Spin-echo-based T2W MRI sequences form the basis of highresolution prostate imaging. These images (in which water has high signal) allow excellent visualization of the margins and internal anatomy of the prostate gland as well as the seminal vesicles, lymph nodes, and bladder (2,3,4). The anatomic detail depicted on T2W sequences forms the foundation of the multiparametric approach to interpretation of prostate MRI (5,6). Examples of the typical prostate anatomy on the T2W sequence are depicted in Figure 35.1. However, T2W MRI has been shown to have poor specificity for prostate cancer when interpreted alone (7,8). Among the reasons for the poor performance is the similarity of the appearance of prostate cancer to commonly seen changes of prostatitis as well as that of benign prostatic hyperplasia (BPH) (6,7). Figures 35.2 and 35.3 depict typical findings at prostate MRI in clinical practice. Recently, there has been near universal adoption of an interpretation approach based on a combination of T2W and two other MRI sequences (DW and DCE) which is termed multiparametric (8). In fact, a standardized reporting language based on specific criteria from these three sequences has been proposed (9).

DW MRI is a technique that allows visualization of the random movement of water molecules (Brownian motion). This allows for a differentiation of tissues that have different levels of cellularity, based on the level of restriction of the natural random motion of water (10). Simple fluids, which have no restriction of this motion, are readily differentiated from the minimally cellular normal peripheral zone of the prostate and from highly cellular prostate cancer (11). An example case of prostate cancer with abnormal signal on DW-MR is depicted in Figure 35.4. Experience in the interpretation of DW-MR is necessary as the differentiation of prostatitis and BPH from cancer can still be problematic, as with the other sequences when read independently (8,9).

The third essential sequence for high-quality prostate MRI, DCE MRI, assesses the vascularity of tissues in the prostate.
In brief, intravenous gadolinium-based contrast agents are administered rapidly through automated bolus injectors and carefully timed MRI sequences are performed. Prostate cancer should be visible as tissue that has a rapid, early enhancement and rapid clearance (washout). Normal prostate tissue more often has a slower peak to the contrast enhancement curve and a more delayed clearance, although BPH can overlap in appearance (9,12). This method can be the most difficult technically and also takes experience in interpretation. Postprocessing technology exists in order to simplify the interpretation, although the addition of another layer of image processing can be yet another source of error; therefore, knowledge of processing artifacts is essential (12). DCE MRI images from the same patient as in Figure 35.4A are included in Figure 35.5.

Although the overall performance of multiparametric MRI for identifying prostate cancer is good, there can still be significant discrepancies between MRI interpretation and prostate biopsy results (8,9,13). A major point to consider is that until recently, most available research evaluating the correlation between MRI and prostate biopsy depended on the assumption that the areas identified on MRI could be well-correlated by transrectal biopsy or prostatectomy specimens. Many experienced researchers would likely acknowledge the major limitations of this approach. Recently, there is a growing body of research which uses more direct means of correlation including MRI-guided biopsy and real-time image fusion between MRI and transrectal ultrasound-guided biopsy (13,14,15). These methods have increased the ability to evaluate the performance of MRI, given the increased confidence of correlation between MRI and biopsy. However, MRI will not identify all or even nearly all prostate cancer. Clinically, the significance of such a reduced sensitivity for cancer has at least two potential outcomes. First, many of the cancers undetected by MRI are lowgrade, clinically insignificant cancers (16). However, failure to identify cancers that would be deemed to be of low clinical significance by MRI may potentially represent a desirable outcome. However, clinically significant cancers can also be undetectable at MRI, leading to potentially deleterious patient outcomes. Therefore, despite the strong ability of MRI to identify cancer, the optimal clinical role of MRI in its diagnosis has not been established to date (17).

It is important to note that there is not a high clinical need for prostate MRI to identify all prostate cancer foci. In fact, the diagnosis of prostate cancer is often not difficult to obtain, given the high incidence of the disease and the high sensitivity of routine prostate biopsy for cancer. Although the exact role of prostate MRI has yet to be universally established, there has been a rapid expansion of the use of this technology based on published reports and anecdotal accounts. In the following sections, potential clinical roles for prostate MRI will be explored. Specifically, postulated clinical information which prostate MRI can supply in several common clinical scenarios will be reviewed.