Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Genitourinar y Imaging • Original Research Park et al. MRI to Predict Biochemical Recurrence After Radical Prostatectomy Genitourinary Imaging Original Research
Prostate Cancer: Role of Pretreatment Multiparametric 3-T MRI in Predicting Biochemical Recurrence After Radical Prostatectomy Jung Jae Park1,2 Chan Kyo Kim1 Sung Yoon Park1 Byung Kwan Park1 Hyun Moo Lee 3 Seong Whi Cho2 Park JJ, Kim CK, Park SY, Park BK, Lee HM, Cho SW
Keywords: multiparametric MRI, prediction, prostatectomy, prostate neoplasm, recurrence DOI:10.2214/AJR.13.11381 Received June 14, 2013; accepted after revision August 28, 2013. 1 Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea. Address correspondence to C. K. Kim ([email protected]). 2 Department of Radiology, Kangwon National University Hospital, Chuncheon, Republic of Korea. 3 Department of Urology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea.
WEB This is a web exclusive article. AJR 2014; 202:W459–W465 0361–803X/14/2025–W459 © American Roentgen Ray Society
OBJECTIVE. The purpose of this study is to retrospectively investigate whether pretreatment multiparametric MRI findings can predict biochemical recurrence in patients who underwent radical prostatectomy (RP) for localized prostate cancer. MATERIALS AND METHODS. In this study, 282 patients with biopsy-proven prostate cancer who received RP underwent pretreatment MRI using a phased-array coil at 3 T, including T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced MRI (DCE-MRI). MRI variables included apparent tumor presence on combined imaging sequences, extracapsular extension, and tumor size on DWI or DCE-MRI. Clinical variables included baseline prostate-specific antigen (PSA) level, clinical stage, and Gleason score at biopsy. The relationship between clinical and imaging variables and biochemical recurrence was evaluated using Cox regression analysis. RESULTS. After a median follow-up of 26 months, biochemical recurrence developed in 61 patients (22%). Univariate analysis revealed that all the imaging and clinical variables were significantly associated with biochemical recurrence (p < 0.01). On multivariate analysis, however, baseline PSA level (p = 0.002), Gleason score at biopsy (p = 0.024), and apparent tumor presence on combined T2WI, DWI, and DCE-MRI (p = 0.047) were the only significant independent predictors of biochemical recurrence. Of the independent predictors, apparent tumor presence on combined T2WI, DWI, and DCE-MRI showed the highest hazard ratio (2.38) compared with baseline PSA level (hazard ratio, 1.05) and Gleason score at biopsy (hazard ratio, 1.34). CONCLUSION. The apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI is an independent predictor of biochemical recurrence after RP. This finding may be used to construct a predictive model for biochemical recurrence after surgery.
R
adical prostatectomy (RP) is an established treatment modality that reduces disease-specific mortality for patients with localized prostate cancer [1, 2], although biochemical recurrence occurs in quite a few patients after surgery. According to a postoperative nomogram predicting long-term outcomes, the 10-year actuarial rate of biochemical recurrence–free survival was approximately 79%, but biochemical recurrence may occur beyond 10 years after RP [3, 4]. Because biochemical recurrence is associated with progression to distant metastases and cancer-specific mortality [5, 6], a prediction of biochemical recurrence before surgery is clinically important in determining primary treatment options, adjuvant therapies, and patient counseling.
To date, several preoperative and postoperative clinical variables have been used to predict biochemical recurrence after RP, but the accuracies of each are limited [7]. The inclusion of novel biomarkers or imaging tools may potentially improve the predictive accuracy. As MRI evolves into a valuable diagnostic method in prostate imaging, pretreatment MRI findings also have been investigated to improve predictive accuracy for biochemical recurrence after primary local treatment. The T stage and the presence of extracapsular extension (ECE) and seminal vesicle invasion (SVI), as determined by MRI, are significant prognostic parameters for predicting biochemical recurrence after primary local treatment, such as RP [8, 9]. More recently, it was found that emerging multiparametric prostate MRI can provide
AJR:202, May 2014 W459
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. functional information along with anatomic information; thus, it has contributed to tumor detection, localization, local staging, and determination of aggressiveness [10–14]. This promising technique may offer incremental value to a prognostic model using clinical variables to predict biochemical recurrence in prostate cancer after RP, with improved accuracy. However, few studies have reported the prognostic value of pretreatment multiparametric MRI findings in predicting biochemical recurrence after RP [8, 15]. The purpose of our study was to retrospectively investigate whether pretreatment multiparametric MRI findings at 3 T, including T2-weighted imaging (T2WI), diffusion-weighted (DWI), and dynamic contrastenhanced MRI (DCE-MRI), can predict biochemical recurrence in patients who undergo RP for localized prostate cancer.
insufficient follow-up to determine biochemical recurrence (n = 22). Five patients were also excluded because of a 3-month or longer interval between pretreatment MRI and surgery. Finally, 282 men (mean [± SD] age, 64 ± 7.2 years; range, 38–88 years) were enrolled in this study (Fig. 1). The mean interval between biopsy and MRI was 28.2 ± 4.9 days (range, 22–50 days), and surgery was performed 3–35 days after MRI (mean, 16 ± 6.2 days). The clinical stage was evaluated with digital rectal examination. Before surgery (mean, 2 ± 2.4 days; range, 1–28 days), baseline PSA level was identified in all patients. After surgery, the patients were followed-up with digital rectal examination and serum PSA level within 1 month, then every 3 months for the first 2 years, and every 6 months from the third year. Biochemical recurrence after RP was defined as an initial serum PSA level greater than or equal to 0.2 ng/mL, with a second confirmatory level of PSA of greater than 0.2 ng/mL [16].
Materials and Methods Patients
Imaging Technique
This retrospective study was approved by our institutional review board, with a waiver of the requirement for informed consent. Between May 2007 and April 2009, 435 consecutive men with biopsy-proven clinically localized prostate cancer were treated with RP. Of these, 372 patients underwent preoperative prostate MRI at 3 T, including T2WI, DWI, and DCE-MRI, in our institution. Of the 372 patients, 85 were excluded for the following reasons: hormone or radiation treatment before surgery (n = 29); imaging artifact or missing sequence (n = 18); missing data, such as Gleason score at biopsy or baseline prostate-specific antigen (PSA) level (n = 16); and
All patients underwent prostate MRI at 3 T (Intera Achieva 3 T or 3TX, Philips Healthcare System) using a phased-array coil. To reduce bowel peristalsis, 20 mg of butyl scopolamine (Buscopan, Boehringer Ingelheim) was administered intramuscularly before imaging. The MRI protocol was composed of routine T1-weighted imaging, T2WI, DWI, and DCE-MRI. After obtaining three plane localizer images, T2-weighted turbo spin-echo images were obtained in the axial, sagittal, and coronal planes. The imaging parameters were as follows: TR/TE, 3800–4700/80– 100; slice thickness, 3 mm; interslice gap, 1 mm; matrix, 512 × 304 or 568 × 341; FOV, 18–20 cm;
Patients who underwent radical prostatectomy n = 435
Preoperative MRI n = 372 Excluded patients n = 90
number of signals acquired, 3; sensitivity encoding factor, 2; number of slices, 21; and acquisition time of each plane, 3 minutes 48 seconds. Axial T1-weighted turbo spin-echo images (4-mm slice thickness) were acquired to assess the lymph nodes and the pelvic bone. Axial DWI was acquired using the single-shot echo-planar imaging technique with the following parameters: TR/TE, 4400–4800/63–75; slice thickness, 3 mm; interslice gap, 1 mm; matrix, 112 × 112–110; FOV, 20 cm; sensitivity encoding factor, 2; number of signals acquired, 4; number of slices, 20; acquisition time, within 1 minute 40 seconds. Diffusion-encoding gradients were applied at b values of 0 and 1000 s/mm 2, along the three orthogonal directions of the motion-probing gradients. The direction of the phase-encoding gradient was from left to right to minimize motion artifacts. Apparent diffusion coefficient (ADC) maps were automatically constructed using the manufacturer’s software. Axial DCE-MRI was obtained using a 3D fastfield echo sequence (TR/TE, 7.4/3.9; flip angle, 25°; matrix, 224 × 179; slice thickness, 4 mm; interslice gap, no; number of signals acquired, 1; FOV, 20 cm; and 11 partitions on a 3D slab). DCEMRI was performed from the apex to the base of the prostate. The 3D volume with 11 partitions was acquired every 5 seconds with 58 repetitions, or every 3 seconds with 60 repetitions. A contrastenhanced series was performed immediately after a bolus injection of gadopentetate dimeglumine (Magnevist, Schering) or gadolinium diethylenetriamine pentaacetic acid (Gadovist, Schering) at a dose of 0.1 mmol/kg body weight and a rate of 2−3 mL/s using a power injector, followed by a 20-mL saline flush. All dynamic datasets were transferred to an independent workstation (ViewForum, Philips Healthcare). After time–signal intensity curves were first fit to a general exponential signal-enhancement model, dynamic images were converted to a five-parameter model: washin rate, washout rate, maximal enhancement, maximal relative enhancement, and time to peak [17]. The dynamic images of the five-parameter model on DCE-MRI were superimposed onto enhanced fast-field echo or T2-weighted images, resulting in color-coded parametric images.
Image Analysis Final inclusion n = 282
Biochemical recurrence absent n = 221
W460
Biochemical recurrence present n = 61
Fig. 1—Flowchart of study design.
All MR images were retrospectively reviewed in consensus by two radiologists (with 6 and 8 years of experience in prostate MRI, respectively) using our PACS (Centricity, GE Healthcare). Patients’ clinical, pathologic, and outcome results were blinded to minimize bias, although readers were aware that all patients had undergone RP for biopsy-proven prostate cancer.
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy By using established morphologic and functional findings of multiparametric prostate MRI [11, 18–22], the readers assigned a radiologically determined tumor stage that was based on the American Joint Committee on Cancer staging system [23]. T1 tumors were not apparent on MRI, and T2 tumors were visible on MRI but were organ confined. T3a and T3b tumors extended outside the prostate capsule and into the seminal vesicles, respectively, and T4 tumors invaded the adjacent structures. The apparent tumor presence was assessed by the readers at three separate sessions, as follows: combined T2WI and DWI, combined T2WI and DCE-MRI, and combined T2WI, DWI, and DCE-MRI in sequence. Each session was separated by 1 month to avoid learning bias. An apparent tumor on combined T2WI, DWI, and DCE-MRI was regarded as present if it was visible on both DWI and DCE-MRI. To determine the probability for the presence of apparent tumor, a 5-point grading scale was adopted, as follows: 1, definitely invisible; 2, probably invisible; 3, possibly visible; 4, probably visible; and 5, definitely visible. Any lesions marked grade 3–5 on the MR images of each combination were regarded as apparent tumors. Tumor size was defined as the longest axial diameter measured on the ADC map image on the combined T2WI and DWI session and also on the acquired image of DCE-MRI on the combined T2WI and DCE-MRI session using the electronic caliper function of our PACS workstation. If there was not an apparent tumor, the size was recorded as zero. A prostate cancer was defined on each MRI as follows [11, 18–22]: a masslike lesion with abnormally low signal intensity compared with that of normal prostate tissue on T2WI, except for an area of high signal intensity on T1-weighted imaging suggestive of postbiopsy hemorrhage; a focal lesion with high signal intensity and a b value of 1000 s/mm2 on DWI, with a low ADC value on the ADC map; or a focal lesion with the presence of an asymmetric wash-in rate, washout rate, maximal enhancement, maximal relative enhancement or time-to-peak on color-coded parametric image, or early arterial enhancement with or without early washout on DCE-MRI. The presence of ECE and SVI was also determined on combined T2WI and DWI and DCE-MRI, using the same grading scale. The ECE was considered present if a tumor revealed asymmetric neurovascular bundle thickening, an irregular bulge in the prostate contour, measurable extracapsular disease, or obliteration of the rectoprostatic angle on T2WI or high wash-in rate or maximal enhancement with or without washout rate at DCE-MRI. The SVI was considered present if a tumor showed contiguous low-signal-intensity extension from the base
TABLE 1: Demographic Characteristics and Postoperative Findings for 282 Patients Who Underwent Radical Prostatectomy Variable Age (y), median ± SD (range) Baseline PSA level (ng/mL), median ± SD (range) Gleason score at biopsy, median (range)
Value 64 ± 7.2 (38–88) 6.59 ± 6.84 (0.26–56.98) 7 (6–9)
Clinical stage (no. of patients) T1c
187
T2a
70
T2b
16
T2c
2
T3a
7
Postoperative findings Gleason score at surgical specimen, median (range)
7 (6–9)
Extracapsular extension, no. (%) of patients
68 (24)
Seminal vesicle invasion, no. (%) of patients
22 (8)
Lymph node metastasis, no. (%) of patients
2 (0.7)
Distant metastasis, no. of patients
0
Note—PSA = prostate-specific antigen.
of the gland into the seminal vesicles or asymmetric decrease in the signal intensity of the seminal vesicles with mass effect on T2WI, or revealed high wash-in rate or maximal enhancement with or without washout rate in the seminal vesicles on DCE-MRI or had low ADC value in the seminal vesicles on DWI.
Statistical Analysis Statistical analysis was performed using PASW statistical software (version 18.0, SPSS). Univariate and multivariate Cox regression analyses were performed for all clinical and imaging variables to determine their relationship with biochemical recurrence. The clinical variables included clinical stage, baseline PSA level, and Gleason score at biopsy. The imaging variables were apparent tumor presence at each combination of MRI sequences, ECE, and tumor size on DWI or DCE-MRI. We did not include SVI as a predictive MRI variable because we observed very few cases (n = 9) on MRI. Two-tailed tests were used to calculate all p values; p < 0.05 was considered statistically significant.
Results After a median follow-up of 26 ± 8.5 months (range, 12–48 months), 61 (22%) of 282 patients had biochemical recurrence after RP. The median interval between biochemical recurrence and RP was 14 ± 4.6 months (range, 3–32 months). The patients’ clinical characteristics and postoperative surgical findings are summarized in Table 1.
The number of patients who exhibited apparent prostate cancer at each combined interpretation session was as follows: 187 patients (66%) on combined T2WI and DWI session, 185 (65%) on combined T2WI and DCE-MRI session, and 145 (51%) on combined T2WI, DWI, and DCE-MRI sessions. The mean tumor size was 10.3 ± 6.7 mm (range, 0–44 mm) on DWI and 11.3 ± 7.2 mm (range, 0–51 mm) on DCE-MRI. The ECE and SVI were shown on MRI for 96 (34%) and 9 (3%) patients, respectively. Table 2 presents the results of univariate and multivariate Cox regression analyses of clinical and imaging variables for predicting biochemical recurrence after RP. On univariate Cox regression analysis, all of the clinical and imaging variables were significantly related to biochemical recurrence (p < 0.01) (Figs. 2 and 3). Multivariate Cox regression analysis revealed that the only significant predictors of biochemical recurrence after RP were baseline PSA level (p = 0.002), Gleason score at biopsy (p = 0.024), and apparent tumor presence on combined T2WI, DWI, and DCE-MRI (p = 0.047). Of the independent predictors, the hazard ratio (HR) of apparent tumor presence on combined T2WI, DWI, and DCE-MRI (HR = 2.38) was higher compared with that of the other predictors (HR = 1.05 and 1.34 for baseline PSA level and Gleason score at biopsy, respectively). ECE at MRI showed a statisti-
AJR:202, May 2014 W461
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. TABLE 2: Results of Univariate and Multivariate Analyses for MRI and Clinical Variables in Predicting Biochemical Recurrence in 282 Patients Who Underwent Radical Prostatectomy
Discussion Our results show that all MRI variables (apparent tumor presence on combined imaging sequences, ECE, and tumor size on DCE-MRI or DWI) and all clinical variables (baseline PSA level, clinical stage, and Gleason score at biopsy) were significantly associated with biochemical recurrence after RP in univariate
analysis. Of the MRI variables, apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI was the only independent indicator in predicting biochemical recurrence in the multivariate analysis. In addition, clinical variables of baseline PSA level and Gleason score at biopsy were also independent indicators of biochemical recurrence in the multivariate analysis. These findings suggest that pretreatment multiparametric MRI findings at 3 T, as a promising indicator, might have the potential to predict biochemi-
cal recurrence after RP in patients with localized prostate cancer. Considering the promising role of multiparametric MRI findings in assessing the detection, localization, staging, or aggressiveness of prostate cancer, we hypothesized that the presence or absence of apparent tumor at each combination of multiparametric MRI sequences might be a significant predictor of patient outcome, because radiologically invisible cancers may represent low tumor burden or tumor with low Gleason score and, thus, a cancer with a lower T stage and better prognosis. Similarly, a previous study [24] found that sparse prostate cancer shows relatively similar signal intensity on T2WI and ADC values similar to those of normal peripheral zone tissue compared with dense tumors. Accordingly, this result suggests that there is a relationship between the histologic features of prostate cancer and visibility on MRI [24]. In the present study, on univariate analysis, apparent tumor presence on combined T2WI and DWI, combined T2WI and DCE-MRI, and combined T2WI, DWI, and DCE-MRI showed a statistical significance in predicting biochemical recurrence after RP. However, on multivariate analysis, apparent tumor presence on combined T2WI, DWI, and DCE-MRI was the only marker predictive of biochemical recurrence after RP. These findings suggest that the combination of T2WI, DWI, and DCE-MRI might improve the accuracy for assessing tumor extent. Although statistical significance was not found, the HR
A
B
C
Univariate Analysis Variable
Multivariate Analysis
HR (95% CI)
p
HR (95% CI)
p
Apparent tumor on T2WI and DWI
4.23 (1.92–9.29)
< 0.001
1.71 (0.49–5.93)
0.358
Apparent tumor on T2WI and DCE-MRI
1.57 (1.17–2.11)
0.003
1.00 (0.46–2.20)
0.989
Apparent tumor on T2WI and DWI and DCE-MRI
4.90 (2.55–9.41)
< 0.001
2.38 (1.03–6.00)
0.047
Tumor size on DWI
1.05 (1.03–1.07)
< 0.001
0.97 (0.91–1.03)
0.360
Tumor size on DCE-MRI
1.04 (1.02–1.06)
< 0.001
1.00 (0.95–1.06)
0.937
Extracapsular extension at MRI
4.00 (2.37–6.74)
< 0.001
1.81 (0.93–3.54)
0.081
Clinical stage
1.71 (1.05–1.10)
< 0.001
1.08 (0.80–1.46)
0.630
Pretreatment PSA level
1.07 (1.05–1.10)
< 0.001
1.05 (1.02–1.08)
0.002
Gleason score at biopsy
1.69 (1.38–2.07)
< 0.001
1.34 (1.05–1.72)
0.024
Note—HR = hazard ratio, T2WI = T2-weighted imaging, DWI= diffusion-weighted imaging, DCE-MRI = dynamic contrast-enhanced MRI, PSA= prostate-specific antigen.
cally borderline significance for predicting biochemical recurrence on multivariate Cox analysis (p = 0.081).
Fig. 2—73-year-old man with pathologic stage T3b prostate cancer, Gleason score of 9 at surgical specimen, and baseline prostate-specific antigen (PSA) level of 12.05 ng/mL. A, Axial T2-weighted turbo spin-echo image shows focal low-signal-intensity lesion that is suspicious for tumor in right peripheral zone. Lesion is associated with irregular contour bulge (arrowhead) and obliteration of rectoprostatic angle at right peripheral zone, representing possible extracapsular extension. B, Axial apparent diffusion coefficient (ADC) map image also shows focal mass (arrows) with decreased ADC value, which is highly suspicious for prostate cancer in right peripheral zone, associated with focal contour bulge (arrowhead) in same location as in (A) representing possibility of extracapsular extension. C, Parametric image of wash-in rate shows asymmetric color-coded lesion (arrows) in site corresponding with site of lesion in (A), associated with focal contour bulge (arrowhead) representing possibility of extracapsular extension. At 5 months after surgery, serum PSA level was elevated to 0.51 ng/mL, which confirmed biochemical recurrence.
W462
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy
A
B
C
Fig. 3—75-year-old man with pathologic stage T2a prostate cancer who had 1.6-cm lesion on left peripheral zone with Gleason score of 7 at surgical specimen and baseline prostate-specific antigen level of 4.18 ng/mL. A and B, On axial T2-weighted turbo spin-echo (A) and axial apparent diffusion coefficient map image (B), there is no focal lesion in prostate. C, Parametric image of wash-in rate shows no asymmetric color-coded lesions (arrows) in left peripheral zone. Patient had no biochemical recurrence for 4 years after surgery.
of the apparent tumor presence on combined T2WI and DWI was higher than the HR on combined T2WI and DCE-MRI. According to a recent study [25], combined T2WI and DWI showed a better diagnostic performance than did combined T2WI and DCEMRI for the detection of prostate cancer. Furthermore, the ADC value of prostate cancer also showed better performance to identify peripheral zone tumors compared with the quantitative parameters derived from DCE-MRI, such as volume transfer constant and extravascular extracellular volume fraction [10]. These findings suggest that DWI has superior diagnostic value for the detection of prostate cancer, compared with DCEMRI, which was consistent with our study. A recent study [26] found that a tumor’s being inapparent at 1.5-T endorectal MRI and MR spectroscopic imaging was not a favorable predictive factor in patients with prostate cancer who select active surveillance, suggesting that the distinction between stage T1 and stage T2 prostate cancer is of questionable prognostic value. Although prostate MRI using an endorectal coil may provide higher resolution and signal-to-noise ratio than the use of phased-array body coil, that study [26] did not include DWI and DCEMRI in the assessment of prostate cancer. The present study performed T2WI, DWI, and DCE-MRI at 3 T, which is a minimal requirement for the multiparametric MRI protocol [21] to detect and localize prostate cancer. Furthermore, another study with a larger population and longer follow-up period showed the relationship between the presence of suspicious prostate lesions on MRI
and increased risk for subsequent upgrade of Gleason score during active surveillance [27]. According to our results, both univariate and multivariate analyses found that apparent tumor presence at each combination of multiparametric MRI sequences was an independent predictor for biochemical recurrence after RP. A recent study [15] reported that tumor apparent presence on DWI before RP was an important factor to predict biochemical recurrence. Park et al. [28] also suggested that a low ADC value of prostate cancer was a significant predictor of biochemical recurrence after RP. This result can be explained by a possible relationship between tumor ADC and aggressiveness. Therefore, aggressive prostate cancer, which has the potential for biochemical recurrence after RP, may be easily distinguished from normal prostate tissue on DWI, possibly explaining our results. Furthermore, another recent study [29] found that MR spectroscopic findings, such as index lesion volume and high-grade voxels, were correlated with biochemical recurrence after RP. These findings suggest that multiparametric MRI findings have the potential to predict biochemical recurrence in patients who undergo RP. Further research is needed to predict patient outcome after treatment of localized prostate cancer using multiparametric MRI. Several previous studies have assessed the utility of MRI and MR spectroscopic imaging to predict biochemical recurrence of prostate cancer after RP [8, 30, 31]. These results suggest that MRI and MR spectroscopic imaging findings of ECE are indepen-
dent significant predictors of biochemical recurrence. More recently, a study reported that the ECE findings of combined T2WI and DWI were a significant predictor of biochemical recurrence after RP [15]. However, our study did not have the statistical power to reveal the prognostic significance of ECE at multiparametric MRI on multivariate analysis (p = 0.081). In our study, an endorectal coil was not used for 3-T MRI, which might have affected the accuracy of ECE prediction. A further study should be performed using endorectal multiparametric MRI at 3 T. To date, several preoperative factors have been identified as predictors of biochemical recurrence after RP in patients with prostate cancer [7]. In our study, the clinical variables of baseline PSA level and Gleason score at biopsy revealed a statistical significance to predict biochemical recurrence after RP in univariate and multivariate analyses, and these clinical variables were independently predictive markers for biochemical recurrence after RP. These results are consistent with those of previous studies [6, 32, 33]. Furthermore, compared with those well-known predictors, the presence of visible tumors on combined multiparametric MRI showed an even higher HR in multivariate analysis. These results suggest that preoperative multiparametric MRI findings may have a significant role in the prediction of patient prognosis. Moreover, it may be incorporated into a prognostic model for prostate cancer to stratify the patients for individualized treatment plans (i.e., selection of active surveillance or watchful waiting, nonsurgical treatment, or RP). In
AJR:202, May 2014 W463
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. addition, our findings are remarkable because even a negative finding for biopsyproven prostate cancer on preoperative MRI has clinical significance regarding a favorable outcome after RP. Our study has several limitations. First, this study was performed retrospectively at a single institution with a relatively small population (n = 282). It is likely that there was selection bias in our study design. Further investigation with a larger number of subjects may reveal more statistically significant pretreatment multiparametric MRI variables for predicting biochemical recurrence after RP. Second, the follow-up period was relatively short (median, 26 months). However, predicting early biochemical recurrence is clinically meaningful, because approximately 60% of biochemical recurrences happen within 2 years after RP [34] and early recurrence may represent an unfavorable prognosis in the natural history of prostate cancer after treatment [5, 35]. Nonetheless, further studies should be conducted with a longer follow-up period to evaluate prostate cancer–specific mortality and recurrence-free survival rates in association with findings of pretreatment multiparametric MRI. Third, a completely blinded study was impossible because it was known which patients underwent RP for biopsy-proven prostate cancer. To minimize bias, other medical information was withheld and image analysis was done in random order with a sufficient time interval. Fourth, our study is limited by the lack of pathologic correlation with MRI findings related to apparent tumor presence, although the purpose of this study was to evaluate an apparent tumor presence determined by the radiologist on multiparametric MRI as an imaging predictor for biochemical recurrence. Fifth, we did not include quantitative imaging parameters derived from DWI or DCE-MRI in our analysis. Although the primary aim of our study was to evaluate the relationship between the apparent tumor presence and biochemical recurrence, exploring quantitative parameters might be necessary to strengthen the results of our study. Finally, although MR spectroscopic imaging findings could help to predict biochemical recurrence, we did not perform MR spectroscopic imaging because it was not routinely performed at our institution because of the relatively long scan time. In conclusion, apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI is an independent predictor of
W464
biochemical recurrence after RP. This finding may be used to construct a predictive model for biochemical recurrence after surgery. References 1. Heidenreich A, Bellmunt J, Bolla M, et al.; European Association of Urology. EAU guidelines on prostate cancer. Part 1. Screening, diagnosis, and treatment of clinically localised disease. Eur Urol 2011; 59:61–71 2. Bill-Axelson A, Holmberg L, Ruutu M, et al.; Scandinavian Prostate Cancer Group Study No. 4. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2005; 352:1977–1984 3. Suardi N, Porter CR, Reuther AM, et al. A nomogram predicting long-term biochemical recurrence after radical prostatectomy. Cancer 2008; 112:1254–1263 4. Stephenson AJ, Scardino PT, Eastham JA, et al. Postoperative nomogram predicting the 10-year probability of prostate cancer recurrence after radical prostatectomy. J Clin Oncol 2005; 23:7005–7012 5. Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progression after PSA elevation following radical prostatectomy. JAMA 1999; 281:1591–1597 6. Freedland SJ, Humphreys EB, Mangold LA, et al. Risk of prostate cancer-specific mortality following biochemical recurrence after radical prostatectomy. JAMA 2005; 294:433–439 7. Shariat SF, Karakiewicz PI, Roehrborn CG, Kattan MW. An updated catalog of prostate cancer predictive tools. Cancer 2008; 113:3075–3099 8. Poulakis V, Witzsch U, de Vries R, et al. Preoperative neural network using combined magnetic resonance imaging variables, prostate-specific antigen, and Gleason score for predicting prostate cancer biochemical recurrence after radical prostatectomy. Urology 2004; 64:1165–1170 9. D’Amico AV, Whittington R, Malkowicz B, et al. Endorectal magnetic resonance imaging as a predictor of biochemical outcome after radical prostatectomy in men with clinically localized prostate cancer. J Urol 2000; 164:759–763 10. Langer DL, van der Kwast TH, Evans AJ, Trachtenberg J, Wilson BC, Haider MA. Prostate cancer detection with multi-parametric MRI: logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast-enhanced MRI. J Magn Reson Imaging 2009; 30:327–334 11. Fütterer JJ, Heijmink SW, Scheenen TW, et al. Prostate cancer localization with dynamic contrast-enhanced MR imaging and proton MR spectroscopic imaging. Radiology 2006; 241:449–458 12. Tempany CM, Zhou X, Zerhouni EA, et al. Staging of prostate cancer: results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques. Radiology 1994;
192:47–54 13. Wang L, Mazaheri Y, Zhang J, Ishill NM, Kuroiwa K, Hricak H. Assessment of biologic aggressiveness of prostate cancer: correlation of MR signal intensity with Gleason grade after radical prostatectomy. Radiology 2008; 246:168–176 14. 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 2011; 197:1382–1390 15. Nishida K, Yuen S, Kamoi K, et al. Incremental value of T2-weighted and diffusion-weighted MRI for prediction of biochemical recurrence after radical prostatectomy in clinically localized prostate cancer. Acta Radiol 2011; 52:120–126 16. Cookson MS, Aus G, Burnett AL, et al. Variation in the definition of biochemical recurrence in patients treated for localized prostate cancer: the American Urological Association Prostate Guidelines for Localized Prostate Cancer Update Panel report and recommendations for a standard in the reporting of surgical outcomes. J Urol 2007; 177:540–545 17. Kim CK, Park BK, Lee HM, Kim SS, Kim E. MRI techniques for prediction of local tumor progression after high-intensity focused ultrasonic ablation of prostate cancer. AJR 2008; 190:1180–1186 18. Kim CK, Choi D, Park BK, Kwon GY, Lim HK. Diffusion-weighted MR imaging for the evaluation of seminal vesicle invasion in prostate cancer: initial results. J Magn Reson Imaging 2008; 28:963–969 19. 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 2007; 42:842–847 20. Hricak H, Choyke PL, Eberhardt SC, Leibel SA, Scardino PT. Imaging prostate cancer: a multidisciplinary perspective. Radiology 2007; 243:28–53 21. Barentsz JO, Richenberg J, Clements R, et al.; European Society of Urogenital Radiology. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22:746–757 22. Choi YJ, Kim JK, Kim N, Kim KW, Choi EK, Cho KS. Functional MR imaging of prostate cancer. RadioGraphics 2007; 27:63–75; discussion, 75–67 23. Edge S, Byrd DR, Compton CC, Fritz AG, Greene FL, Trott A, eds. AJCC cancer staging manual, 7th ed. New York, NY: Springer Healthcare, 2010 24. Langer DL, van der Kwast TH, Evans AJ, et al. Intermixed normal tissue within prostate cancer: effect on MR imaging measurements of apparent diffusion coefficient and T2—sparse versus dense cancers. Radiology 2008; 249:900–908
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy 25. Delongchamps NB, Rouanne M, Flam T, et al. Multiparametric magnetic resonance imaging for the detection and localization of prostate cancer: combination of T2-weighted, dynamic contrastenhanced and diffusion-weighted imaging. BJU Int 2011; 107:1411–1418 26. Cabrera AR, Coakley FV, Westphalen AC, et al. Prostate cancer: is inapparent tumor at endorectal MR and MR spectroscopic imaging a favorable prognostic finding in patients who select active surveillance? Radiology 2008; 247:444–450 27. Fradet V, Kurhanewicz J, Cowan JE, et al. Prostate cancer managed with active surveillance: role of anatomic MR imaging and MR spectroscopic imaging. Radiology 2010; 256:176–183 28. 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 2011; 21:1111–1118 29. Zakian KL, Hricak H, Ishill N, et al. An exploratory study of endorectal magnetic resonance imaging and spectroscopy of the prostate as preoperative predictive biomarkers of biochemical relapse after radical prostatectomy. J Urol 2010; 184:2320–2327 30. Shukla-Dave A, Hricak H, Ishill N, et al. Prediction of prostate cancer recurrence using magnetic resonance imaging and molecular profiles. Clin Cancer Res 2009; 15:3842–3849 31. Fuchsjäger MH, Shukla-Dave A, Hricak H, et al. Magnetic resonance imaging in the prediction of biochemical recurrence of prostate cancer after radical prostatectomy. BJU Int 2009; 104:315–320 32. D’Amico AV, Whittington R, Malkowicz SB, et al. Utilizing predictions of early prostate-specific
antigen failure to optimize patient selection for adjuvant systemic therapy trials. J Clin Oncol 2000; 18:3240–3246 33. Moul JW, Connelly RR, Lubeck DP, et al. Predicting risk of prostate specific antigen recurrence after radical prostatectomy with the Center for Prostate Disease Research and Cancer of the Prostate Strategic Urologic Research Endeavor databases. J Urol 2001; 166:1322–1327 34. Walz J, Chun FK, Klein EA, et al. Nomogram predicting the probability of early recurrence after radical prostatectomy for prostate cancer. J Urol 2009; 181:601–607; discussion, 607–608 35. Freedland SJ, Humphreys EB, Mangold LA, Eisenberger M, Partin AW. Time to prostate specific antigen recurrence after radical prostatectomy and risk of prostate cancer specific mortality. J Urol 2006; 176:1404–1408
AJR:202, May 2014 W465
Genitourinar y Imaging • Original Research Park et al. MRI to Predict Biochemical Recurrence After Radical Prostatectomy Genitourinary Imaging Original Research
Prostate Cancer: Role of Pretreatment Multiparametric 3-T MRI in Predicting Biochemical Recurrence After Radical Prostatectomy Jung Jae Park1,2 Chan Kyo Kim1 Sung Yoon Park1 Byung Kwan Park1 Hyun Moo Lee 3 Seong Whi Cho2 Park JJ, Kim CK, Park SY, Park BK, Lee HM, Cho SW
Keywords: multiparametric MRI, prediction, prostatectomy, prostate neoplasm, recurrence DOI:10.2214/AJR.13.11381 Received June 14, 2013; accepted after revision August 28, 2013. 1 Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea. Address correspondence to C. K. Kim ([email protected]). 2 Department of Radiology, Kangwon National University Hospital, Chuncheon, Republic of Korea. 3 Department of Urology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea.
