DWI visualizes the Brownian molecular motion of water. It has become an important part of oncological imaging, as malignant tumors are typically composed of densely arranged cells in which their numerous membranes limit this Brownian molecular motion [14]. Prostate cancers are thus visualized on highly diffusionweighted images (typically upper b-values of 800-1,000 s/mm2 in prostate imaging) as areas with high signal intensity (Fig. 7a). Diffusion coefficients [apparent diffusion coefficient (ADC)] can be calculated from the diffusion-weighted data. In ADC maps, areas with normal diffusion can then be differentiated as having high signal intensity and those with diffusion restriction as having comparatively low signal intensity (Fig. 7b). Most studies show that DWI is a very useful addition to morphological imaging and can increase sensitivity in particular by 10–25% [15–17]. Since hyperplastic stromal nodules can have pronounced diffusion restriction in BPH, DWI must be evaluated together with morphological imaging (T2w), especially when assessing the central portions of the prostate gland [16].

Studies have increasingly examined the capabilities of DWI with respect to evaluating the aggressiveness of prostate cancer. Such studies show that the ADC value has a negative correlation with the Gleason score, i.e., low ADC values are seen primarily in high-grade, aggressive prostate cancers [15, 18]. Results regarding the ability to differentiate low-grade (Gleason score 3+3=6) and high-grade (Gleason score >4+3=7) tumors in the peripheral zone are very promising, since they may be able to help to better determine patient risk potential [19, 20]. However, since there can still be relevant overlapping of ADC values in different Gleason groups, additional studies are needed to implement this grading potential of DWI in clinical routine. DWI will play an important role in differentiating patients with a low risk from those with a high risk and separating them with respect to management [21].

Contrast-enhanced MRI sequences can be used to assess tissue vascularity and permeability. Gadolinium-containing extracellular T1w contrast agent is typically administered intravenously for this purpose [22]. Fast T1w gradient- echo (GRE) sequences with a temporal resolution of 4–10 s are primarily used in prostate imaging. Measurements should be performed over a period of up to approximately 5 min after contrast agent application [13]. Enhancement can be displayed in the form of curves over time, thus helping to characterize tissues. Prostate cancers are characterized by fast washin (early peak enhancement) and fast washout compared with healthy tissue (type III curve). Enhancement typically increases steadily (type I) in the given measurement time in cancer-free tissue, whereas a curve with a plateau (type II) occurs relatively frequently in both healthy and cancerous tissue [22]. On the basis of enhancement curves, gadolinium concentrations in tissue and tissue-transport constants in the direction of the tumor interstitium (K^trans) and back in the direction of the blood plasma (k_ep) can be calculated using suitable mathematical models (two-compartment Tofts model). Prostate cancers are characterized by an increase in the tissue-transport constants, which can be displayed using color coding in pharmacokinetic parameter maps (Fig. 8) [23]. Available studies do not provide a clear conclusion regarding improvement of prostate cancer detection via DCE. Some studies showed an improvement in diagnostic accuracy of conventional MRI (T2w and T1w) when supplemented by DCE imaging [24, 25]. However, it is problematic in DCE that hyperplastic nodules in BPH can enhance and wash out quickly, like prostate carcinoma foci, and K^trans and k_ep can be accordingly increased, thus limiting sensitivity and specificity, particularly in the central portions of the prostate gland [26, 27]. In addition, inflammation can often have greater vascularity and tissue permeability, which also limits its sensitivity and specificity in detecting prostate cancer. However, local staging of prostate cancer and detection of local tumor relapses after definitive therapy can be significantly improved using dynamic T1w sequences, with good spatial resolution [28, 29].

In 2012, the European Society of Urogenital Radiology (ESUR) created the Magnetic Resonance Prostate Imaging Reporting and Data System (MR PI-RADS) as part of its MRI guidelines for prostate imaging [13]. Based on the Breast Imaging Reporting and Data System (BIRADS), a standardized method for reporting multiparametric MRI of the prostate for detecting prostate cancer, was proposed here. The goal was to standardize image interpretation and to simplify communication between the radiologist and colleagues in other departments.

Based on clearly defined criteria according to PIRADS, every lesion suspicious for tumor within the prostate is assigned a point value between 1 and 5 for every sequence performed as part of multiparametric MRI (consisting of at least T2w, DWI, and DCE). Moreover, a total point value is calculated for every lesion suspicious for tumor [33]. Thus, a statement regarding the probability of the presence of a clinically significant prostate cancer should be possible: a point value of 1 means that a lesion is probably benign, whereas a point value of 5 indicates a high probability of malignancy. The development of PI-RADS and criteria contained therein for the assignment of point values are based on published literature and an expert consensus. Since being published in February 2012, PI-RADS has been evaluated in multiple studies.

The point values of T2w, DWI, and DCE were added to form a total point value for lesions suspicious for tumor. Good and reproducible diagnostic accuracy was documented for the total point values calculated in this manner [33–37].

Standardization of image interpretation in research and the clinical routine is an important milestone that should accelerate the acceptance of multiparametric MRI in the coming years. The results of clinical studies should be easier to compare using PI-RADS. Moreover, PI-RADS makes it possible to formulate guidelines for diagnostic clarification and perhaps even for treating prostate cancer. The present data indicate that a biopsy should be performed in PI-RADS ⩾4 lesions, whereas monitoring via MRI and PSA could be sufficient in lesions with PI-RADS ⩽3.

The standard prostate biopsy is the transrectal ultrasound (TRUS)-guided systematic prostate biopsy. Urological studies in particular show that the detection rate increases as expected with the number of samples so that approximately six samples are taken from each lateral lobe in accordance with the recommendations of the professional associations [38, 39]. However, the entire context of the diagnostic weaknesses of TRUS biopsy only becomes clear with MRI and visualization of prostate cancer [40]. The rate of carcinomas in the ventral portion of the prostate gland, as well as in an extreme lateral position in the peripheral zone and on the apex of the gland, is significant in patients with multiple negative US- guided biopsies. It was shown that patients with a diagnosis of prostate cancer in the ventral portions of the gland have a higher clinical risk than those with a diagnosis of prostate cancer in the peripheral zone [41].