TRUS-guided prostate biopsy is by far the most common and frequently used intervention in relation to prostate cancer, given its advantages of relative ease of use, availability, and low cost. It is the initial diagnostic step in the vast majority of cases to obtain tissue samples for pathological evaluation in men suspected of having prostate cancer. TRUS imaging of the prostate has a long history using gray scale and Doppler techniques, even though newer inversion pulsed harmonics and, now in certain select sites, IV microbubble contrast, and ultrasound elastography are under investigation. The basic TRUS imaging used to guide transrectal needle sampling involves a simple visualization of the gland and subdivision of this gland volume into sextants for systematic biopsies. The division allows for 6–12 cores to be removed in an orderly fashion in an attempt to provide uniform gland sampling. Much debate surrounds the recommended approach that ranges from numbers of cores to locations. In general the standard approach is 12 cores with peripheral/lateral aspects of the gland sampled using an 18-gauge (G) side-cutting biopsy needle. The TRUS-guided biopsy procedure is typically performed in an outpatient setting. The patient usually takes prophylactic antibiotics for 1–2 days prior and after the biopsy. Local anesthesia can be used; it is injected into the periprostatic nerves. A technologist can perform TRUS imaging while the urologist acquires the biopsy samples, but more commonly, a single physician does both. Each biopsy sample is labeled, either right or left, and occasionally, more location-specific details are provided. This is, however, very variable and often leads to difficulty in correlating imaging and pathology findings. Each sample is examined separately, and a diagnosis is made between benign, premalignant, prostatic intraepithelia, neoplasia, and adenocarcinoma. In the case of adenocarcinoma, a Gleason score is assigned, depending upon cell types and degrees of differentiation. The percentage tumor volume in each sample is also recorded. The final report also includes the number of positive samples relative to the total samples harvested.

If one critically assesses TRUS image quality, its overall accuracy for detection, characterization, and locoregional staging of prostate cancer then MRI far surpasses TRUS.

In 1997, the Brigham and Women’s Hospital established the first direct transperineal prostate biopsy program using 0.5 T open MRI scanner and validated this approach in over 50 men [45, 46]. To optimize the biopsy yield, the prostate gland is sampled transperineally, thereby maximizing contact with the PZ that is statistically the most likely location of potential cancer and allowing easy access to all parts of the gland, especially the anterior part. This work established a new approach to prostate biopsies, and many prominent research groups followed suit. The concept was derived from brachytherapy. The needles were introduced freehand using the reference frame of a needle-guidance template commonly used in brachytherapy with grid of holes spaced 5 mm apart. The template was registered to a 0.5 T open MRI scanner using an optical tracking system integrated to the MRI scanner. This registration achieved geometric correlation between the template and patient anatomy, enabling the selection of a hole that will guide a biopsy needle towards the target determined from the MRI scans. By using the template for needle guidance, the needle path was restricted due to the fixed direction of guide holes in the template. In all patients, both targeted and random sampling biopsies were obtained, to allow for the evaluation of the efficacy of targeted sampling and maximizing the diagnostic yield for each patient. To improve needle guidance, navigation software was created by adding specialized modules to 3D Slicer, the powerful open-source visualization and surgical guidance platform [21]. First, target definition and planning functions were added to 3D Slicer. These functions were performed by measuring the coordinates of the suspicious tumors foci and selecting a corresponding hole in the template grid to reach the targets. Target definition was carried out using T2-weighted images after the patient went under general anesthesia. A key feature of the system was volumetric data registration that allows for planning the targets on preoperative high-resolution T2-weighted images mapped onto intraoperative 0.5 T images. With the use of the same volumetric mapping method, patterns like statistical maps or spectroscopic MRI were superimposed on the intraoperative images. Finally, the reconstructed 3D anatomy was blended with all forms of information mentioned above. Due to significant prostate shape changes occurring between preoperative high-field endorectal coil imaging (legs down) and intraoperative 0.5 T imaging (legs up), a deformable registration method based on finite element modeling was developed [47, 48]. The targeting and planning system outlined above also established the tradition of using 3D Slicer as the visualization interface of choice in a long line of experimental image-guided intervention systems.

Efficient and accurate biopsy targeting in the real-time open-MRI environment demanded methods for needle tracking. Needles are visible in MRI due to their susceptibility artifact. Paramagnetic needles when imaged produce field distortions that appear as local regions of signal loss and, in some cases, signal enhancement [49]. The size and location of the artifact change with the material properties of the needle, imaging signal, image resolution, and the relative orientation of the needle, B0 field and the gradient field. In other words, the needle is usually not where it appears to be in the MR image. There is a compromise between imaging speed and quality that impacts needle localization, accuracy, and reliability [50]. The susceptibility artifacts produced by needles in magnets of various field strengths were characterized. Techniques have been proposed for optimizing the visualization of image artifacts and driving the scan plane [51]. Recent work reported on the feasibility of mapping the spatial displacement of the needle susceptibility artifact versus the true needle position [52, 53].

A detailed investigation of needle placement accuracy study concluded that precise and accurate placement of needle may be best achieved by robotic needle placement [54].

New 3 T magnet bore size at 70 cm diameter now allow us to perform transperineal 3T MR-guided prostate biopsy. We at Brigham and Women’s Hospital (BWH) now perform this regularly at 3 T, with the patient in prone position, under intravenous conscious sedation (IVCS) (Figs. 56.9 and 56.10). The perineal template is used and manual needle insertion is performed under MR guidance (Figs. 56.6 and 56.9). We obtain a full multiparametric MR exam several days to weeks prior to the biopsy procedure. These images are used to define targets for sampling. The images are imported into 3D Slicer (Fig. 56.11) and registered to the procedure images with Z-frame, using the method introduced by Fichtinger et al. [18]. This approach provides the unique opportunity to obtain site-specific pathological samples from image abnormalities.

This unique ability to validate the MR findings along with the improvements of MR for detection and characterization of focal disease allows for investigation of image-guided focal therapy. Many centers are using the MR and US capabilities to guide High Intensity Focused Ultrasound (HIFU), cryotherapy, laser or photodynamic therapy. We have completed preclinical trials in MR-guided FUS surgery (Fig. 56.12a, b) and demonstrated our ability to induce focal prostate necrosis. The exciting feature of this approach to focal therapy is the ability to closely monitor the temperature change in near real-time with MR thermometry (Fig. 56.2b). This allows for accurate placement of sonications and immediate feedback of therapeutic result. New clinical trials are underway which will allow carefully assessment of this exciting new approach to focal therapy for prostate cancer have begun and are underway.

To marry the convenience of TRUS with the superior anatomical visualization of MRI in prostate biopsy, fusion of preoperative high-resolution MRI and intraoperative real-time TRUS has been proposed [55–57]. This method critically depends on the assumption of spatially correct co-registration between intraoperative decubitus ultrasound and pre-procedural supine MRI in the presence of extreme tissue motion and deformation, a problem that, at the moment, does not have a clinically robust solution. A sufficiently accurate and robust solution will undoubtedly turn up, but this approach will always suffer from the lack of direct and immediate pathological evidence of whether the registration was indeed correct and the biopsy was executed as planned in MRI. In contrast to the image fusion-based approach, in situ MRI-guided biopsy eliminates spatial uncertainty and doubt from the procedure.

In addition to previously mentioned challenges, MRI-guided robotic biopsy involves more difficulties, each of which we will elaborate on next.