CME. Multiparametric MRI for Prostate Cancer Detection: Performance in Patients With Prostate-Specific Antigen Values Between 2.

Transcription

1 CME JOURNAL OF MAGNETIC RESONANCE IMAGING 00:00 00 (2013) Original Research Multiparametric MRI for Prostate Cancer Detection: Performance in Patients With Prostate-Specific Antigen Values Between 2.5 and 10 ng/ml Antonella Petrillo, MD, 1 * Roberta Fusco, Eng, 1 Sergio V. Setola, MD, 1 Francesco M. Ronza, MD, 1 Vincenza Granata, MD, 1 Mario Petrillo, MD, 1 Guglielmo Carone, MD, 1 Mario Sansone, Eng PhD, 2 Renato Franco, MD, 3 Franco Fulciniti, MD, 3 and Sisto Perdona, MD 4 Purpose: To assess the diagnostic performance of multiparametric MRI (mpmri), in the detection of prostate cancer, including morphologic sequences (mmri), diffusionweighted imaging (DWI), and MR spectroscopy (MRS). Combined morphological and functional MRI scoring systems was used for urological radiological work-up of patients with a prostate-specific antigen (PSA) value 10 ng/ml. Materials and Methods: The study included 136 of 200 consecutive patients with PSA values between 2.5 and 4 ng/ml and an abnormal digital rectal examination (DRE), or patients with PSA values between 4 and 10 ng/ml, independently from DRE. Each patient provided informed consent to undergo at serum free/total PSA ratio (f/t PSA) assay, mmri, MRS, DWI, and transrectal ultrasonography (TRUS) biopsy. The MRI datasets were scored singularly; then mmri and DWI, mmri and MRS data were combined in a coupled score, and finally mmri, DWI, and MRS data were combined in a single score (cmri score). Results: Scores were correlated to negative biopsies and significant/insignificant Gleason score biopsies. Receiveroperator-characteristic curve and McNemar tests were performed. Cancer was diagnosed in 18% of patients. The cmri score showed: (i) the highest sensitivity (0.84) and negative predictive value (0.93); (ii) a significant correlation with Gleason score; and (iii) a statistically different median value between significant and insignificant Gleason score. 1 National Cancer Institute of Naples Fondazione Giovanni Pascale, Department of Diagnostic Imaging, Naples, Italy. 2 University of Naples Federico II, Department of Electrical Engineering and Information, Naples, Italy. 3 National Cancer Institute of Naples Fondazione Giovanni Pascale, Department of Pathology, Naples, Italy. 4 National Cancer Institute of Naples Fondazione Giovanni Pascale, Department of Urology, Naples, Italy. *Address reprint requests to: A.P., I.N.T. G. Pascale, Via Mariano Semmola, Naples, Italy. antonellapetrillo2@gmail.com Received July 17, 2012; Accepted May 16, DOI /jmri View this article online at wileyonlinelibrary.com. Conclusion: The cmri score could identify patients with a PSA10 ng/ml who will have a negative work-up, for its high negative predictive value, and patients at high risk for significant prostate cancer because of its correlation with the Gleason score. Key Words: prostate cancer; MRI; spectroscopy; diffusion-weighted imaging; Gleason score J. Magn. Reson. Imaging 2013;00: VC 2013 Wiley Periodicals, Inc. PROSTATE CANCER DIAGNOSIS still represents a clinical challenge, as currently available diagnostic methods remain suboptimal (1 3). On the other side, there are both overdiagnosis and overtreatment risks (4,5). Prostate biopsy remains the only procedure determining definitive diagnosis. However, the biopsy Gleason grade could undergrade the disease (6 8). Moreover, this is an invasive procedure with a suboptimal sensitivity because there is still 10 20% cancer detection rate on repeat biopsy after an initial negative biopsy (9). Transrectal biopsy is also associated to some complication, particularly urosepsis, with increasing cases of antibiotic resistant infections (10). Infection rate is correlated to the number of biopsy cores (11). In general, a reliable diagnostic test should allow early cancer diagnosis, minimizing the occurrence of unnecessary biopsies, and it should detect significant cancer. MRI provides excellent highcontrast, high-resolution images of the prostate (8). In recent years, functional techniques have been applied to improve the performance of MRI in the diagnosis of prostate cancer (12). MR spectroscopy (MRS) measures prostate metabolites concentrations, particularly choline and citrate, that are respectively increased and reduced in cancer. Diffusion Weighted Image (DWI) provides water diffusion properties based imaging, with reduced water diffusion in highly cellular VC 2013 Wiley Periodicals, Inc. 1

