Dynamic Contrast-enhanced MR Imaging Features of the Normal Central Zone of the Prostate Barry G. Hansford, MD, Ibrahim Karademir, MD, Yahui Peng, PhD, Yulei Jiang, PhD, Gregory Karczmar, PhD, Stephen Thomas, MD, Ambereen Yousuf, MBBS, Tatjana Antic, MD, Scott Eggener, MD, Aytekin Oto, MD Rationale and Objectives: Evaluate qualitative dynamic contrast-enhanced magnetic resonance imaging (MRI) characteristics of normal central zone based on recently described central zone MRI features. Materials and Methods: Institutional review board–approved, Health Insurance Portability and Accountability Act compliant study, 59 patients with prostate cancer, histopathology proven to not involve central zone or prostate base, underwent endorectal MRI before prostatectomy. Two readers independently reviewed T2-weighted images and apparent diffusion coefficient (ADC) maps identifying normal central zone based on low signal intensity and location. Next, two readers drew bilateral central zone regions of interest on dynamic contrast-enhanced magnetic resonance images in consensus and independently recorded enhancement curve types as type 1 (progressive), type 2 (plateau), and type 3 (wash-out). Identification rates of normal central zone and enhancement curve type were recorded and compared for each reviewer. The institutional review board waiver was approved and granted 05/2010. Results: Central zone identified in 92%–93% of patients on T2-weighted images and 78%–88% on ADC maps without significant difference between identification rates (P = .63 and P = .15 and inter-reader agreement (k) is 0.64 and 0.29, for T2-weighted images and ADC maps, respectively). All central zones were rated either curve type 1 or curve type 2 by both radiologists. No statistically significant difference between the two radiologists (P = .19) and inter-reader agreement was k = 0.37. Conclusions: Normal central zone demonstrates either type 1 (progressive) or type 2 (plateau) enhancement curves on dynamic contrastenhanced MRI that can be potentially useful to differentiate central zone from prostate cancer that classically demonstrates a type 3 (washout) enhancement curve. Key Words: Prostate cancer; central zone; dynamic contrast-enhanced magnetic resonance imaging. ªAUR, 2014
M
cNeal’s model of prostate anatomy is now over a quarter century old and widely accepted by pathologists, urologists, and radiologists who are involved in the management of prostate cancer (PCa) (1–3). According to McNeal, the prostate can be divided into three glandular zones (peripheral zone [PZ], transition zone [TZ], and central zone [CZ]) and one nonglandular zone (anterior fibromuscular stroma) based on histology, anatomic location, and embryologic features (4). TZ refers to the bilat-
Acad Radiol 2014; 21:569–577 From the Department of Radiology, University of Chicago Medicine, 5841 South Maryland Ave, MC2026, Chicago, IL 60637 (B.G.H., I.K., Y.P., Y.J., G.K., S.T., A.Y., A.O.); Department of Pathology (T.A.); and Department of Urology, University of Chicago Medicine, Chicago, IL (S.E.). Received October 21, 2013; accepted January 22, 2014. Disclosure: Neither B.G.H. nor his immediate family members have a financial relationship with a commercial organization that may have a direct or indirect interest in the content. This work was supported in part by the US Army Medical Research and Materiel Command Prostate Cancer Research Program through an Idea Development Award (PC093485). Address correspondence to: B.G.H. e-mail: [email protected] ªAUR, 2014 http://dx.doi.org/10.1016/j.acra.2014.01.013
eral regions in the mid to base of the gland, which form two bulges on either side of the urethra that extend superiorly, anteriorly, and laterally from the verumontanum. The CZ is a flattened conical structure located posterior to the TZ and surrounds the ejaculatory ducts (Fig 1). Embryologically, CZ derives from the Wolffian duct whereas TZ is a derivative of the urogenital sinus (5). The CZ accounts for 25% of the total prostate volume, but almost 40% of the epithelium because of its high epithelial-tostromal ratio (6,7). It is most prominent at the base of the prostate and has a conical shape, extending inferiorly down to the level of verumontanum and surrounds the ejaculatory ducts (Fig 1). The glands in the CZ are large (about twice the size of those of the PZ) and have irregular contours where epithelial-covered stromal ridges project into the gland lumens (8). McNeal described the epithelial cells of CZ as being columnar, crowded, with somewhat darker granular cytoplasm than those of the PZ (8). CZ glands show unique Roman bridge architecture with the formation of intraglandular lacunae (7,9,10). The secretory profile of CZ glands is also unique and includes cytoplasmic lactoferrin, tissue 569
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
there are no data in the literature about DCE-MRI characteristics of normal CZ. Thus, the purposes of this study were to determine the detection rate of normal CZ based on the recently described MRI features and to describe the qualitative DCE-MRI characteristics of normal CZ. MATERIALS AND METHODS Study Patients
Figure 1. Coronal prostate diagram adapted from McNeal’s model of prostate anatomy demonstrating the conical-shaped CZ extending from the base of the prostate to the level of the verumontanum. CZ, central zone; DU, distal urethra; E, ejaculatory duct; PZ, peripheral zone; V = verumontanum.
