Accurate detection of the exact location of cancer within the prostate contributes significantly to patient care. It has been found that using any two functional sequences in addition to T2WI improves detection; however the addition of a third functional sequence to improve prostate cancer detection is uncertain [30]. Using all three functional sequences (DWI, DCE-MRI and MRSI) in addition to T2WI has been shown to achieve a positive predictive value of 98 % in experienced hands [31] (Fig. 2).

While detection of prostate cancer using mp-MRI is mostly effective for large and high grade tumors, sensitivity for small Gleason 6 tumors as well as sparse tumors (malignant tissue intermixed with more than 50 % normal PZ tissue) has been shown to be lower [31, 32]. Moreover, careful interpretation of mp-MRI should be done in patients with multiple negative biopsies and increasing PSA levels for tumors in the TZ and those located anteriorly, as these areas are less frequently sampled during TRUS biopsy. T2WI findings supporting the presence of TZ tumors include homogenous low signal intensity, ill-defined margins, lack of capsule, interruption of surgical pseudocapsule (TZ to PZ boundary of low signal intensity), lenticular shape and invasion of the urethra or anterior fibromuscular stroma [33]. Nevertheless, for detection of TZ lesions the sensitivity and specificity of mp-MRI is 88 % and 86 %, respectively [35]. Tumor detection may be hampered by post-biopsy hemorrhage (Fig. 3), leading to over or underestimation of tumor presence and extent. A delay of 6–8 weeks is recommended between biopsy and MRI [35–37]. This prolonged hemorrhage may be attributed to high levels of citrate in the prostate, which has anticoagulant properties [38].

Also, tumor volume estimation by multiparametric MRI has been shown to closely correlate with histopathological findings following radical prostatectomy. Such information about the spatial location along with the tumor volume may become important for evolving concepts of focal treatments with directed brachytherapy.

Characterization of biological aggressiveness of detected lesions is ideal for clinical decision making. An inverse correlation has been widely reported between the Gleason score and ADC values from DWI. The ADC values in addition to characterization of aggressiveness can also be helpful in assessing interval change in a tumor. However, as there is substantial inter-patient variation in ADC values of normal PZ, it has been recently shown that when the background ADC values of benign PZ are taken into consideration, there is a significant improvement in the ability to discriminate among different Gleason scores.

DCE-MRI quantitative parameters have not been shown to correlate with tumor grade or vascular growth factor (VEGF) expression as a molecular marker of angiogenesis [39]. Meanwhile, MRSI has been found to have higher sensitivity of 89.5 % for detection of Gleason 8 and above tumors compared with a sensitivity of 44.4 % for Gleason 6 tumors [40].

The use of endorectal coil in conjunction with pelvic coil regardless of the magnetic field strength has been proven to be superior for the assessment of extraprostatic extension when compared to use of pelvic coil alone (Fig. 4). Criteria for detecting extraprostatic extension includes NVB asymmetry, gross tumor surrounding the NVB, a tumor-capsule interface of more than 1 cm, obliteration of the rectoprostatic angle, irregular or speculated margin and a breach of the capsule with direct evidence of tumor extension. Seminal vesicle invasion is demonstrated on T2WI as focal low signal-intensity seminal vesicle, enlarged low signal-intensity seminal vesicles or ejaculatory ducts, obliteration of the angle between the prostate and SV and direct extension of tumor from prostate base into the SV. Accuracy of MRI in prostate cancer staging has been shown to range from 54 to 93 %, which has led to concerns about inter-observer variability.

MRI and CT have shown equivocal efficacy for lymph node metastasis, with both having low sensitivity. Recently, MRI with lymphotropic superparamagnetic iron oxide nanoparticles has shown considerable promise in occult lymph node metastases detection.

There are several factors which limit the effectiveness of prostate cancer diagnosis by TRUS biopsy.

In a cognitive fusion biopsy, significant lesions that appear on previously reviewed MRI are targeted during TRUS guided biopsy. Such sampling of targeted lesions in addition to the systematic biopsy does appear to yield improved accuracy. However, it is subject to potential human error in overlaying MRI findings over real-time TRUS.

In MRI guided biopsy, targeted sampling of the prostate is done under real-time MRI. This has been made feasible due to the relative decrease in the time required to perform MRI, development of MRI-compatible biopsy needles and targeting mechanisms, and advanced visualization tools for guiding and verifying needle placement.

MRI guided biopsy is performed in low-field strength open systems or in closed-bore 1.5 or 3 T MRI [53, 54, 60, 61]. Low-field strength modalities offer easy access to the patient, while closed-bore systems have a higher SNR allowing increased spatial resolution. Transperineal and transgluteal approaches have been used utilizing the open MRI platform, while a transrectal approach has been used in most studies with a closed-bore MRI platform. Using MRI guidance for prostate cancer detection yields rates of cancer detection ranging from 38 to 59 % [52, 54, 61, 62]. The disadvantages of this method include increased expenses, as two MRI scans are required, and the time required; the reported duration has varied from 30 to 90 min [54].

Robot-assisted MRI guided biopsy is currently under evaluation for use in guided biopsies and it may increase the accuracy of needle placement in the prostate gland. The robot is constructed of nonmagnetic and dielectric materials, which allows it to be fully operational in a magnetic resonance imaging suite. In recent years, various MRI-compatible robotic intervention systems have been introduced. The first commercially available robot assisted MRI guided system (Innomotion; Innomedic, Herxheim, Germany) was recently evaluated in a cadaver study for MRI guided transgluteal biopsies [63].

MRI-US fusion offers an alternative to targeted prostate biopsies. The fusion of two imaging modalities is due to advances in coregistration of previously acquired MRI and real-time TRUS. The techniques for fusion were primarily developed for brachytherapy. The early systems were limited by prostate motion, resulting in loss of accuracy. Since then other techniques have been developed to overcome this problem.

In MRI-Fusion biopsy, first a multiparametric endorectal MRI is done; the studies are then loaded into software that allows the radiologist to mark the prostate gland and the regions of interest for biopsy in different slices and views of the MRI, known as segmentation (Fig. 6). This information is then loaded onto the device. The second step involves real-time MRI-TRUS fusion to create a three-dimensional real-time reconstruction of the prostate on which the aiming and tracking of biopsy site is done. This technique can be done in an outpatient setting under local anesthesia within a few minutes. Currently five devices approved by the FDA are available for MRI-TRUS fusion biopsy. The Artemis device (Eigen, GrassValley, California, USA) has a mechanical arm with ultrasound transducer probe and is capable of tracking and recording biopsy locations.

The real-time fusion on Artemis steps are as follows: