level of serum prostate specific antigen. Preoperative prostate specific antigen levels ranged from 3.6 to 18.9 ng/mL.

Indications for transurethral resection of the prostate were clinical evidence of bladder outflow obstruction due to benign or malignant enlargement of the prostate. Two patients had locally advanced carcinoma of the prostate. The remainder had benign disease with preoperative prostate specific antigen levels ranging from 0.9 to 44.4 ng/mL. Radical cystoprostatectomy was performed because of transitional cell carcinoma of the bladder that had invaded the muscle. Two patients were included in this study, both of whom had preoperative serum prostate specific antigen levels of less than 4 ng/mL. One patient included in the study with symptoms of bladder outflow obstruction due to prostatomegaly underwent open prostatectomy because of the large size of the prostatic adenoma. This patient had a preoperative prostate specific antigen level of 8.6 ng/mL.

Collection of Tissue Specimens
All tissue specimens for MR spectroscopic experiments were collected from fresh surgical specimens prior to fixation in formalin. In patients who underwent open surgery (ie, radical prostatectomy, cystoprostatectomy, and open prostatectomy), the fresh surgical specimen was taken intact to the pathology department, where tissue samples were obtained. In patients who underwent transurethral resection of the prostate, tissue samples were obtained intraoperatively. Tissue samples from open surgical specimens were obtained with a 5-mm dermatologic biopsy punch (Stiefel Laboratories, New South Wales, Australia). The surgical specimen was first stained with India ink (Drawing Ink; Sanford Rotring, Hamburg, Germany) to identify the excision margin. The posterior surface of the prostatic capsule was then incised and retracted to maintain the integrity of the surgical margin and avoid contamination with the ink. Six biopsies were performed in the peripheral zone of the prostate in radical retropubic prostatectomy specimens, while two biopsies were performed in the peripheral zone in cystoprostatectomy specimens. Two biopsies were performed in the external surface of the open prostatectomy specimens, thus allowing sampling of the transitional zone. A single tissue biopsy was performed intraoperatively with a standard monopolar diathermy resectoscope in the transitional zone of patients who underwent transurethral resection of the prostate.

Several tissue samples were collected from each patient in an attempt to obtain both benign and malignant tissue from the same patient for control purposes. Tissue specimens were collected with different methods to sample the zones of the prostate. The complex zonal anatomy led us to believe that a single technique could not be used to obtain tissue from the zones. For each zone, the sampling techniques were standardized. We analyzed only tissues from peripheral and transitional zones in our study. The peripheral zone biopsies were performed with punch biopsy after radical prostatectomy. Transitional zone tissue biopsies were performed in patients who underwent transurethral resection of the prostate with a resectoscope loop biopsy. Visual analysis of the MR spectra clearly indicated that benign prostatic hypertrophy (BPH) and cancer each had a unique biochemical signature, irrespective of biopsy procedure.

Preparation and Storage of Tissue Specimens
All specimens, regardless of how they were obtained, were bisected along the longitudinal axis with a size 22 scalpel (Swann-Morton, Sheffield, England). Half of the specimen was fixed immediately in 10% formalin, and half was placed in 0.3 mL of a solution of phosphate buffered saline and D2O and was snap frozen in liquid nitrogen. Specimens that were placed immediately in formalin were processed and examined histologically by one of the authors (P.R.). Specimens that were placed in the phosphate buffered saline and D2O solution were transported in a Dewar flask to the Institute for Magnetic Resonance Research and stored at -70°C for up to 60 days prior to MR spectroscopic analysis.

MR Spectroscopy
MR data were collected by three authors (P.S., S.M., and U.H.) with a wide-bore MR spectrometer (Avance 360; Bruker, Karsruhe, Germany) operating at 8.5 T and equipped with a standard 5-mm inverse detection hydrogen 1 (1H), carbon 13 probe head. The sample was spun at 20 Hz, and the temperature was maintained at 37°C. The residual water signal was suppressed with selective gated irradiation. The chemical shifts of resonances were referenced to aqueous sodium 3-(trimethylsilyl)-propanesulfonate at 0.00 ppm or internal remaining water at 4.65 ppm. One-dimensional spectra were acquired over a spectral width of 3,597 Hz (10.0 ppm) with a 90° pulse of 6.5–7.0 µsec, 8,192 data points, 256 accumulations, and a relaxation delay of 1 second, which resulted in a pulse repetition time of 2.14 seconds. MR spectroscopic data were processed with computer software (Xwinnmr 2.6; Bruker, Rheinstetten, Germany) as previously described (38). All spectra were analyzed individually with resonances assigned (26,35,39). The resonance intensities were measured with this software for all spectra, and those that provided the clearest distinction between the various tissue categories are described in the Results.

Data Analysis
Analysis of variance was undertaken by using computer software (Genstat 5) to produce P values. Analysis of the MR spectroscopic data was undertaken by comparing two resonance intensity values for stromal versus glandular tissue and by calculating two ratios with four resonance intensity values for benign versus malignant tissue.

Histologic Examination
The initial report and secondary review of histologic data were undertaken by one of the authors (P.R). As described previously, each specimen was divided in half. One half was used for MR spectroscopic analysis, and the other half was kept for a future histologic examination. Routine hospital histologic examination (ie, 2 x 5-µm slices per specimen) was performed on the half of the specimen that was submitted for MR spectroscopy. Tissue preservation, proportion of stromal and glandular BPH, presence and proportion of malignant tissue, and the presence of inflammatory cells were reported. All histologic examinations were performed by physicians who had been blinded to the MR spectroscopic data. Criteria for establishing firm histologic diagnosis of cancer and its precursors followed established guidelines according to the relevant World Health Organization International Histological Tumor Classification System.

When the visual inspection of the MR data yielded a chemical fingerprint that was consistent with malignant findings yet the result of a routine histologic examination was consistent with benign findings, the second half of the tissue specimen was serially sliced. In addition, 23 specimens that were correctly identified as benign with MR spectroscopy were also serially sliced to serve as controls. From the tissues that were serially sliced, five 5-µm slices were taken every 100 µm, and the intervening tissue was discarded. Each set of five sections was placed on a single slide. For routine histologic examination, every fifth slide was examined. Thus, five 5-µm sections were examined every 500 µm. When the first half of a specimen from a patient with adenocarcinoma elsewhere in the prostate demonstrated benign findings at MR spectroscopy, the second half of the specimen was serially sliced and all slices were examined. Thus, in these cases, five 5-µm sections were examined every 100 µm.

The spectral analyses were undertaken by scientists who had been blinded to histologic information and clinical outcomes. Spectral differences were then compared with the histologic findings after both routine and repeat histologic examinations. Criteria for establishing firm histologic diagnosis of cancer and its precursors followed established guidelines according to the relevant World Health Organization International Histological Tumor Classification System.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Seventy-seven tissue specimens from 43 patients who underwent radical prostatectomy (n = 25), retropubic prostatectomy (n = 1), radical cystectomy (n = 2), or transurethral resection of the prostate (n = 15) were analyzed with MR spectroscopy. Samples from the peripheral zone (n = 61) and transitional zone (n = 16) were collected. No specimens from either the central zones or the periurethral zones were included. Six specimens from two patients were excluded from analysis because the patients received preoperative hormonal ablation therapy. Thus, analysis of the MR chemical profile of each histologic variant was performed in 71 samples consisting of 55 peripheral zone specimens and 16 transitional zone specimens.

Histologic Examination with Routine Hospital Procedures
Initial histologic examination of the 71 specimens available for analysis showed adenocarcinoma in two of the transitional specimens and 25 of the peripheral zone specimens. There were 44 specimens with BPH: 30 peripheral specimens and 14 transitional specimens. BPH was subclassified into glandular (>5% glandular elements, n = 25) or stromal (<5% glandular elements, n = 19).

Spectral Characterization of Prostatic Tissue
Typical MR spectra from four histologically determined categories, including adenocarcinoma with 50% of specimen involved, adenocarcinoma with 5% of specimen involved, stromal BPH, and glandular BPH, are shown in Figure 1. The spectrum of tissue containing 50% adenocarcinoma is typical of that seen in other organs. Resonances consistent with citrate (2.5–2.7 ppm); glutamate (2.00–2.04 ppm); creatine (3.0 ppm); a composite of choline, phosphocholine, glycerophosphocholine, and phosphoethanolamine (3.2 ppm); a composite of lysine and lipid (1.7 ppm); and a composite of lipid, lactate, and threonine (1.28–1.33 ppm) were previously identified as having diagnostic potential (24,30,35). A large resonance at 3.1 ppm was also identified as characteristic of glandular BPH and was assigned to spermine, spermidine, and other polyamines (31). Many of these assignments were made or confirmed with two-dimensional spectroscopy (spectra not shown) (38). A summary of resonances assigned and contributing to these spectra are listed in Table 1.



View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. One-dimensional 1H MR spectra of prostate biopsy specimens obtained at 8.5 T (256 acquisitions). A, Adenocarcinoma with 50% of the specimen involved. B, Adenocarcinoma with 5% of the specimen involved. C, Stromal BPH (95% stromal, 5% glandular). D, Glandular BPH (85% glandular, 15% stromal). The distinction between the abnormalities and tissue types are the result of variations of the creatine (3.0 ppm), choline (3.2 ppm), lipid (1.3 ppm), and lysine (1.7 ppm) resonances. See Table 1 for assignments. Acly = acetyl residues; Chol = choline; Cit = citrate; Cre = creatine; Glu/Gln = glutamate, glutamine; Ile = isoleucine; Lac = lactate; Leu = leucine; Lip = lipid; Lys = lysine; PA = spermine, spermidine, polyamines; Thr = threonine; Val = valine.

 

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

  		TABLE 1. Resonances in One-dimensional Proton MR Spectra of Human Prostate Tissue

 
Comparison of Different Histologic Categories
Visual inspection of one-dimensional MR spectra.—The discrimination between glandular BPH and stromal BPH can be made on the basis of the ratio of the intensity of the lysine-containing resonance (1.7 ppm) to the citrate resonance (2.5 ppm). Glandular BPH has an intense citrate resonance in comparison to stromal BPH and thus demonstrates a lower lysine-citrate ratio than does stromal BPH. A well-separated resonance at 3.1 ppm was assigned to spermine and spermidine and is characteristic of glandular tissue (31).

Malignant prostate tissue can be distinguished from stromal BPH by a greater lipid-lysine ratio (1.3:1.7 ppm) (P < .001) and from glandular BPH by a greater lipid-citrate ratio (1.3:2.5 ppm) (P < .001). Malignant tissue demonstrates an elevated choline-creatine ratio (3.2:3.0 ppm) and an elevated glutamate-lysine ratio (2.0:1.7 ppm) compared with benign tissue (P < .001) (Fig 1).

By comparing the resonance intensities at discrete frequencies, prostatic adenocarcinoma is distinguished from stromal BPH by a greater lipid-lysine ratio (1.3:1.7 ppm) (P < .001) and glandular BPH, by a greater lipid-citrate ratio (1.3:2.5 ppm) (P < .001).

As the amount of involvement by adenocarcinoma increases, it appears that the spermine and spermidine resonance (3.1 ppm) decreases. In Figure 1, A, there appears to be no or minimal resonance at 3.1 ppm. In comparison to Figure 1, B, with only 5% involvement by adenocarcinoma, there appears to be a substantial contribution to this 3.1-ppm resonance. These observations were confirmed with two-dimensional spectroscopy (not shown).

Analysis of one-dimensional MR spectra.—Spectra from prostatic adenocarcinoma were characterized by an elevated choline-creatine ratio (3.2:3.0 ppm) (P < .001) and an elevated lipid-lysine ratio (1.3:1.7 ppm) (P < .0001) when compared with spectra from benign specimens. Although the differences between these individual ratios are statistically significant, there is considerable overlap of ratios, resulting in poor separation of BPH from cancer. By combining these ratios, the sensitivity and specificity including all specimens is 100% and 82%, respectively (Table 2). These results are consistent with those published previously (35).


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

  		TABLE 2. Comparison of MR Spectroscopic Results with Histologic Results

 
Reassessment of MR spectra and histologic findings.—When the spectra obtained from BPH specimens (with both stromal and glandular contributions) from patients without cancer elsewhere in the prostate were compared with spectra obtained from specimens from patients with adenocarcinoma, a clear difference became visible. From an analysis of resonance height ratios of choline-creatine (3.2:3.0 ppm) and lipid-lysine (1.3:1.72 ppm), a clear distinction is visible between benign and malignant tissue (Fig 2). By using these chemical ratios, it is possible to distinguish between benign and malignant tissue with a sensitivity of 100% and a specificity of 94% (Table 2).