plasminogen activator, endothelin-1, and numerous lipochrome pigments (7,9,11). With these features, CZ differs from the PZ and TZ and is part way between the features of seminal vesicles and the epithelium of the PZ and TZ (12,13). Despite these distinct characteristics of the CZ and TZ and data from early anatomic studies indicating that the CZ could be identified distinctly from the TZ on T2-weighted magnetic resonance (MR) images (14,15), many publications in the MR literature have commonly grouped CZ and TZ together as the ‘‘central gland’’ (16,17). Recently, Vargas et al. (18) reported that CZ could be distinguished from other prostate zones in most (81%–84%) patients, based on its location, low signal intensity on T2-weighted images, and apparent diffusion coefficient (ADC) maps. Unfortunately, T2-weighted and diffusion-weighted MR (DW-MR) image features of normal CZ overlap with the features of PCa, and thus CZ can be an important pitfall which mimics PCa, especially at the posterior base, where the prostate is comprised mostly of CZ (18). As suggested by Vargas et al. (18), one possible means for separating normal CZ from PCa on magnetic resonance imaging (MRI) is dynamic contrast-enhanced (DCE) MRI. The European Society of Urogenital Radiology (ESUR) guidelines for standardized prostate MRI reporting suggest that two functional MRI techniques together with T2-weighted images provide better characterization than T2-weighted images plus one functional technique (19). These guidelines identified DCE-MRI enhancement curve type as a criterion for scoring DCE-MR images in the tabulation of the Prostate Imaging Reporting and Data System (PI-RADS) score indicating the likelihood of significant PCa (19). Unfortunately, as mentioned by Vargas et al. (18), 570
This retrospective study was conducted with an institutional review board–approved waiver of informed consent and was in compliance with the Health Insurance Portability and Accountability Act. We searched clinical records at our institution to identify consecutive patients who underwent preoperative endorectal multiparametric MRI of the prostate between January 2008 and June 2009 who subsequently underwent radical prostatectomy for treatment of PCa. Preoperative multiparametric endorectal MRI (1.5T or 3T) of the prostate was available in 82 patients. An experienced genitourinary (GU) pathologist (7 years of experience in GU pathology) reviewed histologic specimens of the prostate, and 19 patients (23%) were excluded because of either tumor involvement in a portion of the CZ or presence of cancer at the base of the prostate (Fig 2). Of the remaining 63 patients, four were excluded because of DW-MR image artifacts that affected the prostate base. The final study cohort consisted of 59 patients: mean age, 59.9 years; age standard deviation (SD), 7.0; age range, 43–72; average serum prostate-specific antigen (PSA) level, 8.7 mL/ng; PSA SD, 8.0; and PSA range, 1.7– 40.9. Of this final cohort of 59 patients, eight whole-mount specimens were evaluated with the GU pathologist to evaluate for concordance of MRI findings of CZ with CZ histology. MRI Protocols
All MR examinations were performed with an endorectal coil (Medrad, Warrendale, PA) in combination with a phased array surface coil in 1.5Tor 3T scanners (33 were performed on 3T, 26 were performed on 1.5T) (Excite HD; GE Healthcare, Waukesha, WI, n = 33; Achieva; Philips Healthcare, Eindhoven, The Netherlands, n = 26). Immediately before MR examination, 1 mg glucagon (Lilly, Indianapolis, IN) was injected intramuscularly. We imaged the entire prostate and oriented axial images to be perpendicular to the rectal wall guided by sagittal images, and a parallel imaging factor of 2 was used in all sequences. The following axial, coronal, and sagittal images were obtained: T2-weighted fast spin-echo (FSE) (slice thickness 3 mm), axial T1 FSE, axial freebreathing DW imaging (b = 0, 1000 and 1500 s/mm2), and axial free-breathing DCE-MRI. Field of view was 14–18 cm, and resolution was 0.8 0.8 3 mm. Acquisition of T1-weighted DCE-MRI images (of the entire prostate) started 30 seconds before intravenous administration of 0.1 mmol/kg gadodiamide (Omniscan; GE Healthcare,
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
Figure 2. Flowchart demonstrating stepwise exclusion of patients in the study cohort. ADC, apparent diffusion coefficient; CZ, central zone; DCE-MR, dynamic contrast-enhanced magnetic resonance.
Princeton, NJ) followed by a 20-mL saline flush at a rate of 2.0 mL/s. Temporal resolution range for DCE-MRI was 5– 12 seconds 5–7 minutes for GE, and (3–5 seconds for 7–9 minutes) for Philips. Please see the Appendix for complete imaging protocols. MRI Analysis
Analysis of T2-weighted images and ADC maps. Two radiologists, A and B (9 and 3 years of prostate MRI experience, respectively), retrospectively evaluated the prostate MRI studies using a picture archiving and communication system (Stentor; Phillips Healthcare) with coregistration software. ADC maps were generated from DW-MR images using commercial software (Advantage Windows, version 4.2.3; GE Healthcare; or Philips ViewForum; Philips Healthcare). The two radiologists independently evaluated T2-weighted images (axial and coronal), using methodology similar to that described by Vargas et al., to identify normal CZ (10). First, they identified the ejaculatory ducts and verumontanum. Second, they determined whether normal CZ could be distinguished as a hypointense region between the prostate base and the verumontanum surrounding the ejaculatory ducts. And last, they reviewed ADC maps in conjunction with T2-weighted images and recorded whether the CZ could be identified as a low-signal region in its expected location. Coregistration with T2-weighted images and anatomic landmarks, such as the ejaculatory ducts, was used to help identify the CZ on ADC maps (Figs 3–5). DCE-MRI Analysis
Strictly qualitative DCE-MRI analysis was performed after the evaluation of T2-weighted images and ADC maps. Before
DCE-MRI image analysis, the two radiologists performed a consensus evaluation of T2-weighted images and ADC maps to agree in which patients the CZ could be identified. Next, DCE-MRI images were processed and reviewed on a commercial workstation, allowing coregistration between the different sequences (DynaCAD for Prostate; Invivo, Gainesville, FL). In cases that they could identify the CZ, the two radiologists placed, in consensus, regions of interest (ROIs) on T2-weighted images on both the left and right CZs, which was divided by a line that can be drawn to connect the urethra and verumontanum, to generate qualitative DCE time curves (signal intensity vs. time). With the help of the coregistration software, these ROIs were then transferred to DCE-MRI images, and DCE time curves were generated by the software for each ROI. Next, the radiologists independently determined the curve types in terms of type 1, progressive (enhancement increases over time); type 2, plateau (stable enhancement over time after possibly an initial increase); and type 3, wash-out (early and fast enhancement followed by decreasing enhancement over time) (20). Statistical Analysis
Recorded data included patient age, PSA level, identification of CZ on T2-weighted images (yes or no), identification of CZ on ADC maps (yes or no), and DCE time curve type (types 1–3). Per-patient identification rates of CZ (the left and right CZs were always concordant in terms of identifiable or not identifiable) on T2-weighted images and on ADC maps were calculated for each radiologist, together with binomial exact 95% confidence intervals (21). The binomial exact version of the McNemar test was used to evaluate whether differences between the detection rates of the two radiologists were significant statistically (22). In DCE time curve analysis, 571
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
Figure 3. A 47-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 1 (progressive) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
the left and right CZs of each patient were considered separately because their time curves could differ. Inter-reader agreement was evaluated with the kappa statistic, k, with the following interpretation of the k value: 0.00–0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81– 1.00, almost perfect agreement (23).
RESULTS Identification of Normal CZ on T2-weighted Images and ADC Maps
The CZ identification rates on T2-weighted images and ADC maps are listed in Table 1. The more experienced radiologist (A) detected more CZs on both T2-weighted images and ADC maps. However, these differences were not statistically significant (P = .63 for T2-weighted images, P = .15 for ADC maps). Inter-reader agreement, k, was 0.64 and 0.29 for T2-weighted images and ADC maps, respectively, indicating substantial and fair agreement, respectively (Tables 2 and 3). 572
DCE-MRI Analysis
Consensus evaluation between the two radiologists yielded a total of 52 cases for which the CZ could be identified on both T2-weighted imaging and ADC maps. These 52 cases (a total of 104 left and right CZs) were analyzed for DCE time curves (Table 4). Both radiologists identified all CZ time curves as either type 1 (Fig 3) or type 2 (Figs 4 and 5). Radiologist A: type 1 progressive enhancement (24/104 or 23% of curve types); type 2, plateau enhancement (80/104 or 77% of curve types); and type 3, wash-out (0/104 or 0% of curve types). Radiologist B: type 1 progressive enhancement (17/104 or 16% of curve types); type 2, plateau enhancement (87/104 or 84% of curve types); and type 3, wash-out (0/ 104 or 0% of curve types). There was no statistically significant difference between the two radiologists (P = .19). Inter-reader agreement, k, was 0.37, indicating fair agreement.
DISCUSSION Despite the widespread notion that the CZ cannot be visualized distinctly on MR images, our study supports the recent
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
Figure 4. A 57-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 2 (plateau) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
observation of Vargas et al. (18) that the CZ could be differentiated from other prostate zones on T2-weighted images and ADC maps. One of the early studies describing the prostate anatomy using a 0.3 Tesla MR scanner reported that the CZ was identifiable on T2-weighted images in 31 of 32 men, 25–35 years of age, and in 8 of 23 men, aged $40 years (6). Vargas et al reported the anatomic location (surrounding the ejaculatory ducts from the prostatic base to the verumontanum), left-and-right symmetry on coronal T2-weighted images, and homogenous low signal intensity on T2-weighted images and on ADC maps as the most common MR image features of the CZ. Using these criteria, we were able to identify CZ in 92%–93% of patients on T2weighted images and 78%–88% of patients on ADC maps (mean age, 59.9 years), with CZ identification rates not significantly different between the reviewers. Moderate agreement between observers of different experience levels in our study in the detection of the CZ on T2-weighted images suggests that readers with varying prostate MR experience may fairly reliably identify the CZ. Although the exact reason for relatively low signal intensity of CZ on T2-weighted images and ADC maps is not known, one hypothesis is that the cribriform protrusions of the thick walls of the CZ glands into the lumen forming ‘‘Roman bridge type architecture’’ limit the
relative volume of luminal fluid in the glands. Luminal glandular fluid contributes to high T2 signal and high ADC in normal PZ glands. This relative decrease of luminal fluid, together with a more compact stroma in the CZ, may contribute to the low signal of CZ on T2-weighted images and ADC maps (24). Given the similar MR signal characteristics on T2-weighted images and ADC maps between the CZ and PCa, differentiation of normal CZ from PCa based on these two sequences alone can be problematic, particularly in the region of the posterior prostatic base, an area composed almost entirely of CZ (4). Our DCE time curve analysis showed that normal CZ demonstrated either type 1 (progressive enhancement, 16–23% of CZs), or type 2 (plateau enhancement, 77–84% of CZs), but not type 3 (wash-out, 0% of CZs) time curves, the last of which is typically associated with PCa. DCE-MRI is an important component of the accepted multiparametric prostate MRI paradigm and has been shown to improve detection of clinically important prostate cancer (25–27). In the recently released ESUR prostate MR guidelines, one of the selected major criteria for tabulating a high PI-RADS score, indicative of a high risk of PCa, was the wash-out, type 3, enhancement pattern of a lesion on DCE-MRI, which underscores the importance of qualitative 573
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
Figure 5. A 64-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 2 (plateau) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
TABLE 1. CZ Identification Rates on T2-weighted Images and on ADC Maps
Modality T2-weighted ADC map
Radiologist
CZ Detection Rate (%) [95% Confidence Interval]
A B A B
93 (55/59) [84%–98%] 92 (54/59) [81%–97%] 88 (52/59) [77%–95%] 78 (46/59) [65%–88%]
ADC, apparent diffusion coefficient; CZ, central zone.
DCE-MRI (11). Given the absence of type 3 enhancement time curves in the normal CZ in our study, our findings suggest that qualitative DCE-MRI analysis can be a valuable tool in distinguishing between normal CZ and PCa. CZ cancers are rare, accounting for 0.5%–2.5% of all PCa and 3%–8% of index tumors (28–30). Their behavior is not well understood, partly because of their relatively low incidence and partly because of the general failure to recognize the correct zonal origin in routine pathologic reporting (26). They are usually associated with a high risk of adverse pathologic features (increased propensity for seminal vesicle invasion, high Gleason grade, and high rates of 574
TABLE 2. Agreement in CZ Identification on T2-weighted Images between the Two Radiologists B Radiologist A
Identified Not Total
Identified
Not
Total
53 1 54
2 3 5
55 4 59
CZ, central zone.
extracapsular extension) and early biochemical failure after curative surgery (10,31). Despite these important prognostic differences, because of lack of a clearly delineated boundary between PZ and CZ, (unlike the surgical capsule between TZ and PZ), most pathologists do not routinely recognize tumors as originating in the CZ (31). Given the difficulty of identifying a CZ of origin PCa on histology and the somewhat poorer prognostic implications of a CZ tumor, the identification of CZ involvement by PCa may be a clinically relevant finding that can be worthwhile to actively look for while interpreting MR images. Our study had several limitations. First, our study was subject to the inherent limitations of retrospective design and a
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
TABLE 3. Agreement in CZ Identification on ADC Maps between the Two Radiologists B Radiologist A
Identified
Not
Total
43 3 46
9 4 13
52 7 59
Identified Not Total
ADC, apparent diffusion coefficient; CZ, central zone.