2 2 Petrillo et al. cancer tissues (9). Early promising data suggest that multiparametric MRI (mpmri), including morphologic sequences (mmri), dynamic contrast, DWI, and MRS, may be of additional value for the localization of prostate cancer and its local staging (12). The aim of this study was to evaluate the diagnostic performance of mpmri in the detection of prostate cancer, including mmri, DWI and MRS. This evaluation was performed by means of a combined morphological and functional MRI scoring system used in urological radiological work-up for patients with a prostate specific antigen (PSA) values 10 ng/ml. MATERIAL AND METHODS Patient Selection From November 2009 to July 2010, 200 consecutive male patients were screened to be enrolled in a singlecenter prospective observational study. The inclusion criteria were as follows: PSA values between 2.5 and 4 ng/ml and an abnormal DRE, or PSA values between 4 and 10 ng/ml, independently from DRE. The exclusion criteria were as follows: inability to give informed consent; prior history of prostate cancer; prior pelvic irradiation; previous hormonal or surgical therapy; MRI contraindications (cardiac pacemakers, surgical clips, metallic hip implant); recent rectal surgery; severe hemorrhoids; or latex allergy. All eligible patients underwent morphological and functional MRI and TRUS prostate biopsies. A register included every single patient that decided to be enrolled in this study approved by the Institutional Ethics Committee. Any patient who decided to undergo the MRI examination signed an explicit informed consensus. MRI Data Acquisition The MRI was performed before prostate biopsies. The MRI protocol included insertion of an endorectal coil (EC) (Medrad, Pittsburg, PA), inflated with ml of air, and subsequent imaging acquisition using both a 1.5 Tesla (T) MRI system (Siemens Symphony Tim, Erlangen, Germany) and a four-channel phased-array surface coil coupled to the endorectal coil. The MRI total acquisition time was approximately 40 min. The mpmri included mmri, MRS, and DWI. The mmri included turbo spin echo (TSE) T2-weighted sequences in three perpendicular planes and coronal and transversal TSE T1-weighted sequence. Transversal TSE T2- weighted sequence parameters were as follows: repetition time/echo time (TR/TE), 3800/104 ms (sagittal: 4660/96 ms; coronal: 5000/98 ms); slice thickness, 3 mm/gap 0 mm; flip angle, 180 ; acquisition matrix, (sagittal and coronal ); field of view (FOV), mm 2. Transversal T2-weighted images were acquired with and without fat saturation. TSE T1-weighted were acquired in coronal to visualize lymph nodes, with the following sequence parameters: TR/TE, 550/12 ms; slice thickness, 3 mm/gap 0 mm; flip angle, 150 ; acquisition matrix, ; FOV, mm 2. Transversal TSE T1-weighted sequence parameters were: TR/TE, 706/7.8 ms, slice thickness, 3 mm/gap 0 mm; flip angle, 150 ; acquisition matrix, ; FOV, mm 2.The MRS parameters were: TR/TE, 690/120 ms; flip angle, 90. The volume of interest (VOI) was composed by voxel of mm 3. The transversal echo-planar DWI pulse sequence parameters were: TR/TE, 2700/83 ms; slice thickness, 3.56 mm; flip angle, 90, acquisition matrix; and FOV, mm 2 ; b value ¼ 0, 50, 100, 150, 300, 600, 800 s/mm 2. Antispasmodic drug was not used. Image Analysis All mpmri findings were graded by two radiologists with over 5 years of experience in prostate MRI and blinded to both clinical and biochemical data of the patients. A consensus evaluation method for scoring MRI findings was adopted to reach more reliability in image interpretation (12). The scoring system was evaluated on a per-patient basis, being not achievable at a reasonably accurate evaluation on a per-site basis. A 5-mm diameter was considered as the inferior threshold for including lesions. The mmri was scored using a 0 3 scale for peripheral gland (0 ¼ no abnormality; 1 ¼ geometric hypointense area, with low cancer suspicion; 2 ¼ diffuse non nodular hypointense area, with intermediate cancer suspicion; 3 ¼ nodular hypointense area, with high cancer suspicion) and by using a 0 2 scale for the central gland (0 ¼ heterogeneous well-marginated nodules; 1 ¼ homogeneously hypointense well-marginated nodules; 2 ¼ homogeneously hypointense ill-marginated nodules) (13 16). The MRS raw data were elaborated using Syngo MR- B17 spectroscopy package (Siemens, Erlangen, Germany). The software generated spectra and localized corresponding voxels on superimposed T2-weighted images. The MRS was rated as diagnostic or nondiagnostic if the metabolites spectrum was good, after resolving metabolic resonances, and limited baseline distortions, due to residual water or lipids, were present. For each voxel, the software calculated areas under citrate, choline, and creatine peaks and cholineþcreatine/citrate peak ratio. Diagnostic spectra were considered negative or positive, and scored as 0 or 1, respectively, in relation to cholineþcreatine/citrate ratio threshold of 0.86, as reported in the literature (17). The DWI data and apparent diffusion coefficient (ADC) maps were elaborated using Syngo MR-B17 diffusion package (Siemens, Erlangen, Germany). DWI was considered positive when focal areas, characterized by persistent signal intensity at b-value increase and/or ADC map hypointensity in relation to the adjacent gland, were evident (15). DWI was scored respectively as 0 or 1 if a negative or positive finding was reported. A separate score for each MRI sequence was given for each prostate site. These data were, first, combined in couples (mmri/mrs score and mmri/dwi score), as the sum of every site scores. Finally, a multiparametric combined score (cmri score) was obtained for every site as the sum of scores from every

3 Multiparametric MRI for Prostate Cancer Detection 3 Table 1 Patients Characteristics a Patients with diagnosis Overall of prostate cancer Patients, n Age, years 66,35 (8.4) 64.5 (7.3) PSA, ng/ml 6.8 (2.4) 7.4 (2.1) f/t PSA, % 18.5 (7.3) 18.3 (8.6) Positive biopsy 3 (2.5) cores, n Max % of tumor 40.4 (26.7) in cores, % Gleason score 6.4 (1.5) range 3 10 Combined MRI score 2.4 (1.1) range 0 5 a Values are expressed as a mean (SD). MRI technique. For every derived score, the highest site score was used as patient reference score. Biopsy and Pathological Analysis TRUS biopsy was performed as previously described in literature (18), by an expert urologist with a 10-years experience. A 12-site biopsy scheme (two cores, paramedian and lateral, from base, midgland, and apex, bilaterally) was adopted to cover the peripheral prostate, using a biopsy gun armed with a 18-gauge needle. In patients with a prostate volume >45 cc, four additional cores were acquired from the central gland (superiorly and inferiorly, bilaterally). The urologist was blinded to the MRI findings during the prostatic biopsy. A pathologist with more than 10 years experience in genitourinary pathology evaluated biopsy cores. Gleason scores were considered significant if at least cancer grade 4 was evident. In case of a negative biopsy, the patient was scheduled for a follow-up visit in a year. When high-grade prostatic intraepithelial neoplasia (HG-PIN) or atypical small acinar proliferation (ASAP) were diagnosed, the biopsy was repeated after 3 6 months using the 16-sites (12 peripheral and 4 central sites) scheme. If the repeat biopsy was negative for cancer, the patient was scheduled for a follow-up visit at 1 year. Patients were considered positive at biopsy if cancer was reported in at least one core, independently from the Gleason grade. The highest Gleason grade reported at biopsy was considered as patients reference Gleason grade value. Statistical Analysis Statistical correlation was performed on a per-patient basis correlating patients reference scores derived from MRI datasets with the results of the biopsies, as described above. Receiver-operator-characteristic (ROC) curves were calculated for serum free/total PSA ratio (f/t PSA), mmri/mrs scores, mmri/dwi scores, and cmri scores. For each, ROC area under curve (AUC) was calculated (19). AUCs were compared using nonparametric methods (20) and optimal thresholds were obtained maximizing the Youden index (21). Fishers exact test was used to evaluate statistical significance of decision matrix (DM) tables (22). Sensitivity, specificity, and positive and negative predictive values (PPV and NPV) were calculated for mmri, MRS, DWI, mmri/mrs score, mmri/dwi score, and for cmri score, using DM. Optimal thresholds were used for the last three scores. Sensitivity, specificity, PPV, and NPV were calculated for f/t PSA using the different cutoffs reported in the literature (10%, 20%, and 30%) and the optimal threshold (¼ 2). Matched sample tables and McNemar tests were used to compare diagnostic performance for all techniques and derived scores (23). Spearman s rank correlation coefficient was calculated to evaluate correlation between cmri and Gleason scores (24). Wilcoxon rank sum test for unpaired data was used to evaluate cmri score median values differences between significant and insignificant Gleason score groups. Percentage of negative, significant, and insignificant Gleason score biopsies were calculated in relation to mmri/mrs scores, mmri/dwi scores, and cmri scores. A P value <0.05 was considered significant. All analysis were performed using Statistic Toolbox of Matlab R2007a. RESULTS Overall, 136 patients were enrolled in the study. They completed the study protocol and were included in the analysis (Table 1). PSA values ranged between 4 and 10 ng/ml in 119 patients and between 2.5 and 4 Figure 1. Patient number 19. Transversal TSE T2-weighted sequence (left image; TR/TE, 3800/104 ms; flip angle, 180 ) showed a small ill-defined nodular hypointense area (mmri ¼ 3) in the right peripheral midgland (arrowhead). MRS was negative, while positive findings were evident on ADC maps (right image, arrowhead). cmri score was 4. Gleason score was 8, with 1 positive core and 60% maximum involved core length at biopsy.