WEB This is a web exclusive article. AJR 2014; 202:W459–W465 0361–803X/14/2025–W459 © American Roentgen Ray Society
OBJECTIVE. The purpose of this study is to retrospectively investigate whether pretreatment multiparametric MRI findings can predict biochemical recurrence in patients who underwent radical prostatectomy (RP) for localized prostate cancer. MATERIALS AND METHODS. In this study, 282 patients with biopsy-proven prostate cancer who received RP underwent pretreatment MRI using a phased-array coil at 3 T, including T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced MRI (DCE-MRI). MRI variables included apparent tumor presence on combined imaging sequences, extracapsular extension, and tumor size on DWI or DCE-MRI. Clinical variables included baseline prostate-specific antigen (PSA) level, clinical stage, and Gleason score at biopsy. The relationship between clinical and imaging variables and biochemical recurrence was evaluated using Cox regression analysis. RESULTS. After a median follow-up of 26 months, biochemical recurrence developed in 61 patients (22%). Univariate analysis revealed that all the imaging and clinical variables were significantly associated with biochemical recurrence (p < 0.01). On multivariate analysis, however, baseline PSA level (p = 0.002), Gleason score at biopsy (p = 0.024), and apparent tumor presence on combined T2WI, DWI, and DCE-MRI (p = 0.047) were the only significant independent predictors of biochemical recurrence. Of the independent predictors, apparent tumor presence on combined T2WI, DWI, and DCE-MRI showed the highest hazard ratio (2.38) compared with baseline PSA level (hazard ratio, 1.05) and Gleason score at biopsy (hazard ratio, 1.34). CONCLUSION. The apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI is an independent predictor of biochemical recurrence after RP. This finding may be used to construct a predictive model for biochemical recurrence after surgery.
R
adical prostatectomy (RP) is an established treatment modality that reduces disease-specific mortality for patients with localized prostate cancer [1, 2], although biochemical recurrence occurs in quite a few patients after surgery. According to a postoperative nomogram predicting long-term outcomes, the 10-year actuarial rate of biochemical recurrence–free survival was approximately 79%, but biochemical recurrence may occur beyond 10 years after RP [3, 4]. Because biochemical recurrence is associated with progression to distant metastases and cancer-specific mortality [5, 6], a prediction of biochemical recurrence before surgery is clinically important in determining primary treatment options, adjuvant therapies, and patient counseling.
To date, several preoperative and postoperative clinical variables have been used to predict biochemical recurrence after RP, but the accuracies of each are limited [7]. The inclusion of novel biomarkers or imaging tools may potentially improve the predictive accuracy. As MRI evolves into a valuable diagnostic method in prostate imaging, pretreatment MRI findings also have been investigated to improve predictive accuracy for biochemical recurrence after primary local treatment. The T stage and the presence of extracapsular extension (ECE) and seminal vesicle invasion (SVI), as determined by MRI, are significant prognostic parameters for predicting biochemical recurrence after primary local treatment, such as RP [8, 9]. More recently, it was found that emerging multiparametric prostate MRI can provide
AJR:202, May 2014 W459
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. functional information along with anatomic information; thus, it has contributed to tumor detection, localization, local staging, and determination of aggressiveness [10–14]. This promising technique may offer incremental value to a prognostic model using clinical variables to predict biochemical recurrence in prostate cancer after RP, with improved accuracy. However, few studies have reported the prognostic value of pretreatment multiparametric MRI findings in predicting biochemical recurrence after RP [8, 15]. The purpose of our study was to retrospectively investigate whether pretreatment multiparametric MRI findings at 3 T, including T2-weighted imaging (T2WI), diffusion-weighted (DWI), and dynamic contrastenhanced MRI (DCE-MRI), can predict biochemical recurrence in patients who undergo RP for localized prostate cancer.
insufficient follow-up to determine biochemical recurrence (n = 22). Five patients were also excluded because of a 3-month or longer interval between pretreatment MRI and surgery. Finally, 282 men (mean [± SD] age, 64 ± 7.2 years; range, 38–88 years) were enrolled in this study (Fig. 1). The mean interval between biopsy and MRI was 28.2 ± 4.9 days (range, 22–50 days), and surgery was performed 3–35 days after MRI (mean, 16 ± 6.2 days). The clinical stage was evaluated with digital rectal examination. Before surgery (mean, 2 ± 2.4 days; range, 1–28 days), baseline PSA level was identified in all patients. After surgery, the patients were followed-up with digital rectal examination and serum PSA level within 1 month, then every 3 months for the first 2 years, and every 6 months from the third year. Biochemical recurrence after RP was defined as an initial serum PSA level greater than or equal to 0.2 ng/mL, with a second confirmatory level of PSA of greater than 0.2 ng/mL [16].
Materials and Methods Patients
Imaging Technique
This retrospective study was approved by our institutional review board, with a waiver of the requirement for informed consent. Between May 2007 and April 2009, 435 consecutive men with biopsy-proven clinically localized prostate cancer were treated with RP. Of these, 372 patients underwent preoperative prostate MRI at 3 T, including T2WI, DWI, and DCE-MRI, in our institution. Of the 372 patients, 85 were excluded for the following reasons: hormone or radiation treatment before surgery (n = 29); imaging artifact or missing sequence (n = 18); missing data, such as Gleason score at biopsy or baseline prostate-specific antigen (PSA) level (n = 16); and
All patients underwent prostate MRI at 3 T (Intera Achieva 3 T or 3TX, Philips Healthcare System) using a phased-array coil. To reduce bowel peristalsis, 20 mg of butyl scopolamine (Buscopan, Boehringer Ingelheim) was administered intramuscularly before imaging. The MRI protocol was composed of routine T1-weighted imaging, T2WI, DWI, and DCE-MRI. After obtaining three plane localizer images, T2-weighted turbo spin-echo images were obtained in the axial, sagittal, and coronal planes. The imaging parameters were as follows: TR/TE, 3800–4700/80– 100; slice thickness, 3 mm; interslice gap, 1 mm; matrix, 512 × 304 or 568 × 341; FOV, 18–20 cm;
Patients who underwent radical prostatectomy n = 435
Preoperative MRI n = 372 Excluded patients n = 90
number of signals acquired, 3; sensitivity encoding factor, 2; number of slices, 21; and acquisition time of each plane, 3 minutes 48 seconds. Axial T1-weighted turbo spin-echo images (4-mm slice thickness) were acquired to assess the lymph nodes and the pelvic bone. Axial DWI was acquired using the single-shot echo-planar imaging technique with the following parameters: TR/TE, 4400–4800/63–75; slice thickness, 3 mm; interslice gap, 1 mm; matrix, 112 × 112–110; FOV, 20 cm; sensitivity encoding factor, 2; number of signals acquired, 4; number of slices, 20; acquisition time, within 1 minute 40 seconds. Diffusion-encoding gradients were applied at b values of 0 and 1000 s/mm 2, along the three orthogonal directions of the motion-probing gradients. The direction of the phase-encoding gradient was from left to right to minimize motion artifacts. Apparent diffusion coefficient (ADC) maps were automatically constructed using the manufacturer’s software. Axial DCE-MRI was obtained using a 3D fastfield echo sequence (TR/TE, 7.4/3.9; flip angle, 25°; matrix, 224 × 179; slice thickness, 4 mm; interslice gap, no; number of signals acquired, 1; FOV, 20 cm; and 11 partitions on a 3D slab). DCEMRI was performed from the apex to the base of the prostate. The 3D volume with 11 partitions was acquired every 5 seconds with 58 repetitions, or every 3 seconds with 60 repetitions. A contrastenhanced series was performed immediately after a bolus injection of gadopentetate dimeglumine (Magnevist, Schering) or gadolinium diethylenetriamine pentaacetic acid (Gadovist, Schering) at a dose of 0.1 mmol/kg body weight and a rate of 2−3 mL/s using a power injector, followed by a 20-mL saline flush. All dynamic datasets were transferred to an independent workstation (ViewForum, Philips Healthcare). After time–signal intensity curves were first fit to a general exponential signal-enhancement model, dynamic images were converted to a five-parameter model: washin rate, washout rate, maximal enhancement, maximal relative enhancement, and time to peak [17]. The dynamic images of the five-parameter model on DCE-MRI were superimposed onto enhanced fast-field echo or T2-weighted images, resulting in color-coded parametric images.