TABLE 4. Agreement in DCE Time Curve Analysis between the Two Radiologists B Radiologist A
Type 1 Type 2 Type 3 Total
Type 1
Type 2
Type 3
Total
10 7 0 17
14 73 0 87
0 0 0 0
24 80 0 104
DCE, dynamic contrast-enhanced.
relatively small sample size. Second, our prostate examinations were performed on two different MR machines. This second limitation is probably not particularly problematic because varying temporal resolution in the DCE time curve analysis is less of an issue for qualitative analysis compared to quantitative analysis. Third, we did not use quantitative or semiquantitative DCE-MRI analysis. Because of a large variation in the reported values of quantitative parameters derived from MRI in the literature and in keeping with the recent ESUR prostate MR guidelines, we chose qualitative enhancement time curve analysis because it is the most readily accessible and straightforward methodology. In addition, qualitative DCE-MRI analysis has been used successfully for breast MRI. However, the inherent subjectivity of this qualitative method does make standardization among institutions more challenging and future research using quantitative and semiquantitative methodologies would be a valuable contribution to the literature. Fourth, we did not include any cases of CZ cancer in our study. Although type 3 time curves are known to be frequently associated with PCa regardless of the zonal origin, future studies specifically investigating the DCE-MRI features of CZ cancer will be helpful additions to the literature. And finally, we did not compare CZ histology on wholemount pathology inspection to the MRI findings of CZ for all 59 patients in the final cohort. We did review eight whole-mount prostate pathology specimens with a GU radiologist who confirmed the CZ findings on prostate MRI were concordant with CZ histology. For the remaining cases, we relied on previous radiology literature identifying the CZ on MRI, which was supported by whole-mount pathology evaluation for concordant CZ identification (14,18). In conclusion, our results support the previous observations by Vargas et al. that it is possible to distinguish the CZ from the surrounding prostate gland in the vast majority of patients
undergoing prostate MR based on its low signal on T2weighted images and on ADC maps. Our study shows that the normal CZ demonstrates either type 1 or type 2 enhancement time curves on DCE-MRI, which can be potentially used to differentiate the CZ from PCa.
REFERENCES 1. Villers A, Steg A, Boccon-Gibod L. Anatomy of the prostate: review of the different models. Eur Urol 1991; 20(4):261–268. 2. McNeal JE. Prostate. In: Mills SE, ed. Histology for the pathologist. 3rd edition. Philadelphia: Lippincott Williams and Wilkins, 2007; p923–p963. Chapter 36. 3. Jaffe WI, Kaplan SA, McConnell JD. Epidemiology and pathophysiology of benign prostatic hyperplasia. In: Scardino PT, Slawin KM, eds. Atlas of the prostate. 3rd edition. Hong Kong: Current Medicine LLC, 2006; p2–11. Chapter 1. 4. McNeal JE. The zonal anatomy of the prostate. The prostate 1981; 2: 35–49. 5. Quick CM, Gokden N, Sangoi AR, et al. The distribution of PAX-2 immunoreactivity in the prostate gland, seminal vesicle, and ejaculatory duct: comparison with prostatic adenocarcinoma and discussion of prostatic zonal embryogenesis. Hum Pathol 2010; 41(8):1145–1149. 6. McNeal JE. Normal histology of the prostate. Am J Surg Pathol 1988; 12(8): 619–633. 7. McNeal JE. Cancer volume and site of origin of adenocarcinoma in the prostate: relationship to local and distant spread. Hum Pathol 1992; 23(3):258–266. 8. McNeal JE. Normal and pathologic anatomy of the prostate. Urology 1981; 17(suppl 3):11–16. 9. Reese JH, McNeal JE, Goldenberg SL, et al. Distribution of lactoferrin in the normal and inflamed human prostate: an immunohistochemical study. Prostate 1992; 20(1):73–85. 10. Srodon M, Epstein JI. Central zone histology of the prostate: a mimicker of high grade prostatic intraepithelial neoplasia. Hum Pathol 2002; 33(5): 518–523. 11. Prayer-Galetti T, Rossi GP, Belloni AS, et al. Gene expression and autoradiographic localization of endothelin-1 and its receptors A and B in the different zones of the human prostate. J Urol 1997; 157(6): 2334–2339. 12. Lexander H, Franzen B, Hirschberg D, et al. Differential protein expression in anatomical zones of the prostate. Proteomics 2005; 5(10):2570–2576. 13. Laczko I, Hudson DI, Freeman A, Feneley MR, Masters JR. Comparison of the zones of the human prostate with the seminal vesicle: morphology, immunohistochemistry and cell kinetics. Prostate 2005; 62(3):260–266. 14. Hricak H, Dooms GC, McNeal JE, et al. MR imaging of the prostate gland: normal anatomy. AJR Am J Roentgenol 1987; 148(1):51–58. 15. Sommer FG, McNeal JE, Carrol CL. MR depiction of zonal anatomy of the prostate at 1.5T. J Comput Assist Tomogr 1986; 10(6):983–989. 16. Kim CK, Park BK, Lee HM, et al. 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(12):842–847. 17. Schiebler ML, Tomaszewski JE, Bezzi M, et al. Prostatic carcinoma and benign prostatic hyperplasia: correlation of high resolution MR and histopathologic findings. Radiology 1989; 172(1):131–137. 18. Vargas HA, Akin O, Franiel T, et al. Normal central zone of the prostate and central zone involvement by prostate cancer: clinical and MR imaging implications. Radiology 2012; 262(3):894–902. 19. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22(4):746–757. 20. Verma S, Turkbey B, Muradyan N, et al. Overview of dynamic contrastenhanced MRI in prostate cancer management. AJR Am J Roentgenol 2012; 198(6):1277–1288. 21. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174. 22. Zar JH. Biostatistical analysis. 4th ed. New Jersey: Prentice-Hall, 1999. 23. Riffenburgh RH. Statistics in medicine. 2nd ed. Amsterdam; Boston: Elsevier Academic Press, 2006; 274. 24. Bourne RM, Kurniawan N, Cowin G, et al. Microscopic diffusivity compartmentation in formalin-fixed prostate tissue. Magn Reson Med 2012; 68(2): 614–620.
575
HANSFORD ET AL
25. 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(1):248–254. 26. Van Dorsten FA, Van Der Graaf M, Engelbrecht MR, et al. Combined quantitative dynamic contrast-enhanced MR imaging and (1)H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imaging 2004; 20(2):279–287. 27. Padhani AR, Gapinksi CJ, Macvicar DA, et al. Dynamic contrast-enhanced MRI of prostate cancer: correlation with morphology and tumour stage, histologic grade and PSA. Clin Radiol 2000; 55(2):99–109.
576
Academic Radiology, Vol 21, No 5, May 2014
28. McNeal JE, Redwine EA, Freiha FS, et al. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. Am J Surg Pathol 1988; 12(12):897–906. 29. Cheng L, Jones TD, Pan CX, et al. Anatomic distribution and pathologic characterization of small-volume prostate cancer (
M
cNeal’s model of prostate anatomy is now over a quarter century old and widely accepted by pathologists, urologists, and radiologists who are involved in the management of prostate cancer (PCa) (1–3). According to McNeal, the prostate can be divided into three glandular zones (peripheral zone [PZ], transition zone [TZ], and central zone [CZ]) and one nonglandular zone (anterior fibromuscular stroma) based on histology, anatomic location, and embryologic features (4). TZ refers to the bilat-
Acad Radiol 2014; 21:569–577 From the Department of Radiology, University of Chicago Medicine, 5841 South Maryland Ave, MC2026, Chicago, IL 60637 (B.G.H., I.K., Y.P., Y.J., G.K., S.T., A.Y., A.O.); Department of Pathology (T.A.); and Department of Urology, University of Chicago Medicine, Chicago, IL (S.E.). Received October 21, 2013; accepted January 22, 2014. Disclosure: Neither B.G.H. nor his immediate family members have a financial relationship with a commercial organization that may have a direct or indirect interest in the content. This work was supported in part by the US Army Medical Research and Materiel Command Prostate Cancer Research Program through an Idea Development Award (PC093485). Address correspondence to: B.G.H. e-mail: [email protected] ªAUR, 2014 http://dx.doi.org/10.1016/j.acra.2014.01.013
eral regions in the mid to base of the gland, which form two bulges on either side of the urethra that extend superiorly, anteriorly, and laterally from the verumontanum. The CZ is a flattened conical structure located posterior to the TZ and surrounds the ejaculatory ducts (Fig 1). Embryologically, CZ derives from the Wolffian duct whereas TZ is a derivative of the urogenital sinus (5). The CZ accounts for 25% of the total prostate volume, but almost 40% of the epithelium because of its high epithelial-tostromal ratio (6,7). It is most prominent at the base of the prostate and has a conical shape, extending inferiorly down to the level of verumontanum and surrounds the ejaculatory ducts (Fig 1). The glands in the CZ are large (about twice the size of those of the PZ) and have irregular contours where epithelial-covered stromal ridges project into the gland lumens (8). McNeal described the epithelial cells of CZ as being columnar, crowded, with somewhat darker granular cytoplasm than those of the PZ (8). CZ glands show unique Roman bridge architecture with the formation of intraglandular lacunae (7,9,10). The secretory profile of CZ glands is also unique and includes cytoplasmic lactoferrin, tissue 569
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
there are no data in the literature about DCE-MRI characteristics of normal CZ. Thus, the purposes of this study were to determine the detection rate of normal CZ based on the recently described MRI features and to describe the qualitative DCE-MRI characteristics of normal CZ. MATERIALS AND METHODS Study Patients
Figure 1. Coronal prostate diagram adapted from McNeal’s model of prostate anatomy demonstrating the conical-shaped CZ extending from the base of the prostate to the level of the verumontanum. CZ, central zone; DU, distal urethra; E, ejaculatory duct; PZ, peripheral zone; V = verumontanum.
plasminogen activator, endothelin-1, and numerous lipochrome pigments (7,9,11). With these features, CZ differs from the PZ and TZ and is part way between the features of seminal vesicles and the epithelium of the PZ and TZ (12,13). Despite these distinct characteristics of the CZ and TZ and data from early anatomic studies indicating that the CZ could be identified distinctly from the TZ on T2-weighted magnetic resonance (MR) images (14,15), many publications in the MR literature have commonly grouped CZ and TZ together as the ‘‘central gland’’ (16,17). Recently, Vargas et al. (18) reported that CZ could be distinguished from other prostate zones in most (81%–84%) patients, based on its location, low signal intensity on T2-weighted images, and apparent diffusion coefficient (ADC) maps. Unfortunately, T2-weighted and diffusion-weighted MR (DW-MR) image features of normal CZ overlap with the features of PCa, and thus CZ can be an important pitfall which mimics PCa, especially at the posterior base, where the prostate is comprised mostly of CZ (18). As suggested by Vargas et al. (18), one possible means for separating normal CZ from PCa on magnetic resonance imaging (MRI) is dynamic contrast-enhanced (DCE) MRI. The European Society of Urogenital Radiology (ESUR) guidelines for standardized prostate MRI reporting suggest that two functional MRI techniques together with T2-weighted images provide better characterization than T2-weighted images plus one functional technique (19). These guidelines identified DCE-MRI enhancement curve type as a criterion for scoring DCE-MR images in the tabulation of the Prostate Imaging Reporting and Data System (PI-RADS) score indicating the likelihood of significant PCa (19). Unfortunately, as mentioned by Vargas et al. (18), 570
This retrospective study was conducted with an institutional review board–approved waiver of informed consent and was in compliance with the Health Insurance Portability and Accountability Act. We searched clinical records at our institution to identify consecutive patients who underwent preoperative endorectal multiparametric MRI of the prostate between January 2008 and June 2009 who subsequently underwent radical prostatectomy for treatment of PCa. Preoperative multiparametric endorectal MRI (1.5T or 3T) of the prostate was available in 82 patients. An experienced genitourinary (GU) pathologist (7 years of experience in GU pathology) reviewed histologic specimens of the prostate, and 19 patients (23%) were excluded because of either tumor involvement in a portion of the CZ or presence of cancer at the base of the prostate (Fig 2). Of the remaining 63 patients, four were excluded because of DW-MR image artifacts that affected the prostate base. The final study cohort consisted of 59 patients: mean age, 59.9 years; age standard deviation (SD), 7.0; age range, 43–72; average serum prostate-specific antigen (PSA) level, 8.7 mL/ng; PSA SD, 8.0; and PSA range, 1.7– 40.9. Of this final cohort of 59 patients, eight whole-mount specimens were evaluated with the GU pathologist to evaluate for concordance of MRI findings of CZ with CZ histology. MRI Protocols
All MR examinations were performed with an endorectal coil (Medrad, Warrendale, PA) in combination with a phased array surface coil in 1.5Tor 3T scanners (33 were performed on 3T, 26 were performed on 1.5T) (Excite HD; GE Healthcare, Waukesha, WI, n = 33; Achieva; Philips Healthcare, Eindhoven, The Netherlands, n = 26). Immediately before MR examination, 1 mg glucagon (Lilly, Indianapolis, IN) was injected intramuscularly. We imaged the entire prostate and oriented axial images to be perpendicular to the rectal wall guided by sagittal images, and a parallel imaging factor of 2 was used in all sequences. The following axial, coronal, and sagittal images were obtained: T2-weighted fast spin-echo (FSE) (slice thickness 3 mm), axial T1 FSE, axial freebreathing DW imaging (b = 0, 1000 and 1500 s/mm2), and axial free-breathing DCE-MRI. Field of view was 14–18 cm, and resolution was 0.8 0.8 3 mm. Acquisition of T1-weighted DCE-MRI images (of the entire prostate) started 30 seconds before intravenous administration of 0.1 mmol/kg gadodiamide (Omniscan; GE Healthcare,
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
Figure 2. Flowchart demonstrating stepwise exclusion of patients in the study cohort. ADC, apparent diffusion coefficient; CZ, central zone; DCE-MR, dynamic contrast-enhanced magnetic resonance.
Princeton, NJ) followed by a 20-mL saline flush at a rate of 2.0 mL/s. Temporal resolution range for DCE-MRI was 5– 12 seconds 5–7 minutes for GE, and (3–5 seconds for 7–9 minutes) for Philips. Please see the Appendix for complete imaging protocols. MRI Analysis
Analysis of T2-weighted images and ADC maps. Two radiologists, A and B (9 and 3 years of prostate MRI experience, respectively), retrospectively evaluated the prostate MRI studies using a picture archiving and communication system (Stentor; Phillips Healthcare) with coregistration software. ADC maps were generated from DW-MR images using commercial software (Advantage Windows, version 4.2.3; GE Healthcare; or Philips ViewForum; Philips Healthcare). The two radiologists independently evaluated T2-weighted images (axial and coronal), using methodology similar to that described by Vargas et al., to identify normal CZ (10). First, they identified the ejaculatory ducts and verumontanum. Second, they determined whether normal CZ could be distinguished as a hypointense region between the prostate base and the verumontanum surrounding the ejaculatory ducts. And last, they reviewed ADC maps in conjunction with T2-weighted images and recorded whether the CZ could be identified as a low-signal region in its expected location. Coregistration with T2-weighted images and anatomic landmarks, such as the ejaculatory ducts, was used to help identify the CZ on ADC maps (Figs 3–5). DCE-MRI Analysis
Strictly qualitative DCE-MRI analysis was performed after the evaluation of T2-weighted images and ADC maps. Before
DCE-MRI image analysis, the two radiologists performed a consensus evaluation of T2-weighted images and ADC maps to agree in which patients the CZ could be identified. Next, DCE-MRI images were processed and reviewed on a commercial workstation, allowing coregistration between the different sequences (DynaCAD for Prostate; Invivo, Gainesville, FL). In cases that they could identify the CZ, the two radiologists placed, in consensus, regions of interest (ROIs) on T2-weighted images on both the left and right CZs, which was divided by a line that can be drawn to connect the urethra and verumontanum, to generate qualitative DCE time curves (signal intensity vs. time). With the help of the coregistration software, these ROIs were then transferred to DCE-MRI images, and DCE time curves were generated by the software for each ROI. Next, the radiologists independently determined the curve types in terms of type 1, progressive (enhancement increases over time); type 2, plateau (stable enhancement over time after possibly an initial increase); and type 3, wash-out (early and fast enhancement followed by decreasing enhancement over time) (20). Statistical Analysis
Recorded data included patient age, PSA level, identification of CZ on T2-weighted images (yes or no), identification of CZ on ADC maps (yes or no), and DCE time curve type (types 1–3). Per-patient identification rates of CZ (the left and right CZs were always concordant in terms of identifiable or not identifiable) on T2-weighted images and on ADC maps were calculated for each radiologist, together with binomial exact 95% confidence intervals (21). The binomial exact version of the McNemar test was used to evaluate whether differences between the detection rates of the two radiologists were significant statistically (22). In DCE time curve analysis, 571
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
Figure 3. A 47-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 1 (progressive) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
the left and right CZs of each patient were considered separately because their time curves could differ. Inter-reader agreement was evaluated with the kappa statistic, k, with the following interpretation of the k value: 0.00–0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81– 1.00, almost perfect agreement (23).
RESULTS Identification of Normal CZ on T2-weighted Images and ADC Maps
The CZ identification rates on T2-weighted images and ADC maps are listed in Table 1. The more experienced radiologist (A) detected more CZs on both T2-weighted images and ADC maps. However, these differences were not statistically significant (P = .63 for T2-weighted images, P = .15 for ADC maps). Inter-reader agreement, k, was 0.64 and 0.29 for T2-weighted images and ADC maps, respectively, indicating substantial and fair agreement, respectively (Tables 2 and 3). 572
DCE-MRI Analysis
Consensus evaluation between the two radiologists yielded a total of 52 cases for which the CZ could be identified on both T2-weighted imaging and ADC maps. These 52 cases (a total of 104 left and right CZs) were analyzed for DCE time curves (Table 4). Both radiologists identified all CZ time curves as either type 1 (Fig 3) or type 2 (Figs 4 and 5). Radiologist A: type 1 progressive enhancement (24/104 or 23% of curve types); type 2, plateau enhancement (80/104 or 77% of curve types); and type 3, wash-out (0/104 or 0% of curve types). Radiologist B: type 1 progressive enhancement (17/104 or 16% of curve types); type 2, plateau enhancement (87/104 or 84% of curve types); and type 3, wash-out (0/ 104 or 0% of curve types). There was no statistically significant difference between the two radiologists (P = .19). Inter-reader agreement, k, was 0.37, indicating fair agreement.
