1　Introduction

       Based on the GLOBOCAN 2012 estimates, prostate cancer (PCa) is the second most commonly diagnosed cancer and the fifth most common cause of cancer death in men worldwide in 2012^[1]. An early detection of the PCa is an important part to have an effective treatment of PCa^[2].

       For over a decade, magnetic resonance imaging (MRI) has been used in PCa diagnosis with varying degrees of success^[3]. Earlier emphasis of prostate MRI relied primarily on morphologic and signal changes present on the conventional MRI, including T1 and T2-weighted images, which gave relatively poor sensitivity and specificity for detecting PCa^[4,5]. There are some new technical advancement that lead to the development of anatomical and functional MRI techniques that potentially lead to increased sensitivity, specificity, and accuracy of detection and characterization of disease processes^{[4, 6]}. Of all functional MR imaging techniques, diffusion weighted imaging (DWI) MRI is the most practical and simple in its use^[7]. It has several advantages such as: not requiring exogenous contrast material, simple to process, and it requires less time and less technologist training to perform^[8]. However it has the disadvantages of being susceptible to motion and to magnetic field inhomogeneities^[7]. This article will talk and focus on DWI, its basics, and progress of the DWI application in diagnosing PCa.

2　The basics diffusion of DWI

       DWI is a method to functionally assess tissue and organs of the body by analyzing their cellular structure, that relies on the concept process of the Brownian molecular motion. DWI uses the differences in the motion of water molecules in extracellular and intracellular fluid and vascular fluids to produce image contrast, with no need for exogenous contrast materials^[9]. The degree of water diffusion restriction in biologic tissue is oppositely related to the tissue cellularity and the integrity of the cell membranes; the degree of water diffusion is more restricted in tissues that has a high cellular density because of the presence of many intact cell membranes^{[3, 10]}. When measurements of the diffusion are being performed, the water diffusion direction along the three orthogonal directions of the magnet (phase select, frequency select, and slice select) can be assessed independently by applying diffusion gradients in each of these directions. DWI that is the sum of the directionally acquired DWI is known as trace or index DWI^[11,12]. The relative change in DWI signal intensity at different b-values can be used to characterize tissues on the basis of differences in water diffusion. Thus, subjective visual assessment of the relative tissue signal changes on DWI using multiple b-values may be useful for tumor detection, tumor characterization and assessment of treatment response^[12] .

       Diffusion weighted images taken with at least two different b-values allow for the calculation of the apparent diffusion coefficient (ADC), which can quantify the water diffusion values in the tissue with reflecting the net displacement of water molecules per time (mm²/s) in a given volume element (voxel)^[13] . Water motion sensitivity of DWI is determined by the b-value (mm²/s), which reflects the influence of the diffusion sensitizing gradients and can be altered by changing gradient amplitude, gradient duration and time interval between the paired gradients^[12,14]. For a useful interpretation, DWI needs to be performed using at least two b-values: b=0 mm²/s and b=100 to 1000 mm²/s^[12]. DWI performed with b=0 mm²/s is equal to a T2-weighted sequence. At lower b-values (200 mm²/s or less), the calculated ADC is influenced by tissue perfusion and water diffusivity. Increasing the b-value over 200 mm²/s reduces the effect of perfusion.

       ADC maps of the entire imaged volume can be generated automatically on most MRI scanners and workstations. Average ADC value is determined by drawing an electronic region of interest (ROI) on an ADC map image generated on the scanner. The changes in ADC are inversely correlated with the changes in tissue cellularity: Decreased ADC values compared to normal tissue indicate restricted diffusion. Conversely, increased ADC values suggest increased diffusivity^[12].

