Genitourinary Imaging Original Research

Genitourinary Imaging Original Research Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved Park et al. ADC in Prostate Cancer Genitourinary Imaging Original Research Sung Yoon Park 1, Chan Kyo Kim 1 Byung Kwan Park 1 Ghee Young Kwon 3 Park SY, Kim CK, Park BK, Kwon GY Keywords: apparent diffusion coefficient (ADC), comparative study, diffusion-weighted MRI (DWI), prostate DOI:1.1/AJR.13.111 Received August, 13; accepted after revision January 1, 1. 1 Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, Ilwon-dong, Gangnam-gu, Seoul, Korea 13-71. Address correspondence to C. K. Kim (chankyokim@skku.edu). Present address: Department of Radiology and Research Institute of Radiological Science, Severance Hospital, Yonsei University College of Medicine, Yonsei-ro, Seodaemun-gu, Seoul, Korea 1-7. 3 Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea. WEB This is a web exclusive article. AJR 1; 3:W7 W9 31 3X/1/33 W7 American Roentgen Ray Society Comparison of Apparent Diffusion Coefficient Calculation Between Two-Point and Multipoint b Value Analyses in Prostate Cancer and Benign Prostate Tissue at 3 T: Preliminary Experience OBJECTIVE. The purpose of this study was to prospectively evaluate the reliability and variability of apparent diffusion coefficient (ADC) calculations between two-point and multipoint b value analyses in prostate cancer and benign prostate tissue. SUBJECTS AND METHODS. Forty-eight consecutive patients with suspected prostate cancer underwent diffusion-weighted MRI (DWI) at 3 T followed by surgery. DWI was examined under different b values. ADC maps were generated by two different methods: two-point b values ( and 1 s/mm ) and multipoint b values (, 1, 3, 7, and 1 s/mm ). Two independent readers measured ADC in the cancers, benign peripheral zone and transition zone, and obturator internus muscle. Statistical analyses were performed using the intraclass correlation coefficient (ICC), correlation of variation (CV), Bland-Altman test, and paired Student t test. RESULTS. The intermethod ADC calculation revealed excellent reliability for all tissues in both readers: cancer (ICC =.979.91), transition zone (.99.993), peripheral zone (.99.99), and obturator internus muscle (.97.97). In both readers, the variability of the intermethod ADC calculation was.9 3.9% CV in cancer, 1.1 1.% CV in the transition zone, 1.3 1.9% CV in the peripheral zone, and..% CV in the obturator internus muscle. For interreader variability, the CVs of ADC calculation for two-point versus multipoint b value analyses in all tissues were 7.1 9.% versus 7.1 9.1%. CONCLUSION. For estimating ADC values on 3-T DWI of the prostate, two-point b value analysis seems to present excellent correlation with multipoint b value analysis, with little error in accuracy. R ecent advances in MRI enable clinical applications of functional imaging techniques in prostate cancer, which could offer improved diagnostic performance in detection and localization, staging, posttreatment follow-up, biopsy guidance, and assessment of tumor aggressiveness [1, ]. Of these functional imaging techniques, diffusionweighted MRI (DWI) is a novel noninvasive imaging technique that has potential in evaluating prostate cancer, from the detection of cancer to the prediction of pathologic grading or even early prediction of prognosis [3 ]. However, in many recent studies of prostate DWI, a variety of imaging parameters have been used, which has led to conflicting results for cancers and benign prostatic tissues in quantitative analyses []. DWI depends on the microscopic mobility of water. Water diffusion values in tissue can be quantified by calculating the appar- ent diffusion coefficient (ADC); the degree of restriction of water diffusion in biologic tissue is inversely related to the tissue cellularity and the integrity of cell membranes [, 7]. For the basic ADC calculation on DWI, at least two b values are required (for any body