Genitourinary Imaging Original Research

Transcription

1 Genitourinary Imaging Original Research Roy et al. Detection of Prostate Cancer Recurrence With Different Functional MRI Sequences Genitourinary Imaging Original Research Catherine Roy 1 Fatah Foudi 1 Jeanne Charton 1 Michel Jung 2 Hervé Lang 3 Christian Saussine 3 Didier Jacqmin 3 Roy C, Foudi F, Charton J, et al. Keywords: functional MRI, MRI, prostate neoplasm, recurrence DOI: /AJR Received April 21, 2012; accepted after revision July 31, Department of Radiology B, University Hospital of Strasbourg, New Civil Hospital, 1 Pl de l hopital BP 426, Strasbourg, Cedex 67091, France. Address correspondence to C. Roy (catherine.roy@chru-strasbourg.fr). 2 Department of Radiotherapy, Centre Paul Strauss, Centre de Lutte Contre le Cancer, Strasbourg, France. 3 Department of Urology, University Hospital of Strasbourg, New Civil Hospital, Strasbourg, France. WEB This is a Web exclusive article. AJR 2013; 200:W361 W X/13/2004 W361 American Roentgen Ray Society Comparative Sensitivities of Functional MRI Sequences in Detection of Local Recurrence of Prostate Carcinoma After Radical Prostatectomy or External-Beam Radiotherapy OBJECTIVE. The aim of this retrospective study was to determine the respective accuracies of three types of functional MRI sequences diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) MRI, and 3D 1 H MR spectroscopy (MRS) in the depiction of local prostate cancer recurrence after two different initial therapy options. MATERIALS AND METHODS. From a cohort of 83 patients with suspicion of local recurrence based on prostate-specific antigen (PSA) kinetics who were imaged on a 3-T MRI unit using an identical protocol including the three functional sequences with an endorectal coil, we selected 60 patients (group A, 28 patients who underwent radical prostatectomy; group B, 32 patients who underwent external-beam radiation) who had local recurrence ascertained on the basis of a transrectal ultrasound-guided biopsy results and a reduction in PSA level after salvage therapy. RESULTS. All patients presented with a local relapse. Sensitivity with T2-weighted MRI and 3D 1 H-MRS sequences was 57% and 53%, respectively, for group A and 71% and 78%, respectively, for group B. DCE-MRI alone showed a sensitivity of 100% and 96%, respectively, for groups A and B. DWI alone had a higher sensitivity for group B (96%) than for group A (71%). The combination of T2-weighted imaging plus DWI plus DCE-MRI provided a sensitivity as high as 100% in group B. CONCLUSION. The performance of functional imaging sequences for detecting recurrence is different after radical prostatectomy and external-beam radiotherapy. DCE-MRI is a valid and efficient tool to detect prostate cancer recurrence in radical prostatectomy as well as in external-beam radiotherapy. The combination of DCE-MRI and DWI is highly efficient after radiation therapy. Three-dimensional 1 H-MRS needs to be improved. Even though it is not accurate enough, T2-weighted imaging remains essential for the morphologic analysis of the area. T he diagnosis of local recurrence of a prostate carcinoma after radical prostatectomy (RP) or external-beam radiotherapy is a real challenge. At present, the diagnosis of relapse is based mainly on a prostate-specific antigen (PSA) level above a threshold or the PSA kinetic value and is called biochemical failure. However, biochemical failure is not synonymous with local recurrence in the prostatic bed. It can be due to distant metastases, local disease, or both. Only rarely, a biochemical recurrence is accompanied by a detectable mass at digital rectal examination. The existence of salvage procedures in cases of local recurrence is necessary to localize the recurrence as soon as biochemical failure occurs. Unfortunately, until recently, local recurrence of early-stage prostate cancer had a reputation of being difficult to detect by imaging, as well as by transrectal ultrasoundguided biopsy, because, in the initial phase of relapse, the cancer volume is low. CT is not widely used for the detection of local recurrence because of the low accuracy of this technique [1]. Immunoscintigraphy and 11 C-choline computed PET [2] have been introduced as innovative imaging modalities for the detection of disease relapse, but their role is still incompletely defined. Because of its high contrast and spatial resolution, especially with an endorectal coil, MRI represents a promising technique for accurate evaluation of prostatic fossa after prostatectomy [3 6]. It has been shown that, using a 3-T magnet, the sensitivity is improved [7 9] for smaller size lesions. In addition, the recent development of functional MRI techniques has yielded promising results in detecting and localizing primary and recurrent prostate cancer AJR:200, April 2013 W361

