Original Research

Combined Diffusion-Weighted and DynamicContrast-Enhanced MRI for Prostate CancerDiagnosis—Correlation With Biopsy andHistopathology

Piotr Kozlowski, PhD,1–3* Silvia D. Chang, MD,1,3,5 Edward C. Jones, MD,1,4,7

Kenneth W. Berean, MD,4,6 Henry Chen,1 and S. Larry Goldenberg, MD1,2,8

Purpose: To determine whether the combination of diffu-sion-weighted (DW) and dynamic contrast-enhanced (DCE)MRI provides higher diagnostic sensitivity for prostate can-cer than each technique alone.

Materials and Methods: Fourteen patients with a clinicalsuspicion of prostate cancer underwent endorectal MRI ona 1.5T scanner prior to transrectal ultrasound (TRUS)-guided biopsies. The average values of the apparent diffu-sion coefficient (ADC, calculated from b-values of 0 and600), Ktrans, ve, maximum gadolinium (Gd) concentration,onset time, mean gradient, and maximum enhancementwere determined. Correlation with histology was based onbiopsy (six patients) and prostatectomy specimen (eightpatients) results. The Tukey-Kramer test was used for sta-tistical analysis.

Results: The average values of all MRI parameters, exceptve and maximum Gd concentration, showed significant dif-ferences between tumor and normal prostate. The sensitiv-ity and specificity values were respectively 54% (35–72%)

and 100% (95–100%) for the ADC data, and 59% (39–77%)and 74% (63–83%) for the DCE data. When both ADC andDCE results were combined, the sensitivity increased to87% (68–95%) and specificity decreased to 74% (62–83%).

Conclusion: All but two DW- and DCE-MRI parametersshowed significant differences between tumor and normalprostate. Combining both techniques provides better sen-sitivity, with a small decrease in specificity.

Key Words: diffusion-weighted MRI; dynamic contrast-en-hanced MRI; prostate cancer; endorectal coli; histopathol-ogyJ. Magn. Reson. Imaging 2006;24:108–113.© 2006 Wiley-Liss, Inc.

MAGNETIC RESONANCE IMAGING (MRI) has beenused to diagnose prostate cancer, with varying success,for over a decade (1). Although the development of en-dorectal coils significantly improved the image qualityrequired for accurate diagnosis (2), the sensitivity andspecificity of this technique have been variable (3) andits diagnostic value for staging prostate cancer remainsunder investigation (4). Therefore, several MRI-basedtechniques have been investigated recently with theaim to improve tumor delineation and thus providemore reliable staging and the ability to predict the gradeof tumor. Two such techniques—dynamic contrast-en-hanced (DCE) MRI and diffusion-weighted (DW) MRI—have shown particularly promising results.

DCE-MRI has been shown to significantly improvetissue characterization, and both the dynamic param-eters of the contrast enhancement (e.g., onset time, rateof enhancement, etc.) and the pharmacokinetic model-ing parameters have been used to differentiate betweencancer and normal tissue in the breast (5), bladder (6),and bone (7). DCE-MRI has been quite successful indiagnosing and staging prostate cancer in the periph-eral zone (PZ) (8–12); however, there appears to be someoverlap in the enhancement patterns between the tu-mor and the transitional zone (10). This technique isalso useful for identifying equivocal capsular penetra-tion to assess the involvement of the seminal vesicles

1Prostate Centre, Vancouver General Hospital, Vancouver, British Co-lumbia, Canada.2Department of Surgery, University of British Columbia, Vancouver,British Columbia, Canada.3Department of Radiology, University of British Columbia, Vancouver,British Columbia, Canada.4Department of Pathology and Laboratory Medicine, University of Brit-ish Columbia, Vancouver, British Columbia, Canada.5Department of Radiology, Vancouver General Hospital, Vancouver,British Columbia, Canada.6Department of Pathology and Laboratory Medicine, University of Brit-ish Columbia Hospital, Vancouver, British Columbia, Canada.7Department of Pathology and Laboratory Medicine, Vancouver Gen-eral Hospital, Vancouver, British Columbia, Canada.8Department of Surgery, Vancouver General Hospital, Vancouver, Brit-ish Columbia, Canada.Contract grant sponsor: Canadian Prostate Cancer Research Initiative.Presented in part at the 13th Annual Meeting of ISMRM, Miami Beach,FL, USA, 2005.*Address reprint requests to: P.K., Prostate Centre at VGH, 2660 OakSt., Vancouver, B.C., V6H 3Z6 Canada.E-mail: pkozlow@interchange.ubc.caReceived June 3, 2005; accepted March 30, 2006.DOI 10.1002/jmri.20626Published online 9 June 2006 in Wiley InterScience (www.interscience.wiley.com).

JOURNAL OF MAGNETIC RESONANCE IMAGING 24:108–113 (2006)

© 2006 Wiley-Liss, Inc. 108

(9), and can be used to monitor neoadjuvant hormonaltherapy (13). There are, however, conflicting reports asto whether DCE-MRI correlates well with histologicalgrades of prostate carcinoma (8,10,14).

In addition to characterizing tissue based on protondensity (PD) and relaxation properties (T1 and T2), MRIis capable of measuring molecular diffusion of water invivo. Molecular diffusion is a process in which mole-cules move along random paths. When a strong mag-netic field gradient is present, this random motion re-sults in irreversible dephasing of the MR signal, makingdiffusion a dominant source of contrast in the MR im-age (15). Diffusion of water molecules in tissue is af-fected by the presence of cellular structures that pro-vide barriers to free movement, and can therefore beused to characterize these structures and their changesin pathology (16). Early applications of DW-MRI in can-cer diagnosis have been limited to brain tumors, largelybecause of the extreme sensitivity of this technique tobulk motion. However, recent advances in MR hard-ware development, particularly fast gradient coils, cre-ated the possibility of using ultrafast MRI techniques ina standard clinical setting. DW-MRI studies of focalliver masses showed lower apparent diffusion coeffi-cient (ADC) values in malignant masses as compared tobenign masses and cysts (17). In another study, DW-MRI showed lower ADC values in mucin-producing tu-mors of the pancreas and pseudocysts as compared toserous cysts and cerebrospinal fluid (18). This tech-nique was also used to show a significant perfusioncontribution to the ADC values in the abdominal organsand hepatic lesions (19). Within last few years severalstudies have been published in which DW-MRI wasapplied to prostate cancer diagnosis (20–23).

The studies noted above showed that both DCE- andDW-MRI provide important information about the tu-mor microenvironment that may help to improve thediagnostic capability of traditional T2-weighted MRI.The purpose of the current pilot study was to investi-gate whether further gains could be made by combiningthe two techniques. Specifically, we correlated bothDCE- and DW-MRI techniques with histology to test thehypothesis that the combination of these techniqueswould provide higher diagnostic sensitivity than eachtechnique alone.

MATERIALS AND METHODS

Patient Selection and Histology

The study was approved by the institutional humanethics board, and all participants gave signed consentprior to entering the study. Fourteen patients with ahigh clinical suspicion for prostate cancer (elevated se-rum prostate-specific antigen (PSA) and/or prostatenodule detected during digital rectal examination) withno prior treatment were recruited for this prospectivestudy. The subjects underwent the MRI examinationprior to transrectal ultrasound (TRUS)-guided biopsies.

TRUS biopsies of the prostate were performed on anAcuson 128XP/10 scanner (Acuson Computed Sonog-raphy, Mountain View, CA, USA). The patients wereexamined with grayscale imaging in the axial and sag-

ittal planes with a 5 MHz transrectal probe. Prophylac-tic antibiotics and local anesthetic were administered topatients before the prostate biopsies were performed.Systematic octant biopsies were obtained from the fol-lowing locations in the PZ: left base, right base, leftlateral mid-gland, left medial mid-gland, right medialmid-gland, right lateral mid-gland, left apex, and rightapex. In one case a repeat biopsy was obtained from asuspicious area noted on the MR images, which con-firmed the MR finding.

The radical prostatectomy specimens, obtained fromeight of 14 patients with biopsy-proven carcinoma,were all dissected and examined histopathologically ina uniform manner. The external surfaces were inkedand the seminal vesicles were amputated. Apical andbladder neck tissues were removed as 0.5-cm-thick tis-sue doughnuts. The remainder of the prostate glandwas then dissected with serial transverse cuts perpen-dicular to the posterior capsule, at 0.4–0.6 cm inter-vals, from inferior to superior. The entire prostate glandwas submitted for histological examination on hema-toxylin and eosin (H&E)-stained slides. Each transverseslice was mapped for prostatic adenocarcinoma, main-taining the identity of the right, left, anterior, and pos-terior locations. The amount of carcinoma in the tissuesection, by surface area, was microscopically semi-quantitatively scored as follows: 0 � no carcinoma; 1 ��10% carcinoma; 2 � 10–25% carcinoma; 3 � 25–75%carcinoma; and 4 � �75% carcinoma. In addition, tu-mors identified with both biopsies and radical prosta-tectomy were assigned a Gleason score (24).

MRI Examinations

All MRI experiments were carried out on a 1.5T GESigna Horizon clinical MRI scanner (General Electric,Milwaukee, WI, USA) equipped with an echo-speed self-shielded gradient set (max. gradient strength � 23mT/m, max. rise time � 120 mT/m/msec). MRI signalswere acquired with an integrated endorectal pelvicphased-array coil (Medrad, Pittsburgh, PA, USA). Fastspin-echo (FSE) T2-weighted images (TR � 4450 msec,TE � 96 msec, FOV � 14 cm, slice thickness � 5 mmwith no gap, 256 � 192 matrix, four averages) wereacquired in axial and coronal orientations to provideanatomical details of the prostate gland. The slicethickness of 5 mm was chosen for the T2-weighted im-ages to allow direct comparison with the DW and DCEimages. Positions of 10 axial slices were then selected tocover the entire gland and were used for the DW- andDCE-MRI scans.

DW-MRI data were acquired using a single-shot FSEsequence developed at the University of California–SanFrancisco (FOV � 24 cm, slice thickness � 5 mm withno gap, BW � 31 kHz, TR � 8000 msec, TE � 87.5msec, flip angle � 90°, echo train length � 128, 128 �128 matrix, 16 averages, b-value � 600 sec/mm2).Four sets of images were acquired in every scan with nodiffusion gradient and with the diffusion gradient in thex, y, and z directions.

DCE-MRI was performed using a multislice fastspoiled gradient recalled (FSPGR) sequence (FOV � 24cm, slice thickness � 5 mm with no gap, TE � 3 msec,

DW- and DCE-MRI in Prostate Cancer 109

TR � 18.5 msec (T1) or 120 msec (PD), flip angle � 25°(T1) or 8° (PD), 256 � 128 matrix). PD images wereacquired prior to the injection of the contrast agent toallow calculations of the T1 values and subsequentlythe contrast agent concentrations in the tissue (25). Aseries of T1-weighted images were acquired prior to andfollowing a bolus injection of Gadodiamide (Omniscan,Nycomed-Amersham, Norway; dose: 0.1 mmol/kg in-jected within 10 seconds followed by a 20-mL flush ofnormal saline). Three baseline images were acquiredjust prior to the injection and an additional 42 wereobtained following the injection, with a time resolutionof 22 seconds per time point. The total time the patientsspent in the magnet during the MRI examinations wasapproximately 60 minutes.

Data Processing

All MRI data were processed offline with a softwarepackage developed in-house using Matlab (The Math-Works Inc., Natick, MA, USA).

ADC values were calculated from two DW-MR imagesacquired with b � 0 and 600 sec/mm2 (26). The ADCmaps were reconstructed by calculating ADC values ineach pixel of each slice. The average ADC values of thetumors were calculated from regions of interest (ROIs)defined as the hypointense areas in the PZ on ADCmaps, since previous studies showed lower ADC valuesin tumors as compared to normal PZ (21,22,27–29). Themean diffusivity was calculated as an average of ADCx,ADCy, and ADCz.

From the DCE-MRI data time-signal intensity param-eters, as defined in Ref. 10, were calculated in eachpixel of each slice to produce maps of the maximumenhancement (ME), onset time (OT), and mean gradient(MG). In addition, a two-compartment pharmacokineticmodel (30) was used to extract Ktrans, extravascularextracellular space (ve) and maximum tissue gadolin-ium (Gd) concentration from the time-contrast agentconcentration curves (31). The fitting was performed ineach pixel of each slice within ROIs that included theprostate gland. The criteria for fit acceptance were that1) the average difference between the measured and thefitted data was less than 0.5; 2) 0.0 � Ktrans � 5.0[mL/mL/min]; and 3) 0.0 � ve � 1.0 (32). Pixels that didnot meet all of these criteria were excluded from thecalculation of the average parameter values, and wereset to zero for display purposes on the parametricmaps. The average values of DCE parameters of thetumors were calculated from the ROIs defined as thehyperintense areas in the PZ on the Ktrans maps. TheseROIs corresponded to the fast-enhancing areas in theGd-DTPA concentration maps.