region or organ being imaged), and this technique has been performed using most clinical systems. The application of a greater number of b values will improve the accuracy of the calculated ADC [7]. The disadvantage of using multiple b values is an associated increase in scanning time and decrease in signal-to-noise ratio (SNR). Thus, it may not be practical in many clinical settings. A few investigations to optimize DWI protocols for clinical use in the prostate have been reported that have focused on the optimization of the b value to improve the detection of prostate cancer [, 9]. No standardized method for ADC calculation in the prostate is currently available, although AJR:3, September 1 W7

Park et al. Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved the European Society of Urogenital Radiology (ESUR) prostate MR guidelines 1 recommended the use of multipoint b value analysis in prostate DWI; however, this recommendation was not sufficiently validated [1]. To our knowledge, no studies have been reported that investigate the ADC difference between two-point and multipoint b value analyses in the prostate at 3-T MRI. If the ADC values derived from two-point b value analysis could be closely correlated with those derived from multipoint b value analysis, the former may be sufficient for clinical usage in evaluating prostate cancer. Therefore, the purpose of our study was to prospectively evaluate the reliability and variability of ADC calculations between two-point and multipoint b value analyses in prostate cancer and benign prostate tissue. Subjects and Methods Patients The study protocol was approved by our institutional review board and informed consent was obtained from each participant. Between September 11 and March 1, 7 consecutive patients who were referred for MRI of the prostate gland and met the following inclusion criteria were prospectively enrolled in our study (Fig. 1): clinically suspected prostate cancer, such as positive findings on digital rectal examination or elevated serum prostate-specific antigen (PSA) level (. ng/ml) and prostate DWI at 3 T with a wide range of b values (, 1, 3, 7, and 1 s/mm ) before biopsy or surgery. Among the 7 patients who underwent DWI, we excluded nine patients for the following reasons: poor image quality due to the motion artifact (n = ), no surgery (n = ), and radiation and hormonal therapy before the MRI examination (n = 1). Thus, a total of patients (mean age [± SD],.7 ±. years; range, 3 years) with surgically confirmed prostate cancers were finally included in our study for image analysis. Fig. 1 Flowchart shows selection of study group. DRE = digital rectal examination, PSA = prostatespecific antigen, DWI = diffusion-weighted MRI, ADC = apparent diffusion coefficient. MRI Protocols All patients underwent imaging 3 weeks after transrectal ultrasound (TRUS) guided biopsy and before radical prostatectomy using a 3-T MRI scanner (Intera Achieva 3TX, Philips Healthcare) equipped with a six-channel phasedarray coil (Cardiac Sense, Philips Healthcare). All patients underwent both conventional T-weighted imaging and DWI. Before MRI, mg of butyl scopolamine (Buscopan, Boehringer Ingelheim) was injected intramuscularly to suppress bowel peristalsis; no bowel preparation was performed. T-weighted turbo spin-echo images were acquired in three orthogonal planes (axial, sagittal, and coronal). T-weighted imaging parameters were as follows: TR range/te range, 33 3/ 1; slice thickness, 3 mm; interslice gap, 1 mm; 31 matrix; FOV, cm; number of signals acquired, 3; sensitivity encoding (SENSE) factor, ; number of slices, 1; and acquisition time of each plane, minutes seconds. Axial DWI was performed using the single-shot echo-planar imaging technique with the following imaging parameters: TR/TE, /; slice thickness, 3 mm; interslice gap, 1 mm; matrix, 1 11; FOV, cm; SENSE factor, ; number of signals acquired, ; number of slices, 1; and acquisition time, 7 minutes 31 seconds. The phase-encoding gradient moved from left to right to minimize motion artifacts on the prostate image. Diffusionencoding gradients were applied at b values of, 1, 3, 7, and 1 s/mm along the three orthogonal main directions of the motion probing gradients. Two ADC maps were generated by two different methods using a monoexponential model: two-point b values of and 1 s/mm (hereafter referred to as ADC 1 ) and multipoint b values of, 1, 3, 7, and 1 s/mm (hereafter referred to as ADC 1 3 7 1 ). In the multipoint b value method, ADC values were calculated using the least-square exponential fitting of all b value data within the defined range of b values. The equation used to calculate the ADC value was as follows: ADC = ln (S / S ) / b, where S is the signal intensity of no diffusion gradients and b is the b value. After data acquisition, all images were transferred to the workstation for analysis with manufacturer-supplied software. All MR images were also archived using a PACS (PathSpeed Workstation, GE Healthcare). Patients who underwent prostate MRI and met the following inclusion criteria (n = 7) Positive findings on DRE or elevated serum PSA (. ng/ml) Prostate DWI at 3 T with different b values (, 1, 3, 7, and 1 s/mm ) prior to biopsy or surgery Exclusion (n = 9) Poor image quality due to motion artifacts (n = ) No surgical confirmation (n = ) History of previous radiation and hormone therapies for prostate cancer (n = 1) Patients with both ADC with b values of and 1 s/mm and ADC with b values of, 1, 3, 7, and 1 s/mm maps and after surgical confirmation (n = ) Histopathologic Analysis All surgical specimens were marked using ink and were fixed overnight in 1% buffered formalin. Axial step sections were acquired at 3-mm intervals in a plane vertical to the prostatic urethra. All slides prepared from the tissue slices were reviewed by an experienced pathologist without knowledge of the MRI findings. The tumor size, volume, location, and Gleason score were recorded on a standardized diagram of a histopathologic tumor map. The standardized diagram was divided into the base, midgland, and apex in each lobe. Image Analysis Two readers with 7 years and 1 year of experience in prostate MRI who were aware of the pathologic results assessed all MR images independently. The ADC calculations were performed using ROI in the cancer, benign peripheral zone, and benign transition zone. As the standard of reference, ADC calculation was also performed in the left obturator internus muscle. The typical MRI findings suggesting cancer foci were as follows: a focal area of low signal intensity relative to the high-signal-intensity background of the benign prostatic tissues on T-weighted images and a focal hyperintense area at b = 1 s/mm of DWI with a low ADC value at the corresponding area compared with that of the surrounding benign prostatic tissue on the ADC maps [11 13]. In accordance with the MRI criteria of prostate cancer, an ellipsoid ROI was drawn manually within a cancerous area on ADC 1 3 7 1 first, and the ROI of ADC 1 3 7 1 was subsequently copied onto ADC 1 at the same axial plane, using software (ViewForum Workstation R.1, Philips Healthcare). For each benign peripheral zone and transition zone and for obturator internus muscle tissues, the same processes of ADC measurements were performed. Cancer localization was performed using the six sectors (base, midgland, and apex of the right and left prostate gland) with the correlation W AJR:3, September 1

