Multiparametric MRI of prostate cancer: an update on state-of-the-art techniques and their performance in detecting and localizing prostate cancer

Abstract

Magnetic resonance (MR) examinations of men with prostate cancer are most commonly performed for detecting, characterizing, and staging the extent of disease to best determine diagnostic or treatment strategies, which range from biopsy guidance to active surveillance to radical prostatectomy. Given both the exam's importance to individual treatment plans and the time constraints present for its operation at most institutions, it is essential to perform the study effectively and efficiently. This article reviews the most commonly employed modern techniques for prostate cancer MR examinations, exploring the relevant signal characteristics from the different methods discussed and relating them to intrinsic prostate tissue properties. Also, a review of recent articles using these methods to enhance clinical interpretation and assess clinical performance is provided. J. Magn. Reson. Imaging 2013;37:1035-1054. © 2013 Wiley Periodicals, Inc.

Figure 1. MRI Localizer Imaging for Assessment of Endorectal Coil Placement

Sagittal localizing single-shot fast-spin echo (SSFSE) images to verify positioning of the endorectal coil prior to initiating diagnostic scanning. The top of the base (B) and the bottom of the apex (A) of the prostate are annotated for clarity. (a) Initially, the endorectal coil is low-lying, with its superior border lying posterior to the middle of the prostate gland (red arrow). The double-headed arrow highlights the area of signal hyperintensity, which only covers about two-thirds of the prostate in (a). (b) After repositioning of the coil, accurate coil placement is seen, as the entire prostate (red arrow) now shows a hyperintense signal, with the double-headed arrow now including the span of the entire prostate gland.

Figure 2. Prostate Hemorrhage following Biopsy Complicates Cancer Detection on T2-Weighted Imaging

A 52 year-old man with PSA 5.8, Gleason 5+4=9, T2cN0M0 prostate cancer. (a) An axial T1-weighted image of the prostate showing possible areas of hemorrhage, which have a hyperintense signal (major hemorrhagic regions are starred). (b) An axial T2-weighted image of the prostate at the same level showing diffuse T2 signal hypointensity in the peripheral zone, indicating either tumor or hemorrhage. The white arrow in both (a) and (b) in the lateral region of the right side of the prostate shows an area suspicious for tumor, as there is T2-weighted signal hypointensity without T1-weighted signal hyperintensity.

Figure 3. T2-Weighted Images of Critical Findings in Prostate Cancer

(a) In a 61 year-old man with PSA 3.7, Gleason 3+3 =6, T2cN0M0 disease, an axial view of the prostate presenting a diffuse, capsule-contained peripheral zone (PZ) lesion with a “chalky” appearance. Notice how the PZ, usually hyperintense and clearly demarcated from the central gland (CG) which it surrounds, is now heterogeneous. (b, c) A different, 56 year-old man with PSA 9.3, Gleason 3+4=7, T3bN0M0 disease shows right-sided seminal vesicle invasion (white arrows) by tumor in axial (b) and sagittal (c) images. In the axial image (b), compare the hypointense, diseased right SV (white arrow) to the normal, hyperintense SV on the left (arrowhead). (d) In a 60 year-old man with PSA 134, Gleason 3+4=7, T4N0M0 disease, an axial image of the base of the prostate showing right-sided extracapsular extension (arrow) at 7 o'clock with probable involvement of the right SV (arrowheads). (e) In a 65 year-old man with PSA 40.1, Gleason 5+4=9, T4N1M1a disease, an axial view of the prostate displaying a large posterolateral right peripheral zone lesion (arrow) at 7 o'clock which invades the right neurovascular bundle. This lesion starkly contrasts with the normal NVB on the left (arrowhead). (f) From the same patient as (e), significant pelvic lymphadenopathy is present in an axial view at the level of the seminal vesicles, including this 1.5 × 1.3-cm, right posterior obturator node (arrow). (g) In a 75 year-old man with PSA 11.4, Gleason 4+4 = 8, T3aN0M1b prostate cancer, an axial view of the spine at L5 demonstrating a 1.5 × 1.1-cm metastatic bone lesion.

Figure 4. Diffusion-Weighted Imaging of Prostate Cancer

A 58 year-old man with PSA 16.4, Gleason 4+3=7, T3aN0M0 prostate cancer. On a previous biopsy of this patient, the right side of the prostate was graded as Gleason 4+3=7 (an intermediate grade). (a) A DW image showing a dominant lesion (boxed), evidenced by the increased signal, found on the posterior right side at 7 o'clock. (b) Concordantly, the ADC map, generated from the DWI data, shows a darkened signal, indicating restricted diffusion at the same level and region as in (a). (c) For comparison, the T2-weighted image at the same level, showing a hypointense lesion (boxed) in the same region of the gland as the dominant lesion seen in (a) and (b). The ADC value in the region of tumor is 680 × 10^-6 mm²/s, while the ADC value in the less suspicious contralateral side of the prostate is 1150 × 10^-6 mm²/s. Thus, both morphologic and quantitative diffusion imaging data support the biopsy finding of prostate cancer on the right side of the gland.

Figure 5. Dynamic Contrast-Enhanced Imaging of Prostate Cancer

The following images are from a 61 year-old man diagnosed with PSA 4.1, Gleason 4+4 = 8, T2cN0M0 prostate cancer. (a) An axial T2-weighted image of the prostate with an unsuspicious, high signal intensity region of PZ prostate (outlined in green). (b) An axial T2-weighted image of the prostate at a different level with a focal region of hypointense signal in the PZ suspicious for prostate cancer (outlined in red). (c) A K^trans (forward volume transfer constant) map at the same level as the unsuspicious region in (a), showing low enhancement (blue indicates low enhancement) in the ROI. (d) In contrast, a K^trans map for the same level as the suspicious region in (b) reveals focal enhancement (yellow and red indicate higher levels of enhancement). (e) A v_e (the fraction of extracellular extravascular space) map for the unsuspicious region, indicating no increased fraction relative to the rest of the prostate. (f) The v_e map for the suspicious region in (b) shows lower signal (light blue) in the ROI compared to the background prostate gland, as expected in prostate cancer. (g) Kinetic curves of gadolinium concentration versus time for the normal (green) and tumor (red) regions of interest. Overall, the region suspicious for tumor demonstrates lower signal intensity on T2-weighted imaging, higher K^trans, lower v_e, and greater enhancement (higher maximum concentration of contrast reached, as seen in the kinetic curves) as compared to the normal region of interest.

Figure 6. MRS Imaging of Prostate Cancer

Images and post-image processing data from a man with biopsy-diagnosed prostate cancer. (a) An axial T2-weighted image of a prostate divided into voxels for MRS imaging analysis. (b) The MR spectra for choline+creatine (the first dominant “composite” peak from the left of each spectrum) and citrate (the second dominant peak from the left). The blue voxel incorporating unsuspicious central gland on T2-weighted imaging (a) shows a normal spectrum along with a normal (Cho+Cr)/Cit ratio, with much more citrate than choline+creatine present (see spectrum also boxed in blue in (b)). Alternatively, the red voxel contains peripheral zone prostate which has T2 hypointensity, making it suspicious for cancer (notably, the associated T1 image was negative for hemorrhage). The voxel's associated spectrum (boxed in red in (b)) shows a higher (Cho+Cr)/Cit ratio (approximately equal peaks) than seen in the unsuspicious blue voxel. This indicates a possible region of tumor in the red voxel's region, especially given the suspicious hypointense signal on T2-weighted imaging.