View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Two-directional plot of MR spectroscopic ratios of tissue specimens reported to contain adenocarcinoma and those reported to contain BPH from patients without cancer elsewhere in the prostate gland. The choline-creatine ratio (3.2:3.0 ppm) is plotted on the y axis; the lipid-lysine ratio (1.3:1.7 ppm), on the x axis. Diagnosis was determined with routine histologic analysis of the tissue specimen analyzed with MR spectroscopy. There is a clear distinction between the two abnormalities. {bullet} = adenocarcinoma, {diamond} = BPH but no cancer elsewhere in prostate.

 
A two-directional plot of resonance height ratios of choline-creatine (3.2:3.0 ppm) and lipid-lysine (1.3:1.7 ppm) is shown in Figure 3, where the 26 BPH specimens from patients with cancer elsewhere in the prostate are added to those shown in Figure 2. This distinction led us to conclude that these BPH specimens should be placed into two groups. Seven specimens were classified with MR spectroscopy as malignant and 19 were classified as benign. Thus, seven tissue specimens that were histologically benign and obtained from patients with cancer elsewhere in the prostate were classified as malignant with this analysis of MR spectra.



View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Added to the information presented in Figure 2 are the data from patients with histologically identified BPH but also with cancer elsewhere in the prostate. Marked on the two-directional plots are numbers 1-3; these are the specimens in which the histologic results were corrected after serial slicing and examination. Yellow diamond = BPH but no cancer elsewhere, green square = histologically corrected after serial sectioning, red triangle = BPH and cancer elsewhere in the prostate, blue circle = adenocarcinoma.

 
The clue to this puzzle is demonstrated in Figure 4, where two spectra are compared with histologic slides. Both tissue specimens were harvested from the same patient with prostate cancer. The MR spectra are similar, and both are indicative of malignancy; however, histologic analysis of specimen B clearly indicates BPH when based on a routine hospital histologic examination. In contrast, the histologic analysis of specimen A indicates malignancy.



View larger version (79K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. A, 1H MR spectrum from tissue histologically identified as adenocarcinoma. B, 1H MR spectrum from tissue histologically identified as glandular BPH from the same patient as in A. Histologic slides of each of these specimens are shown under the spectra. The MR spectra are similar, and both indicate malignancy, yet results of a routine hospital histologic examination indicate that specimen B is glandular BPH. (Hematoxylin-eosin stain; original magnification, x150.)

 
Subsequently, all 26 specimens from patients with adenocarcinoma elsewhere in the prostate and considered to be benign at routine hospital histologic examination were reexamined by using serial step slicing as previously described. The step slicing was performed on the control half of the specimen, which was not examined with MR spectroscopy. The results of histologic study were reviewed blind.

Of the seven specimens that were classified as malignant with MR spectroscopic criteria, two were reclassified as adenocarcinoma (specimens 1 and 2 on Fig 3), and one was reclassified as high-grade prostatic intraepithelial neoplasia (specimen 3 on Fig 3). With serial step slicing, the detection rate of carcinoma was increased to two (8%) of 26 cases. The diagnosis of BPH remained unchanged for the remaining four specimens.

With the revised histologic results, the sensitivity and specificity of MR spectroscopy in distinguishing BPH and adenocarcinoma of the prostate were 97% and 87%, respectively (Table 2).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Proton MR spectroscopy of human prostatic tissue can depict subtle chemical differences associated with different cellular components and different pathologic processes. Benign stromal hyperplasia, benign glandular hyperplasia, and adenocarcinoma of the prostate can be distinguished with a very high degree of accuracy. It is clear from this study, however, that for MR spectroscopy to be developed as a routine clinical tool, the data need to be compared with the results of a specialized, nonstandard histologic examination.