ADC in Prostate Cancer TABLE 1: Results of Apparent Diffusion Coefficient (ADC) Values Between Two-Point and Multipoint b Value Analyses for Both Readers ADC Value ( 1 3 mm /s) Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved Parameter Cancer between typical MRI findings and the surgical specimen. When the MRI-suspected cancer and surgically proven cancer were regarded as located in the same sector, ADC calculation in the cancer was performed. When two or more cancer foci were visible, the ADC value was measured only in the largest lesion after image-pathology correlation. For all patients in our study with surgically proven cancers, at least one cancer focus was visible at the corresponding sector of the ADC map and was analyzed on both the ADC 1 and ADC 1 3 7 1 maps. The ADC calculations in the benign peripheral zone and transition zone were carried out on both sides of the area Right Transition Left Transition Right Peripheral outside the radiologically suspected and surgically proven cancerous area. For all patients, the baseline PSA level, Gleason score, and cancer size on the surgical specimen were investigated. Statistical Analysis Statistical analysis was performed using PASW statistical software, version 1. (SPSS) and Med- Calc, version 1.3. (MedCalc Software). For assessing the difference in the mean ADC values of each tissue (cancer, peripheral zone, transition zone, and obturator internus muscle between ADC 1 and ADC 1 3 7 1 ), the paired Student t test was applied. Left Peripheral Obturator Internus Muscle Reader 1 Two-point. ±.17 1.37 ±.19 1.33 ±.1 1. ±.7 1. ±.19. ±.1 Multipoint.1 ±.17 1.39 ±. 1.3 ±.1 1.9 ±. 1. ±.19.1 ±.1 p a. <.1 <.1.99 <.1.3 Reader Two-point.7 ±.17 1.33 ±.1 1.3 ±.1 1. ±. 1.7 ±.1.1 ±.1 Multipoint.7 ±.1 1.3 ±.1 1.3 ±.1 1. ±.3 1.71 ±..3 ±.1 p a. <.1 <.1.1 <.1 <.1 Note Data are mean ± SD. a Comparison between two-point and multipoint b value analyses. TABLE : Results of Reliability and Variability in Apparent Diffusion Coefficient Calculation Between Two-Point and Multipoint b Value Analyses in Both Readers Parameter Intermethod ICC Intermethod CV (%) Intermethod Difference (%) Reader 1 Cancer.979 (.9.9) 3.9.7 (. to 1.9) Right transition zone.99 (.9.99) 1. 1.1 (.9 1.1) Left transition zone.991 (.9.99) 1.1. (.39 1.3) Right peripheral zone.99 (.99.99) 1.. (.1 to.9) Left peripheral zone.991 (.93.99) 1.7.9 (. 1.) Obturator internus muscle.97 (.9.91). 1.3 (.33.3) Reader Cancer.91 (.97.99).9. (. to.) Right transition zone.993 (.97.99) 1.37 1. (.9 1.7) Left transition zone.99 (.91.99) 1. 1. (.7 1.7) Right peripheral zone.99 (.9.99) 1.9.7 (. 1.3) Left peripheral zone.99 (.9.99) 1.3. (.37 1.17) Obturator internus muscle.97 (.9.9). 1. (..) Note Data in parentheses are 9% CI. ICC = intraclass correlation coefficient, CV = correlation of variation. For evaluating the intermethod reliability and variability of ADC values between ADC 1 and ADC 1 3 7 1, the intraclass correlation coefficient (ICC) and correlation of variation (CV), respectively, were used. The ICC values of.. indicated a poor reliability of ADC calculation,.1. indicated fair,.1. indicated moderate,.1. indicated good, and.1 1. indicated excellent reliability. The definition of CV was as follows: CV (%) = 1 (SD / mean). The percentage differences between ADC 1 and ADC 1 3 7 1 were also plotted using a Bland Altman test for both readers. For assessing the interreader variability between two independent readers, the CV was applied. In addition, a relationship between cancer ADC value and Gleason score was investigated on each ADC 1 and ADC 1 3 7 1 map using Spearman rank correlation, Two-tailed tests were used to calculate all p values, and p values <. were considered statistically significant. Results The mean baseline PSA of patients with surgically proven cancers was 3.3 ng/ ml (range,. 71.3 ng/ml). On histopathologic examination, the mean size of the cancer was 3.3 cm (range,.. cm). The cancers had Gleason scores of (n = 9), 7 (n = ), (n = 9), and 9 (n = ). Quantitative Analysis Using Two Different Methods Table 1 presents the results of ADC values in all tissues between two-point and multipoint b value analyses for both readers. For the two methods for both readers, the mean ADC values of the cancers were significantly lower than those of the benign peripheral AJR:3, September 1 W9

