When John Adams first described a case of prostate cancer in a 59-year-old man in 1853, the disease was considered rare because of the short life expectancy at the time. Today, prostate cancer represents 11% of all cancers in the European Union, and 9% of all cancer deaths. The management and imaging in prostate cancer remains challenging. Although standard MRI sequences have been incorporated into practice to identify whether or not disease is organ confined, we describe how a relatively new imaging technique termed diffusion-weighted MRI (DW-MRI) has evolving clinical utility in the management of prostate cancer, including the potential as a non-invasive biomarker for tumour aggressiveness.
PRINCIPLES OF DW-MRI
DW-MRI first gained wide acceptance in the assessment of acute stroke, where this technique highlighted sites of ischaemia, previously undetectable on routine morphological T1- and T2-weighted MRI scans. As a result of technological advances, the application of this technique has more recently been applied to other systems in the body, including organs such as the liver, breast and prostate. Although the mechanisms used to generate the images are beyond the scope of this review, we briefly discuss the basic principles of DW-MRI to enable an understanding in daily clinical practice.
DW-MRI provides information associated with the molecular movement of water in biological tissues. The 3D diffusion of water within the body is not truly random because natural barriers impede diffusion. Within tissues, these barriers are cell membranes. In the case of a highly cellular tissue, an increased number of cell membranes may significantly impede the motion of extracellular water. Consequently, parameters derived from DW-MRI indirectly provide information regarding the cellularity of a given tissue.
In DW-MRI, images are generated by applying a diffusion-sensitizing gradient, the strength of this gradient can be altered by changing the ‘b-value’ on the scanner. Higher b-values indicate higher diffusion weighting, and the typical b-values used for prostate imaging varies in the range 0–1500 s/mm2. A b= 0 image indicates no diffusion-weighting has been applied. In a high b-value DW-MRI image, tumour tissue (i.e. an area with impeded diffusion) will often appear to have a higher signal intensity compared to the native tissue from which it arises (Fig. 1).
Performing DW-MRI using two or more b-values allows for the calculation of a quantitative parameter known as the apparent diffusion coefficient (ADC). Mathematically, the ADC is the slope of the line that describes the relationship between the logarithm of the measured signal intensity and the b-value. In the evaluation of prostate cancers, both DW-MI and ADC maps are important for disease assessment.
Conventional MRI for prostate cancer typically includes high resolution T2-weighted images (T2WI). These images allow for detailed morphological assessment of the prostate gland and the tumour because tumours arising from the peripheral zone typically appear dark compared to the normally bright peripheral gland. Images obtained by DW-MRI are acquired in addition to T2WI. The DW-MRI images are acquired very rapidly (few minutes only) without the use of contrast media or specialized hardware.
There are as yet no standardized DW-MRI techniques and the evolving technology has meant that investigators continue to push the boundaries on DW-MRI. DW-MRI is most commonly performed using the so called ‘echo planar imaging’ based techniques. Image acquisitions with b-values of 800–1000 s/mm2 are most commonly used. The choice of b-value is in part limited by signal-to-noise ratio and varies with the make of scanner. On some systems, a high b-value of up to 1500 s/mm2 can be achieved. The type of coils used also varies. An endorectal coil provides a superior signal-to-noise ratio compared to a pelvic phased array coil but causes reduced patient compliance and increased susceptibility artefacts. Diffusion-weighted imaging is likely to perform better with a 3T MRI system as a result of a better signal-to-noise ratio compared to a 1.5T system.
Although DW-MRI is relatively quick to perform, it is sensitive to a range of artefacts. For these reasons, the imaging sequences need to be optimized, often with the help of specialists, to ensure that high-quality images are consistently attained.
Image interpretation requires the review of DW-MRI images with the T2WI and the ADC maps. Prostate cancer typically shows high signal intensity as a result of impeded diffusion on high b-value images and returns a low ADC value (i.e. darker on ADC maps). ADC maps may be more useful in evaluating prostate cancer as a result of ‘T2 shine through’ from the normal high signal peripheral zone on the DW-MRI images. On source DW-MRI images, tumour is usually not appreciable on b-values <600 s/mm2.
The wide reported range of ADCs value for normal prostate relates to the use of different b-values, field strength, imaging protocols and MRI systems. However, the mean ADC is highest in the peripheral zone (1.54–2.99 × 10−3 mm2/s) followed by the central gland (0.9–2.14 × 10−3 mm2/s) and then prostate cancer (0.8–1.66 × 10−3 mm2/s) . Prostate cancer shows a lower ADC than normal prostate but with a degree of overlap. The lower ADC values in prostate cancers correlates with higher cellularity on histology .