2 Roy et al. [3]. Previous series published in this field have studied one or two functional sequences for one type of treatment (RP or external-beam radiotherapy) in the detection of prostate cancer recurrence [10 14]. To our knowledge, none of them presented an analysis of all sequences now available on the same 3-T MRI unit for the two therapeutic options [3]. The aim of this retrospective study was to determine the accuracy of the three types of functional MRI, including diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) T1-weighted MRI, and the metabolic ratio with 3D 1 H MR spectroscopy (MRS), to detect local prostate cancer recurrence for patients treated with two different therapeutic options. Materials and Methods Study Population This retrospective single-institution study received approval by our local ethical committee. No written informed consent was necessary. Between January 2009 and July 2011, 83 patients were referred to our MRI department for suspicion of local recurrence, after two different therapy options (RP and external-beam radiotherapy), of a localized prostate cancer. Twenty-three patients were excluded because they did not have histologic confirmation of recurrence. We selected 60 consecutive patients who had histologic proof of recurrence. Among them, two groups of patients were defined according to initial therapy: 28 patients (mean age, 65 years; range, years) had undergone RP (group A), and 32 patients (mean age, 67 years; range, years) had received external-beam radiation (group B). In group A, the pretreatment characteristics of patients were negative surgical margins; initial mean (± SD) PSA value of 5.8 ± 2.2 ng/ml; pathologic stages of pt2a (11 patients), pt2b (eight patients), pt2c (two patients), and pt3a (seven patients); and Gleason score of 6 (16 patients) or 7 (12 patients). In group B, the pretreatment characteristics of patients were initial mean PSA value of 13.5 ± 3.2 ng/ ml, clinicopathologic stage T2c (21 patients) or T3 (11 patients), and Gleason score of 7 for all of them. The suspicion of a local recurrence was based on biochemical data derived from PSA kinetics. The threshold used for biologic relapse was in agreement with the current consensus. After RP, a PSA level above 0.2 ng/ml indicates a residual or recurrent disease [1]. After external-beam radiotherapy, we considered a PSA increase by 2 ng/ml or more above the nadir as biochemical relapse, in accordance with previously published criteria [1]. No patient had received previous or neoadjuvant hormonal treatment. At the time of biochemical failure, the mean PSA level was 0.98 ng/ml (range, ng/ml) for surgical treatment and 3.6 ng/ ml (range, ng/ml) for radiation therapy. The mean interval between functional MRI and initial treatment was 55 months (range, months). The mean interval between suspected relapse based on biochemical failure and functional MRI was 18 days. Imaging Methods MRI examinations were performed on a 3-T MRI unit (Achieva, Philips Healthcare) by using an endorectal coil (ecoil, Medrad) in association with an external 3 3 channel cardiac coil. A standard set of three conventional orthogonal T2-weighted fast spin-echo sequences was first acquired. Then, axial 2D DWI and 3D 1 H-MRS sequences were performed with orientation identical to that of the axial T2-weighted images (Table 1). DWI was obtained by using fast single-shot spin-echo echo-planar imaging with b values of 0 and 1000 s/mm 2. The diffusion images were reconstructed for each pixel by using the dedicated postprocessing software (diffusion calculation, version 6.2, Achieva, Philips Healthcare), and an apparent diffusion coefficient (ADC) map was automatically generated to obtain a monoexponential fit of the data. The mean ADC value at a b value of 1000 s/ mm 2 with SD was calculated by placing manually on the ADC map a region of interest strictly encompassing the whole area of the suspected zone. Three-dimensional 1 H-MRS was performed with the point-resolved spatially localized spectroscopy technique with water and lipid suppression. Metabolic ratio maps of the ratio of choline plus polyamine plus creatine to citrate were generated, with a multivoxel analysis composed by 25 voxels per slice. The voxel size was mm 3. Finally, axial 3D gradient-echo DCE-MRI was performed with 30 consecutive dynamic sequences, with a resolution time of 6 seconds, during an IV bolus injection of contrast agent (1.0 mmol/ml/kg of body weight; Dotarem, Guerbet) with a power injector at a rate of 3.5 ml/s. The start of the gadolinium injection and the start of the dynamic sequence were simultaneous, to obtain a noninjected acquisition for subtracted images, which were generated by an automated algorithm during the acquisition (Table 1). The total examination time was about 50 minutes, including time for coil placement, patient positioning, and localization of the prostate. MRI Data Analysis The analysis of MRI examinations was independently performed by two radiologists: a fellow involved in this field and an experienced radiologist with 15 years of experience in pelvic MRI. They were aware that patients had been diagnosed with biochemical failure but were unaware of all other clinical or histopathologic findings, as well as the initial MRI report. Each reader recorded the interpretation results (side, localization, and type of sequence) on a diagram and saved it in a different folder. Differences in assessment of each topic (localization for four cases and sequences for 12 cases) were resolved by means of consensus with a third radiologist. The method for reviewing was as follows with anonymized datasets: each sequence s images alone, T2-weighted imaging with each functional sequence added in turn, T2-weighted imaging plus two functional sequences, and all sequences together. Each reading session consisted of reviewing each set randomly. The interval between the reading sessions was 1 week. The total time for checking all data was 3 months. Axial images of DWI at a b value of 1000 s/mm 2, ADC maps, T2-weighted sequences, and T1-weighted DCE sequences were merged together on the same level by our software package (Achieva, Philips Healthcare), which automatically indicates the exact location across all images. On a T2-weighted sequence, a nodule or area was considered suspicious if the signal intensity was slightly hyperintense in the prostatic fossa after RP or was hypointense inside the prostate in cases of radiation treatment. On an ADC map, a nodule was considered positive for malignancy if the cutoff value was equal to or lower than mm 2 /s [15, 16]. For spectroscopic analysis, a voxel was considered positive for malignancy for a ratio value threshold up to 0.86 [16, 17] or when only a choline peak was detected [18]. Gray-scale DCE sequences were visualized for all sections of the gland at the first five time points where early enhancement was detected. For DCE data, because the definition of tumoral tissue is not uniform, we considered semiquantitative parameters by placing regions of interest in an early enhanced area. Of the multiple parameters provided by our software, we chose the visual system with a qualitative analysis of time signal intensity curves, as described by Engelbrecht et al. [19], such as high maximum relative enhancement (> 100%) and short time to peak, with high wash-in rate and washout slope of the curve, to diagnose malignant tissue. Correlations The standard of reference for local recurrence was ascertained on the basis of a transrectal ultrasound guided biopsy result and a reduction in PSA level after salvage therapy. All transrectal ultrasound-guided biopsies were performed by radiologists involved in this field. For group A, a complete evaluation of the prostatic fossa was performed before biopsy using internal landmarks (i.e., external sphincter, vesicourethral anastomosis, and posteri- W362 AJR:200, April 2013

3 Detection of Prostate Cancer Recurrence With Different Functional MRI Sequences TABLE 1: Parameters Used for Different MRI Sequences Parameter Fast Spin-Echo MRI Diffusion-Weighted Imaging or wall of the bladder) to better localize the suspicious MRI findings on the initial report. By using these criteria, the suspicious zone was projected as accurately as possible on sonogram, and transrectal ultrasound-guided biopsy was used to remove two cores from the suspicious zone. In addition to those samples, biopsies were performed according to the standard scheme (i.e., three cores from each side of the vesicourethral anastomosis). For group B, biopsies were performed using the standard technique, consisting of 14 random laterally directed cores, in addition to two cores from the highly suspicious area based on the initial MRI report. All samples were thoroughly labeled by the operator according the localization of the sample. A tumor site on each MRI scan was considered to match histologic findings by an experienced radiologist if the tumor was present in the same region of the prostate as that indicated in the pathologic reports. The presence of malignancy was confirmed by a positive transrectal ultrasound-guided biopsy from the prostatic fossa after RP (28 patients) or in the suspected intraprostatic zone on MRI after radiation therapy (32 patients). Positive biopsy sites correlated with the same anatomic localization suspected and identified on MRI, except for a second suspected site in two patients after RP. However, we have considered those localizations as positive, because their aspect on MRI was quite similar to that of zones with positive biopsy. All patients presented with a reduction in PSA level after salvage therapy: external-beam radiotherapy in 28 patients in group A and high-intensity focused ultrasound or hormone therapy in 32 patients in group B. Differences of sensitivity between any two MRI methods were tested using the MacNemar test. The Fisher test was used to determine significant differences in sensitivity of multiparametric MRI. Statistical analyses were performed using J software (version 11.5 for Windows, SPSS). A p value less than 0.05 was considered as indicative of statistically significant differences. Results In all patients treated with external-beam radiotherapy, the entire treated prostatic gland had signal intensity changes manifesting as moderate diffusely decreased signal intensity of the prostatic tissue, especially of the peripheral zone, and loss of the zonal anatomy. A low-signal-intensity rim around the gland was frequently observed. All our patients were considered as having recurrent disease. Because no patients were negative for recurrence, only sensitivity was calculated. There were no false-positive cases after the consensus with the third radiologist. 3D 1 H MR Spectroscopy Dynamic Contrast-Enhanced 3D Gradient-Echo T1-Weighted MRI TR/TE 4200/ / /140 13/4.6 Flip angle ( ) Matrix size (mm) interpolated to interpolated to interpolated to 512 Orientation Three orthogonal planes Axial Axial Axial Fat suppression No Yes Yes Yes Parallel imaging acceleration factor FOV (mm) Echo-train length Echo-planar imaging 73 No. of excitations Slice number Slice thickness/gap (mm) 3.5/0.4 4/0 13/0 3.5/1 Acquisition time 3 min 25 s 3 min 54 s 9 min 25 s 4 min/30 s Note Dashes indicate that there are no parameters for this type of sequence. Group A Local recurrence was most common around the perianastomotic site, occurring in 21 cases. Other localizations were at the level of extracted seminal vesicles in 11 cases and in the retrovesical area in two cases. Four patients had two localizations. One patient had three lesions in the perivesical zone and inside the bladder wall. One patient had recurrent disease at the level of retained prostatic tissue. A total of 34 suspicious areas were detected. Considering isolated T2-weighted sequences, 19 abnormal zones were found to be suspicious for cancer on visual analysis in 16 patients (Figs. 1 and 2 and Table 2). On isolated DWI, all suspicious zones appeared as a focal hyperintense lesion at a b value of 1000 s/ mm 2, with a clear hypointense aspect on the ADC map. The mean ADC value in those suspected areas was ± mm 2 /s (range, to mm 2 /s). Overall, 23 abnormal zones were detected in 21 patients. With a cutoff value of mm 2 /s, 20 patients were positive. The ADC value was quite similar in cases of several masses in the same patient. Isolated DCE sequences were positive, with a typical aspect of the curve for 32 nodules in 28 patients. All patients were positive, but in cases of multiple nodules, two of them presented a less typical curve. On 3D 1 H-MRS, there was a significant choline peak without a citrate peak outside the noise level in one voxel or two separated voxels in 15 patients. Only one zone was detected in 13 patients. Sensitivity was highest for multiparametric T2-weighted imaging plus DCE-MRI followed by, in decreasing order, T2-weighted imaging plus DCE-MRI plus DWI. These differences were statistically significant only for comparisons of sensitivity between T2-weighted and DCE imaging (p = ) and T2-weighted plus DCE imaging (p = 0.003) or T2-weighted imaging plus DCE-MRI plus DWI (p = ). With regard to findings on T2-weighted sequences, the mean diameter of soft-tissue nodules was 9 mm (range, 5 14 mm). Group B Local recurrence was found more often in the peripheral zone (22 patients) than in the transitional zone (10 patients). Two patients AJR:200, April 2013 W363

4 Roy et al. Signal Intensity ( 10 3 ) were considered to have two localizations. A total of 34 abnormal tissue areas were detected (Table 2 and Fig. 3). Considering isolated T2-weighted sequences, only 25 abnormal zones were suspected in 23 patients on the basis of a lower signal intensity than in the surrounding tissue. Considering isolated DWI sequences, 31 patients had an abnormal zone at the visual evaluation. Those zones appeared as a focal hyperintense lesion at a b value of 1000 s/mm 2 on DWI, with a low-intensity focal area relative to that of the surrounding benign tissues on the ADC map. The mean ADC value in areas of suspected malignancy was ± mm 2 /s (range, to mm 2 /s), versus ± mm 2 /s in the surrounding tissue. With the cutoff value of mm 2 /s, all 31 patients were positive. Thirty-two abnormal zones were visualized. TABLE 2: Sensitivity for Different Isolated Sequences or in Combination in Groups A and B Group, Nodule or Patient Time (s) MRI DWI DCE-MRI 3D 1 H-MRS A D MRI Plus DCE-MRI MRI Plus DWI Isolated DCE sequences were positive on visual evaluation, with a typical aspect of the semiquantitative curve for 32 nodules in 31 patients. On 3D 1 H-MRS, choline and citrate were clearly detected in certain voxels. There was a significant choline peak with a ratio over 0.8 in one voxel or two separated voxels in 25 zones for 24 patients. No metabolites were found outside the signal noise in the remaining prostatic tissue. The combination of two functional sequences that is, DWI and DCE-MRI in addition to T2-weighted imaging provided a sensitivity of 100%. It was followed by, in decreasing order, T2-weighted imaging plus DWI. Indeed, these differences were not statistically signif- MRI Plus 3D 1 H-MRS MRI Plus DWI Plus DCE-MRI MRI Plus DWI Plus DCE-MRI Plus 3D 1 H-MRS Group A Nodules (n = 34) 19 (56) 22 (65) 32 (94) 17 (50) 33 (97) 22 (65) 18 (53) 32 (94) 25 (74) Patients (n = 28) 16 (57) 20 (71) 28 (100) 15 (54) 28 (100) 20 (71) 16 (57) 28 (100) 20 (71) Group B Nodules (n = 34) 25 (74) 32 (94) 32 (94) 25 (74) 31 (91) 32 (94) 15 (44) 34 (100) 26 (76) Patients (n = 32) 23 (72) 31 (97) 31 (97) 25 (78) 31 (97) 31 (97) 23 (72) 32 (100) 24 (75) Note Data are no. (%). Group A included 28 patients with 34 suspicious areas. Group B included 32 patients with 34 suspicious areas. The difference between combination T2-weighted imaging plus diffusion-weighted imaging (DWI) and T2-weighted imaging plus DWI plus dynamic contrast-enhanced (DCE) MRI was statistically significant for Group A. MRS = MR spectroscopy. B Fig year-old man with prostate-specific antigen (PSA) level progression that reached 1.3 ng/ml after radical prostatectomy (pt2 with positive surgical margins). A, Axial T2-weighted image of prostatectomy bed shows slightly hyperintense mass (arrow) (maximum diameter, 8 mm) up to right lateral part of anastomosis. B, Apparent diffusion coefficient map of diffusion-weighted image shows hypointense artifact (arrow) due to surgical clips. C, Dynamic contrast-enhanced subtracted image shows area of intense enhancement (arrow) corresponding to lesion. D, Time-intensity curve obtained by positioning region of interest inside this area shows typical pattern of malignancy. No metabolic peak was registered at spectroscopy. Reduction of PSA level occurred after salvage radiation treatment. C W364 AJR:200, April 2013

5 Detection of Prostate Cancer Recurrence With Different Functional MRI Sequences icant only for comparisons of sensitivity between T2-weighted imaging plus DCE-MRI and T2-weighted imaging plus DCE-MRI plus DWI (p = 0.002). With regard to findings on T2-weighted sequences, the mean diameter of suspicious nodules revealed at MRI was 8.2 mm (range, 4 15 mm). Discussion Serial PSA level measurements are the standard biomarker for detecting recurrence of prostate cancer, with different definitions for recurrence according to the therapy used. Signal Intensity ( 10 3 ) Time (s) A The existence of salvage procedures has made it necessary to localize local recurrence very early. Cross-sectional imaging modalities (CT, morphologic MRI, and ultrasound) have been used, but each of them is poorly sensitive for detecting a small-sized relapse. Transrectal ultrasound guided biopsy of the prostatic fossa was used as the initial test for the identification of local recurrence. This strategy has not been clearly defined, neither for the location of the cores nor for their number. In fact, the most consistent achievement of a diagnosis of a local recurrence with transrectal ultrasound D Fig year-old man with suspicion of recurrence 15 months after radical prostatectomy (pt3a) (prostate-specific antigen level [PSA], 0.9 ng/ml). A, Axial T2-weighted image shows small moderately hyperintense nodule (arrow) 6 mm posterior to vesicourethral anastomosis. B, Apparent diffusion coefficient (ADC) map of diffusion-weighted image shows clearly hypointense nodule (arrow) in spite of surgical material (arrowhead). Average ADC was mm 2 /s. C, Dynamic contrast-enhanced subtracted image shows area of intense enhancement corresponding to nodule (arrow) and clips artifacts (arrowhead). D, Time-intensity curve obtained by region of interest shows typical pattern of malignancy. E, Spectroscopic analysis shows high choline (Cho) peak only at posterior part of perianastomosis area. Pathologic proof was obtained by transrectal ultrasound-guided biopsy. Cit = citrate, Cr = creatine. B guided biopsy is in patients with a PSA level higher than 4 ng/ml [20]. The sensitivity of PSA level remains low (40 71% for PSA levels < 1 ng/ml) [21]. In agreement with data from others, recurrences in our cohort after RP have been found anywhere along the surgical cavity, but most often at the level of the perianastomotic area [21]. In our patients treated by external-beam radiotherapy, recurrences were intraprostatic, found most often in the peripheral zone. In both groups, we obtained the worst results with isolated T2-weighted and 3D 1 H- C E AJR:200, April 2013 W365

6 Roy et al. Signal Intensity ( 10 3 ) Time (s) A D E MRS sequences. Early reports on T2-weighted sequences with the use of an endorectal coil for detecting recurrence after RP have been promising, with sensitivity and specificity close to 100% [4, 22] for recurrent lesions from 1.4 to 1.9 cm (range, cm) in diameter. More recently, however, the values have been lower (sensitivity, 48 61%; specificity, 52 82%), probably because of the small size and less-advanced stage of recurrence and larger study sample size [10, 11]. Our results, with a sensitivity of 55%, are in agreement. Because small lesions are less hyperintense than large masses compared with muscle on T2-weighted images, as a result of partial volume effects, they can be difficult to differentiate from fibrous scar in the prostate fossa or in fibrous bands of seminal vesicle remnants [22, 23]. The morphologic changes in the prostate after external-beam radiotherapy include inflammation, glandular atrophy, fibrosis, and prostate shrinkage [24]. These changes result in diffusely decreased signal intensity of the prostatic stroma and loss of the normal zonal anatomy, causing difficulty in differentiating recurrence from irradiated normal tissue [12] (Fig. 3). The low sensitivity in the present study is in concordance with that found in other studies. Fig year-old man treated by external-beam radiotherapy 4 years before who presented with suspected local recurrence and increased prostate-specific antigen (PSA) level (4.5 ng/ml). A, Axial T2-weighted image shows diffusely less intense signal in peripheral zone and nodular focal hypointense zone (arrow) of 7 mm on right. B, Apparent diffusion coefficient (ADC) map of diffusion-weighted image shows more clearly hypointense nodule (arrow). Average ADC was mm 2 /s. C, Dynamic contrast-enhanced subtracted image shows area of intense enhancement corresponding to nodule (arrow). D and E, Regions of interest were placed (D), and time-intensity curve (E) thus generated shows typical pattern of malignancy characterized by smaller time to peak and higher peak enhancement values with early downward slope of curve (washout) (1), compared with normal contralateral zone (2). F, Spectroscopic analysis showed high choline (Cho) levels in morphologic abnormality with increased ratio (2.81) of choline plus creatine (Cr) to citrate (Cit). Recurrent carcinoma was histologically validated at transrectal ultrasound guided biopsy. B C F W366 AJR:200, April 2013

7 Detection of Prostate Cancer Recurrence With Different Functional MRI Sequences The lower performance of 3D 1 H-MRS in group A may reflect a partial volume effect due to the voxel size. The use of 3D multivoxel spectroscopy would improve spatial resolution. Sciarra et al. [25] reported a sensitivity of 71 84% and a specificity of 83 88% in the detection of local recurrence after RP. However, despite using a similar sequence, our results are disappointing because citrate could not be measured. After removal of prostatic tissue, a metabolite ratio is difficult to calculate, which was the case for our data. Furthermore, surgical clips in the anastomotic area induce field inhomogeneities and susceptibility artifacts, which can preclude successful spectroscopic measurements [26]. After external-beam radiotherapy, residual prostate cancer can still be identified by a relative increase in the ratio of choline plus creatine to citrate or by an increase in the choline peak with no detectable citrate [26]. Using these criteria, good correlations between spectroscopic data and biopsy findings have been reported [25, 27]. However, some benign glands can exhibit high levels of choline after external-beam radiotherapy and cause falsepositive findings [27]. Therefore, it remains unclear which metabolic criteria should be used to differentiate benign from malignant areas in an irradiated prostate. Our sensitivity value for this isolated sequence agreed with those reported from other studies [4, 12]. Semiquantitative DCE parameters were used because they are relatively simple to achieve and have been shown to work equally well compared with complex quantitative methods [28 30]. We therefore used the visual aspect of the enhancement curve and then the semiquantitative parameters evaluation. Our sensitivity results are in agreement with recent literature, which strongly suggest that MRI combined with DCE-MRI can detect and localize local recurrences soon after RP, in patients with PSA levels less than 2 ng/ml, because the tumoral tissue enhances earlier and faster than fibrosis or remnants of normal prostate in the prostatectomy fossa. The potential perturbation by surgical clips does not seem to be a limitation. However, no cutoff values have been mentioned [25, 26]. Exceptional residual benign prostatic hypertrophic nodules can mimic recurrence and, therefore, need to undergo a T2-weighted sequence for clarification [11, 13]. After external-beam radiotherapy, recurrent tissue can be recognized as early-enhancing homogeneous nodules that contrast well with the surrounding tissue, in which the enhancement was homogeneous, slow, and low probably caused by radiation-induced fibrosis and vascular damage [7, 13, 29]. The utility of DWI in the detection of local recurrence after RP has not been systematically investigated [31 35]. Our results are clearly better for the group that underwent external-beam radiotherapy than for the group that underwent RP. One part of the explanation is the intensity of artifacts due to surgical clips, which are present in the prostatic fossa. The other part could be the small amount of tumor within a voxel, as also noted by Reinsberg et al. [15]. Even if our voxel size was quite similar to that of a DCE sequence, the level of signal and contrast was clearly higher with gadolinium injection (Fig. 1). After external-beam radiotherapy, our sensitivity was quite high with the endorectal coil. In experimental models, normal prostatic tissue that regenerated after thermal injury or radiation therapy showed an increase in the ADC values toward normal [34]. Therefore, sites of residual or recurrent disease continue to show restricted diffusion and theoretically can be detected by DWI. To the best of our knowledge, only two studies have been published on the value of