4 4 Petrillo et al. Figure 2. Patient number 21. On transversal T2-weighted sequence (left image; TR/TE, 3800/104 ms; flip angle, 180 ), a diffuse hypointense area (mmri ¼ 2), particularly in the right peripheral midgland (arrowhead), was evident. MRS was positive, and corresponding ADC map showed hypointense areas (right image, arrowhead). cmri score was 4. Gleason score was 8, with 4 positive cores and 60% maximum involved core length at biopsy. ng/ml in 17 patients. The mmri was scored 0, 1, 2, and 3 in 55, 27, 49, and 5 patients, respectively. The MRS and DWI were positive in 79 and 76 patients, respectively. Prostate biopsy showed cancer in 23 patients, HG- PIN in 2 patients, HG-PINþASAP in 3 patients, and ASAP in 19 patients. Repeat biopsy was performed in 18/24 patients with HG-PIN or ASAP at the first biopsy. At re-biopsy, cancer was diagnosed in 2 cases and ASAP was confirmed in other 3 cases, while 13 patients were negative. Overall, 25 of 136 patients (18%) had a diagnosis of prostate cancer (Table 1; Figs. 1 and 2). The Gleason score was significant in 12/25 cases (48%). All significant Gleason scores (cancer grade 4) were 7. After clinical evaluation and discussion with the patients about available treatments, only nine patients underwent prostatectomy. Figure 3 shows ROC curves and AUC. Optimal thresholds were 2 for mmri/mrs scores, mmri/dwi scores, and cmri scores and 13% for f/t PSA. The AUC were 0.67, 0.69, 0.74, and 0.58, respectively. The cmri AUC was greater than mmri/mrs (P ¼ 0.07) and mmri/dwi AUCs (P ¼ 0.04), and it was significantly greater than f/t PSA AUC (P < 0.01). Fishers exact tests showed statistical significance for each DM table with a P value always < Table 2 reports diagnostic performance of every MRI technique either singularly or combined in scores and f/t PSA diagnostic performance using different thresholds. All scores showed statistical superiority compared with f/ t PSA and with mmri, MRS, DWI considered singularly. The cmri score showed the highest value of sensibility and NPV among the scores, but no statistical superiority. Using a cutoff value of 2 (yielded cutoff in ROC analysis), the cmri score showed the highest sensitivity (0.84) and NPV (0.93) with a 0.50 specificity and a 0.27 PPV. cmri score produced four false negatives (FN), all showing a Gleason score <7 on a single biopsy core and a core length tumor involvement <50%. A significant correlation was obtained by comparing the cmri score to the Gleason score (Spearman s rank correlation coefficient ¼ 0.70, P < 0.05). Wilcoxon rank sum test showed that significant and insignificant Gleason score tumors had a different cmri score median value (P < 0.01). Percentage of negative biopsies and significant and insignificant Gleason score biopsies in relation to mmri/mrs scores, mmri/dwi scores, and cmri score is reported in Table 3. Significant Gleason score tumors showed a cmri score >2 in 75% of cases, while 23% of insignificant Gleason score tumors had a cmri score > 2 (Table 3). No significant Gleason score tumors showed a cmri score < 2. DISCUSSION Patients with a PSA value <10 ng/ml carry a significant risk to have cancer. In current clinical practice these patients usually undergo a complete diagnostic evaluation, including TRUS biopsy. TRUS shows a low accuracy in cancer detection (9), and with this clinical approach a high number of biopsies are performed also in negative patients (2). An initial biopsy technique recommended scheme consists of at least 8 to 12 cores. Despite the introduction of extended biopsies until a saturation scheme, the initial cancer Figure 3. ROC curves and AUC values for f/t PSA and mmri/dwi score, mmri/mrs score, and cmri score. The rectangles mark the couple of sensitivities and specificities correspondent to the optimal thresholds calculated maximizing Youden Index.

5 Multiparametric MRI for Prostate Cancer Detection 5 Table 2 MRI and f/t PSA: Diagnostic Performance MRI Sensitivity (%) Specificity (%) PPV (%) NPV (%) mmri 76 (19/25) 44 (49/111) 23 (19/81) 89 (49/55) MRS 76 (19/25) 46 (51/111) 24 (19/79) 89 (51/57) DWI 64 (16/25) 46 (51/111) 21 (16/76) 85 (51/60) mmri/mrs score 80 (20/25) 61 (68/111) 32 (20/62) 92 (68/74) mmri/dwi score 72 (18/25) 57 (63/111) 27 (18/65) 89 (63/71) cmri score 84 (21/25) 50 (56/111) 28 (21/76) 93 (56/60) f/t PSA cutoff ¼ 10% 8 (2/25) 95 (105/111) 25 (2/8) 82 (105/128) cutoff ¼ 13% 32 (8/25) 88 (98/111) 38 (8/21) 85 (98/115) cutoff ¼ 20% 72 (18/25) 39 (43/111) 21 (18/86) 86 (43/50) cutoff ¼ 30% 96 (24/25) 5 (6/111) 19 (24/129) 86 (6/7) detection did not significantly improve, although adverse events and complications increased, such as fever, pain, hematospermia/hematuria, positive urine cultures, and rarely sepsis, were reported in approximately 0.4% of cases (25). Therefore, it would be desirable to achieve a reliable diagnostic test which allows early cancer detection, and contemporarily reduces the number of negative biopsies (6). With this in mind, we evaluated the diagnostic performance of mpmri in detection of prostate cancer, including mmri, DWI, and MRS. In our study, the cmri score showed higher sensitivity and higher NPV than either single techniques (mmri, DWI, and MRS), or their combinations in couples (mmri/mrs score and mmri/dwi score). Moreover, it was superior to f/t PSA. The cmri score was also better in differentiating negative biopsies and significant or insignificant Gleason score positive biopsies. A significant correlation was observed between cmri score and Gleason score, with all significant Gleason score tumors showing a cmri score 2. Thus, the cmri score could reduce negative biopsies in patients with a PSA 10 ng/ml because of its high NPV, identifying patients who will have a negative work-up. Moreover, prostate cancer detection by means of biopsy could be improved if additional cores were obtained according MRI findings, as suggested in the literature (26). Finally, cmri score could identify patients at high risk for cancer. Many studies investigated MRI performances for the diagnosis of prostate cancer. Because mmri alone showed a wide variability of sensitivity and specificity (9), different techniques, such as MRS and DWI, were proposed (13,27). Combining mmri with MRS improves MRI specificity (28), particularly when using an integrated data analysis (29). A recent metaanalysis on combined mmri-mrs (27) reported that this test has an overall sensitivity and specificity of 68% and 85%, respectively, with higher sensitivity (74%), but lower specificity (78%), in high risk patients. The MRS performance was correlated to tumor size and grade, with a better detection rate in higher grade cancer (30,31). Combined mmri-mrs was included in clinical models to predict insignificant cancer (14), but its role is still debated (12,27), with some studies, such as the prospective multiinstitutional study performed by the American College of Radiology Imaging Network (ACRIN), reporting less encouraging results (32). However, significant differences exist between our study and the ACRIN trial. The latter included only patients who underwent surgery after prostate biopsy, with a reported PSA values range of ng/ml and a median Gleason score of 7. Moreover, 39 patients enrolled in the ACRIN trial had palpable disease at DRE. Therefore, the ACRIN study evaluated patients with tumors more advanced than our patients. This resulted in a better diagnostic performance for cancer detection of mmri alone and reduced the benefit of the combination of MRS with mmri. Combining DWI with mmri significantly improves sensitivity in cancer detection (15,33,34). Haider et al (33) found 81% sensitivity and 84% specificity, while Lim et al (15) reported 88% sensitivity and 88% specificity. Low ADC values were associated with a high Gleason score, and a better detection rate (35). In our study, the number of detected cancers (18%) was in line with the diagnostic rate reported in the literature for similar work-up protocols (2). Contrary to most of the studies present in literature (15,27,33,36), we evaluated a ng/ml PSA range population that represents a clinical challenge for the urologists because of the low, but not negligible, tumor prevalence. Our population characteristics probably affected cmri score specificity, which Table 3 Percentage of Both Negative Biopsies and Positive Biopsies With a Significant or Insignificant Gleason Score in Relation to mmri/mrs Score, mmri/dwi Score, and cmri Score Negative (%) Insignificant (%) Significant (%) mmri/mrs score <2 61 (68/11) 39 (5/13) 0 (0/12) ¼ 2 32 (35/111) 46 (6/13) 58 (7/12) >2 7 (8/111) 15 (2/13) 42 (5/12) mmri/dwi score <2 57 (63/111) 46 (6/13) 8 (1/12) ¼ 2 18 (20/111) 46 (6/13) 17 (2/12) >2 25 (28/111) 8 (1/13) 75 (9/12) cmri score <2 51 (56/111) 31 (4/13) 0 (0/12) ¼ 2 22 (24/111) 46 (6/13) 25 (3/12) >2 28 (31/111) 23 (3/13) 75 (9/12)

6 6 Petrillo et al. resulted lower than that reported in the main studies present in literature, despite a comparable sensitivity and PPV (15,27,33,36). Correlating our results with those reported by similar studies performed in a population with PSA between 4 and 10 ng/ml is not reliable, because to our knowledge there are few reports on this topic (37 42) presenting different inclusion criteria and study design (prospective/retrospective), different demographic characteristics (particularly on cancer incidence), different MRI multiparametric techniques (including or not DWI, MRS, dynamic MRI), and different analysis of data (per region-basis, per halfgland basis, per patient basis). Moreover, many of these studies do not report Gleason scores derived from biopsy. In the study of Girometti et al (37), which presented a similar study design and the same patients ethnicity of our study, a comparable cancer incidence was reported (18% versus 19%). However, comparison of results is not reliable, because they reported data from every performed technique only on a per-region analysis, while a per-patient analysis was performed only on the overall MRI exam, including dynamic technique. In the study by Tamada et al (38) and by Kubota et al (39) prostate cancer incidence was very different than in our study, respectively, 70% and 33.5%. In comparison to the study of Tamada et al (38), in our study, a similar DWI sensitivity on a per-patient basis was obtained (64% versus 69%), with a lower specificity and PPV, but a higher NPV (85% versus 54%). Studies on MRS in patients with PSA between 4 and 10 ng/ml were performed by Kumar et al, who reported interesting results in patients undergoing TRUS biopsy on suspicious voxels at spectroscopy (40) and in patients with a negative biopsy (41). On the contrary, a recent study by Vargas et al (42) on MRS compared with prostatectomy in clinically lowrisk prostate cancer reported a low sensitivity. Therefore, further studies on these patients using standardized study designs and inclusion criteria are necessary to identify possible underlying bias and to derive robust results on this topic. A further investigation on a larger population is needed to confirm our results, probably including dynamic imaging to improve MRI results (36,43). In our study, statistical superiority of cmri score in respect to other scores was not demonstrated, because a higher number of patients with cancer would have been necessary to obtain a statistical power 80% (44). Moreover, MRI was compared with TRUS biopsy. Even if TRUS biopsy is reported to have a low diagnostic accuracy, it remains the only current diagnostic technique for the diagnosis of prostate cancer. The purpose of our study was to evaluate the diagnostic performance of the scoring system for patients work-up, where TRUS biopsy is the gold standard. When surgery is used as standard reference in a study, a large selection bias is introduced because many patients can be selected for nonsurgical treatment or active surveillance (32), as occurred in our population. According to the literature, we adopted a consensus evaluation method for scoring MRI findings to have more reliability in image interpretation (12). We evaluated the scoring system on a per-patient basis because a reasonably accurate evaluation on a per-site basis was not achievable. As reported (26), during needle biopsy, the path does not usually correspond to any MRI plane. Moreover, biopsy tumor localization is affected by cores classification according to the needle entry site, without considering the real needle path (45). Furthermore, in our study, the urologist was blinded to the MRI findings during the prostatic biopsy. Further investigation should be performed to obtain reslicing and registration of MR images to reliably correlate MRI findings to TRUS biopsy. A few study limitations need to be recognized. A limit is the lack of follow-up data, because a 10 20% cancer detection rate is reported at repeat biopsy (2). However, as already described in the literature (18), patients with PSA < 10 ng/ml undergo repeat biopsy a year after