Image Analysis Final inclusion n = 282
Biochemical recurrence absent n = 221
W460
Biochemical recurrence present n = 61
Fig. 1—Flowchart of study design.
All MR images were retrospectively reviewed in consensus by two radiologists (with 6 and 8 years of experience in prostate MRI, respectively) using our PACS (Centricity, GE Healthcare). Patients’ clinical, pathologic, and outcome results were blinded to minimize bias, although readers were aware that all patients had undergone RP for biopsy-proven prostate cancer.
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy By using established morphologic and functional findings of multiparametric prostate MRI [11, 18–22], the readers assigned a radiologically determined tumor stage that was based on the American Joint Committee on Cancer staging system [23]. T1 tumors were not apparent on MRI, and T2 tumors were visible on MRI but were organ confined. T3a and T3b tumors extended outside the prostate capsule and into the seminal vesicles, respectively, and T4 tumors invaded the adjacent structures. The apparent tumor presence was assessed by the readers at three separate sessions, as follows: combined T2WI and DWI, combined T2WI and DCE-MRI, and combined T2WI, DWI, and DCE-MRI in sequence. Each session was separated by 1 month to avoid learning bias. An apparent tumor on combined T2WI, DWI, and DCE-MRI was regarded as present if it was visible on both DWI and DCE-MRI. To determine the probability for the presence of apparent tumor, a 5-point grading scale was adopted, as follows: 1, definitely invisible; 2, probably invisible; 3, possibly visible; 4, probably visible; and 5, definitely visible. Any lesions marked grade 3–5 on the MR images of each combination were regarded as apparent tumors. Tumor size was defined as the longest axial diameter measured on the ADC map image on the combined T2WI and DWI session and also on the acquired image of DCE-MRI on the combined T2WI and DCE-MRI session using the electronic caliper function of our PACS workstation. If there was not an apparent tumor, the size was recorded as zero. A prostate cancer was defined on each MRI as follows [11, 18–22]: a masslike lesion with abnormally low signal intensity compared with that of normal prostate tissue on T2WI, except for an area of high signal intensity on T1-weighted imaging suggestive of postbiopsy hemorrhage; a focal lesion with high signal intensity and a b value of 1000 s/mm2 on DWI, with a low ADC value on the ADC map; or a focal lesion with the presence of an asymmetric wash-in rate, washout rate, maximal enhancement, maximal relative enhancement or time-to-peak on color-coded parametric image, or early arterial enhancement with or without early washout on DCE-MRI. The presence of ECE and SVI was also determined on combined T2WI and DWI and DCE-MRI, using the same grading scale. The ECE was considered present if a tumor revealed asymmetric neurovascular bundle thickening, an irregular bulge in the prostate contour, measurable extracapsular disease, or obliteration of the rectoprostatic angle on T2WI or high wash-in rate or maximal enhancement with or without washout rate at DCE-MRI. The SVI was considered present if a tumor showed contiguous low-signal-intensity extension from the base
TABLE 1: Demographic Characteristics and Postoperative Findings for 282 Patients Who Underwent Radical Prostatectomy Variable Age (y), median ± SD (range) Baseline PSA level (ng/mL), median ± SD (range) Gleason score at biopsy, median (range)
Value 64 ± 7.2 (38–88) 6.59 ± 6.84 (0.26–56.98) 7 (6–9)
Clinical stage (no. of patients) T1c
187
T2a
70
T2b
16
T2c
2
T3a
7
Postoperative findings Gleason score at surgical specimen, median (range)
7 (6–9)
Extracapsular extension, no. (%) of patients
68 (24)
Seminal vesicle invasion, no. (%) of patients
22 (8)
Lymph node metastasis, no. (%) of patients
2 (0.7)
Distant metastasis, no. of patients
0
Note—PSA = prostate-specific antigen.
of the gland into the seminal vesicles or asymmetric decrease in the signal intensity of the seminal vesicles with mass effect on T2WI, or revealed high wash-in rate or maximal enhancement with or without washout rate in the seminal vesicles on DCE-MRI or had low ADC value in the seminal vesicles on DWI.
Statistical Analysis Statistical analysis was performed using PASW statistical software (version 18.0, SPSS). Univariate and multivariate Cox regression analyses were performed for all clinical and imaging variables to determine their relationship with biochemical recurrence. The clinical variables included clinical stage, baseline PSA level, and Gleason score at biopsy. The imaging variables were apparent tumor presence at each combination of MRI sequences, ECE, and tumor size on DWI or DCE-MRI. We did not include SVI as a predictive MRI variable because we observed very few cases (n = 9) on MRI. Two-tailed tests were used to calculate all p values; p < 0.05 was considered statistically significant.
Results After a median follow-up of 26 ± 8.5 months (range, 12–48 months), 61 (22%) of 282 patients had biochemical recurrence after RP. The median interval between biochemical recurrence and RP was 14 ± 4.6 months (range, 3–32 months). The patients’ clinical characteristics and postoperative surgical findings are summarized in Table 1.
The number of patients who exhibited apparent prostate cancer at each combined interpretation session was as follows: 187 patients (66%) on combined T2WI and DWI session, 185 (65%) on combined T2WI and DCE-MRI session, and 145 (51%) on combined T2WI, DWI, and DCE-MRI sessions. The mean tumor size was 10.3 ± 6.7 mm (range, 0–44 mm) on DWI and 11.3 ± 7.2 mm (range, 0–51 mm) on DCE-MRI. The ECE and SVI were shown on MRI for 96 (34%) and 9 (3%) patients, respectively. Table 2 presents the results of univariate and multivariate Cox regression analyses of clinical and imaging variables for predicting biochemical recurrence after RP. On univariate Cox regression analysis, all of the clinical and imaging variables were significantly related to biochemical recurrence (p < 0.01) (Figs. 2 and 3). Multivariate Cox regression analysis revealed that the only significant predictors of biochemical recurrence after RP were baseline PSA level (p = 0.002), Gleason score at biopsy (p = 0.024), and apparent tumor presence on combined T2WI, DWI, and DCE-MRI (p = 0.047). Of the independent predictors, the hazard ratio (HR) of apparent tumor presence on combined T2WI, DWI, and DCE-MRI (HR = 2.38) was higher compared with that of the other predictors (HR = 1.05 and 1.34 for baseline PSA level and Gleason score at biopsy, respectively). ECE at MRI showed a statisti-
AJR:202, May 2014 W461
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. TABLE 2: Results of Univariate and Multivariate Analyses for MRI and Clinical Variables in Predicting Biochemical Recurrence in 282 Patients Who Underwent Radical Prostatectomy
Discussion Our results show that all MRI variables (apparent tumor presence on combined imaging sequences, ECE, and tumor size on DCE-MRI or DWI) and all clinical variables (baseline PSA level, clinical stage, and Gleason score at biopsy) were significantly associated with biochemical recurrence after RP in univariate
analysis. Of the MRI variables, apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI was the only independent indicator in predicting biochemical recurrence in the multivariate analysis. In addition, clinical variables of baseline PSA level and Gleason score at biopsy were also independent indicators of biochemical recurrence in the multivariate analysis. These findings suggest that pretreatment multiparametric MRI findings at 3 T, as a promising indicator, might have the potential to predict biochemi-
cal recurrence after RP in patients with localized prostate cancer. Considering the promising role of multiparametric MRI findings in assessing the detection, localization, staging, or aggressiveness of prostate cancer, we hypothesized that the presence or absence of apparent tumor at each combination of multiparametric MRI sequences might be a significant predictor of patient outcome, because radiologically invisible cancers may represent low tumor burden or tumor with low Gleason score and, thus, a cancer with a lower T stage and better prognosis. Similarly, a previous study [24] found that sparse prostate cancer shows relatively similar signal intensity on T2WI and ADC values similar to those of normal peripheral zone tissue compared with dense tumors. Accordingly, this result suggests that there is a relationship between the histologic features of prostate cancer and visibility on MRI [24]. In the present study, on univariate analysis, apparent tumor presence on combined T2WI and DWI, combined T2WI and DCE-MRI, and combined T2WI, DWI, and DCE-MRI showed a statistical significance in predicting biochemical recurrence after RP. However, on multivariate analysis, apparent tumor presence on combined T2WI, DWI, and DCE-MRI was the only marker predictive of biochemical recurrence after RP. These findings suggest that the combination of T2WI, DWI, and DCE-MRI might improve the accuracy for assessing tumor extent. Although statistical significance was not found, the HR
A
B
C
Univariate Analysis Variable
Multivariate Analysis
HR (95% CI)
p
HR (95% CI)
p
Apparent tumor on T2WI and DWI
4.23 (1.92–9.29)
< 0.001
1.71 (0.49–5.93)
0.358
Apparent tumor on T2WI and DCE-MRI
1.57 (1.17–2.11)
0.003
1.00 (0.46–2.20)
0.989
Apparent tumor on T2WI and DWI and DCE-MRI
4.90 (2.55–9.41)
< 0.001
2.38 (1.03–6.00)
0.047
Tumor size on DWI
1.05 (1.03–1.07)
< 0.001
0.97 (0.91–1.03)
0.360
Tumor size on DCE-MRI
1.04 (1.02–1.06)
< 0.001
1.00 (0.95–1.06)
0.937
Extracapsular extension at MRI
4.00 (2.37–6.74)
< 0.001
1.81 (0.93–3.54)
0.081
Clinical stage
1.71 (1.05–1.10)
< 0.001
1.08 (0.80–1.46)
0.630
Pretreatment PSA level
1.07 (1.05–1.10)
< 0.001
1.05 (1.02–1.08)
0.002
Gleason score at biopsy
1.69 (1.38–2.07)
< 0.001
1.34 (1.05–1.72)
0.024
Note—HR = hazard ratio, T2WI = T2-weighted imaging, DWI= diffusion-weighted imaging, DCE-MRI = dynamic contrast-enhanced MRI, PSA= prostate-specific antigen.