DWI in the detection of recurrence after external-beam radiotherapy with a 3-T MRI system. The combination of DCE-MRI with either MRS or T2-weighted imaging was advocated for a better diagnostic performance than with either technique alone. MRS has the disadvantages of longer acquisition time and the need for additional software [32], which lead to increased cost and decreased throughput. Finally, MRS suffers from a lack of spatial resolution. Even if MRS possesses acceptable accuracy and diagnostic performance, we can postulate that, in the daily work flow, this sequence is not necessary and must not be included in the routine protocol. Improvements in accuracy have also been shown in the follow-up of patients with prostate cancer by combining T2-weighted imaging, DCE-MRI, and DWI [3, 30], but the performance of functional imaging varies with the primary treatment. To minimize time for data acquisition, we suggest that, after RP, DCE-MRI as an isolated functional sequence is sufficient to detect small recurrence but that, after external-beam radiotherapy, the sensitivity between DWI and DCE- MRI is similar, which is in agreement with the findings of Kim et al. [35]. Therefore, because DWI requires a short imaging time, without the need for IV contrast material and has relatively simple postprocessing requirements, it is an alternative to DCE-MRI. In cases of RP with low PSA level, where lesion size may be small, DWI may be less useful. This study has a certain number of limitations. First, all the suspicious areas were considered to be positive for recurrence with pathologic proof in only 66 of 68 nodules. These nodules were typical on MRI and could have been missed during the biopsy procedure because of their less common location (i.e., retrovesical space and bladder wall). The use of PSA level reduction may have been inadequate because it can indicate local or systemic recurrence. However, in all our cases presented, the PSA level was low, which usually indicates local than systemic recurrence. Also, we did not consider the fact that a low PSA level increase can be observed after radiation therapy without recurrent disease. Second, we did not have in our material false-positive cases, after reviewing in consensus for some disagreement between readers. With DCE-MRI in particular, there was no intense enhancement in other zones in the surgical local area after RP or in the peripheral zone after external-beam radiotherapy that did not correlate with positive biopsy. On the other hand, our correlation was done with transrectal ultrasound guided biopsy, which is obviously less pertinent than a surgical biopsy, and only positive samples can be taken into account. Third, in semiquantitative analyses, the standard of reference used for spectroscopic analysis and time-intensity curves of DCE-MRI applicable to untreated glands may not be applicable to treated glands. Although this is a short series for each group, we have used the same protocol for all making comparison acceptable. Finally, a memory effect during the evaluating process is likely, so that a lesion identified on one MRI technique will then be more easily identified on others, thus increasing lesion detection sensitivity. In conclusion, the DCE-MRI technique is a valid and efficient tool to detect small-sized prostate cancer recurrence after RP and after external-beam radiotherapy. The combination of DCE-MRI and DWI is highly efficient after external-beam radiotherapy. Three-dimensional 1 H-MRS needs to be improved when used on a 3-T unit with an endorectal coil. Despite its lower sensitivity, T2-weighted imaging remains essential for the morphologic analysis. References 1. Boccon-Gibod L, Djavan WB, Hammerer P, et al. Management of prostate-specific antigen relapse AJR:200, April 2013 W367

8 Roy et al. in prostate cancer: a European consensus. Int J Radiology 2004; 233: normal prostate gland. Histopathology 1997; Clin Pract 2004; 58: Haider MA, Chung P, Sweet J, et al. Dynamic 30: Schillaci O, Calabria F, Tavolozza M, et al. Influ- contrast-enhanced magnetic resonance imaging 25. Sciarra A, Panebianco V, Salciccia S, et al. Role of ence of PSA, PSA velocity and PSA doubling time for localization of recurrent prostate cancer after dynamic contrast-enhanced magnetic resonance on contrast-enhanced (18) F-choline PET/CT de- external beam radiotherapy. Int J Radiat Oncol (MR) imaging and proton MR spectroscopic im- tection rate in patients with rising PSA after radical prostatectomy. Eur J Nucl Med Mol Imaging 2012; 39: De Visschere PJ, De Meerleer GO, Fütterer JJ, Villeirs GM. Role of MRI in follow-up after focal therapy for prostate carcinoma. AJR 2010; 194: Silverman JM, Krebs TL. MR imaging evaluation with a transrectal surface coil of local recurrence of prostatic cancer in men who have undergone radical prostatectomy. AJR 1997; 168: Sella T, Schwartz LH, Swindle PW, et al. Suspected local recurrence after radical prostatectomy: endorectal coil MR imaging. Radiology 2004; 231: Sala E, Eberhardt SC, Akin O, et al. Endorectal MR imaging before salvage prostatectomy: tumor