Control ADC and DCE parameter values were calcu-lated from ROIs defined in the areas that were deemednormal based on both T2-weighted MR images and his-tology results. T2-weighted images, ADC, and Gd-DTPAconcentration maps were coregistered using controlpoint registration with four manually defined controlpoints identifying anatomical features of the prostategland visible on all three images. For each control pointpair, a normalized cross-correlation calculation wasperformed on an 11 � 11 pixel area centered around the

control point. The control point was then moved to thepixel of highest correlation to optimize positioning.From the optimized control points, 3 � 3 transforma-tion matrices were calculated and applied to the differ-ent image types to compensate for translation and ro-tation. ROIs encompassing normal PZ and centralgland (CG) were defined on T2-weighted images andthen transferred to the parametric maps to calculatethe average values for all MRI parameters.

Correlation with histology was based on the results ofbiopsy and (if available) the radical prostatectomy re-sults. The locations of the tumors in MRI data weremapped to standard octant biopsy sites (MRI slices 1–3correspond to the base of the gland, 4–7 to the mid-gland, and 8–10 to the apex). Locations within the baseand apex were specified as left or right. For the mid-gland four locations were specified as the left mid-lat-eral, left mid-medial, right mid-medial, and right mid-lateral. Since our experimental setup did not allow for adirect comparison between histological slides and MRimages, histological results obtained from the entireglands following radical prostatectomy were alsomapped to the standard octant biopsy sites in a similarfashion.

Sensitivity and specificity were assessed based on thepresence or absence of tumor in the standard octantbiopsy sites. The presence of tumor was confirmed bythe biopsy and/or histology results, i.e., the locationswhere the histology showed the presence of the tumorwere considered positive even when biopsy was nega-tive for that site. The diffusion data were consideredpositive for the presence of tumor in a specific biopsysite if the ADC map showed a hypointense area in thatsite. Similarly, the DCE data were considered positive ifthe Ktrans map showed a hyperintense area in the bi-opsy site. The combined DW and DCE data were con-sidered positive for the presence of tumor if the resultsof either technique were positive. The hypo- (hyper-)intense areas were defined as the areas with lower(higher) intensity as compared to the intensity withinthe PZ in the same map. This assessment was carriedout visually by two independent readers, and the re-sults from both readers were identical.

Statistical analysis of differences between prostatecarcinoma (PCa), benign PZ, and benign CG was carriedout using a Tukey-Kramer test.

RESULTS

The average age of the 14 subjects in the study was 60.3years (range � 48–71 years), and the average value ofserum PSA was 9.4 ng/mL (range � 4.3–46 ng/mL).Biopsies acquired from four subjects were negative forcarcinoma and MRI data acquired from these subjectsdid not show any abnormalities. Data from these sub-jects were included in the overall analysis. Biopsiesfrom the remaining 10 subjects were positive for carci-noma, and eight of these subjects underwent radicalprostatectomy. Out of 32 tumors identified by biopsyand histology, 25 had a Gleason score of 7, and sevenhad Gleason score of 6.

Figure 1 shows typical MRI data acquired from one ofthe subjects. The reconstructed ADC map shows a hy-

110 Kozlowski et al.

pointense area in the right PZ, while the reconstructedKtrans map shows a hyperintense area in the same re-gion. A biopsy obtained from the right lateral mid-gland(i.e., the section corresponding to this area) was positive

for PCa. Note that the areas identified on the ADC andKtrans maps cover slightly different regions of the rightPZ. This seemed to be the case for the majority of thetumors.

The average values of all MRI parameters from PCa,normal PZ, and normal CG are presented in Table 1.The PCa values presented in the table were determinedonly from areas that were confirmed as prostate cancerby both MRI and histology. All MRI parameters, with theexception of extravascular extracellular space (ve) andmaximum concentration of Gd-DTPA, showed signifi-cant differences between cancer and benign PZ. In ad-dition, the average values of ADC, Ktrans, and meangradient within PCa were significantly different frombenign CG. Only the average value of ADC showed asignificant difference between normal PZ and CG.

When correlated with histology, the ADC data showed54% (35–72% at 95% confidence level) sensitivity and100% (95–100% at 95% confidence level) specificity.The DCE data showed a similar sensitivity of 59% (39–77% at 95% confidence level) but a lower specificity of74% (63–83% at 95% confidence level). When both theADC and DCE results were combined, the sensitivityincreased to 87% (68–95% at 95% confidence level)while specificity decreased to 74% (62–83% at 95%confidence level).

DISCUSSION

The results of our pilot study confirm the notion thateither DW- or DCE-MRI can be used for noninvasiveprostate cancer diagnosis, albeit with somewhat limitedsensitivity. The ADC values measured in our study inthe normal PZ and CG, and PCa generally correspondwell with previously published values (21,23,33,34).The only exception is a study by Gibbs et al (20), whichreported much higher ADC values compared to otherstudies. However, that study, unlike the majority ofother studies (including ours), did not use an endorec-tal RF coil for signal reception, and therefore the signal-to-noise ratio (SNR) may have played a role in the muchhigher ADC values reported. We validated our DW-MRImethod with phantoms, and used signal averaging toensure sufficient SNR on DW-MR images with the dif-fusion-sensitizing gradient switched on.

With the relatively low b-value employed in our study(b � 600 sec/mm2), the measured ADC primarily re-flects the diffusion coefficient of extracellular water(20). It has been shown that low ADC values in braintumors are associated with high cellular density (35),and one may expect a similar dependence in prostatetumors. Therefore, one can speculate that the differ-ences in ADC values between normal PZ, normal CG,and PCa are likely related to differences in the volumeand tortuosity of extracellular space between thesestructures. The normal prostate gland consists of anetwork of water-rich ducts and acini supported bystroma. The morphology of ducts and acini is similar inthe PZ and CG; however, the stroma is much looser inthe PZ than in the CG, resulting in larger extracellularspace (36). This is likely the main reason for higher ADCvalues in the PZ compared to the CG. In the PCa loosestroma is replaced by densely packed malignant epithe-

Figure 1. a: Axial cross-section through the mid-gland portionof a prostate gland. The left column displays a T2-weighted image,the central column displays an ADC map, and the right columndisplays a Ktrans map. ROIs in the bottom row correspond to lowADC values (cyan; average ADC � 0.001103 mm2/sec) and highKtrans values (magenta; average Ktrans � 1.472 min–1). Note thatthe two ROIs do not cover exactly the same area. b: RelativeT1-weighted image intensity (top) and Gd-DTPA concentration(bottom) vs. time following the bolus arrival time. Asterisks rep-resent average values from the ROI defined as a high-intensityarea on the Ktrans map (magenta ROI in part a). Open trianglesrepresent the average values from the ROI encompassing thecontralateral PZ. The solid line represents the model fit to thecontrast agent concentration curve. DCE parameters from the“magenta” ROI vs. contralateral PZ are as follows: Ktrans � 1.472min–1 vs. 0.492 min–1, ve � 0.430 vs. 0.371, maximum Gd-DTPAconcentration � 0.347 mM vs. 0.277 mM, maximum enhance-ment � 1.580 vs. 1.219, onset time � 0.594 min vs. 0.623 min,and mean gradient � 1.949 vs. 0.867.

DW- and DCE-MRI in Prostate Cancer 111

lial cells, resulting in a significant decrease in extracel-lular space and thus a lower ADC. Lower leakage spaceof contrast medium in the PCa compared to the PZ andCG appears to support this hypothesis, although thedifferences are not statistically significant.

DCE-MRI has been quite successful in diagnosingand staging prostate cancer. The significant image in-tensity enhancement of DCE-MRI has been attributedto increased microvessel density and increased perme-ability of the vessels (10,11), both of which are charac-teristics of tumor tissue. The values of the DCE param-eters measured in our study generally correspond wellwith those published previously (8–12), although thetime-signal intensity parameters (i.e., maximum en-hancement, mean gradient, and onset time) depend onthe specific pulse sequence settings, and therefore it issomewhat difficult to make direct comparisons betweendifferent studies. However, our findings that all DCEparameters, with the exception of extracellular ex-travascular space and maximum concentration of Gd-DTPA, differ significantly between the PCa and normalPZ, correspond well with previous reports.

In this study we correlated the histology and MRIresults by mapping them to a standard octant biopsymap. This procedure was uniformly carried out for allcases, even when the radical prostatectomy data wereavailable. We did not use T2-weighted MR images todefine tumor location, as this is considered to be some-what unreliable (4). Instead, we used T2-weighted MRimages to identify normal PZ and CG; however, we al-ways confirmed by histology that the chosen locationindeed corresponded to normal tissue.

The sensitivity values of both the DW- and DCE-MRItechniques measured in our study were similar (54%(35–72%) and 59% (39–77%), respectively), while thespecificity of DW-MRI was much higher than that of theDCE technique (100% (95–100%) vs. 74% (63–83%)).The perfect specificity of DW-MRI in our study is likelyrelated to the minimal presence or lack of scar tissue inthe PZ. Such scar tissue, which may result from needlebiopsy, would also have densely packed cells with littleextracellular space available for water to diffusethrough, and thus the ADC values in the scar tissuewould likely be similar to those measured in the tumor.

To our knowledge, this is the first report on the use ofcombined DW- and DCE-MRI in correlation with histol-ogy for prostate cancer diagnosis. Noworolski et al (34)

used combined DW- and DCE-MRI before; however,their main goal was to separate highly ductal tissues(PZ and glandular BPH) from less ductal tissues (CGand stromal BPH), and they did not provide any directcorrelation between MRI and histology in their study.Our results show that the sensitivity of combined DW-and DCE-MRI is significantly improved over that ofeach technique alone, with a small decrease in speci-ficity. This suggests that both techniques provide com-plementary information about the morphology and pa-thology of prostatic tissue. Indeed, in the majority ofcases we noticed that the areas of low ADC did notentirely coincide with the fast-enhancing areas in DCE-MRI (see Fig. 1a). This is likely because the areas of lowADC are associated with high cellular density withlower extracellular space (35), whereas high contrastenhancement is attributed to areas with high microves-sel density and leaky vasculature (10,11) characteristicof tumor tissue. Since both types of tissue are presentwithin tumor, the combined ADC and DCE-MRI tech-niques are more likely to provide accurate cancer de-tection, as was found in our study.

It is reasonable to expect that the combined DW- andDCE-MRI techniques may provide a more reliable wayto noninvasively grade tumors, as compared to theseand other MRI techniques alone. Indeed, tumors ofhigher Gleason grade appear to have increased cellulardensity (24), which will likely result in lower ADC val-ues. Higher-grade tumors also have fewer ducts, andsince Noworolski et al (34) suggested that combinedDW- and DCE-MRI can distinguish between highly duc-tal and less ductal tissues within the prostate (34), onewould expect this technique to improve tumor grading.The range of Gleason scores in our study was limited to6 and 7, with the majority of tumors having a score of 7.This prevented any correlation between MRI parame-ters and the Gleason score. One of the limitations of thecurrent study is the small number of patients (N � 14)involved. A larger study group of patients with a muchbroader range of tumor grades, and a direct correlationbetween histological sections and corresponding MRimages are required to test whether a correlation be-tween the MRI parameters and the Gleason score ex-ists. The ideal diagnostic technique would be highlyspecific for malignancy and highly sensitive for biolog-ically significant cancers. We believe that the combined

Table 1Average Values (Mean � SD) of Mean Diffusivity and DCE MRI Parameters

ADC(10–3mm2/second)

Ktrans (minutes–1) veMax. Conc.

(mM)Max. Enh. OT (minutes) Mean. Grad.

PZ 1.5730 � 0.2700 0.598 � 0.560 0.382 � 0.148 0.274 � 0.115 1.218 � 0.245 0.741 � 0.133 1.008 � 0.659CG 1.3750 � 0.1670c 0.596 � 0.288 0.443 � 0.084 0.314 � 0.069 1.431 � 0.190 0.733 � 0.090 1.004 � 0.311PCa 0.9930 � 0.1580a,b 1.263 � 0.544a,b 0.326 � 0.131 0.257 � 0.099 1.463 � 0.218a 0.644 � 0.089d 2.297 � 1.023a,b

aPCa significantly different than PZ (P � 0.001).bPCa significantly different than CG (P � 0.001).cCG significantly different than PZ (P � 0.05).dPCa significantly different than PZ (P � 0.05).ADC � mean diffusivity, Max. Conc. � maximum concentration of Gd-DTPA, Max. Enh. � maximum enhancement, OT � onset time, MeanGrad. � mean gradient, PZ � normal peripheral zone (DW MRI: N � 18, DCE MRI: n�24), CG � normal central gland (N � 15), PCa �tumour (DW MRI: N � 16, DCE MRI: N � 24).

112 Kozlowski et al.

DW- and DCE-MRI techniques may fulfill this require-ment.

In conclusion, all of the experimental parametersmeasured with DW- and DCE-MRI, with the exceptionof the extravascular extracellular space and maximumconcentration of Gd-DTPA, showed significant differ-ences between tumor and normal PZ. The combinationof both techniques provides improved sensitivity with asmall decrease in specificity.

ACKNOWLEDGMENT

The authors thank Dr. Don Vigneron from the Univer-sity of California–San Francisco for providing the DWsingle-shot FSE sequence.

REFERENCES1. Engelbrecht MR, Jager GJ, Laheij RJ, Verbeek AL, van Lier HJ,

Barentsz JO. Local staging of prostate cancer using magnetic res-onance imaging: a meta-analysis. Eur Radiol 2002;12:2294–2302.

2. Schnall MD, Imai Y, Tomaszewski J, et al. Prostate cancer: localstaging with endorectal surface coil MR imaging. Radiology 1991;178:797–802.

3. Jager GJ, Ruijter ET, van der Kaa CA, et al. Local staging of prostatecancer with endorectal MR imaging: correlation with histopathol-ogy. AJR Am J Roentgenol 1996;166:845–852.

4. Jager GJ, Severens JL, Thornbury JR, et al. Prostate cancer stag-ing: should MR imaging be used?—a decision analytic approach.Radiology 2000;215:445–451.

5. Gibbs P, Liney GP, Lowry M, Kneeshaw PJ, Turnbull LW. Differen-tiation of benign and malignant sub-1 cm breast lesions usingdynamic contrast enhanced MRI. Breast 2004;13:115–121.

6. Tekes A, Kamel I, Imam K, et al. Dynamic MRI of bladder cancer:evaluation of staging accuracy. AJR Am J Roentgenol 2005;184:121–127.

7. Montemurro F, Russo F, Martincich L, et al. Dynamic contrastenhanced magnetic resonance imaging in monitoring bone metas-tases in breast cancer patients receiving bisphosphonates and en-docrine therapy. Acta Radiol 2004;45:71–74.

8. Turnbull LW, Buckley DL, Turnbull LS, Liney GP, Knowles AJ.Differentiation of prostatic carcinoma and benign prostatic hyper-plasia: correlation between dynamic Gd-DTPA-enhanced MR imag-ing and histopathology. J Magn Reson Imaging 1999;9:311–316.

9. Barentsz JO, Engelbrecht M, Jager GJ, et al. Fast dynamic gado-linium-enhanced MR imaging of urinary bladder and prostate can-cer. J Magn Reson Imaging 1999;10:295–304.

10. Padhani AR, Gapinski CJ, Macvicar DA, et al. Dynamic contrastenhanced MRI of prostate cancer: correlation with morphology andtumor stage, histological grade and PSA. Clin Radiol 2000;55:99–109.

11. Muramoto S, Uematsu H, Kimura H, et al. Differentiation of pros-tate cancer from benign prostate hypertrophy using dual-echo dy-namic contrast MR imaging. Eur J Radiol 2002;44:52–58.

12. Noworolski SM, Henry RG, Vigneron DB, Kurhanewicz J. Dynamiccontrast-enhanced MRI in normal and abnormal prostate tissuesas defined by biopsy, MRI, and 3D MRSI. Magn Reson Med 2005;53:249–255.

13. Padhani AR, MacVicar AD, Gapinski CJ, et al. Effects of androgendeprivation on prostatic morphology and vascular permeabilityevaluated with MR imaging. Radiology 2001;218:265–374.

14. Jager GJ, Ruijter ET, van de Kaa CA, et al. Dynamic TurboFLASHsubtraction technique for contrast-enhanced MR imaging of theprostate: correlation with histopathologic results. Radiology 1997;203:645–652.

15. Le Bihan D, Breton E, Lallemand D, et al. MR imaging of intravoxelincoherent motions: application to diffusion and perfusion in neu-rologic disorders. Radiology 1986;161:401–407.

16. Le Bihan D. Molecular diffusion nuclear magnetic resonance im-aging. Magn Reson Q 1991;7:1–30.

17. Namimoto T, Yamashita Y, Sumi S, Tang Y, Takahashi M. Focalliver masses: characterization with diffusion-weighted echo-planarMR imaging. Radiology 1997;204:739–744.

18. Yamashita Y, Namimoto T, Mitsuzaki K, et al. Mucin-producingtumor of the pancreas: diagnostic value of diffusion-weighted echo-planar MR imaging. Radiology 1998;208:605–609.

19. Yamada I, Aung W, Himeno Y, Nakagawa T, Shibuya H. Diffusioncoefficients in abdominal organs and hepatic lesions: evaluationwith intravoxel incoherent motion echo-planar MR imaging. Radi-ology 1999;210:617–623.

20. Gibbs P, Tozer DJ, Liney GP, Turnbull LW. Comparison of quanti-tative T2 mapping and diffusion-weighted imaging in the normaland pathologic prostate. Magn Reson Med 2001;46:1054–1058.

21. Issa B. In vivo measurement of the apparent diffusion coefficient innormal and malignant prostatic tissues using echo-planar imag-ing. J Magn Reson Imaging 2002;16:196–200.

22. Chan I, Wells 3rd W, Mulkern RV, et al. Detection of prostate cancerby integration of line-scan diffusion, T2-mapping and T2-weightedmagnetic resonance imaging; a multichannel statistical classifier.Med Phys 2003;30:2390–2398.

23. Hosseinzadeh K, Schwarz SD. Endorectal diffusion-weighted imag-ing in prostate cancer to differentiate malignant and benign periph-eral zone tissue. J Magn Reson Imaging 2004;20:654–661.

24. Montironi R, Mazzuccheli R, Scarpelli M, et al. Gleason grading ofprostate cancer in needle biopsies or radical prostatectomy speci-mens: contemporary approach, current clinical significance andsources of pathology discrepancies. BJU Int 2005;95:1146–1152.

25. Parker GJ, Suckling J, Tanner SF, et al. Probing tumor microvas-cularity by measurement, analysis and display of contrast agentuptake kinetics. J Magn Reson Imaging 1997;7:564–574.

26. Hosseinzadeh K, Read K, DiBiase S, Gullapalli R. Combined use ofdiffusion-weighted imaging (DWI) and MR spectroscopic imaging(MRSI) in the detection of prostate cancer for disease-targeted ther-apy. In: Proceedings of the 9th Annual Meeting of ISMRM, Glasgow,Scotland, 2001. Abstract 2343.

27. Gibbs P, Pickles MD, Sreenivas M, Knowles A, Turnbull LW. Diffu-sion tensor imaging of the normal and pathologic prostate at 3T. In:Proceedings of the 13th Annual Meeting of ISMRM, Miami Beach,FL, USA, 2005. Abstract 1940.

28. Xu JG, Humphrey PA, Kibel AS, Narra V, Ackerman JJ, Song S-K.In vivo detection and localization of prostate carcinoma using DTI.In: Proceedings of the 13th Annual Meeting of ISMRM, MiamiBeach, FL, USA, 2005. Abstract 2125.

29. Haker SJ, Szot Barnes A, Maier SE, Tempany CM, Mulkern RV.Diffusion tensor imaging for prostate cancer detection: preliminaryresults from a biopsy-based assessment. In: Proceedings of the13th Annual Meeting of ISMRM, Miami Beach, FL, USA, 2005.Abstract 2126.

30. Tofts PS, Berkowitz B, Schnall MD. Quantitative analysis of dy-namic Gd-DTPA enhancement in breast tumors using a permeabil-ity model. Magn Reson Med 1995;33:564–568.

31. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parametersfrom dynamic contrast-enhanced T(1)-weighted MRI of a diffusabletracer: standardized quantities and symbols. J Magn Reson Imag-ing 1999;10:223–232.

32. Galbraith SM, Lodge MA, Taylor NJ, et al. Reproducibility of dy-namic contrast-enhanced MRI in human muscle and tumors: com-parison of quantitative and semi-quantitative analysis. NMRBiomed 2002;15:132–142.

33. Vigneron DB, Xu D, Chen AP, Swanson MG, Kurhanewicz J. Dif-fusion tensor imaging of the prostate using single-shot fast spinecho. In: Proceedings of the 10th Annual Meeting of ISMRM, Ho-nolulu, HI, USA, 2002. Abstract 457.

34. Noworolski SM, Chen AP, Vigneron DB, Kurhanewicz J. Assess-ment of prostatic ductal volume using combined dynamic contrast-enhanced MRI and diffusion MRI. In: Proceedings of the 11th An-nual Meeting of ISMRM, Toronto, Canada, 2003. Abstract 1464.

35. Lyng H, Haraldseth O, Rofstad EK. Measurement of cell densityand necrotic fraction in human melanoma xenografts by diffusionweighted magnetic resonance imaging. Magn Reson Med 2000;43:828–836.

36. McNeal JE. Anatomy and normal histology of the human prostate.In: Foster CS, Bostwick DG, editors. Pathology of the prostate.Philadelphia: WB Saunders; 1998. p 19–34.

DW- and DCE-MRI in Prostate Cancer 113