cally borderline significance for predicting biochemical recurrence on multivariate Cox analysis (p = 0.081).
Fig. 2—73-year-old man with pathologic stage T3b prostate cancer, Gleason score of 9 at surgical specimen, and baseline prostate-specific antigen (PSA) level of 12.05 ng/mL. A, Axial T2-weighted turbo spin-echo image shows focal low-signal-intensity lesion that is suspicious for tumor in right peripheral zone. Lesion is associated with irregular contour bulge (arrowhead) and obliteration of rectoprostatic angle at right peripheral zone, representing possible extracapsular extension. B, Axial apparent diffusion coefficient (ADC) map image also shows focal mass (arrows) with decreased ADC value, which is highly suspicious for prostate cancer in right peripheral zone, associated with focal contour bulge (arrowhead) in same location as in (A) representing possibility of extracapsular extension. C, Parametric image of wash-in rate shows asymmetric color-coded lesion (arrows) in site corresponding with site of lesion in (A), associated with focal contour bulge (arrowhead) representing possibility of extracapsular extension. At 5 months after surgery, serum PSA level was elevated to 0.51 ng/mL, which confirmed biochemical recurrence.
W462
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy
A
B
C
Fig. 3—75-year-old man with pathologic stage T2a prostate cancer who had 1.6-cm lesion on left peripheral zone with Gleason score of 7 at surgical specimen and baseline prostate-specific antigen level of 4.18 ng/mL. A and B, On axial T2-weighted turbo spin-echo (A) and axial apparent diffusion coefficient map image (B), there is no focal lesion in prostate. C, Parametric image of wash-in rate shows no asymmetric color-coded lesions (arrows) in left peripheral zone. Patient had no biochemical recurrence for 4 years after surgery.
of the apparent tumor presence on combined T2WI and DWI was higher than the HR on combined T2WI and DCE-MRI. According to a recent study [25], combined T2WI and DWI showed a better diagnostic performance than did combined T2WI and DCEMRI for the detection of prostate cancer. Furthermore, the ADC value of prostate cancer also showed better performance to identify peripheral zone tumors compared with the quantitative parameters derived from DCE-MRI, such as volume transfer constant and extravascular extracellular volume fraction [10]. These findings suggest that DWI has superior diagnostic value for the detection of prostate cancer, compared with DCEMRI, which was consistent with our study. A recent study [26] found that a tumor’s being inapparent at 1.5-T endorectal MRI and MR spectroscopic imaging was not a favorable predictive factor in patients with prostate cancer who select active surveillance, suggesting that the distinction between stage T1 and stage T2 prostate cancer is of questionable prognostic value. Although prostate MRI using an endorectal coil may provide higher resolution and signal-to-noise ratio than the use of phased-array body coil, that study [26] did not include DWI and DCEMRI in the assessment of prostate cancer. The present study performed T2WI, DWI, and DCE-MRI at 3 T, which is a minimal requirement for the multiparametric MRI protocol [21] to detect and localize prostate cancer. Furthermore, another study with a larger population and longer follow-up period showed the relationship between the presence of suspicious prostate lesions on MRI
and increased risk for subsequent upgrade of Gleason score during active surveillance [27]. According to our results, both univariate and multivariate analyses found that apparent tumor presence at each combination of multiparametric MRI sequences was an independent predictor for biochemical recurrence after RP. A recent study [15] reported that tumor apparent presence on DWI before RP was an important factor to predict biochemical recurrence. Park et al. [28] also suggested that a low ADC value of prostate cancer was a significant predictor of biochemical recurrence after RP. This result can be explained by a possible relationship between tumor ADC and aggressiveness. Therefore, aggressive prostate cancer, which has the potential for biochemical recurrence after RP, may be easily distinguished from normal prostate tissue on DWI, possibly explaining our results. Furthermore, another recent study [29] found that MR spectroscopic findings, such as index lesion volume and high-grade voxels, were correlated with biochemical recurrence after RP. These findings suggest that multiparametric MRI findings have the potential to predict biochemical recurrence in patients who undergo RP. Further research is needed to predict patient outcome after treatment of localized prostate cancer using multiparametric MRI. Several previous studies have assessed the utility of MRI and MR spectroscopic imaging to predict biochemical recurrence of prostate cancer after RP [8, 30, 31]. These results suggest that MRI and MR spectroscopic imaging findings of ECE are indepen-
dent significant predictors of biochemical recurrence. More recently, a study reported that the ECE findings of combined T2WI and DWI were a significant predictor of biochemical recurrence after RP [15]. However, our study did not have the statistical power to reveal the prognostic significance of ECE at multiparametric MRI on multivariate analysis (p = 0.081). In our study, an endorectal coil was not used for 3-T MRI, which might have affected the accuracy of ECE prediction. A further study should be performed using endorectal multiparametric MRI at 3 T. To date, several preoperative factors have been identified as predictors of biochemical recurrence after RP in patients with prostate cancer [7]. In our study, the clinical variables of baseline PSA level and Gleason score at biopsy revealed a statistical significance to predict biochemical recurrence after RP in univariate and multivariate analyses, and these clinical variables were independently predictive markers for biochemical recurrence after RP. These results are consistent with those of previous studies [6, 32, 33]. Furthermore, compared with those well-known predictors, the presence of visible tumors on combined multiparametric MRI showed an even higher HR in multivariate analysis. These results suggest that preoperative multiparametric MRI findings may have a significant role in the prediction of patient prognosis. Moreover, it may be incorporated into a prognostic model for prostate cancer to stratify the patients for individualized treatment plans (i.e., selection of active surveillance or watchful waiting, nonsurgical treatment, or RP). In
AJR:202, May 2014 W463
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
Park et al. addition, our findings are remarkable because even a negative finding for biopsyproven prostate cancer on preoperative MRI has clinical significance regarding a favorable outcome after RP. Our study has several limitations. First, this study was performed retrospectively at a single institution with a relatively small population (n = 282). It is likely that there was selection bias in our study design. Further investigation with a larger number of subjects may reveal more statistically significant pretreatment multiparametric MRI variables for predicting biochemical recurrence after RP. Second, the follow-up period was relatively short (median, 26 months). However, predicting early biochemical recurrence is clinically meaningful, because approximately 60% of biochemical recurrences happen within 2 years after RP [34] and early recurrence may represent an unfavorable prognosis in the natural history of prostate cancer after treatment [5, 35]. Nonetheless, further studies should be conducted with a longer follow-up period to evaluate prostate cancer–specific mortality and recurrence-free survival rates in association with findings of pretreatment multiparametric MRI. Third, a completely blinded study was impossible because it was known which patients underwent RP for biopsy-proven prostate cancer. To minimize bias, other medical information was withheld and image analysis was done in random order with a sufficient time interval. Fourth, our study is limited by the lack of pathologic correlation with MRI findings related to apparent tumor presence, although the purpose of this study was to evaluate an apparent tumor presence determined by the radiologist on multiparametric MRI as an imaging predictor for biochemical recurrence. Fifth, we did not include quantitative imaging parameters derived from DWI or DCE-MRI in our analysis. Although the primary aim of our study was to evaluate the relationship between the apparent tumor presence and biochemical recurrence, exploring quantitative parameters might be necessary to strengthen the results of our study. Finally, although MR spectroscopic imaging findings could help to predict biochemical recurrence, we did not perform MR spectroscopic imaging because it was not routinely performed at our institution because of the relatively long scan time. In conclusion, apparent tumor presence on combined T2WI, DWI, and DCE-MRI of pretreatment MRI is an independent predictor of
W464
biochemical recurrence after RP. This finding may be used to construct a predictive model for biochemical recurrence after surgery. References 1. Heidenreich A, Bellmunt J, Bolla M, et al.; European Association of Urology. EAU guidelines on prostate cancer. Part 1. Screening, diagnosis, and treatment of clinically localised disease. Eur Urol 2011; 59:61–71 2. Bill-Axelson A, Holmberg L, Ruutu M, et al.; Scandinavian Prostate Cancer Group Study No. 4. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2005; 352:1977–1984 3. Suardi N, Porter CR, Reuther AM, et al. A nomogram predicting long-term biochemical recurrence after radical prostatectomy. Cancer 2008; 112:1254–1263 4. Stephenson AJ, Scardino PT, Eastham JA, et al. Postoperative nomogram predicting the 10-year probability of prostate cancer recurrence after radical prostatectomy. J Clin Oncol 2005; 23:7005–7012 5. Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progression after PSA elevation following radical prostatectomy. JAMA 1999; 281:1591–1597 6. Freedland SJ, Humphreys EB, Mangold LA, et al. Risk of prostate cancer-specific mortality following biochemical recurrence after radical prostatectomy. JAMA 2005; 294:433–439 7. Shariat SF, Karakiewicz PI, Roehrborn CG, Kattan MW. An updated catalog of prostate cancer predictive tools. Cancer 2008; 113:3075–3099 8. Poulakis V, Witzsch U, de Vries R, et al. Preoperative neural network using combined magnetic resonance imaging variables, prostate-specific antigen, and Gleason score for predicting prostate cancer biochemical recurrence after radical prostatectomy. Urology 2004; 64:1165–1170 9. D’Amico AV, Whittington R, Malkowicz B, et al. Endorectal magnetic resonance imaging as a predictor of biochemical outcome after radical prostatectomy in men with clinically localized prostate cancer. J Urol 2000; 164:759–763 10. Langer DL, van der Kwast TH, Evans AJ, Trachtenberg J, Wilson BC, Haider MA. Prostate cancer detection with multi-parametric MRI: logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast-enhanced MRI. J Magn Reson Imaging 2009; 30:327–334 11. Fütterer JJ, Heijmink SW, Scheenen TW, et al. Prostate cancer localization with dynamic contrast-enhanced MR imaging and proton MR spectroscopic imaging. Radiology 2006; 241:449–458 12. Tempany CM, Zhou X, Zerhouni EA, et al. Staging of prostate cancer: results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques. Radiology 1994;
192:47–54 13. Wang L, Mazaheri Y, Zhang J, Ishill NM, Kuroiwa K, Hricak H. Assessment of biologic aggressiveness of prostate cancer: correlation of MR signal intensity with Gleason grade after radical prostatectomy. Radiology 2008; 246:168–176 14. 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 2011; 197:1382–1390 15. Nishida K, Yuen S, Kamoi K, et al. Incremental value of T2-weighted and diffusion-weighted MRI for prediction of biochemical recurrence after radical prostatectomy in clinically localized prostate cancer. Acta Radiol 2011; 52:120–126 16. Cookson MS, Aus G, Burnett AL, et al. Variation in the definition of biochemical recurrence in patients treated for localized prostate cancer: the American Urological Association Prostate Guidelines for Localized Prostate Cancer Update Panel report and recommendations for a standard in the reporting of surgical outcomes. J Urol 2007; 177:540–545 17. Kim CK, Park BK, Lee HM, Kim SS, Kim E. MRI techniques for prediction of local tumor progression after high-intensity focused ultrasonic ablation of prostate cancer. AJR 2008; 190:1180–1186 18. Kim CK, Choi D, Park BK, Kwon GY, Lim HK. Diffusion-weighted MR imaging for the evaluation of seminal vesicle invasion in prostate cancer: initial results. J Magn Reson Imaging 2008; 28:963–969 19. 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 2007; 42:842–847 20. Hricak H, Choyke PL, Eberhardt SC, Leibel SA, Scardino PT. Imaging prostate cancer: a multidisciplinary perspective. Radiology 2007; 243:28–53 21. Barentsz JO, Richenberg J, Clements R, et al.; European Society of Urogenital Radiology. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22:746–757 22. Choi YJ, Kim JK, Kim N, Kim KW, Choi EK, Cho KS. Functional MR imaging of prostate cancer. RadioGraphics 2007; 27:63–75; discussion, 75–67 23. Edge S, Byrd DR, Compton CC, Fritz AG, Greene FL, Trott A, eds. AJCC cancer staging manual, 7th ed. New York, NY: Springer Healthcare, 2010 24. Langer DL, van der Kwast TH, Evans AJ, et al. Intermixed normal tissue within prostate cancer: effect on MR imaging measurements of apparent diffusion coefficient and T2—sparse versus dense cancers. Radiology 2008; 249:900–908
AJR:202, May 2014
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 05/31/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
MRI to Predict Biochemical Recurrence After Radical Prostatectomy 25. Delongchamps NB, Rouanne M, Flam T, et al. Multiparametric magnetic resonance imaging for the detection and localization of prostate cancer: combination of T2-weighted, dynamic contrastenhanced and diffusion-weighted imaging. BJU Int 2011; 107:1411–1418 26. Cabrera AR, Coakley FV, Westphalen AC, et al. Prostate cancer: is inapparent tumor at endorectal MR and MR spectroscopic imaging a favorable prognostic finding in patients who select active surveillance? Radiology 2008; 247:444–450 27. Fradet V, Kurhanewicz J, Cowan JE, et al. Prostate cancer managed with active surveillance: role of anatomic MR imaging and MR spectroscopic imaging. Radiology 2010; 256:176–183 28. 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 2011; 21:1111–1118 29. Zakian KL, Hricak H, Ishill N, et al. An exploratory study of endorectal magnetic resonance imaging and spectroscopy of the prostate as preoperative predictive biomarkers of biochemical relapse after radical prostatectomy. J Urol 2010; 184:2320–2327 30. Shukla-Dave A, Hricak H, Ishill N, et al. Prediction of prostate cancer recurrence using magnetic resonance imaging and molecular profiles. Clin Cancer Res 2009; 15:3842–3849 31. Fuchsjäger MH, Shukla-Dave A, Hricak H, et al. Magnetic resonance imaging in the prediction of biochemical recurrence of prostate cancer after radical prostatectomy. BJU Int 2009; 104:315–320 32. D’Amico AV, Whittington R, Malkowicz SB, et al. Utilizing predictions of early prostate-specific
antigen failure to optimize patient selection for adjuvant systemic therapy trials. J Clin Oncol 2000; 18:3240–3246 33. Moul JW, Connelly RR, Lubeck DP, et al. Predicting risk of prostate specific antigen recurrence after radical prostatectomy with the Center for Prostate Disease Research and Cancer of the Prostate Strategic Urologic Research Endeavor databases. J Urol 2001; 166:1322–1327 34. Walz J, Chun FK, Klein EA, et al. Nomogram predicting the probability of early recurrence after radical prostatectomy for prostate cancer. J Urol 2009; 181:601–607; discussion, 607–608 35. Freedland SJ, Humphreys EB, Mangold LA, Eisenberger M, Partin AW. Time to prostate specific antigen recurrence after radical prostatectomy and risk of prostate cancer specific mortality. J Urol 2006; 176:1404–1408
AJR:202, May 2014 W465