Benign hyperplasia of the prostate affects both the glandular and stromal compartments of the gland. Glandular epithelial cells in the prostate produce citrate, and we have confirmed previous reports (28,29) that citrate levels are reduced in stromal BPH compared to glandular BPH, as measured by the resonance height ratio (1.7:2.5 ppm). As seen in Figure 1, A, and Figure 1, C, citrate is present in relatively low levels in both stromal BPH and tissue with adenocarcinoma, and is therefore not a unique marker of healthy tissue. This observation may explain some of the reported overlap between stromal BPH and adenocarcinoma with use of in vivo MR spectroscopy that relies on citrate levels as discriminatory (25,29). A direct comparison of the choline-citrate ratio from this database with those reported from the published in vivo results is inappropriate, since the in vivo data were obtained with a spin-echo experiment, whereas our data were obtained with a simple 90° pulse sequence.

In contrast to other studies (25,29), our study has demonstrated that spectra from prostate adenocarcinoma tissue are also characterized by an elevation of the lipid-lysine ratio (1.3:1.7 ppm). To the best of our knowledge, this has not been previously reported. When specimens from patients without cancer elsewhere in the prostate that were histologically confirmed as BPH were studied with MR spectroscopy and compared with specimens that were histologically confirmed as adenocarcinoma, an excellent correlation was made with histologic analysis. This correlation provided a sensitivity of 100% and a specificity of 94%. While authors of previous studies compared concentrations of single chemicals to distinguish BPH from adenocarcinoma, this comparison produced poor results because of substantial overlap of chemical concentrations between histologic variants.

When MR spectroscopy was used to analyze histologically diagnosed BPH in patients with carcinoma elsewhere in the prostate gland, the correlation was less than impressive, with seven specimens showing MR spectroscopy characteristics consistent with malignancy. In fact, three of these specimens had adenocarcinoma (n = 2) or high-grade prostatic intraepithelial neoplasia (n = 1) at further histologic examination of the adjacent specimen. Thus, MR spectroscopic examination of these samples was more sensitive than routine histologic examination. An increased detection rate of cancer with nonroutine step slice histologic analysis has also been reported in other organs where the detection rate increased 8% with serial step slicing of lymph nodes (40).

Among the remaining four specimens with positive MR spectra, the adjacent tissue specimens showed no morphologically visible malignant changes. These findings could be explained by very small areas of malignant change that were missed at step slice histologic examination or by the control specimen not being representative of the MR spectroscopic specimen. Alternatively, we are unable to exclude the possibility that these specimens may represent cells that underwent early genetic, and hence biochemical, changes associated with malignant transformation but did not manifest this change morphologically. MR spectroscopy may have depicted cells that were already committed to becoming malignant, yet appeared normal with light microscopy. Another explanation is that MR spectroscopy may have depicted biochemical changes associated with a field change phenomenon within the prostate, similar to what has been reported for other organs (37).

Although histopathologic examination remains the standard in the assessment and classification of human prostatic disease, the ability of MR spectroscopy to aid in the characterization of prostatic lesions on the basis of their biochemical composition may provide additional diagnostic and prognostic information as an adjunct to routine histologic assessment. MR spectroscopy is rapid; less than 5 minutes are required to collect the data, and the entire process may be automated. Once appropriate statistical classification strategy classifiers are developed from databases with correct disease, MR spectroscopy of the biopsy specimen may become an independent modality to aid in the determination of the patient’s diagnosis and prognosis in less than 15 minutes.

Two issues are pertinent to the discrepancies between MR spectroscopic results and histologic results. The first issue is whether disease that is indolent or aggressive was missed at routine hospital histologic examination. If a small focus of aggressive disease has been missed, the patient’s life has been placed in jeopardy. With a more complete understanding of the MR chemistry of prostate disease (not compromised by lipid suppression), correct and accurate diagnosis will be possible. The second issue is the use of an incorrect database to generate a classifier that will aid in the correct prediction of the disease process on the basis of the MR spectroscopic analysis.

The additional chemical information available with the high field strength of 8.5 T in the biopsy specimens also suggests that newly available high-field-strength whole-body magnets will facilitate a more accurate means of in vivo prostate assessment. Here, it is of utmost importance that there be a clear documentation of the MR spectroscopic fingerprint of each of the abnormalities known to exist and of healthy tissue from each of the anatomic zones of the prostate. This will allow assessment of the disease process with in vivo endorectal MR spectroscopic imaging. In addition, the use of a shorter echo time in vivo may allow a more accurate analysis because phase errors due to J coupling are minimized. This could result in more accurate staging of prostate cancer than is presently possible.