When interpreting the images, there should be an awareness of the potential pitfalls with this modality (i.e. motion, susceptibility artefacts, image distortion and noise). For example, susceptibility artefacts mean than diagnostic DW-MRI may not be reliably obtained in patient with hip prostheses.
Several studies have shown that DW-MRI improves sensitivity and specificity in the diagnosis of prostate cancer on MRI by increasing conspicuousness of the tumour on the DW-MRI images or ADC map [1,3]. Improved tumour localization has implications for focal therapy including radiotherapy planning in terms of escalating the dose to the tumour relative to surrounding normal tissues.
However, the ability to detect lesion is dependent on the site of the tumour, its composition and size. Tumours <5 mm are difficult to detect and inflammatory processes in the prostate can have lower ADC values and produce false positive appearances of cancer . One of the most challenging areas in which to detect a tumour is the transitional zone. At this site, a benign nodule can have low ADC values, and may mimic sites of tumour. However, T2WI with ADC maps has been shown to improve the detection of tumour foci in the transitional zone .
Haemorrhage into the prostate gland or seminal vesicles can result in low signal foci on T2-weighted sequences and a high signal on T1WI. These areas can mimic sites of tumour and lead to false positives when interpreting MRI. Therefore, a delay of 6–8 weeks after biopsy has been recommended for prostate MRI examinations. Haemorrhage may decrease ADC values in benign tissues and affect the tissue contrast on DW-MRI between cancer and benign tissue and so reduce diagnostic accuracy. Furthermore, haemorrhage may increase imaging distortion as a result of susceptibility artefacts. Despite this, DW-MRI has been reported to be slightly more sensitive than T2WI in detecting tumour within areas of haemorrhage .
DW-MRI can help to localize tumour in patients with a previous failed TRUS biopsy. In a study of 43 patients with previous negative TRUS biopsy and persistently elevated PSA before repeat biopsy, prostate cancer was detected in 17 (40%) patients, and was more predominant in the transitional zone (76%) than in the peripheral zone (24%); of the 17 cancers detected on DW-MRI, only six lesions were visible on conventional T2WI . Another study conducted in 68 patients found that DW-MRI had the highest positive predictive value compared to T2-weighted, dynamic contrast-enhanced and 3D-spectroscopy MRI. Logistic regression showed the probability of cancer in a segment increasing by 12-fold when T2WI and DW-MRI were both suspicious compared to both being non-suspicious 
The most exciting potential role for DW-MRI in prostate cancer is the possibility of identifying tumour aggressiveness. Currently, the Gleason score on a biopsy specimen is the standard method for histological grading of prostate cancer. Unfortunately, the Gleason score has many pitfalls (i.e. sampling errors and comprises an invasive test). The principle behind the effect of tumour grade on diffusion is the change in cellular architecture. Altered gland formations such as medullary or solid patterns are more often found with increased histological grade. Therefore, tumours of higher Gleason grade are likely to show impeded water diffusion.
A number of studies have shown a correlation with Gleason score and ADC values [2,3]. A significant difference in ADC values in patients with low vs intermediate risk disease has been reported  and the baseline ADC was shown to be an independent predictor for both adverse repeat biopsy findings and time to radical treatment in a cohort of patients on active surveillance . In a retrospective study of 158 men who underwent radical prostatectomy for prostate cancer, 30 patients (19%) had biochemical relapse. Univariate analysis showed that tumour ADC, Gleason score, serum PSA, greatest percentage of cancer in biopsy core, percentage of positive cores in all biopsy cores and tumour volume were all significantly related to biochemical relapse. However, multivariate analysis identified tumour ADC as the only independently predictive factor . Thus, an ADC value derived from DW-MRI appears to be a potentially important prognostic disease marker.
PRIMARY T STAGE
The generally lower spatial resolution of DW-MRI compared to T2WI means that DW-MRI may not provide improved staging information in terms of extracapsular tumour extension. However, DW-MRI may be helpful for the assessment of seminal vesicle involvement (Fig. 2). Seminal vesicles are usually fluid rich glands; they appear to be of high signal on T2WI, low on DW-MRI and have high ADC values. A recent study showed seminal vesicle invasion as areas of higher signal on DW-MRI, with significantly lower ADC values compared to normal seminal vesicles . DW-MRI has also been reported to detect bladder wall invasion from prostate cancer. A retrospective study showed the mean ADC for urinary bladder wall invasion was significantly lower (0.963 × 10−3 mm2/s) compared to normal bladder wall ADC (1.517 × 10−3 mm2/s) .