the first negative biopsy. Therefore, we compared mpmri with first work-up biopsy to evaluate if MRI can better differentiate between patients at low or high risk for cancer in this preliminary workup phase. The cmri score significantly correlated with the biopsy Gleason score even if biopsy can downgrade the tumor in up to 27% of cases because of an incomplete tumor sampling, according to the literature (46). In this study, we considered only a morphological evaluation of DWI/ADC map to derive the score from DWI and to combine it with scores from the other MRI techniques. Further studies could be performed to evaluate quantitative functional parameters from DWI data. In conclusion, according to our preliminary experience, the proposed scoring system allowed to combine mmri, MRS, and DWI datasets. The cmri score was superior in term of sensitivity and NPV, and significantly correlated to biopsy Gleason grade. Morphological and functional MRI using cmri score could have a role in patients with a PSA 10 ng/ml, identifying those patients who will have a negative work-up and those patients at high risk for a significant Gleason score cancer of the prostate. ACKNOWLEDGMENTS We thank Maura C. Tracey, BSN, RN, for her thorough editing of this article. Drs. Petrillo and Fusco have contributed equally at manuscript. REFERENCES 1. Cooperberg MR, Lubeck DP, Meng MV, Mehta SS, Carroll PR. The changing face of low-risk prostate cancer: trends in clinical presentation and primary management. J Clin Oncol 2004;22: Djavan B, Remzi M, Schulman CC, Marberger M, Zlotta AR. Repeat prostate biopsy: who, how and when? A review. Eur Urol 2002;42: Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate cancer among men with a prostate specific antigen level 4.0 ng per millilitre. N Engl J Med 2004;350: Draisma G, Boer R, Otto SJ, et al. Lead times and over detection due to prostate-specific antigen screening: estimates from the

7 Multiparametric MRI for Prostate Cancer Detection 7 European Randomized Study of Screening for Prostate Cancer. J Natl Cancer Inst 2003;95: Walsh PC, DeWeese TL, Eisenberger MA. Clinical practice. Localized prostate cancer. N Engl J Med 2007;357: Zackrisson B, Aus G, Bergdahl S, et al. The risk of finding focal cancer (less than 3 mm) remains high on re-biopsy of patients with persistently increased prostate specific antigen but the clinical significance is questionable. J Urol 2004;17: Klotz L. Active surveillance for prostate cancer: patient selection and management. Curr Oncol 2010;17: Yu KK, Hricak H. Imaging prostate cancer. Radiol Clin North Am 2000;38: Choi YJ, Kim JK, Kim N, Kim KW, Choi EK, Cho KS. Functional MR imaging of prostate cancer. Radiographics 2007;27: Sanders A, Buchan N. Infection-related hospital admissions after transrectal biopsy of the prostate. ANZ J Surg 2013;83: Utrera NM, Alvarez MB, Polo JM, Sanchez AT, Martinez JP, Gonzalez RD. Infectious complications after transrectal ultrasoundguided prostatic biopsy. Analysis of our experience. Arch Esp Urol 2011;64: Dickinson L, Ahmed HU, Allen C. Magnetic resonance imaging for the detection, localisation, and characterisation of prostate cancer: recommendations from a European Consensus Meeting. Eur Urol 2011;59: Kim JK, Jang YJ, Cho G. Multidisciplinary functional MR imaging for prostate cancer. Korean J Radiol 2009;10: Shukla-Dave A, Hricak H, Kattan MW, et al. The utility of magnetic resonance imaging and spectroscopy for predicting insignificant prostate cancer: an initial analysis. BJU 2007;99: Lim HK, Kim JK, Kim KA, Cho KS. Prostate cancer: apparent diffusion coefficient map with T2-weighted images for detection-a multireader study. Radiology 2009;250: Akin O, Sala E, Moskowitz CS, et al. Transition zone prostate cancers: features, detection, localization, and staging at endorectal MR imaging. Radiology 2006;239: 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 ( cm3) spatial resolution. Radiology 1996;198: Perdona S, Di Lorenzo G, Autorino R, et al. Combined magnetic resonance spectroscopy and dynamic contrast-enhanced imaging for prostate cancer detection. Urol Oncol 2011 [Epub ahead of print]. 