DISCUSSION Despite the widespread notion that the CZ cannot be visualized distinctly on MR images, our study supports the recent
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
Figure 4. A 57-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 2 (plateau) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
observation of Vargas et al. (18) that the CZ could be differentiated from other prostate zones on T2-weighted images and ADC maps. One of the early studies describing the prostate anatomy using a 0.3 Tesla MR scanner reported that the CZ was identifiable on T2-weighted images in 31 of 32 men, 25–35 years of age, and in 8 of 23 men, aged $40 years (6). Vargas et al reported the anatomic location (surrounding the ejaculatory ducts from the prostatic base to the verumontanum), left-and-right symmetry on coronal T2-weighted images, and homogenous low signal intensity on T2-weighted images and on ADC maps as the most common MR image features of the CZ. Using these criteria, we were able to identify CZ in 92%–93% of patients on T2weighted images and 78%–88% of patients on ADC maps (mean age, 59.9 years), with CZ identification rates not significantly different between the reviewers. Moderate agreement between observers of different experience levels in our study in the detection of the CZ on T2-weighted images suggests that readers with varying prostate MR experience may fairly reliably identify the CZ. Although the exact reason for relatively low signal intensity of CZ on T2-weighted images and ADC maps is not known, one hypothesis is that the cribriform protrusions of the thick walls of the CZ glands into the lumen forming ‘‘Roman bridge type architecture’’ limit the
relative volume of luminal fluid in the glands. Luminal glandular fluid contributes to high T2 signal and high ADC in normal PZ glands. This relative decrease of luminal fluid, together with a more compact stroma in the CZ, may contribute to the low signal of CZ on T2-weighted images and ADC maps (24). Given the similar MR signal characteristics on T2-weighted images and ADC maps between the CZ and PCa, differentiation of normal CZ from PCa based on these two sequences alone can be problematic, particularly in the region of the posterior prostatic base, an area composed almost entirely of CZ (4). Our DCE time curve analysis showed that normal CZ demonstrated either type 1 (progressive enhancement, 16–23% of CZs), or type 2 (plateau enhancement, 77–84% of CZs), but not type 3 (wash-out, 0% of CZs) time curves, the last of which is typically associated with PCa. DCE-MRI is an important component of the accepted multiparametric prostate MRI paradigm and has been shown to improve detection of clinically important prostate cancer (25–27). In the recently released ESUR prostate MR guidelines, one of the selected major criteria for tabulating a high PI-RADS score, indicative of a high risk of PCa, was the wash-out, type 3, enhancement pattern of a lesion on DCE-MRI, which underscores the importance of qualitative 573
HANSFORD ET AL
Academic Radiology, Vol 21, No 5, May 2014
Figure 5. A 64-year-old patient with prostate cancer. Axial (a) and coronal (b) T2-weighted images and ADC map (c) show symmetric, hypointense, homogeneous signal intensity at the prostatic base representing the normal CZ (arrows). Both readers identified the normal CZ. Qualitative dynamic contrast-enhanced MR enhancement time curve generated from a region of interest in the CZ (d) (x-axis, time in seconds; y-axis, image intensity) demonstrates a type 2 (plateau) time curve identified by both readers. ADC, apparent diffusion coefficient; CZ, central zone; MR, magnetic resonance; PZ, peripheral zone.
TABLE 1. CZ Identification Rates on T2-weighted Images and on ADC Maps
Modality T2-weighted ADC map
Radiologist
CZ Detection Rate (%) [95% Confidence Interval]
A B A B
93 (55/59) [84%–98%] 92 (54/59) [81%–97%] 88 (52/59) [77%–95%] 78 (46/59) [65%–88%]
ADC, apparent diffusion coefficient; CZ, central zone.
DCE-MRI (11). Given the absence of type 3 enhancement time curves in the normal CZ in our study, our findings suggest that qualitative DCE-MRI analysis can be a valuable tool in distinguishing between normal CZ and PCa. CZ cancers are rare, accounting for 0.5%–2.5% of all PCa and 3%–8% of index tumors (28–30). Their behavior is not well understood, partly because of their relatively low incidence and partly because of the general failure to recognize the correct zonal origin in routine pathologic reporting (26). They are usually associated with a high risk of adverse pathologic features (increased propensity for seminal vesicle invasion, high Gleason grade, and high rates of 574
TABLE 2. Agreement in CZ Identification on T2-weighted Images between the Two Radiologists B Radiologist A
Identified Not Total
Identified
Not
Total
53 1 54
2 3 5
55 4 59
CZ, central zone.
extracapsular extension) and early biochemical failure after curative surgery (10,31). Despite these important prognostic differences, because of lack of a clearly delineated boundary between PZ and CZ, (unlike the surgical capsule between TZ and PZ), most pathologists do not routinely recognize tumors as originating in the CZ (31). Given the difficulty of identifying a CZ of origin PCa on histology and the somewhat poorer prognostic implications of a CZ tumor, the identification of CZ involvement by PCa may be a clinically relevant finding that can be worthwhile to actively look for while interpreting MR images. Our study had several limitations. First, our study was subject to the inherent limitations of retrospective design and a
Academic Radiology, Vol 21, No 5, May 2014
DCE-MR NORMAL CENTRAL ZONE OF PROSTATE
TABLE 3. Agreement in CZ Identification on ADC Maps between the Two Radiologists B Radiologist A
Identified
Not
Total
43 3 46
9 4 13
52 7 59
Identified Not Total
ADC, apparent diffusion coefficient; CZ, central zone.
TABLE 4. Agreement in DCE Time Curve Analysis between the Two Radiologists B Radiologist A
Type 1 Type 2 Type 3 Total
Type 1
Type 2
Type 3
Total
10 7 0 17
14 73 0 87
0 0 0 0
24 80 0 104
DCE, dynamic contrast-enhanced.
relatively small sample size. Second, our prostate examinations were performed on two different MR machines. This second limitation is probably not particularly problematic because varying temporal resolution in the DCE time curve analysis is less of an issue for qualitative analysis compared to quantitative analysis. Third, we did not use quantitative or semiquantitative DCE-MRI analysis. Because of a large variation in the reported values of quantitative parameters derived from MRI in the literature and in keeping with the recent ESUR prostate MR guidelines, we chose qualitative enhancement time curve analysis because it is the most readily accessible and straightforward methodology. In addition, qualitative DCE-MRI analysis has been used successfully for breast MRI. However, the inherent subjectivity of this qualitative method does make standardization among institutions more challenging and future research using quantitative and semiquantitative methodologies would be a valuable contribution to the literature. Fourth, we did not include any cases of CZ cancer in our study. Although type 3 time curves are known to be frequently associated with PCa regardless of the zonal origin, future studies specifically investigating the DCE-MRI features of CZ cancer will be helpful additions to the literature. And finally, we did not compare CZ histology on wholemount pathology inspection to the MRI findings of CZ for all 59 patients in the final cohort. We did review eight whole-mount prostate pathology specimens with a GU radiologist who confirmed the CZ findings on prostate MRI were concordant with CZ histology. For the remaining cases, we relied on previous radiology literature identifying the CZ on MRI, which was supported by whole-mount pathology evaluation for concordant CZ identification (14,18). In conclusion, our results support the previous observations by Vargas et al. that it is possible to distinguish the CZ from the surrounding prostate gland in the vast majority of patients
undergoing prostate MR based on its low signal on T2weighted images and on ADC maps. Our study shows that the normal CZ demonstrates either type 1 or type 2 enhancement time curves on DCE-MRI, which can be potentially used to differentiate the CZ from PCa.
REFERENCES 1. Villers A, Steg A, Boccon-Gibod L. Anatomy of the prostate: review of the different models. Eur Urol 1991; 20(4):261–268. 2. McNeal JE. Prostate. In: Mills SE, ed. Histology for the pathologist. 3rd edition. Philadelphia: Lippincott Williams and Wilkins, 2007; p923–p963. Chapter 36. 3. Jaffe WI, Kaplan SA, McConnell JD. Epidemiology and pathophysiology of benign prostatic hyperplasia. In: Scardino PT, Slawin KM, eds. Atlas of the prostate. 3rd edition. Hong Kong: Current Medicine LLC, 2006; p2–11. Chapter 1. 4. McNeal JE. The zonal anatomy of the prostate. The prostate 1981; 2: 35–49. 5. Quick CM, Gokden N, Sangoi AR, et al. The distribution of PAX-2 immunoreactivity in the prostate gland, seminal vesicle, and ejaculatory duct: comparison with prostatic adenocarcinoma and discussion of prostatic zonal embryogenesis. Hum Pathol 2010; 41(8):1145–1149. 6. McNeal JE. Normal histology of the prostate. Am J Surg Pathol 1988; 12(8): 619–633. 7. McNeal JE. Cancer volume and site of origin of adenocarcinoma in the prostate: relationship to local and distant spread. Hum Pathol 1992; 23(3):258–266. 8. McNeal JE. Normal and pathologic anatomy of the prostate. Urology 1981; 17(suppl 3):11–16. 9. Reese JH, McNeal JE, Goldenberg SL, et al. Distribution of lactoferrin in the normal and inflamed human prostate: an immunohistochemical study. Prostate 1992; 20(1):73–85. 10. Srodon M, Epstein JI. Central zone histology of the prostate: a mimicker of high grade prostatic intraepithelial neoplasia. Hum Pathol 2002; 33(5): 518–523. 11. Prayer-Galetti T, Rossi GP, Belloni AS, et al. Gene expression and autoradiographic localization of endothelin-1 and its receptors A and B in the different zones of the human prostate. J Urol 1997; 157(6): 2334–2339. 12. Lexander H, Franzen B, Hirschberg D, et al. Differential protein expression in anatomical zones of the prostate. Proteomics 2005; 5(10):2570–2576. 13. Laczko I, Hudson DI, Freeman A, Feneley MR, Masters JR. Comparison of the zones of the human prostate with the seminal vesicle: morphology, immunohistochemistry and cell kinetics. Prostate 2005; 62(3):260–266. 14. Hricak H, Dooms GC, McNeal JE, et al. MR imaging of the prostate gland: normal anatomy. AJR Am J Roentgenol 1987; 148(1):51–58. 15. Sommer FG, McNeal JE, Carrol CL. MR depiction of zonal anatomy of the prostate at 1.5T. J Comput Assist Tomogr 1986; 10(6):983–989. 16. Kim CK, Park BK, Lee HM, et al. 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(12):842–847. 17. Schiebler ML, Tomaszewski JE, Bezzi M, et al. Prostatic carcinoma and benign prostatic hyperplasia: correlation of high resolution MR and histopathologic findings. Radiology 1989; 172(1):131–137. 18. Vargas HA, Akin O, Franiel T, et al. Normal central zone of the prostate and central zone involvement by prostate cancer: clinical and MR imaging implications. Radiology 2012; 262(3):894–902. 19. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22(4):746–757. 20. Verma S, Turkbey B, Muradyan N, et al. Overview of dynamic contrastenhanced MRI in prostate cancer management. AJR Am J Roentgenol 2012; 198(6):1277–1288. 21. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174. 22. Zar JH. Biostatistical analysis. 4th ed. New Jersey: Prentice-Hall, 1999. 23. Riffenburgh RH. Statistics in medicine. 2nd ed. Amsterdam; Boston: Elsevier Academic Press, 2006; 274. 24. Bourne RM, Kurniawan N, Cowin G, et al. Microscopic diffusivity compartmentation in formalin-fixed prostate tissue. Magn Reson Med 2012; 68(2): 614–620.
575
HANSFORD ET AL
25. 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(1):248–254. 26. Van Dorsten FA, Van Der Graaf M, Engelbrecht MR, et al. Combined quantitative dynamic contrast-enhanced MR imaging and (1)H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imaging 2004; 20(2):279–287. 27. Padhani AR, Gapinksi CJ, Macvicar DA, et al. Dynamic contrast-enhanced MRI of prostate cancer: correlation with morphology and tumour stage, histologic grade and PSA. Clin Radiol 2000; 55(2):99–109.
576
Academic Radiology, Vol 21, No 5, May 2014
28. McNeal JE, Redwine EA, Freiha FS, et al. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. Am J Surg Pathol 1988; 12(12):897–906. 29. Cheng L, Jones TD, Pan CX, et al. Anatomic distribution and pathologic characterization of small-volume prostate cancer (