3　Clinical application of prostate DWI

3.1　Qualitative analysis of DWI

       DWI images can be assessed qualitatively by visual inspection. Water molecules with a large degree of motion (e.g., in blood vessels) have less signal attenuation (remain hyperintense or bright) at small b-values (e.g., b=50—100 mm²/s) and greater signal attenuation (become hypointense or dark) at large b-values. Water molecules motion that slow and restricted (e.g., in PCa) have less signal attenuation (remain hyperintense or bright) at large b-values (e.g., b=500—1000 mm²/s). Overall, the higher the b-value, the more sensitive the sequence is to diffusion effects. Higher b-values (e.g., b=1000) are also optimum for background signal suppression^[12,14,15] .

       However, one of the disadvantages of qualitative assessment at index DWI is that the signal intensity depends on both water mobility and T2 relaxation time (also known as "T2 shine-through" effect). "T2 shine-through" effect can result in high signal intensity on high b-value DWI images without restricted diffusion. This may result in image interpretation errors, particularly if the ADC maps are not examined. ADC maps, which are actually a quantitative measure of tissue diffusivity, can also be visually inspected. Tissues with restricted diffusion appear hypointense on ADC maps/images, while they remain hyperintense on DWI. However, "T2 shine through" effect will result in a hyperintense signal in a tissue on both DWI as well as ADC maps^[3,12] . This effect can be sometimes reduced by the choice of an appropriate TE (a short one) and b-value (a large one), but it cannot be easily avoided. For prostate tissues, b-values >1000 s/mm² are occasionally needed to decrease the effects of "T2-shine through" ^[16].

       In clinical practice, however, qualitative assessment of DWI during imaging interpretation by radiologists is of critical importance because quantitative analysis of DWI requires additional time. A few published articles on DWI at 3 T or 1.5 T showed that for qualitative assessment of PCa, the peripheral zone (PZ) and transition zone (TZ) showed low signal intensity relative to benign prostate tissues on ADC maps, which indicates that DWI could have incremental value relative to conventional T2-weighted imaging^[17]. Moreover, these authors used ADC maps to predict localized PCa instead of index DWI to avoid the "T2-shine through" effect. A recent study reported the additional utility of DWI when used together with T2-weighted imaging at 3 T for predicting PCa localization in 68 tumors^[17]. In this study, the overall sensitivity and positive predictive value of T2-weighted imaging plus DWI were 84% and 86%, respectively, whereas those of T2-weighted imaging alone were 66% and 63%, respectively (P<0.05).

3.2　Quantitative analysis of DWI

       Quantitative analysis of DWI can be performed by calculating the ADC. The logarithm of tissue relative signal intensity (signal decay) on the y-axis against the b-values on the x-axis results in a line (exponential function). The slope of this line represents the ADC. ADC is a quantitative measure of tissue diffusivity and is expressed in (×10^-3) as mm²/s. This graphical fit can be improved by using multiple b-values to reduce error involved in the calculation and monoexponential and multiexponential modeling of signal decay^[9]. Quantitative analysis has advantages over visual qualitative analysis in that it is independent of magnetic field strength and can overcome the effects of "T2 shine-through" ^[12]. A simple method to detect T2 shine-through is to use the "exponential image" formula, where an increased signal ratio [DWI signal intensity (at b=X)/unweighted signal intensity (at b=0)] suggests true restricted diffusion^[12]. The ADC is calculated for each pixel and displayed as a parametric map. An average ADC value can be obtained by drawing an electronic ROI, as mentioned above. Certain tissue characteristics, such as increased cellularity and ischemia, are known to be associated with low ADC values due to restricted diffusion of water^[12] .

       PCa is histologically characterized by a higher cellular density than the normal prostate tissue, with replacement of the normal glandular tissue; thus, it is expected to show a more restricted diffusion of water molecules, compared with normal prostatic gland^[18]. Which will cause the PCa lesion displayed as a lower intensity area in the ADC maps, compared to the normal prostate gland areas^[19]. The ADC value of BPH is significantly lower than that of normal central gland (CG) of the prostate^[20]. This can be cause by CG tissue in patients with BPH may contain more stroma tissue and accordingly less glandular components than healthy CG tissue.

       Some studies using ADC value to differentiate the PCa in different region of the prostate. For example, Kim et al^[21], reported that according to receiver operating characteristic analysis for the prediction of PCa, an ADC cutoff value of 1.67 × 10^-3 mm²/s had 0.97 area under the curve (AUC) in PZ cancer. For the prediction of TZ cancer, an ADC cutoff value of 1.61 × 10^-3 mm²/s showed 0.92 AUC. Some other studies using ADC value to differentiate the PCa lesion with other prostate abnormalities like BPH and prostatitis. For example, Liu et al^[20], reported that according to receiver operating characteristic analysis for discriminating PCa from BPH, the AUC of ADC was 0.92 at a cutoff of 0.91 × 10^-3 mm²/s. When making a distinction between PCa and prostatitis, the AUC of ADC was 0.99 at a cutoff of 1.13 × 10^-3 mm²/s.

       Several authors have demonstrated that malignant lesions (range, 0.49±0.13-1.66±0.32 × 10^-3 mm²/s) have approximately 20 to 60% lower ADC values than noncancerous tissue (range, 1.26±0.27—2.19±0.24 × 10^-3 mm²/s) in the PZ of the prostate, depending on patient population characteristics and technical issues^{[4,5,21,22,23,24,25,26,27,28,29,30,31,32,33,34]}. Based on the literature, DWI alone has a respectively sensitivity and specificity of 81.0%—94.0% and 72.2%—91.0% with a cut-off ADC value of 1.45—1.67 × 10^-3 mm²/s for tumor detection, at 1.5 or 3 T ^{[21, 26, 32]}. The variation of ADC value that shown above may be related to the strength of the diffusion gradient (300—1000 s/mm²), technical parameters utilized, and the magnetic field (1.5 or 3 T) used^[35]. Although the mean ADC values of PCa DWI were significantly lower than the benign prostatic tissues, there are some overlaps found between the PCa and benign prostatic tissues. Thus, the use of ADC values alone could result in the misdiagnosis of PCa.

3.3　Choices of b-value in monoexponential model

       The b-value is one of the most important parameters that can affect PCa detection capability^[36]. To obtain more accurate information about prostate tumors, an appropriate b-value is essential for producing high-quality ADC maps that affect the accuracy of ADC measurements and visual imaging interpretations. However, there are conflicting opinions as to the optimal b-value for tumor detection^[37,38]. Beside choosing a b-value of 0 and a second one in the 600-1000 s/mm² range, some study have been used more than two b-values^[39,40] and upper values greater than 1000 s/mm² for analysis as well^[40]. Different choices of b-values as well as different underlying diffusion models^[39] will generally followed by variations in the resulting absolute ADC values. In addition, there is also evidence that the choice of b-values has a significant influence on visual analysis, in particular on the lesion delineation and visual ADC contrast. Katahira et al^[41] and Metens et al^[37], reported improved utility of acquired DWI images using an ultra-high b-value of 2000 s/mm² compared those with using a standard b-value of 1000 s/mm², whereas Kitajima et al^[38], observed no benefit of ADC maps obtained with the higher b-value. Rosenkrantz et al^[42], showed that using a b-value of 2000 s/mm² compared with a b-value of 1000 s/mm² resulted in improved tumor sensitivity and higher tumor-to-peripheral zone contrast on the DWI images, whereas performance of the ADC maps corresponding to the two b-values was similar. For standardization, it is crucially important to identify an optimal b-value of prostate DWI for detection of PCa^[41]. A b-value of 1500 s/mm² has not been reported as frequently as those of 1000 or 2000 s/mm². Metens et al^[37], reported that the contrast and image quality were improved in b=1500 s/mm² images compared to b=1000 or 2000 s/mm², and Wang et al^[11], concluded DWI images and ADC map using b=1500 s/mm² should be considered more effective than those with b=2000 or b=1000 s/mm² for detecting PCa.

3.4　Intravoxel incoherent motion (IVIM) model

       The IVIM model, introduced by Le Bihan et al^[43] in 1988, the water molecules motion due to blood microcirculation in the capillary network (perfusion) has a similar impact on the MRI signal intensity as their motion caused by the molecular diffusion, that utilized by low b values. IVIM DWI obtained by using biexponential decay function^[44,45]. The study of Shinmoto et al^[44], showed that IVIM DWI parameters are significantly different between PCa and PZ. IVIM DWI may offer additional information for tissue characterization in the prostate gland. Thus, in the study performed by Dopfert et al^[45], shown that compared to ADC, IVIM method still has lower diagnostic performance for PCa detection.

4　Limitations of prostate DWI

       Although prostate DWI can be useful in evaluating patients with PCa, current prostate DWI at both 1.5 and 3 T still has several limitations to overcome^[9]. First, the lack of standardization with respect to technical parameters utilized, such as pulse sequence type, TR, TE, b-values utilized, gradient direction and modeling method of signal decay, are major challenges to the reproducibility and reliability of DWI and measured ADC values. In many institutions, various methods are applied for prostate DWI using various b factors, which result in various ADC values being reported for PCa. Second, even though prostate DWI has been used in a clinical setting to assess PCa, only a few studies investigating its reproducibility have been published. Third, DWI has inherent flaws, such as imaging distortions and susceptibility artifacts. These flaws make assessing the therapeutic responses after hormonal or radiation therapy in PCa challenging. To overcome or decrease the impact of these problems, more advanced software and hardware need to be developed. Fourth, DWI plays limited role in the local staging of disease, primarily due to its low spatial resolution compared to conventional spin-echo techniques. Fifth, more in vivo studies are required to determine and clarify the pathologic changes related to the features observed at DWI.

5　Future directions of prostate DWI

       In terms of future research directions, 3 T DWI should be used to evaluate PCa using biexponential analysis that use b-values higher than 1000 s/mm², diffusion tensor imaging (DTI), and diffusion-weighted whole-body imaging with background body signal suppression (DWIBS). Theoretically, monoexponential analysis, which assumes an inverse linear logarithmic relationship between signal intensity and b-values, could fail to differentiate between the fast and slow components of diffusion. The fast diffusion component is studied at lower b-values. The slow diffusion component is studied at relatively higher b-values (b> 1000 s/mm²)^[39]. The biexponential method of diffusion analysis theoretically eliminates the effects of perfusion, reflecting tissue diffusion characteristics to a closer degree^[39]. Further studies are required to show biexponential analysis has significant benefit over monoexponential analysis in clinical practice. DWIBS is a recently introduced application of DWI that is performed using a STIR EPI diffusion-weighted technique with background suppression^[46,47]. At 3 T, DWIBS potentially offers higher SNR because SNR increases linearly with increasing field strength^[48]. However, larger susceptibility-induced image distortion and signal intensity losses, stronger blurring artifacts, and more pronounced motion artifacts degraded the imaging quality at 3 T. More studies are needed to determine the potential role of DWIBS in PCa. In vivo diffusion tensor imaging (DTI) of the prostate gland is feasible, including that at 3 T, based on the fact that the motion of water molecules is not truly random, but follows the orientation of the tissue structure (anisotropy)^[49,50]. PCa theoretically disrupts this normal anisotropy. Limitations of DTI include variation of noise, long imaging times and specific absorption rate limits^[49]. More studies are required to evaluate the role of 3 T DTI in the clinical evaluation of PCa.

6　Conclusion

       DWI is a promising and useful noninvasive imaging method that provides qualitative and quantitative information related to tumor cellularity, tissue structure and the integrity of the cellular membrane in PCa. It could help us to improve the PCa detection and localization which also could decrease the excessive biopsy rate. It also may be helpful in differentiating PCa from other benign prostate abnormalities like BPH and prostatitis. But there are still some limitations of this method that needed to overcome, especially the standardization and optimization of the various technical parameters to have a better accuracy and universal standard value which needed to be further researched to establish it. In the clinical settings, DTI, and DWIBS at 3 T have a valuable potential to better evaluate the PCa; however, further studies are still required.