In conclusion, proton MR spectroscopy of prostate tissue specimens provides an objective and independent evaluation of the disease. Stromal BPH, glandular BPH, and adenocarcinoma each have a unique chemical fingerprint. Analysis of the MR spectroscopic data allows distinction between BPH and adenocarcinoma with a sensitivity of 97% and a specificity of 87%. With use of routine hospital histologic examinations, microscopic foci of carcinoma will be missed in 8% of the specimens. Furthermore, histologic data are inadequate for correlation with MR spectroscopic data. Step slicing is required to validate the MR spectroscopic data.

Citrate and choline alone are not accurate markers for distinguishing histologic variants of prostatic tissue. In our study, prostate carcinoma and prostatic intraepithelial neoplasia—which were missed with routine histologic examination but were subsequently confirmed with nonstandardized histologic step slice examination—were correctly identified with analysis of the MR spectroscopic data.

MR spectroscopic databases with detailed nonstandard histologic examination need to be analyzed with statistical classification methods to provide a robust analysis, potentially providing an accuracy that is likely to approach 100%.


   ACKNOWLEDGMENTS
 
We thank T. S. Reeve, CBE, AC, FRACS, FRACMA, for continued support and guidance and Roger Bourne, PhD, and Sinead Doran, BSc, for their help in preparing the manuscript.


   FOOTNOTES
 
Abbreviation: BPH = benign prostatic hypertrophy

Author contributions: Guarantors of integrity of entire study, C.M., P.R.; study concepts, C.M., M.K., C.L., P.R.; study design, C.M., C.L., P.M., M.K.; literature research, P.S., S.M.; experimental studies, P.S., S.M., U.H., P.R.; data acquisition, P.S., S.M., P.R.; data analysis/interpretation, U.H., P.S., S.M., C.M.; statistical analysis, P.S., S.M.; manuscript definition of intellectual content, P.S., S.M., P.R., U.H., C.L., C.M.; manuscript editing, P.R., C.M., M.K.; manuscript preparation, revision/review, and final version approval, all authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. van der Hoek R. Malignant neoplasms of the prostate: Disease Registers Unit Australia: Australian Institute of Health and Welfare, 1995.
  2. Hirst GH, Ward JE, Del Mar CB. Screening for prostate cancer: the case against. Med J Aust 1996; 164:285-288.[Medline]
  3. Greenlee RT, Murray T, Bolden S, Wingo PA. Cancer statistics, 2000. CA Cancer J Clin 2000; 50:7-33.[Abstract]
  4. McNeal JE. Structure and pathology of the prostate. In: Goland M, ed. Normal and abnormal growth of the prostate. Springfield, Ill: Thomas, 1975; 750:55-65.
  5. McNeal JE. Normal and pathologic anatomy of prostate. Urology 1981; 17:11-16.[CrossRef][Medline]
  6. McNeal JE. Normal histology of the prostate. Am J Surg Pathol 1988; 12:619-633.[CrossRef][Medline]
  7. Aihara M, Wheeler TM, Ohori M, Scardino PT. Heterogeneity of prostate cancer in radical prostatectomy specimens. Urology 1994; 43:60-67.[CrossRef][Medline]
  8. Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 1974; 111:58-64.[Medline]
  9. McNeal JE. Normal anatomy of the prostate and changes in benign prostatic hypertrophy and carcinoma. Semin Ultrasound CT MR 1988; 9:329-334.[Medline]
  10. McNeal JE, Redwine EA, Freiha FS, Stamey TA. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. Am J Surg Pathol 1988; 12:897-906.[Medline]
  11. Stamey TA, McNeal JE, Yemoto CM, Sigal BM, Johnstone IM. Biological determinants of cancer progression in men with prostate cancer. JAMA 1999; 281:1395-1400.[Abstract/Free Full Text]
  12. Scardino PT. The Gordon Wilson lecture: natural history and treatment of early stage prostate cancer. Trans Am Clin Climatol Assoc 2000; 111:201-241.[Medline]
  13. Delikatny EJ, Russell P, Hunter JC, et al. Proton MR and human cervical neoplasia: ex vivo spectroscopy allows distinction of invasive carcinoma of the cervix from carcinoma in situ and other preinvasive lesions. Radiology 1993; 188:791-796.[Abstract/Free Full Text]
  14. Rutter A, Hugenholtz H, Saunders JK, Smith IC. Classification of brain tumors by ex vivo 1H NMR spectroscopy. J Neurochem 1995; 64:1655-1661.[Medline]
  15. Lean CL, Delbridge L, Russell P, et al. Diagnosis of follicular thyroid lesions by proton magnetic resonance on fine needle biopsy. J Clin Endocrinol Metab 1995; 80:1306-1311.[Abstract]
  16. Lean CL, Newland RC, Ende DA, Bokey EL, Smith IC, Mountford CE. Assessment of human colorectal biopsies by 1H MRS: correlation with histopathology. Magn Reson Med 1993; 30:525-533.[Medline]
  17. Mackinnon WB, Huschtscha L, Dent K, Hancock R, Paraskeva C, Mountford CE. Correlation of cellular differentiation in human colorectal carcinoma and adenoma cell lines with metabolite profiles determined by 1H magnetic resonance spectroscopy. Int J Cancer 1994; 59:248-261.[Medline]
  18. Wallace JC, Raaphorst GP, Somorjai RL, et al. Classification of 1H MR spectra of biopsies from untreated and recurrent ovarian cancer using linear discriminant analysis. Magn Reson Med 1997; 38:569-576.[Medline]
  19. Mackinnon WB, Barry PA, Malycha PL, et al. Fine-needle biopsy specimens of benign breast lesions distinguished from invasive cancer ex vivo with proton MR spectroscopy. Radiology 1997; 204:661-666.[Abstract/Free Full Text]
  20. Barry P, Wadstrom C, Falk G, et al. What is the value of 1H MRS in detecting early malignant changes? In: Giuli R, ed The oesophagrogastric junction. Montrouge, France: John Libbey Eurotext, 1998; 1122-1127.
  21. Soper R, Himmelreich U, Painter D, et al. Pathology of hepatocellular carcinoma and its precursors using proton magnetic resonance and statistical classification strategies. Pathology 2002; 34:417-422.[CrossRef][Medline]
  22. Lean CL, Somorjai RL, Smith ICP, Russell P, Mountford CE. Accurate diagnosis and prognosis of human cancers by proton MRS and a three stage classification strategy. In: Webb G, eds. Annual reports on NMR spectroscopy. London, England: Academic Press, 2002; 71-111.
  23. Mountford CE, Somorjai RL, Malycha P, et al. Diagnosis and prognosis of breast cancer by magnetic resonance spectroscopy of fine needle aspirates analyzed using a statistical classification strategy. Br J Surg 2001; 88:1234-1240.[CrossRef][Medline]
  24. Kurhanewicz J, Vigneron DB, Hricak H, et al. Prostate cancer: metabolic response to cryosurgery as detected with 3D H-1 MR spectroscopic imaging. Radiology 1996; 200:489-496.[Abstract/Free Full Text]
  25. Kurhanewicz J, Vigneron DB, Hricak H, Narayan P, Carroll 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]
  26. Yacoe ME, Sommer G, Peehl D. In vitro proton spectroscopy of normal and abnormal prostate. Magn Reson Med 1991; 19:429-438.[Medline]
  27. Cornel EB, Heerschap A, Smits GA, Oosterhof GO, Debruyne FM, Schalken JA. Magnetic resonance spectroscopy detects metabolic differences between seven Dunning rat prostate tumor sublines with different biological behavior. Prostate 1994; 25:19-28.[Medline]
  28. Kurhanewicz J, Dahiya R, Macdonald JM, Chang LH, James TL, Narayan P. Citrate alterations in primary and metastatic human prostatic adenocarcinomas: 1H magnetic resonance spectroscopy and biochemical study. Magn Reson Med 1993; 29:149-157.[Medline]
  29. Kurhanewicz J, Vigneron DB, Nelson SJ, et al. Citrate as an in vivo marker to discriminate prostate cancer from benign prostatic hyperplasia and normal prostate peripheral zone: detection via localized proton spectroscopy. Urology 1995; 45:459-466.[CrossRef][Medline]
  30. Lynch MJ, Nicholson JK. Proton MRS of human prostatic fluid: correlations between citrate, spermine, and myo-inositol levels and changes with disease. Prostate 1997; 30:248-255.[CrossRef][Medline]
  31. van der Graff M, Schipper RG. 1H MRS of prostatic tissue focussed on the detection of spermine, a possible biomarker of malignant behaviour in prostate cancer (abstr) In: Book of abstracts: Society of Magnetic Resonance in Medicine 1998. Berkeley, Calif: Society of Magnetic Resonance in Medicine, 1998; 613.
  32. 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]
  33. Wefer AE, Hricak H, Vigneron DB, et al. Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology. J Urol 2000; 164:400-404.[CrossRef][Medline]
  34. 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]
  35. Hahn P, Smith ICP, Leboldus L, Littman C, Somorjai RL, Bezabeh T. The classification of benign and malignant human prostate tissue by multivariate analysis of H-1 magnetic resonance spectra. Cancer Res 1997; 57:3398-3401.[Abstract/Free Full Text]
  36. Doran S, Falk G, Somorjai R, et al. Pathology of Barrett’s esophagus by proton magnetic resonance spectroscopy and a statistical classification strategy. Am J Surg 2003; 185:232-238.
  37. Polacarz S, Darne J, Sheridan E, Ginsberg R, Sharp F. Endocervical carcinoma and precursor lesions: c-myc expression and the demonstration of field changes. J Clin Pathol 1991; 44:896-899.[Abstract/Free Full Text]
  38. Delikatny EJ, Hull WE, Mountford CE. The effect of altering time domains and window functions in two-dimensional proton COSY spectra of biological specimens. J Magn Reson 1991; 94:463-473.
  39. Swindle PW. Distinguishing adenocarcinoma of the prostate from benign prostatic hyperplasia using proton magnetic resonance spectroscopy. Thesis Australia: University of Sydney, 2000.
  40. Sabel MS, Zhang P, Barnwell JM, Winston JS, Hurd TC, Edge SB. Accuracy of sentinel node biopsy in predicting nodal status in patients with breast carcinoma. J Surg Oncol 2001; 77:243-246.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 RadiologyHome page
J. F. Faria, S. M. Goldman, J. Szejnfeld, H. Melo, C. Kater, P. Kenney, M. P. Huayllas, G. Demarchi, V. V. Francisco, C. Andreoni, et al.
Adrenal Masses: Characterization with in Vivo Proton MR Spectroscopy Initial Experience
Radiology, December 1, 2007; 245(3): 788 - 797.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
D. J. A. Margolis, J. M. Hoffman, R. J. Herfkens, R. B. Jeffrey, A. Quon, and S. S. Gambhir
Molecular Imaging Techniques in Body Imaging
Radiology, November 1, 2007; 245(2): 333 - 356.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
V. Kundra, P. M. Silverman, S. F. Matin, and H. Choi
Imaging in Oncology from The University of Texas M. D. Anderson Cancer Center: Diagnosis, Staging, and Surveillance of Prostate Cancer
Am. J. Roentgenol., October 1, 2007; 189(4): 830 - 844.
[Abstract] [Full Text] [PDF]

Home page
 CA Cancer J ClinHome page
D. A. Torigian, S. S. Huang, M. Houseni, and A. Alavi
Functional Imaging of Cancer with Emphasis on Molecular Techniques
CA Cancer J Clin, July 1, 2007; 57(4): 206 - 224.
[Abstract] [Full Text] [PDF]

Home page
 JCOHome page
A. G. Sorensen
Magnetic Resonance As a Cancer Imaging Biomarker
J. Clin. Oncol., July 10, 2006; 24(20): 3274 - 3281.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
H. Li, K. Sugimura, Y. Kaji, Y. Kitamura, M. Fujii, I. Hara, and M. Tachibana
Conventional MRI capabilities in the diagnosis of prostate cancer in the transition zone.
Am. J. Roentgenol., March 1, 2006; 186(3): 729 - 742.
[Abstract] [Full Text] [PDF]

Home page
 JAOA: Journal of the American Osteopathic AssociationHome page
N. Shah, A. Sattar, M. Benanti, S. Hollander, and L. Cheuck
Magnetic Resonance Spectroscopy as an Imaging Tool for Cancer: A Review of the Literature
J Am Osteopath Assoc, January 1, 2006; 106(1): 23 - 27.
[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 Swindle, P.
Right arrow Articles by Mountford, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Swindle, P.
Right arrow Articles by Mountford, C.

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