STAGING LYMPH NODES
Although size and morphology have been traditionally used to determine malignant nodal infiltration, it is well known that a small node may harbour disease and nodes may be reactively enlarged. DW-MRI has been shown to be of potential value for evaluating metastatic lymph nodes in a wide range of tumour types. Using fusion of DW-MRI with T2WI can improve the identification of pelvic lymph nodes compared to T2WI alone . The improved nodal identification may aid treatment planning and further nodal characterization. Recently, a small retrospective study in patients with prostate cancer found a significant difference between the mean ADC value of malignant and benign nodes . However, larger and prospective studies are needed to evaluate this further.
Several recent studies have shown some promise for the role of DWI in detecting bone metastases [14,15]. In a study comparing short tau inversion recovery and T1WI, positron emission tomography/CT and DW-MRI to detect bone metastases from prostate tumours, DW-MRI was found to be equal, if not superior, to short tau inversion recovery/T1W images, and equally as effective as 11C choline positron emission tomography/CT . A study of whole body DW-MRI in patients with breast or prostate cancer found that more metastases could be identified with DW-MRI than skeletal scintigraphy . Thus, DW-MRI, when extended to the whole body, may provide comprehensive staging of the patient with prostate cancer.
ASSESSING RESPONSE/DETECTING RECURRENCE
ADC values may be useful as an imaging biomarker for monitoring therapeutic response of prostate cancer to treatment (Fig. 3). In a study of 46 patients, MRI performed at 3T showed the mean ADC value of the tumour increased after radiotherapy, whereas the mean ADC values of benign peripheral zones and benign transition zones were significantly decreased compared to those before radiotherapy .
The treated prostate gland shows a diffuse low signal of the prostate with less distinct zonal anatomy on T2WI, which makes it very difficult to identify recurrent tumour. Although serum PSA can monitor disease status, it is unable to distinguish local and distant recurrent disease and, in this respect, DW-MRI may be of value (Fig. 4). A recent study evaluated DW-MRI in predicting locally recurrent prostate cancer in patients with biochemical failure after radiation therapy. DW-MRI with T2WI increased the sensitivity compared to T2WI alone . The mean ADC values between recurrent cancer and benign irradiated tissue were significantly different . A separate study showed that combining T2WI with DW-MRI was more specific than dynamic contrast enhanced (DCE)-MRI in predicting local tumour progression after high frequency ultrasound ablation. However, DCE-MRI was shown to be more sensitive than T2WI and DW-MRI .
In patients with metastatic bone disease, DW-MRI is also showing substantial promise for the assessment of treatment response . Evaluating treatment effects in bones is notoriously difficult using conventional imaging. An increase in tumour ADC has been corroborated with a decrease in serum PSA levels in bone metastases treated by antiandrogen therapy .
FUTURE OF DWI IN PROSTATE CANCER
The technology related to DW-MRI is rapidly evolving. There is a move toward better quality imaging with a reduction in artefacts, imaging at higher b-values and advanced image evaluation (e.g. bi-exponential ADC analysis). Diffusion tensor imaging is a more sophisticated form of DW-MRI that allows the determination of direction as well as the magnitude of water diffusion, and may provide data about the microstructural changes in tissues. Preliminary data suggest that d tensor imaging may be of value in characterizing prostatic tissues .
In this review, we have mainly dealt with the application of DW-MRI in conjunction with conventional T2WI to the prostate. The evidence available so far suggests that the addition of DW-MRI to T2WI is of value in many clinical scenarios, although larger prospective studies will help to further validate these observations. The prostate may additionally be imaged using other modern sequences, including DCE-MRI and proton magnetic resonance spectroscopic imaging, as well as their outputs combined, providing the so-called multiparametric MRI. The role of DW-MRI compared to other multiparametric MRIs remains to be defined.
Over recent years, there has been an explosion of interest in evaluating the potential of DW-MRI in prostate cancer. Recent technical developments have added to the momentum of its use in clinical practice. There is increasing evidence that DW-MRI improves the sensitivity and specificity of prostate cancer detection, and studies have recently reported success in using DW-MRI in staging as well as identifying tumour aggressiveness. DW-MRI is also showing substantial promise as a response biomarker for both local and metastatic disease. The role of DW-MRI in combination with other MRI techniques for multiparametric assessment of disease needs to be defined further.
Dr D. M. Koh, Dr N. van As and Dr S. A. Sohaib are supported by RMH/ICR NIHR Biomedical Research Centre Funding.