localization and staging. Radiology 2006; 238: Kim CK, Park BK, Park W, Kim SS. Prostate MR imaging at 3T using a phased-arrayed coil in predicting locally recurrent prostate cancer after radiation therapy: preliminary experience. Abdom Imaging 2010; 35: Fütterer JJ, Engelbrecht MR, Jager GJ, et al. Prostate cancer: comparison of local staging accuracy of pelvic phased-array coil alone versus integrated endorectal-pelvic phased-array coils local staging accuracy of prostate cancer using endorectal coil MR imaging. Eur Radiol 2007; 17: Ocak I, Bernardo M, Metzger G, et al. Dynamic contrast-enhanced MRI of prostate cancer at 3 T: a study of pharmacokinetic parameters. AJR 2007; 189:849; [web]w192 W Casciani E, Polettini E, Carmenini E, et al. Endorectal and dynamic contrast-enhanced MRI for detection of local recurrence after radical prostatectomy. AJR 2008; 190: Cirillo S, Petracchini M, Scotti L, et al. Endorectal magnetic resonance imaging at 1.5 tesla to assess local recurrence following radical prostatectomy using T2-weighted and contrast-enhanced imaging. Eur Radiol 2009; 19: Coakley FV, Teh HS, Qayyum A, et al. Endorectal MR imaging and MR spectroscopic imaging for locally recurrent prostate cancer after external beam radiation therapy: preliminary experience. Biol Phys 2008; 70: Rouvière O, Valette O, Grivolat S, et al. Recurrent prostate cancer after external beam radiotherapy: value of contrast-enhanced dynamic MRI in localizing intraprostatic tumor correlation with biopsy findings. Urology 2004; 63: Reinsberg SA, Payne GS, Riches SF, et al. Combined use of diffusion-weighted MRI and 1 H MR spectroscopy to increase accuracy in prostate cancer detection. AJR 2007; 188: Portalez D, Gauthier R, Leanchi P, et al. Prospective comparison of T2w-MRI and dynamic contrast-enhanced MRI, 3D-MR spectroscopic imaging or diffusion MRI in repeat TRUS-guided biopsies. Eur Radiol 2010; 20: Fütterer JJ, Heijmink SW, Scheenen TW, et al. Prostate cancer localization with dynamic contrast-enhanced MR imaging and proton MR spectroscopic imaging. Radiology 2006; 241: 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: Engelbrecht MR, Huisman HJ, Laheij RJ, et al. Discrimination of prostate cancer from normal peripheral zone and central gland tissue by using dynamic contrast-enhanced MR imaging. Radiology 2003; 229: Deliveliotis C, Manousakas T, Chrisofos M, Skolarikos A, Delis A, Dimopoulos C. Diagnostic efficacy of transrectal ultrasound-guided biopsy of the prostatic fossa in patients with rising PSA following radical prostatectomy. World J Urol 2007; 25: Leventis AK, Shariat SF, Slawin KM. Local recurrence after radical prostatectomy: correlation of US features with prostatic fossa biopsy findings. Radiology 2001; 219: Sella T, Schwartz LH, Hricak H. Retained seminal vesicles after radical prostatectomy: frequency, MRI characteristics, and clinical relevance. AJR 2006; 186: Allen SD, Thompson A, Sohaib SA. The normal post-surgical anatomy of the male pelvis following radical prostatectomy as assessed by magnetic resonance imaging. Eur Radiol 2008; 18: Sheaff MT, Baithun SI. Effects of radiation on the aging in the detection of local recurrence after radical prostatectomy for prostate cancer. Eur Urol 2008; 54: Casciani E, Bertini L, Gualdi G. Re: Sciarra A, Panebianco V, Salciccia S, et al. Role of dynamic contrast-enhanced magnetic resonance (MR) imaging and proton MR spectroscopic imaging in the detection of local recurrence after radical prostatectomy for prostate cancer. (letter) Eur Urol 2009; 55:e2 e3; author reply, e4 e5 27. Westphalen AC, Coakley FV, Roach M, McCulloch CE, Kurhanewicz J. Locally recurrent prostate cancer after external beam radiation therapy: diagnostic performance of 1.5T endorectal MR imaging and MR spectroscopic imaging for detection. Radiology 2010; 256: Franiel T, Hamm B, Hricak H. Dynamic contrastenhanced magnetic resonance imaging and pharmacokinetic models in prostate cancer. Eur Radiol 2011; 21: Rouvière O, Vitry T, Lyonnet D. Imaging of prostate cancer local recurrences: why and how? Eur Radiol 2010; 20: Isebaert S, De Keyzer F, Haustermans K, et al. Evaluation of semi-quantitative dynamic contrast-enhanced MRI parameters for prostate cancer in correlation to whole-mount histopathology. Eur J Radiol 2011; 81:e217 e Tan CH, Wang J, Kundra V. Diffusion weighted imaging in prostate cancer. Eur Radiol 2011; 21: Van As N, Charles-Edwards E, Jackson A, et al. Correlation of diffusion weighted MRI with whole mount radical prostatectomy specimens. Br J Radiol 2008; 81: Sato C, Naganawa S, Nakamura T, et al. Differentiation of noncancerous tissue and cancer lesions by apparent diffusion coefficient values in transition and peripheral zones of the prostate. J Magn Reson Imaging 2005; 21: Chen J, Daniel BL, Diederich CJ, et al. Monitoring prostate thermal therapy with diffusionweighted MRI. Magn Reson Med 2008; 59: Kim CK, Park BK, Lee HM. Prediction of locally recurrent prostate cancer after radiation therapy: incremental value of 3T diffusion-weighted MRI. J Magn Reson Imaging 2009; 29: W368 AJR:200, April 2013