Park et al. Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved zone and transition zone (p <.1). For the ADC calculation between the two methods, the mean ADC values of the cancers for both readers did not significantly differ (reader 1, p =.; reader, p =.); the mean ADC values on both sides of the transition zone and peripheral zone and obturator internus muscle significantly differed for both readers (p <.) except for the right peripheral zone in reader 1 (p =.99). However, the mean difference in ADC values between the two methods ranged from.1 to. 1 3 mm /s in all tissues. A TABLE 3: Results of Interreader Variability of Apparent Diffusion Coefficient Calculation in Two-Point and Multipoint b Value Analyses Parameter Cancer Right Transition Interreader CV (%) Left Transition Right Peripheral C B Fig. 3-year-old man with prostate-specific antigen level of. ng/ml and Gleason score of 7. A, On H and E stained section, cancer area (black dashed line) is noted at right peripheral zone. B and C, On apparent diffusion coefficient (ADC) map with b = and 1 s/mm (ADC 1 ) (B) and ADC map with b =, 1, 3, 7, and 1 s/mm (ADC 1 3 7 1 ) (C), cancer (arrow) manifests as focal area of low ADC value at corresponding site of pathologic map. ADC values of cancer on ADC 1 and ADC 1 3 7 1 maps are.3 and.9 1 3 mm /s, respectively. Left Peripheral Obturator Internus Muscle Two-point. 7.91 7.1.3.7 9. Multipoint.3 7.7 7.1.3.3 9.1 Note CV = correlation of variation. Table presents the results of reliability and variability in ADC calculation between two-point and multipoint b value analyses for both readers. Between the two methods, the reliability of ADC calculation was excellent for all tissues: cancer (ICC,.979.91), transition zone (.99.993), peripheral zone (.99.99), and obturator internus muscle (.97.97) for both readers (Fig. ). The intermethod variability of ADC calculation ranged from.9% to 3.9% CV in the cancer, 1.1% to 1.% CV in the transition zone, 1.3% to 1.9% CV in the peripheral zone, and.% to.% CV in the obturator internus muscle for both readers. For the agreement between the two methods, the Bland-Altman plot showed mean ADC difference of.7% for the cancer, 1.1% for the right transition zone,.% for the left transition zone,.% for the right peripheral zone,.9% for the left peripheral zone, and 1.3% for the obturator internus muscle for reader 1 (Fig. 3) and.% for the cancer, 1.% for the right transition zone, 1.% for the left transition zone,.7% for the right peripheral zone,.% for the left peripheral zone, and 1.% for the obturator internus muscle for reader (Fig. ). Interreader Variability The variability of ADC calculation between the two readers in the two-point and multipoint b value analyses is summarized in Table 3. In the two-point method, the CVs of all tissues ranged from 7.1% to 9.%, and in the multipoint method, the CVs of all tissues ranged from 7.1% to 9.1%. W9 AJR:3, September 1

ADC in Prostate Cancer Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved 1 1 9..7 7. 1 1... 1. 1. 1. 1. Average of Cancer ADC A 3 3. 1. 1. 3 1. 1.1 1. 1.3 1. 1. 1. 1.7 Average of Left Transition ADC C.1.9. 1. 1. 1. 1. 1... Average of Left Peripheral ADC 7. 3 1 1.1 1 3.3. 1. 1. 1. 1. 1.. Average of Right Transition ADC B 3.9..9 1. 1... Average of Right Peripheral ADC D 1. 1.3...7..9 1. 1.1 Average of Obturator Internus Muscle ADC E Fig. 3 Agreement of apparent diffusion coefficient (ADC) calculation. A F, Bland-Altman plots show agreement for cancers (A), both transition zones (B and C), both peripheral zones (D and E), and obturator internus muscle (F) between analyses using two b values and multiple b values for reader 1; x-axes show ADCs measured at ADC map, and y-axes show differences between two b value and multiple b value analyses as percentage of mean. Solid lines indicate mean differences. Dashed lines indicate 9% limits of agreement. F AJR:3, September 1 W91

Park et al. 1.1 Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved 1 1 9.. 7.... 1. 1. 1. Average of Cancer ADC A. 1..1. 1. 1. 1. 1. 1.. Average of Left Transition ADC C 3.. 1.9 1. 1. 1. 1.... Average of Left Peripheral ADC 3 1 1 1 1. 1. 1. 1. 1. 1. 1... Average of Right Transition ADC B.1.7.7 1. 1. 1. 1.... Average of Right Peripheral ADC D. 1..9....7..9 1. 1.1 Average of Obturator Internus Muscle ADC E Fig. Agreement of apparent diffusion coefficient (ADC) calculation. A F, Bland-Altman plots show agreement for cancers (A), both transition zones (B and C), both peripheral zones (D and E), and obturator internus muscle (F) between analyses using two b values and multiple b values for reader ; x-axes show ADCs measured at ADC map, and y-axes show differences between two b value and multiple b value analyses as percentage of mean. Solid lines indicate mean differences. Dashed lines indicate 9% limits of agreement. F W9 AJR:3, September 1

ADC in Prostate Cancer Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved Correlation Between Cancer ADC and Gleason Score Cancer ADC values were negatively correlated with Gleason score at both ADC 1 (ρ =.33, p =.) and ADC 1 3 7 1 maps (ρ =.3, p =.1). Discussion At 3-T DWI, our study compared twopoint and multipoint b value analyses in patients with prostate cancer in terms of intermethod reliability and variability and interreader variability for the calculation of ADC values in the cancer, normal prostate tissues, and normal pelvic muscle. In the results of our study, the intermethod reliability for ADC calculation between the two methods was excellent in all tissues, including prostate cancers, normal prostate tissues, and normal muscle. The CVs of the ADC values were less than % in all tissues, including the cancers, normal prostate tissues, and obturator internus muscle. Moreover, the mean percentage differences of ADC values between the two methods in all tissues were less than 1.3% for reader 1 and less than 1.% for reader. These findings suggest that two-point b value analysis for ADC calculation at 3-T DWI that allows reduction in acquisition time may be an alternative to multipoint b value analysis in daily clinical practice, with little decrease in accuracy. In our study, the mean ADC values of the two-point method significantly differed from those of the multipoint method for all tissues except for cancer for both readers and right peripheral zone for reader 1. We assume that the reason for this result may be the different number of b values used for the ADC calculation (two b values for ADC 1 and five b values for ADC 1 3 7 1 ). When a larger number of b value data was used for ADC calculation, the variability of ADCs that resulted from the image noise on DWI at each b value may have been cancelled out, whereas the variability in signal intensity values at the two-point b value is directly reflected in the ADC value calculated using the two-point b value method. In our study, the mean difference was minimal (.1. 1 3 s/mm ), which might be imperceptible for diagnosing prostate cancer from benign tissue between the two methods in clinical practice. Recently, studies have shown that ADC of prostate cancer increases after radiation or hormone therapy. Thus, cancer ADC may have potential as a biomarker for assessing early therapeutic response [1 1]. Even with only a minimal difference in ADC between different b value methods in our study, we recommend that the calculation of cancer ADC after radiation or hormonal therapy should be performed under the same b value method because the minimal ADC difference may affect the interpretation for therapeutic response. For determining whether ADC calculations using different b value methods are interchangeable for evaluating therapeutic response, more evidence is needed. With 3-T MRI, which can offer increased SNR compared with 1.-T MRI, the application of a high b value, such as 1 s/ mm, could provide more sensitive and accurate information of the prostate. However, DWI at 3 T is still affected by magnetic susceptibility effects, resulting in spatial distortion and signal loss, which are amplified as the b value increases. Controversy remains regarding the choice of optimal b values for evaluating prostate cancer [, 19 ]. The ADC represents the perfusion and diffusion effects. The degree of perfusion bias in ADC calculation increases with the volume fraction of flow and decreases the b value range. Thus, the minimum b value threshold to suppress perfusion effects will depend on the vascular properties of tissues; however, for most applications, a lower b value of 1 1 s/mm is probably adequate. Recently, the ESUR recommended three b values of, 1, and 1 s/mm as the minimal requirement for evaluating the prostate on DWI [1]; the choice of three b values as the ESUR recommended enables calculation of perfusion-insensitive ADC values in the prostate. Likewise, our study used multiple b values of, 1, 3, 7, and 1 s/mm. However, further studies using an ultra-high b value of > 1 s/mm may be needed. In our study, cancer ADC values were negatively related to Gleason score on two different types of ADC maps, and the correlation degree was similar between two different types of ADC maps (ADC 1, ρ =.33; ADC 1 3 7 1, ρ =.3). These findings were consistent with other studies that have shown the relationship between cancer ADC and aggressiveness [3,, 3]. Our study had several limitations. First, our study calculated ADCs using the monoexponential model even though the biexponential model is able to better explain the characterization of prostate tissue, including cancer and benign tissues [ ]. However, despite its potential benefits, the biexponential model has not been commonly used for DWI analyses in clinical practice because the software for biexponential model based analysis is not widely available and the analysis requires a wide range of b values for reliable data fitting, especially at a range of lower b values. Furthermore, the biexponential analysis of DWI requires increased scanning time. Thus, in our study, we calculated ADCs using the monoexponential model, which is widely available and is provided by most commercial MRI vendors. Second, our study did not evaluate reproducibility between two-point and multipoint b value analyses. To be used as a clinical tool for the diagnosis and assessment of therapeutic response to treatment, the calculation of ADC value should be reproducible. Further validated studies are needed. Third, in our study, the number of patients was relatively small, which might result in small differences in the ADC values for prostate tissues. Finally, our study did not evaluate the difference of ADC calculation in different MRI scanners and selection of b values on the ADC. In conclusion, our preliminary results suggest that for estimating ADC values at prostate 3-T DWI, two-point b value analysis seems to show excellent correlation with multipoint b value analysis, with little decrease in accuracy. Our study indicates that ADC calculation in the prostate may be performed adequately using a two-point b value analysis that enables reduction in scanning time. References 1. Kim CK, Park BK, Han JJ, Kang TW, Lee HM. Diffusion-weighted imaging of the prostate at 3 T for differentiation of malignant and benign tissue in transition and peripheral zones: preliminary results. J Comput Assist Tomogr 7; 31:9. Hoeks CM, Barentsz JO, Hambrock T, et al. Prostate cancer: multiparametric MR imaging for detection, localization, and staging. Radiology 11; 1: 3. Vargas HA, Akin O, Franiel T, et al. Diffusionweighted endorectal MR imaging at 3 T for prostate cancer: tumor detection and assessment of aggressiveness. Radiology 11; 9:77 7. Park SY, Kim CK, Park BK, Lee HM, Lee KS. Prediction of biochemical recurrence following radical prostatectomy in men with prostate cancer by diffusion-weighted magnetic resonance imaging: initial results. Eur Radiol 11; 1:1111 111. Kim CK, Park BK, Kim B. Diffusion-weighted MRI at 3 T for the evaluation of prostate cancer. AJR 1; 19:11 19. Hambrock T, Somford DM, Huisman HJ, et al. Relationship between apparent diffusion coeffi- AJR:3, September 1 W93

Park et al. cients at 3.-T MR imaging and Gleason grade in weighted and dynamic contrast-enhanced MR tate: effect of signal-to-noise ratio. Magn Reson peripheral zone prostate cancer. Radiology 11; imaging. Radiology 1; 7:71 73 Med Sci ; 7:93 99 9:3 1 1. Nemoto K, Tateishi T, Ishidate T. Changes in dif- 1. Kitajima K, Takahashi S, Ueno Y, et al. Clinical 7. Padhani AR, Liu G, Koh DM, et al. Diffusion- fusion-weighted images for visualizing prostate utility of apparent diffusion coefficient values ob- weighted magnetic resonance imaging as a cancer cancer during antiandrogen therapy: preliminary tained using high b-value when diagnosing pros- Downloaded from www.ajronline.org by 1.1.3.3 on /7/1 from IP address 1.1.3.3. Copyright ARRS. For personal use only; all rights reserved biomarker: consensus and recommendations. Neoplasia 9; 11:1 1. Metens T, Miranda D, Absil J, Matos C. What is the optimal b value in diffusion-weighted MR imaging to depict prostate cancer at 3T? Eur Radiol 1; :73 79 9. Katahira K, Takahara T, Kwee TC, et al. Ultrahigh-b-value diffusion-weighted MR imaging for the detection of prostate cancer: evaluation in 1 cases with histopathological correlation. Eur Radiol 11; 1:1 19 1. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 1. Eur Radiol 1; :7 77 11. Kim CK, Park BK, Lee HM, Kwon GY. Value of diffusion-weighted imaging for the prediction of prostate cancer location at 3T using a phased-array coil: preliminary results. Invest Radiol 7; : 7 1. Lim HK, Kim JK, Kim KA, Cho KS. Prostate cancer: apparent diffusion coefficient map with T-weighted images for detection a multireader study. Radiology 9; :1 11 13. Oto A, Kayhan A, Jiang Y, et al. Prostate cancer: differentiation of central gland cancer from benign prostatic hyperplasia by using diffusion- results. Urol Int 1; :1 1. Takayama Y, Kishimoto R, Hanaoka S, et al. ADC value and diffusion tensor imaging of prostate cancer: changes in carbon-ion radiotherapy. J Magn Reson Imaging ; 7:1331 133 1. Song I, Kim CK, Park BK, Park W. Assessment of response to radiotherapy for prostate cancer: value of diffusion-weighted MRI at 3 T. AJR 1; 19:[web]W77 W 17. Park SY, Kim CK, Park BK, et al. Early changes in apparent diffusion coefficient from diffusionweighted MR imaging during radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 1; 3:79 7 1. Foltz WD, Wu A, Chung P, et al. Changes in apparent diffusion coefficient and T relaxation during radiotherapy for prostate cancer. J Magn Reson Imaging 13; 37:99 91 19. Nam H, Huh SJ, Park W, et al. Prognostic significance of MRI-detected bladder muscle and/or serosal invasion in patients with cervical cancer treated with radiotherapy. Br J Radiol 1; 3: 73. Kitajima K, Kaji Y, Kuroda K, Sugimura K. High b-value diffusion-weighted imaging in normal and malignant peripheral zone tissue of the pros- tate cancer using 3 tesla MRI: comparison between ultra-high b-value ( s/mm ) and standard high b-value (1 s/mm ). J Magn Reson Imaging 1; 3:19. Ohgiya Y, Suyama J, Seino N, et al. Diagnostic accuracy of ultra-high-b-value 3.-T diffusionweighted MR imaging for detection of prostate cancer. Clin Imaging 1; 3: 31 3. Oto A, Yang C, Kayhan A, et al. Diffusionweighted and dynamic contrast-enhanced MRI of prostate cancer: correlation of quantitative MR parameters with Gleason score and tumor angiogenesis. AJR 11; 197:13 139. Riches SF, Hawtin K, Charles-Edwards EM, de Souza NM. Diffusion-weighted imaging of the prostate and rectal wall: comparison of biexponential and monoexponential modelled diffusion and associated perfusion coefficients. NMR Biomed 9; :31 3. Mulkern RV, Barnes AS, Haker SJ, et al. Biexponential characterization of prostate tissue water diffusion decay curves over an extended b-factor range. Magn Reson Imaging ; :3. Shinmoto H, Oshio K, Tanimoto A, et al. Biexponential apparent diffusion coefficients in prostate cancer. Magn Reson Imaging 9; 7:3 39 W9 AJR:3, September 1