19. Obuchowski NA. ROC analysis. AJR Am J Roentgenol 2005;184: DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44: Schisterman EF, Perkins NJ, Liu A, Bondell H. Optimal cut-point and its corresponding Youden Index to discriminate individuals using pooled blood samples. Epidemiology 2005;16: Joseph L, Reinhold C. Statistical inference for proportions. AJR Am J Roentgenol 2005;184: Hawass NE. Comparing the sensitivities and specificities of two diagnostic procedures performed on the same group of patients. Br J Radiol 1997;70: Dendukuri N, Reinhold C. Correlation and regression. AJR Am J Roentgenol 2005;185: National Cancer Institute (NCI). Available at gov/cancertopics/alphalist. Accessed February 15, Cirillo S, Petracchini M, Della Monica P, et al. Value of endorectal MRI and MRS in patients with elevated prostate-specific antigen levels and previous negative biopsies to localize peripheral zone tumours. Clin Radiol 2008;63: Umbehr M, Bachmann LM, Held U. Combined magnetic resonance imaging and magnetic resonance spectroscopy imaging in the diagnosis of prostate cancer: a systematic review and metaanalysis. Eur Urol 2009;55: Vilanova JC, Barcelo J. Prostate cancer detection: magnetic resonance (MR) spectroscopy imaging. Abdom Imaging 2007;32: Westphalen AC, Coakley FV, Qayyum A, et al. Peripheral zone prostate cancer: accuracy of different interpretative approaches with MR and MR spectroscopic imaging. Radiology 2007;246: Carroll PR, Coakley FV, Kurhanewicz J. Magnetic resonance imaging and spectroscopy of prostate cancer. Rev Urol 2006;8: Zakian KL, Sircar K, Hricak H, et al. Correlation of proton MR spectroscopic imaging with Gleason score based on step-section pathologic analysis after radical prostatectomy. Radiology 2005; 234: Weinreb JC, Blume JD, Coakley FV, et al. Prostate cancer: sextant localization at MR imaging and MR spectroscopic imaging before prostatectomy results of ACRIN prospective multiinstitutional clinicopathologic study. Radiology 2009;251: Haider MA, van der Kwast TH, Tanguay J, et al. Combined T2- weighted and diffusion weighted MRI for localization of prostate cancer. AJR Am J Roentgenol 2007;189: Tanimoto A, Nakashima J, Kohno H, Shinmoto H, Kuribayashi S. Prostate cancer screening: the clinical value of diffusion weighted imaging and dynamic MR imaging in combination with T2- weighted imaging. J Magn Reson Imaging 2007;25: Woodfield CA, Tung GA, Grand DJ, Pezzullo JA, Machan JT, Renzulli JF Jr. Diffusion-weighted MRI of peripheral zone prostate cancer: comparison of tumor apparent diffusion coefficient with Gleason score and percentage of tumor on core biopsy. AJR Am J Roentgenol 2010;194: Franiel T, Stephan C, Erbersdobler A, et al. Areas suspicious for prostate cancer: MR-guided biopsy in patients with at least one transrectal US-guided biopsy with a negative findingmultiparametric MR imaging for detection and biopsy planning. Radiology 2011;259: Girometti R, Bazzocchi M, Como G, et al. Negative predictive value for cancer in patients with gray-zone PSA level and prior negative biopsy: preliminary results with multiparametric 3.0 Tesla MR. J Magn Reson Imaging 2012;36: Tamada T, Sone T, Higashi H, et al. Prostate cancer detection in patients with total serum prostate-specific antigen levels of 4 10 ng/ml: diagnostic efficacy of diffusion-weighted imaging, dynamic contrast enhanced MRI, and T2-weighted imaging. AJR Am J Roentgenol 2011;197: Kubota Y. Kamei S, Nakano M, Ehara H, Deguchi T Tanaka O. The potential role of prebiopsy magnetic resonance imaging combined with prostate-specific antigen density in the detection of prostate cancer. Int J Urol 2008;15: Kumar V, Jagannathan NR, Kumar R, et al. Transrectal ultrasound-guided biopsy of prostate voxels identified as suspicious of malignancy on three-dimensional (1)H MR spectroscopic imaging in patients with abnormal digital rectal examination or raised prostate specific antigen level of 4 10 ng/ml.nmr Biomed 2007;20: Kumar R, Nayyar R, Kumar V, et al. Potential of magnetic resonance spectroscopic imaging in predicting absence of prostate cancer in men with serum prostate-specific antigen between 4 and 10 ng/ml: a follow-up study. Urology 2008;72: Vargas HA, Akin O, Shukla-Dave A, et al. Performance characteristics of MR imaging in the evaluation of clinically low-risk prostate cancer: a prospective study. Radiology 2012;265: Futterer JJ, Heijmink SW, Scheenen TW, et al. Prostate cancer localization with dynamic contrast enhanced MR imaging and proton MR spectroscopic imaging. Radiology 2006;241: Petrie A, Sabin C. Medical statistics at a glance. UK: Blackwell Science; p Schulte RT, Wood DP, Daignault S, Shah RB, Wei JT. Utility of extended pattern prostate biopsies for tumor localization: pathologic correlations following radical prostatectomy. Cancer 2008; 113: Berglund RK, Masterson TA, Vora KC, Eggener S, Eastham JA, Guillonneau BD. Pathological upgrading and upstaging with immediate repeat biopsy in patients eligible for active surveillance. J Urol 2008;180: