Editorial Manager(tm) for Topics in Magnetic Resonance Imaging Manuscript Draft

Transcription

1 Editorial Manager(tm) for Topics in Magnetic Resonance Imaging Manuscript Draft Manuscript Number: Title: MR Guided Prostate Interventions Article Type: Review Article Section/Category: Keywords: MR guided prostate therapy prostate brachytherapy prostate biopsy Corresponding Author: Mr Steven John Haker, PhD Corresponding Author's Institution: Brigham and Women's Hospital First Author: Steven John Haker, PhD Order of Authors: Steven John Haker, PhD; Robert V Mulkern, PhD; Joseph R Roebuck, MD, PhD; Agnieska Szot Barnes, MD; Simon DiMaio, PhD; Nobuhiko Hata, PhD; Clare M.C. Tempany, MD Manuscript Region of Origin: UNITED STATES Abstract: We review our experience using an open 0.5 Tesla MR interventional system(gehc Signa SP/i) to guide procedures in the prostate. This system allows access to the patient and real time MR imaging simultaneously, and has made it possible to perform prostate biopsy and brachytherapy under MR guidance. We review MR imaging of the prostate and its use in targeted therapy, and describe our use of image processing methods such as image registration to further facilitate precise targeting. We describe current developments with a robot assist system being developed to aid radioactive seed placement.

2 Manuscript (All Manuscript Text Pages, including Title Page, References and Figure Legends) MR Guided Prostate Interventions Steven J. Haker, PhD Robert V. Mulkern, PhD Joseph R. Roebuck, MD, PhD Agnieska Szot Barnes, MD Simon DiMaio, PhD Nobuhiko Hata, PhD Clare M.C. Tempany, MD Brigham and Women's Hospital Department of Radiology 75 Francis Street Boston, MA Phone: ext Fax: Acknowledgements: The authors wish to thank Robert A. Cormack PhD, Jerome P. Richie MD, Anthony V. D'Amico MD PhD, Angela Kanan RN, Daniel Kacher MS and Janice Fairhurst RT for their professional support. This work is supported by NIH grants R01AG19513 and 1U41RR A2. J. Roebuck is supported by NIH grant 5R25CA

3 Abstract We review our experience using an open 0.5 Tesla MR interventional unit (GEHC Signa SP/i) to guide procedures in the prostate. This system allows access to the patient and real time MR imaging simultaneously, and has made it possible to perform prostate biopsy and brachytherapy under MR guidance. We review MR imaging of the prostate and its use in targeted therapy, and describe our use of image processing methods such as image registration to further facilitate precise targeting. We describe current developments with a robot assist system being developed to aid radioactive seed placement. 2

4 Introduction MR guided interventions are now a well established part of clinical care in many centers around the world. They are performed in a range of clinical conditions and using a range of different systems, devices and approaches. In this article we will review our experience using an open 0.5T MR interventional unit (GEHC Signa SP/i system) to guide procedures in the prostate for both cancer diagnosis and therapy. Prostate cancer is the most common non-cutaneous cancer and the second most common cause of cancer death in American men. For 2004, it has been estimated that 230,110 new cases were diagnosed in the US (1). With the increase in the number of older people in the US (the babyboomers), it is estimated that 450,000 new cases will be diagnosed in (1). The lifetime risk of disease is 16.6% for Caucasian and 18.1% for African-Americans and a lifetime risk of death of 3.5% and 4.3% respectively. The age-adjusted mortality rose an estimated 39% from 1985 to Combined with an aging population, these factors have made prostate cancer a major medical and socioeconomic problem. Prostate cancer remains the key disease for which this imaging is used. The incidence of prostate cancer is high with approx 230,000 men to be diagnosed in the US this year. With the aging of the baby boomers, that number is estimated to increase to over 450,000 by the year The other important fact about prostate cancer is that only approx 4-8% of men with the disease, will actually die of it. It is often thought to be a "normal" cancer of aging, as it is estimated that 80% of men who are 80 years of age will have it. The management of men with clinically localized prostate cancer is one of the most debated and controversial topics in all of clinical medicine today. Basically there are 4 treatment choices: surgery (radical prostatectomy), radiation therapy- either external beam or brachytherapy and observation of "watchful-waiting". Some would correctly argue that the latter is actually not a therapy but is included here as imaging may play a large role in this group. Traditionally, imaging has not been used much by surgeons, due to its perceived lack of sensitivity and more importantly specificity for detecting extraglandular disease. This approach is currently undergoing a relative sea change and with good reason. Despite major advances in screening, increased PSA measurements, detection of cancer and development of nomograms for staging, between 22% and 50% of patients thought to have organ-confined disease have extra-glandular disease at pathology (2,3,4). Magnetic resonance imaging (MRI) of the prostate, especially when performed at 1.5 T (Tesla) with combined endorectal (ER) and pelvic phased array (PPA) coils, provides high resolution images with sufficient anatomic detail to be useful in prostate cancer staging and the determination of extraprostatic disease with up to 82% accuracy. (5, 6). Contrast-enhanced imaging and magnetic resonance spectroscopic imaging (MRSI) are also useful in distinguishing between normal and cancerous tissues. The vascularity of prostate cancers may be increased relative to normal tissue and therefore detected as T1 signal enhancement following intravenous injection of MR contrast agents, such as Gadolinium-DTPA, with relative peak enhancement, time-to-peak and wash-out being of great importance in distinguishing and characterizing cancer. (7-10). Added value comes from MRSI where metabolic differences can distinguish between cancer and healthy tissues. Normal prostate metabolism is characterized by high citrate and low choline/creatine levels, while these ratios are reversed in cancerous tissue (11). Prostate MRSI spectra conveying metabolic information are superimposed on MR anatomic images allowing for precise localization of the tumor. Prostate imaging is now moving towards using higher field 3 Tesla scanners, which provide higher signal-tonoise ratio (SNR) images and allows for better visualization of prostatic substructures and increased spectroscopic resolution. 3

5 Magnetic Resonance Methods for Prostate Examinations When prostate cancer is diagnosed, generally through biopsy, staging examinations are performed to guide patient management. A primary question that then arises is whether the disease is confined to the prostate gland or whether there is extracapsular extension, seminal vesicle invasion, lymph node involvement, or metastases. It is at this stage that MR has proven useful for obtaining information which influences prostate cancer staging within the tumor/nodal/metastases (TNM) scoring framework (12). Various MR-based methods have been devised and are playing an increasingly significant role in patient management decisions. The same methods are also proving capable of determining suspect target tissue within the prostate for guiding local therapies and/or biopsies. Here we discuss these MR-based methods, some standard and some more experimental, currently being used to help with prostate cancer staging and localization. Conventional Imaging The current state-of-the-art hardware for prostate MR generally involves the use of 1.5T scanners together with ER and PPA coils to maximize SNR. Software for spectroscopy and diffusion imaging as well as power injectors for controlled contrast agent administration would largely complete the list for comprehensive prostate imaging. Given only the most basic imaging hardware and software configuration, the conventional MR method for prostate cancer staging is T2WI using a fast spin echo (FSE) sequence and T1WI using either a spin-echo or gradient echo sequence (13-29). The typical exam generally begins with a low resolution/large field-of-view (FOV) rapid gradient echo based 3- plane localizer that serves to a) ensure proper ER coil placement, b) provide guidance for choosing high resolution oblique axial T1WI and T2WI and sagittal and coronal T2WI slice locations, and c) localize the retroperitoneum for subsequent lower resolution T1WI performed using only the PPA coil. The money scans in this case are high resolution (i.e., 12 cm FOV) oblique axial T2-weighted FSE scans with typical effective echo time (ETE) values in the 70 to 100ms range and repetition time (TR) values greater than 2s. These images are used to differentiate the peripheral zone (PZ) from the central gland (CG) and to accurately estimate the transverse and anteroposterior dimensions of the entire gland. In healthy prostate of younger men, considerable differentiation of zonal anatomy is achieved with this approach. The peripheral zone (PZ) is well differentiated from the central gland (CG) by its relative T2-hyperintensity. As men approach the age where prostate cancer becomes more prevalent, benign prostatic hyperplasia (BPH) results in enlargement of the CG where considerable signal intensity variations are observed due to mixtures of stromal and glandular BPH. These naturally occurring, benign T2-signal inhomogeneities within the CG are due to differences in the stromal (low T2-signal) vs. glandular (high T2-signal) tissue and make the identification of prostate cancer within the CG problematic if not impossible from T2WI alone (26, 27). The PZ, however, where approximately some 75% of adenocarcinomas present, remains bright in the absence of disease while cancer causes a decrease in T2-signal that may be focal or diffuse. The hypointensity within the PZ on T2WI is the primary marker for identifying suspected prostate cancer. Unfortunately, reduced T2-signal can also arise from hemorrhage, inflammatory tissue, and treatment effects. T1WI acquired with spin echo or gradient echo sequences with short TR and TE values do not show much contrast in the gland but serve to demonstrate areas of hemorrhage which frequently result from biopsy and may appear dark on T2WI (24). The primary role of the T1WI in evaluation of the gland itself is to rule out hemorrhage as the underlying cause when a focus of abnormally low signal is seen on T2WI. See Figure 1. The high resolution oblique axial T1WI and T2WI sequences should be on the order of 3mm thick, preferably with no gap between slices, should cover the entire gland from apex to base, and should also include the base of the bladder, seminal vesicles, neurovascular bundles, and membranous urethra to assess possible extra-capsular spread. High resolution oblique sagittal and coronal T2WI is used to correlate signal differences seen on axial T1WI and T2W2 in three dimensions and to accurately estimate the cephalocaudal dimension of the entire gland. The three anatomic dimensions are used to calculate the estimated total gland volume, 4

6 which is used in radiation therapy dose calculations. Following high resolution imaging in three planes, the ER coil is removed and lower resolution (i.e., 24 cm FOV) T1WI is performed to assess for metastases to the retroperitoneum or to the bones in the pelvis, both of which affect the TNM staging score. A number of reports over the last decade describe specific imaging findings for different staging tasks such as the detection of extracapsular extension (ECE), seminal vesicle invasion (SVI), tumor location, and tumor volume estimates from T2WI (13-29). Technical improvements such as ER and PPA coil, FSE sequences, and receiver coil sensitivity corrections developed over the last 15 years have steadily reported improved sensitivities and specificities for these tasks. There remains, however, a general understanding that T1WI and T2WI alone is prone to inaccuracies when assessing the locations and volumes of prostate tumors. For example, in one of the first studies which utilized all the modern technical advances, Hricak et al. reported a sensitivity of 78% and a specificity of 55% for determining tumor location (20). Furthermore, it is understood that staging from imaging alone is prone to wide variability among readers with different levels of experience (17). In efforts to improve cancer detection and localization, additional MR-based methods have been developed and applied to the prostate, including MRSI, dynamic contrast enhancement (DCE) imaging, and quantitative diffusion imaging, as we now review in turn. Magnetic Resonance Spectroscopic Imaging While conventional MRI relies on making images from the high concentration of water and lipid protons in the body, MRSI involves detecting the low concentration, low molecular weight metabolites also present in biological tissue. These metabolite signals are typically some 10,000 times weaker than water proton signals, so the water signal and strong signals from lipid protons in fat deposits near the prostate must be suppressed as much as possible in order to appreciate them. The weakness of metabolite signals dictates the size of the voxels that must be sampled in typical MRSI formats. Specifically, while voxels for conventional MRI are on the order of 0.002ml, 3D MRSI studies sample voxels on the order of 0.5ml (14, 16, 21, 30, 31). The metabolite signals of interest within the prostate primarily arise from citrate (Cit), creatine (Cr) and choline (Cho) which resonate at approximately 2.6 parts per million (ppm), 3.05ppm and 3.2ppm, respectively. In practice, the Cr and Cho signals often appear as a single peak due to field inhomogeneity line broadening of the individual resonances. Furthermore, while Cho and Cr are singlet resonances, the Cit signal arises from two coupled protons which, at 1.5T, represent a strongly coupled AB system (32-34). As such the spectral shape of the Cit resonance depends in a complicated manner on the pulse sequence timings. Quantum mechanical calculations (33) have been used to optimize sequence timings for the Cit signal and a typical double echo point resolved echo spectroscopy (PRESS) sequence utilizes echo times on the order of 135ms at 1.5T, simplifying spectral interpretation and measurement of peak areas. The state-of-the-art technique currently used for performing MRSI is a modified PRESS sequence in which three orthogonal slice selective pulses are used to solicit an echo from a selected PRESS Box chosen to encompass the gland, usually using axial T2-weighted FSE images for positioning. Spatial saturation pulses are applied about the box to help reduce lipid contamination from neighboring lipid deposits. The spin echoes are collected in the absence of spatial encoding gradients so that they contain the spectral information from the metabolites. The spatial encoding of the echoes occurs prior to each echo readout with standard 3D phase encoding chemical shift imaging methods (35, 36). Even with very small 3D encoding matrices scan times become quite long due to the need for phase encoding along all three spatial directions. For example, we utilize an 8 x 8 x 16 encoding matrix with a 1s TR so that the scan time is slightly over 17 minutes. Following such an acquisition, fairly sophisticated software for spectral processing and viewing individual spectra from each voxel in the prostate is required. Figure 2 shows how a single slice (one of 8) in the 3D MRSI acquisition may 5

7 be viewed with such software as available from the manufacturer (General Electric Medical Systems, Milwaukee, WI). Note the grid of voxels overlaid upon the corresponding T2-weighted FSE image so that the anatomical location of each spectrum, shown to the right in the figure, can be rapidly assessed. The basic feature of MRS for prostate cancer identification is that distinctly different spectral patterns arise from voxels containing substantial cancer and voxels containing healthy PZ. The presence of PC primarily announces itself in the spectrum by a rise in the signal from the Cho resonance at 3.2 parts per million and a reduction in the signal from Cit at 2.6ppm. For example, in Figure 2 there are several voxels in the bottom row of the grid, along the PZ, which show a larger Cho signal than citrate signal, unlike the spectra in the other rows where citrate signal is larger. The underlying biology behind the PC associated metabolite signal changes has been reviewed by Kurhanewicz et al. (14) and there has been interesting work suggesting a correlation between MRS visible metabolic changes and the biological aggressiveness of cancer (37, 38). In practice, the most straightforward approach to gauging, from MRS, whether a voxel is cancerous is to examine the spectral pattern and determine if the (Cho + Cr) peak is larger than the Cit peak. Such a qualitative impression of the spectral pattern, which takes into account the overall spectral quality, is useful at this stage. Then, if feasible given sufficient SNR and lack of artifact (lipid contamination, poor water suppression, etc.), a numerical measure of the ratio of the combined Cho + Cr signals (integrating these two peaks at 3.2ppm and 3.0ppm together as they are often unresolved under typical experimental conditions at 1.5T) over the peak area of the Cit signal (a complex AB spectral pattern) around 2.6ppm can be assessed. The ratio (Cho + Cr)/Cit from the measured peak areas may be used to classify voxels into suspected cancer vs. non-cancerous tissue. For example, in a study of 85 patients with biopsy proven cancer, Kurhanewicz et al. reported a (Cho + Cr)/Cit ratio of 2.10 ± 1.30 (mean ± SD) in cancer regions vs ± 0.12 in non-cancerous peripheral zone and vs ± 21 in BPH (30). The lack of substantial overlap among the (Cho + Cr)/Cit ratios between the cancerous and non-cancerous prostate tissues is what drives the high specificity associated with MRSI as now reported in a number of studies which have used histological correlations as gold standards (21, 30, 31). The various ways in which the improvements afforded by combining MRI with MRSI for prostate cancer studies has been outlined by Kurhanewicz et al., allowing us to briefly summarize items of potential clinical significance (14, 16). The items are i) improved localization and volume measurement of prostate cancer, ii) improved prediction of extracapsular extension and seminal vesicle invasion, iii) noninvasive assessment of cancer aggressiveness and iv) a non-invasive assessment of cancer response to therapy. The unique metabolic information of MRSI can contribute to each of the items on this list, though low signal-to-noise ratios, poor spatial resolution, lipid contamination, spatially varying water suppression, and susceptibility artifacts affecting the metabolite signals limit the overall utility of MRSI in its present form. Advances will continue to be made with MRSI including the use of higher field strengths and even liquid perfluorocarbon, rather than air, filled balloons to hold the ER coil in place and reduce tissue/air susceptibility effects (39). Dynamic Contrast-Enhanced (DCE) MR Rapid gradient echo imaging performed serially after contrast agent injection allows for measurement of signal intensity changes accompanying the wash-in and wash-out phases of the agent into the tissue. Vascularity and vessel permeability differences between different tissues can lead to differences in the overall time courses of the signal intensity changes and so allow for a means of differentiating tissue types. Using T1-weighted gradient echo images of the prostate, several groups have now explored this avenue towards differentiating prostate cancer from healthy prostate tissue (10, 29, 40-44). Typically 4 to 6 slices are acquired every 6 to 10 s with common 2D-FT gradient echo imaging sequences. Serial imaging over the course of several minutes then allows for the extraction of signal intensity vs. time curves to be plotted from the different tissue types such as suspected cancer regions, healthy peripheral zone, and central gland tissues such as stromal and glandular 6

8 BPH. A recent study by Noworolski et al. utilized both T2-weighted images, 3D MRSI and available biopsy results to identify suspected cancer, healthy PZ, and regions of predominantly stromal vs. glandular BPH (40). A power injector was used to provide a bolus of Gd-DTPA 90 s after scan initiation at a rate of 2ml/s for a final dose of 0.1ml/kg. As with previous studies, the individual time courses from these different regions showed qualitatively similar patterns but with different quantitative aspects. Namely, all curves show an increase in signal as the Gd washes into the tissue followed by a gradual decrease as the Gd washes out. The initial rate of increase, or the initial enhancement slope as calculated from the curves, was lowest in normal PZ, highest in stromal BPH and similar in glandular BPH and cancer. The maximum enhancement measured as a percentage relative to baseline was highest in stromal BPH and lowest in normal PZ. Cancer had higher percent enhancement than glandular BPH with the difference nearly achieving statistical significance in this study of 25 patients (40). Suspected cancer regions also showed faster wash-out rates than healthy PZ, as reported previously. The DCE parameters extracted from this study did show overlap across patients that was only partially removed when normalized using muscle as the control tissue. Despite this, the study, along with previous DCE studies (29 10, 41-44) suggest that DCE may be of value in helping confirm regions of cancer when used in conjunction with MRI and MRSI findings and also may prove useful for assessing CG tumors where both MRI and MRSI have difficulty distinguishing stromal BPH from CG tumor. Diffusion Imaging The magnetic field gradients used for imaging in MRI may also be used to sensitize signal intensities to the diffusive motions of water molecules. This so-called diffusion imaging technique has been particularly useful in the detection of acute stroke (45, 46). Most diffusion imaging studies are carried out using a basic spin echo preparation sequence in which a pair of balanced gradients are placed about the refocusing pulse to sensitize the signal to water diffusion (47). The amplitude, duration and leading edge separation of the two gradient pulses defines the so-called b-factor, b, through which water molecules are sensitized to diffusion, resulting in a signal attenuation generally assumed to be a simple monoexponential decay exp(-bd). Images are usually acquired using three different diffusion sensitization directions at one b-factor, typically 750s/mm 2, and one image is acquired without any, or only a small, diffusion sensitization. Absolute measures of the diffusion coefficient D, can then be made by forming a ratio of the images acquired at high and low b-factors, so as to remove T2- and T1-weighting and allowing for the formation of a pure diffusion image or what we will refer to as a D map. In practice, at least three diffusion sensitization directions are used so that effects of preferential water diffusion, as in white matter fiber tracts, may be averaged out to obtain a rotationally invariant measure of the water diffusion coefficient. The most common diffusion imaging strategy utilizes the diffusion preparation period followed by a rapid single-shot echo planar imaging readout (48). This serves to reduce the effects of motion and, of course, hasten the acquisition at the expense of the usual signal-to-noise loss and geometric distortions associated with EPI methods. Diffusion imaging in the body has been less widely utilized than in the brain for several technical reasons, including its sensitivity to motion. With endorectal coil examinations, however, the prostate is relatively fixed in place so that diffusion imaging is feasible and several preliminary studies (49-52) have now demonstrated that the tissue diffusion coefficient D, may serve as a marker for prostate cancer. In a fairly recent report by Hosseinzadeh and Schwarz (49), 10 men with TRUS-guided biopsy proven prostate cancer were studied with a single-shot echo planar imaging diffusion sequence using endorectal coil and pelvic phase array technology. The values of D were evaluated in the peripheral zone in regions that were classified, on the basis of T2-weighted imaging and available biopsy, as either cancerous or non-cancerous. The mean D values in the regions deemed cancerous were 22% smaller than the mean D values from benign regions and the difference between the means was statistically significant (p<0.01, unpaired, one-tailed t-test). An earlier study, also performed with single-shot diffusion methodology but without the benefit of an endorectal coil, found 7

9 similar results (50). Namely, Issa measured a 26% decrease for D values in cancerous tissue when compared to non-cancerous tissue as distinguished from available imaging and biopsy results (50). The difference between the means was statistically significant (p=0.0005, paired t-test). In general the D values in cancer and peripheral zone. Our own work, performed with a technique known as line scan diffusion imaging (LSDI) (53) revealed similar decreases in the diffusion coefficient within the PZ that were suspected of being cancerous (52). D values for suspected prostate cancer ranged from 0.9 to 1.4µm 2 /ms while healthy PZ typically had D values in the 1.3 to 2.0µm 2 /msec range. The use of LSDI for whole gland D mapping has advantages over EPI methods including much less geometric distortion and minimal susceptibility and chemical shift artifacts. The disadvantage is an overall slower scan time, though whole gland imaging with quantitative LSDI methods can be performed in approximately 5 minutes with voxel volumes on the order of 0.02ml. The lower two images of Figure 3 show the T2-weighted image and an LSDI D map (left and right, respectively) of a prostate cancer patient. There is subtle hypointensity in the central portion of the PZ on the T2-weighted image which appears as a decreased area of signal intensity in the D map. Elevated (Cho + Cr)/Cit spectral patterns shown in the upper two images and spectra from this region make the likelihood of cancer in this region difficult to ignore. Interventional MR The prostate program at our center was established following the successful MR guided neurosurgical program, using the same 0.5T Signa SP open magnet, which allows access to the patients and real time imaging simultaneously. The development of an interventional MR therapy (MRT) system has made it possible to perform prostate biopsy and brachytherapy under MR guidance. Even at lower field strength than routinely used for prostate cancer imaging (i.e., 0.5T versus 1.5T), MRI provides images of good quality for target visualization, as well as identification of the urethra and rectum. Computer software has been developed to provide dosimetry analysis used for both treatment planning and monitoring based on intraoperative MR images (54). Image processing methods adapted from brain surgery are available to further facilitate precise radiation delivery to the prostate gland while sparing surrounding tissues. Currently, treatment delivery with a robot assist system is being developed and tested to improve radioactive seed placement. One great strength of MRI is its sensitivity to temperature changes (55-57). This sensitivity allows for monitoring the delivery of several thermal energy treatments, including radiofrequency (RF) and laser therapy for brain tumors, and cryotherapy and high-intensity focused ultrasound surgery (FUS) for breast, prostate, liver, and uterine lesions. Currently, MRI is a very useful guidance method for cryotherapy because it allows monitoring of the size of the ice ball formed in multiple dimensions (58). Intraoperative MR images are used to depict the slow expansion of the ice ball as well as tissue damage caused by the freezing process. MR-guided FUS is a very promising method for noninvasive cancer treatment. While other minimally invasive therapies require direct insertion of special probes to reach the tumor, this method uses a high-intensity ultrasound beam focused on the target lesion (as seen on MR) without disruption of skin and other tissues. FUS is based on the use of acoustic energy and its secondary thermal effect, which causes thermal coagulation of the target tissue. As early as 1955 it was clinically shown that FUS was capable of destroying mammalian tissues (59). Broad use of this treatment method has been hindered by lack of appropriate image guidance techniques for tumor targeting and most important, real-time monitoring of temperature changes. The introduction of MR-guidance provided an excellent method for monitoring treatment planning and delivery with direct temperature mapping (using MR phase-contrast techniques), as well as post-treatment confirmation of necrotic tissue changes (20). Using MRI, the physician can identify the target lesions and temperature change during treatment delivery. A special transducer moves from one spot to another according to a pre-treatment plan until the entire volume is treated. To date this method has been successfully used in the 8

10 treatment of breast fibroadenomas, cancers and uterine leiomyomata (60-62). Application of this treatment in prostate cancer, liver lesions and brain tumors is currently under investigation. Prostate Cancer Diagnosis Prostate cancer is most commonly diagnosed by transrectal ultrasound-guided needle biopsy (TRUS), prompted by either an elevated PSA level or a palpable nodule at digital rectal examination (DRE). At this time, over million prostate needle biopsies are performed per year in the United States. Many men face a negative biopsy result and a rising PSA level. It is estimated that nearly 25 million men have had at least one negative prostate biopsy. What a dilemma this is for both the patient and the doctor: is it truly negative or should the PSA be aggressively pursued with repeat biopsies? Current techniques offer no answer, with the recent American Joint Committee on Cancer staging handbook stating there are substantial limitations in the ability of digital rectal examination and TRUS to define precisely the size or local extent of prostate cancer (63). In regard to therapy, as prostate cancer generally affects older men and is not always fatal or rapidly growing, a safe and side effect free method of cancer control would be highly desirable. Many men undergo radical surgery for minimal disease and end up with devastating side effects, which significantly reduce the quality of their lives. Therefore, the potential societal impact of developing sensitive prostate biopsy and effective image-guided prostate cancer therapy is great. The sextant biopsy technique, a systematic, spatially distributed set of six biopsy cores first described in 1989, when combined with TRUS guidance, is now the minimum standard of care for the diagnosis of prostate cancer (64, 65). TRUS guidance does not target specific lesions, but rather samples six representative locations in the gland. Using this technique, biopsy of men with PSA values in the range of 4-10 ng/ml, has generally resulted in a detection rate of 20% -30% (66, 67). Numerous studies have shown that TRUS-guided prostate biopsy misses cancer in at least 20% of cases (68, 69, 70). Still others have found that TRUS biopsies are limited by low sensitivity of 60% (71, 72). TRUS suffers from a sampling error, with an 8-30% failure rate in the detection of lesions that are palpable at DRE (73, 74). The gloomy consequence is that in more than 20% of cancers, at least 2 biopsy procedures are required to diagnose the tumor, which amounts to 200,000 cases annually in the US. Other studies have suggested that an optimal systematic approach might require 8 or 10 biopsy cores (65, 75). In fact some groups advocate a saturation biopsy approach (76). Saturation techniques obtain a greater number (typically over 20 sites) of systematically distributed biopsy cores to increase the volume of sampled tissue. Although increasing the number of random TRUS biopsy samples yields a marginal improvement in detection rate (69, 73, 74), only a very small fraction of the gland is sampled with even a dozen needle placements. Thus, it is unlikely that simply increasing the number of samples taken will solve the problem. Improvements in target identification before needle biopsy offers a potential solution. Perrotti et al. have used the ER coil assisted MRI findings of suspected tumor foci to guide the placement of needles during TRUS-guided biopsy (77). They found that the accuracy of the TRUS-guided biopsy aided by MRI was improved to 67%. Also important, the handheld TRUS probe applies variable normal force on the prostate through the rectal wall, thereby causing dynamically changing deformation and displacement of the prostate and surrounding tissue, thus further reducing the accuracy of needle placement. Altogether, TRUS has a poor track record in imaging focal cancer but, even if the cancer-target could be visualized, there is no guarantee that it can be sampled with a conventional handheld biopsy needle. The subject of needle biopsy of the prostate is one of much debate and needs further research. An accurate and controllable image-guided biopsy method, capable of demarcating targets within the gland most likely to contain tumor and successfully positioning the sampling needle within them, could result in a diminished false negative rate of needle biopsy. The process of screening for prostate cancer is further complicated by controversy over the definition of significant disease requiring therapy. Conservative management or watchful waiting may be 9

11 reasonable for lower grade, localized prostate cancer (78). Analysis of prostate cancer volumes suggests that tumors with a volume of less than 0.5ml are unlikely to be clinically significant (79). Currently there are no reliable methods to detect biologically aggressive tumors and there are many men who have indolent tumors detected that will never reach clinical significance. The ideal non-invasive diagnostic test for prostate cancer would limit the number of patients subjected to needle biopsy, by providing limited, targeted biopsy of the image defined significant lesions. More specifically, since the benefit of therapy in low grade, low volume tumors of the prostate is uncertain, this diagnostic test should selectively detect moderate to high grade, medium to large volume cancers. While no single imaging test has this ability today, it is probable that it will require multiple clinical parameters and multiple imaging modalities to reach this important goal. MR-Guided Prostate Biopsy In addition to being an excellent method for guiding prostate cancer therapy, MR imaging also appears to be useful for guiding diagnostic biopsy (80). Similar to its use in therapy, metabolic information from spectroscopy and dynamic contrast MR data can be combined with routine MR images to allow precise tumor targeting. Our group has adapted an interventional MR system to perform MR-guided prostate biopsy (80, 81). This transperineal technique does not use endorectal devices and provides an excellent diagnostic alternative for patients who have undergone rectal surgeries and in whom US-guided procedure is impossible to perform. An additional group of men who can benefit from MR-guided procedure are those with persistently rising PSA values and prior negative US-guided biopsies. Preliminary results of this method in facilitating prostate cancer diagnosis are promising (81). One of the unique aspects of this approach is the interactive imaging provided by using the 3D Slicer, as reported by Hata et al. (80), which facilitates T2 imaging in near-real time. D Amico et al. (82) reported results of the procedure from two MRI-targeted lesions in a patient who could not undergo US-guided procedure because of previous rectal surgery. Several transurethral biopsies yielded negative results in this patient. Following MR-guided biopsy, cancer was confirmed. Procedure Prior to the procedure, each patient undergoes endorectal coil MRI using a 1.5 T imaging system. The T1 and T2-weighted and contrast-enhanced images are collected, and multivoxel spectroscopy may be performed. Using this information the radiologist identifies biopsy targets. Patient positioning for the procedure and initial preparation is similar to MRBT except that an endorectal obturator may not be used in some cases with previous rectal surgery. Subsequently, T2-weighted images are collected at 3.5 mm intervals in a 0.5 Tesla interventional magnet. We first perform sextant sampling of the gland under MRI guidance in a way similar to TRUS, that is, we target the six locations on the gland. We sample the PZ at the apex, base, and midgland bilaterally. All specimens are labeled, sent to pathology and handled in the standard clinical fashion. For biopsy localization, each sextant of the gland can be targeted as they are under ultrasound guidance. We image in the coronal plane and divide the gland into three areas - apex, base, and midgland. We sample the gland using a transperineal approach not the transrectal approach used in TRUS-guided biopsy of the prostate. Thus, the needles enter the posterior gland and travel in a trajectory through the PZ, increasing the probability that all samples will be taken from the PZ. Realtime imaging with FGR sequences, and registered T2W images are be used to confirm that the samples are being taken from the PZ. Initial pre-biopsy 0.5T T2W MR images are analyzed by a radiologist for the presence of any focal lesion, defined as any focal area of low signal intensity in the PZ. The information from preprocedure and intraprocedure images are registered (as described below) and correlated, and target lesions are 10

12 identified in addition to the sextant locations. Computer software is used to calculate appropriate coordinates on the perineal template for transperineal needle insertion, as well as needle insertion depth. All target locations along with sextant biopsies of the PZ from the right and left apex, mid-gland and base are sampled using MR-compatible 18 gauge biopsy guns. This procedure is currently under general anesthesia as a day surgical procedure. It is well tolerated and offers a second line biopsy approach in selected patients. Intra-Procedural Real-Time Image Display and Guidance The MR scanner provides 2D real-time (every six seconds) FGR images of the prostate during the procedure. This allows visualization of the prostate gland, rectum, bladder, catheter and needle position. However, the FGR images do not provide adequate contrast for the visualization of the substructure of the gland, namely the peripheral zone (PZ) and suspicious targets. These sub-structures, critical for precise biopsy sampling, are provided by T2W imaging. Axial and coronal T2W images are obtained after the patient has been correctly positioned. These data sets typically consist of 20 images, with mm pixel size, 5 mm slice thickness, and 256 matrix size. Since these quality T2 images cannot be obtained in real time, we use the following technique to simulate real-time T2 imaging. As each 2D real-time FGR image is obtained, either the axial or coronal T2W image data set is resampled, pixel-by-pixel, at the same spatial locations as those of the FGR image. To limit interpolation artifacts, the particular choice of which T2W data set is sampled depends on whether the FGR image is closer in orientation to axial or coronal. As each simulated real-time T2W image is obtained, alternating views of the most recent FGR and T2W images are presented on a monitor in the bore of the scanner. A 1-2 second delay between the alternating images is used. In this way, the radiologist may see the needle artifact and its position relative to the PZ and suspicious foci. Using the described method to combine FGR and T2W images, we have successfully guided the biopsy needle to suspicious foci and assured proper placement of the needle within the peripheral zone. This is now used in all cases. See Figures 4 and 5. Using image registration methods, we can also provide for views of imaging taken before the day of surgery. Registration is the process of bringing two or more image data sets, possibly of different modalities or taken at different times, into spatial alignment. In this case, pre-operative prostate imaging must be brought into alignment with MR imaging taken in the operating room. In this way, all imaging can be used to aid in the determination of targets as well as guidance during treatment. The literature devoted to image registration is too vast to be summarized here. There are numerous techniques that have been developed for rigid as well as non-rigid registration using single or multiple imaging modalities. We have recently published a survey paper (83), which provides an overview of the field. A major difficulty in registration comes from the shape changes that can occur in soft tissues between imaging sessions. We have found, for example, that significant prostate shape changes occur between pre-operative 1.5T endorectal coil imaging, in which the patient is supine, and intraoperative 0.5T imaging, during which the patient is in the lithotomy position. This shape change is likely the result of changes in patient position and rectal filling necessitated by the procedures. We have performed a quantitative analysis of this deformation, reported in (84). We have also performed (85, 86) a study to develop and evaluate a finite element based non-rigid registration system. In this study, our goal was to register pre-operative MR imaging to intra-operative imaging obtained during MR-guided prostate brachytherapy. We registered 0.5 Tesla (T) intra-operative magnetic resonance images of the prostate with pre-treatment 1.5T images. The pre-operative 1.5T MR imaging was conducted with the patient supine, using an endorectal coil, while intra-operatively, the patient was in the lithotomy position with a rectal obturator in place. The method was based on an extension of an algorithm we developed for brain registration (86, 87). It is suitable for application in prostate matching, since it models basic biomechanical properties soft tissue. In summary, the method involves the following steps: 1) A 3D tetrahedral model of the entire prostate is created from segmented pre-operative 1.5T images; 2) the boundary surface of the capsule is extracted from this 11

13 tetrahedral mesh and is registered to a corresponding capsule surface obtained from intra-operative images; 3) the surface point matches from step 2 are used as boundary conditions when solving a finite element-based system of equations which models the volumetric deformation field within the gland; and, 4) the volumetric deformation field from the previous step is used to interpolate preoperative imaging data. Given two prostate capsule surfaces, obtained from segmentations of pre-operative and intraoperative images, we map them to a sphere using a conformal surface warping technique. Then, using the vertices of a standard triangulation of the sphere, we effectively re-triangulate the original prostate surfaces in a consistent manner, providing a point-by-point registration of the surfaces. We use a spherical triangulation which is the boundary of the high-quality tetrahedral meshing of the unit ball. The tetrahedral mesh within the prostate surface defines the set of finite elements used to discretize a variational form of a partial differential equation describing the bio-mechanics of the prostate. The volumetric deformation field is found by solving for the displacement that minimizes an associated energy. The finite element technique reduces the problem to the solution of a sparse system of linear equations that can be solved for the unknown displacements at each interior vertex. After creating a dataset of manually segmented glands from images obtained in 10 MR-guided brachytherapy cases, we conducted a set of experiments to assess our hypothesis that the proposed registration system can significantly improve the quality-of-matching of total gland, central gland, and peripheral zone over rigid registration alone. The results showed that the method provided a statistically significant improvement in the quality of registration. Specifically, it raised the Dice similarity coefficient, a measure of volume agreement ranging from 0.0 (no agreement) to 1.0 (exact agreement) from pre-matched coefficients of 0.81, 0.78, and 0.59 for TG, CG and PZ, respectively, to 0.94, 0.86 and The volumes of CG and PZ remained constant before and after the registration, indicating that the method maintained the biomechanical topology of the prostate. This registration technique has been successfully tested numerous times in an operating room setting, and is now part of our standard procedure during MR-guided prostate biopsy. See Figure 6. In addition to registration of pre-operative imaging, information on the likely spatial distribution of tumor locations can be registered to the patient intra-operatively and used for needle guidance. We have begun a collaboration with Drs. Shen and Davatzikos of the University of Pennsylvania to use an optimized prostate biopsy strategy based on the statistical distribution of prostate cancer observed in a large number of patients (88,89). Drs. Shen and Davatzikos created a digital repository of 260 surgically resected prostates, with complete and digitally recorded anatomy and histopathology for each specimen. They also developed a processing front-end engine for the digital data. Among its many appealing features, this engine can take a 3D model of an arbitrary prostate as an input and superimpose a statistical map over the model, and mark the biopsy locations and paths to these locations that are expected to produce maximum yield, based on statistical analysis. Prostate Cancer Treatment Established therapeutic approaches for managing prostate cancer include one or a combination of radical prostatectomy (RP), external beam radiation therapy (EBRT), Brachytherapy (BT), hormonal therapy, or deferred treatment, so called watchful-waiting (90). Appropriate treatment is tailored for the individual patient and the clinical outcome strongly depends on the stage of the cancer. The clinical efficacy is gauged by the probability that the cancer will not recur in the prostate after treatment and by the probability of normal tissue complications. For several decades, the definitive treatment for low-risk prostate cancer was RP or EBRT. During the 90 s the technique of TRUSguided brachytherapy underwent improvement and today is nearly equivalent to RP and EBRT, in terms of both 5-7 year local control rates (91-95) and 10-year disease-free survival rates (96-99). 12

14 Prostate cancer is a multifocal disease requiring the treatment of the entire gland; however, it is also highly desirable to selectively target all suspicious and proven cancerous areas with higher dose (so called dose escalation ). This approach may increase the likelihood of stopping the cancer from spreading outside the prostate, without risking overdosing neighboring healthy tissues. In contemporary practice, indiscriminate high-dose treatment of areas where cancer is unlikely to be present may cause significant adverse side effects. One such high-risk area is the urethra, which is in the center of the gland, in the transitional/ peri-urethral zone. If the urethra is over-radiated, severe complications such as stricture and fistula may occur, which are for the most part avoidable with good image guidance (100, 101). Intensity Modulated Radiation Therapy (IMRT) is a technique using variable radiation field intensities in a manner that delivers the desired dose distribution. When used properly it allows the radiation oncologist to deliver a higher radiation dose to target, without exceeding normal tissue tolerances, than is possible with conventional external beam methods. This offers the possibility that biologically relevant regions, as defined by advanced MR, PET or other imaging, can be targeted for increased dose in order to improve the therapeutic ratio. In order to do this, tools must be developed to bring the imaging information from the diagnostic imaging devices into the realm of radiation oncology, which is almost universally CT, or X-ray based. EBRT planning requires accurate delineation of the biologically relevant tissue, in the coordinate system defined by the patient position in the treatment machine. This requires tools for multimodal image registration including deformable registration. External beam treatments are generally delivered in multiple fractions over a number of days. In order to increase dose to a target without damaging neighboring normal structures the radiation therapist should be able to determine the location of the relevant structures on a daily basis. This represents an area for the incorporation of diagnostic information in the verification of daily patient setup based on imaging available at the treatment machine. MR-Guided Brachytherapy Brachytherapy is potentially superior to external beam radiotherapy in that there is no entrance or exit dose. It therefore has the potential to achieve sharp demarcation between irradiated volume and healthy structures, and thus to achieve superior tumor control with significantly reduced morbidity and side effects compared to other treatment modalities and techniques (96, 102). Although recent literature reports relatively favorable rates of long-term complications (96, 97, ), adverse side effects still often follow treatment. Side effects include rectal irritation and ulceration, incontinence, and impotence; ultimately related to the inadvertent delivery of radiation dose to the rectum, urethra, and bladder. Since MRI images can identify the tumor, the prostate border and the rectal wall, it can allow a more careful guided delivery of the radiation. As MRI imaging techniques improve, we hope to identify more cancer, assess its biological activity, guide and monitor the changes induced by therapy, and track needle placement in real time. Altogether, the evolution of brachytherapy has shown that improved local control can improve global control of the disease and also reduce global toxicity, while still treating the entire gland. Patient Selection The patient selection criteria for the MR-guided brachytherapy program in our institution are: clinical stage T1cNXM0 (according to AJCC), PSA less than 10 ng/ml, biopsy Gleason score not more than 3 + 4, low cancer volume, and endorectal MRI demonstrating organ confined disease. Patients with prior transurethral resection of the prostate (TURP) are excluded. We do not exclude men with larger volume prostates, as pubic arch interference can be avoided in this approach. All patients undergo endorectal coil MRI for prostate cancer staging prior to the treatment. A radiologist assesses prostate gland volume, tumor location and volume, the presence or absence of extraglandular disease, seminal vesicle invasion (SVI), and possible spread to pelvic lymph nodes or bones. 13

15 Procedure This multidisciplinary procedure uses many different computer, imaging, and technical skills and therefore requires the cooperation of specialists from various medical and nonmedical fields, including radiation oncologists, medical physicists, radiologists, anesthesiologists, urologists, nurses, radiology technologists, and computer scientists. For the procedure the patient is placed in an open configuration 0.5 T Signa SP MR scanner (GE Medical Systems, Milwaukee WI) in the lithotomy position. The patient is positioned on the table between two magnets with vertically oriented open space for easier access to the patient during the treatment. A Foley catheter is inserted, skin prepared, the template for needle guidance placed against the patient s perineum and secured, and a rectal obturator is inserted. T2-weighted MR images are acquired in axial, coronal, and sagittal planes. The radiologist uses the T2-weighted images to identify the peripheral zone (PZ), urethra, and anterior rectal wall on each axial MR slice. These are then outlined using the 3D Slicer surgical simulation software designed and operated by members of the Surgical Planning Laboratory (SPL) at Brigham and Women s Hospital in Boston and the Massachusetts Institute of Technology. The 3D Slicer is free, open-source software for two- and three-dimensional display, registration, and segmentation of medical images. Pretreatment planning, as well as calculation of the MRI-based peripheral zone as a clinical target volume (CTV), is then performed by the medical physicists using designated planning software (54). The number of I-125 seeds per catheter and the depth of catheter insertion are calculated. The physicians then insert each preloaded catheter in the prostate gland. After every catheter insertion, axial gradient-echo MR images are obtained in real-time and compared to the catheter s expected location according to the plan. Dose volume histograms (DVH) for the CTV, anterior rectal wall, and urethra are recalculated, adjustment of the catheter placement is performed when necessary, and seeds are deposited. Approximately six weeks after the procedure, MRI and CT imaging of the prostate is performed to identify the location of radioactive seeds and calculate final DVHs. Since seeds can be well visualized on CT images, and the underlying anatomy is better depicted on MR images, MR-CT fused images are used to calculate dose distribution to the surrounding tissues. Outcomes Long term biochemical outcomes have been compared for similar patients over similar time frames between MR-guided brachytherapy and radical prostatectomy by D Amico et al (91). At 5-years, PSA control was 95% for brachytherapy and 93% for RP patients (median follow-up was 3.95 years and 4.2 years for brachytherapy and RP patients, respectively). The percentage of positive prostate biopsies was found to be a significant predictor of the time to post-brachytherapy PSA failure. Shortterm toxicity following MR-guided brachytherapy was rare, and no patient reported gastrointestinal or sexual dysfunction during the first month after treatment (108). Acute Urinary Retention (AUR) was observed in 12% of men within 24 hours of removal of the Foley catheter and was self-limiting within 1 to 3 weeks. MR-determined prostate volume, transitional zone (TZ) volume, and total number of seeds were found to be significant predictors of AUR on univariable analysis. The TZ volume was the only significant predictor of AUR on multivariable logistic regression analysis. The authors concluded that benign prostatic hyperplasia (BPH) that results in larger TZ volume, is the most important predictor of AUR. No urinary incontinence was seen at a median follow-up of 14 months (from 9 months to 2 years) (109). MR-guided brachytherapy is a new approach, thus there is only one report to date summarizing long-term toxicity (110). Albert et al. (110) found low incidence of rectal bleeding (8%) and no urethral strictures at a median follow-up of 2.8 years (0.5 to 5 years). While ED reached 82%, two-thirds of the patients reported good erectile function after sildenafil (Viagra). No radiation cystitis was estimated at 4 years after MRBT. Quality of life (QoL) outcomes collected using a previously validated questionnaire (111) are currently being assessed, and early reports indicate that MR-guided prostate brachytherapy has better symptomatic outcomes than the conventional US- 14

16 guided approach (J. Talcott, personal communication). Current research projects will study further the radiation dose distribution to vital organs and its impact on side effects. Image segmentation techniques are used to identify those important organs on endorectal coil MR images. Radiation dose to the organs can then be correlated with changes in patient-reported QoL. Future directions Robotics In collaboration with the National Institute of Advanced Industrial Science and Technology (AIST, Japan), we have developed a first-generation interventional robotic-assistant for needle trajectory guidance during image-guided percutaneous therapy in an open-mri system ( ). While prior work has established the technical and clinical feasibility for MRI-guided biopsy, the manual method of needle placement has remained unchanged. To our frustration, the use of a fixed, needle template guide with holes spaced 5mm apart, significantly limits needle placement resolution and constrains needle orientation. In addition, template registration and the manual computation and transcription of coordinates are prone to human error. We have developed a system that integrates an interactive planning interface with an MR-compatible robotic assistant that acts as a dynamic needle guide for precise, yet flexible needle placement. In collaboration with the National Institute of Advanced Industrial Science and Technology (AIST) in Japan, we have developed a MRI-compatible robotic manipulator for percutaneous interventions in an open-mr unit. Its development has marked the first comprehensive analysis of material and mechatronic compatibility of powered MRI-compatible surgical robots (114). The device consists of a three-degree-of-freedom Cartesian positioning stage and a two-degree-of-freedom orienting mechanism mounted above the surgeon's head in the open MRI magnet. Two long, rigid arms reach into the surgical space and form a parallel linkage for manipulating an acrylic needle holder, or guide. The five motion stages are driven by ultrasonic motors (Shinsei USR-60N) attached to lead screws and motion is measured by optical encoders with 10 µm resolution (Encoder Technology, Cottonwood, AZ). A Flashpoint sensor is attached to the needle holder to provide independent redundant encoding. This robotic device has been integrated with a software planning interface (built into the 3D Slicer) and a tracking and control system for percutaneous interventions in the prostate under MR-guidance (115). The physician interacts with the planning interface to specify a set of desired needle trajectories based on anatomical structures and lesions observed in the patient's MR images. All image-space coordinates are computed and used to position the needle guide automatically, thus avoiding the limitations of the traditional fixed template guide. Once the needle holder is in position, the robot remains stationary while the physician manually inserts the needle through the guide and into the tissue with real-time imaging for monitoring progress. This system has been tested extensively in phantoms. We have evaluated the MRI compatibility of the robot and all components in a 0.5T intra-operative MRI scanner and have observed no adverse effects on image quality. In fact, the robot caused less field distortion than the patient's body. In phantom trials, we used soft PVC models of the prostate and its surrounding tissues. Glass beads embedded in the prostate model were visible in the images and served as targets for tests. With the image-guided system, we were able to place the needle tip within two millimeters of the target points during experiments (115). New Imaging Modalities and Higher Field Strengths As we expand the capabilities of our MR-guided prostate therapy program and respond to the clinical needs of men with prostate disease it is clear that there are several major challenges ahead. As newer imaging modalities will be introduced such as PET with C11, we will require improved non-rigid registration methods which accommodate the multiple modalities. Also as the MR imaging techniques 15

17 and cancer detection rates improve with 3T and new pulse sequences, we will need to change the paradigm of interventions in low-mid field units into closed bore high field systems. This will require new devices and methods to allow access to the prostate. One solution is to adapt the field of MR and surgical robotic assist devices into the pelvis and closed bore MR units. Another direction of research is MR guided focused ultrasound (FUS) for prostate cancer treatment. US guided FUS or HIFU has been in clinical use for several years in Europe and elsewhere. MR guided FUS can be applied to the prostate for focal tumor ablation and early applicator design and testing is underway. References 1. Cancer facts and figures. Atlanta GA American Cancer Society Partin, A.W., et al., Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. Jama, (18): p Hull, G.W., et al., Cancer control with radical prostatectomy alone in 1,000 consecutive patients. J Urol, (2 Pt 1): p Han, M., et al., Biochemical (prostate specific antigen) recurrence probability following radical prostatectomy for clinically localized prostate cancer. J Urol, (2): p Comet-Batlle J, Vilanova-Busquets JC, Saladie-Roig JM, Gelabert-Mas A, Barcelo-Vidal C. The value of endorectal MRI in the early diagnosis of prostate cancer. Eur Urol Aug;44(2):201-7; discussion Hricak H, White S, Vigneron D, Kurhanewicz J, Kosco A, Levin D, et al. Carcinoma of the prostate gland: MR imaging with pelvic phased-array coils versus integrated endorectal-pelvic phased-array coils. Radiology 1994;193: Barentsz JO, Engelbrecht M, Jager GJ, Witjes JA, de LaRosette J, van Der Sanden BP, Huisman HJ, Heerschap A. Fast dynamic gadolinium-enhanced MR imaging of urinary bladder and prostate cancer. J Magn Reson Imaging. 1999, Sep;10(3): Huisman HJ, Engelbrecht MR, Barentsz JO. Accurate estimation of pharmacokinetic contrastenhanced dynamic MRI parameters of the prostate. J Magn Reson Imaging Apr;13(4): Engelbrecht MR, Huisman HJ, Laheij RJ, Jager GJ, van Leenders GJ, Hulsbergen-Van De Kaa CA, de la Rosette JJ, Blickman JG, Barentsz JO. Discrimination of prostate cancer from normal peripheral zone and central gland tissue by using dynamic contrast-enhanced MR imaging. Radiology Oct;229(1): Epub 2003 Aug Engelbrecht MR, Huisman HJ, Laheij RJ, Jager GJ, van Leenders GJ, Hulsbergen-Van De Kaa CA, de la Rosette JJ, Blickman JG, Barentsz JO. Discrimination of prostate cancer from normal peripheral zone and central gland tissue by using dynamic contrast-enhanced MR imaging. Radiology Oct;229(1): Epub 2003 Aug Coakley FV, Qayyum A, Kurhanewicz J. Magnetic resonance imaging and spectroscopic imaging of prostate cancer. J Urol Dec;170(6 Pt 2):S69-75; discussion S Presti JC, Jr. Prostate cancer: Assessment of risk using digital rectal examination, tumor grade, prostate-specific antigen, and systematic biopsy. Radiol Clin North Am 2000;38: Yu KK, Hricak H. Imaging prostate cancer. Radiol Clin North AM 2000;38: Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer. J Magn Reson Imag 2002;16: Yu KK, Hricak H, Alagappan R, Chernoff DM, Bacchetti P, Zaloudek CJ. Detection of extracapsular extension of prostate carcinoma with endorectal and phased-array coil MR imaging: multivariate feature analysis. Radiology 1997; 202: Kurhanewicz J, Vigneron DB, Males RG, Swanson MG, Yu KK, Hricak H. The prostate: MR imaging and spectroscopy. Present and Future. Radiol Clin North Am 2000;38:

18 17. May F, Treumann T, Dettmar P, Hartung R, Breul J. Limited value of endorectal magnetic resonance imaging and transrectal ultra-sonography in the staging of clinically localized prostate cancer. BJU Int 2001;87: D'Amico AV, Whittington R, Malcowicz SB, et al. Critical analysis of the ability of endorectal coil magnetic resonance imaging to predict pathological stage, margin status, and postoperative prostate-specific antigen failure in patients with clinically organ-confined prostate cancer. J Clin Oncol 1996;14: D'Amico AV, Whittington R, Schnall M, et al. The impact of the inclusion of endorectal coil magnetic resonance imaging in a multivariate analysis to predict clinically unsuspected extraprostatic prostate cancer. Cancer 1995;75: Hricak H, White S, Vigneron D, et al., Carcinoma of the prostate gland: MR imaging with pelvic phased-array coils versus integrated endorectal-pelvic phased array coils. Radiology 1994;193: Coakley FV, Kurhanewicz J, Lu Y, Jones KD, Swanson MG, Chang SD, Carroll PR, Hricak H, Prostate cancer tumor volume: Measurement with endorectal MR and MR spectroscopic imaging. Radiology 2002;223: Jager GJ, Rijter ETG, van da Kaa C, de la Rosette, Oosterhof GON, Thornbury JR, Barentsz JO. Local staging of prostate cancer with endorectal MR imaging: Correlation with histopathology. AJR 1996;166: Tempany CM, Zhou X, Zerhouni EA, et al., Staging of prostate cancer: Results of radiology oncology group project comparison of three MR imaging techniques. Radiology 1994;192: White S, Hricak H, Forstner R, Kurhanewicz J, Vigneron DB, Zaloudek CJ, Weiss JM, Narayan P, Carroll PR. Prostate cancer: Effect of postbiopsy hemorrhage on interpretation of MR images. Radiology 1995;195: Schiebler ML, Schnall MD, Pollack HM, Lenkinski RE, Tomaszewski JE, Wein AJ, Whittington R, Rauschning W, Kressel HY. Current role of MR imaging in the staging of adenocarcinoma of the prostate. Radiology 1993;189: Grossfeld GD, Coakley FV. Benign prostatic hyperplasia: Clinical overview and value of diagnostic imaging. Radiol Clin North Am 2000;38: Ling D, Lee JK, Heiken JP, et al. Prostatic carcinoma and benign prostatic hyperplasia: Inability of MR imaging to distinguish between the two diseases. Radiology 1986;158: van Dorsten FA, van der Graaf M, Engelbrecht MRW, van Leenders GJLH, Verhofstad A, Rijpkema M, de la Rosette JJMCH, Barentsz JO, Heerschap A. Combined quantitative dynamic contrast-enhanced MR imaging and 1H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imag 2004;20: van Dorsten FA, van der Graaf M, Engelbrecht MRW, van Leenders GJLH, Verhofstad A, Rijpkema M, de la Rosette JJMCH, Barentsz JO, Heerschap A. Combined quantitative dynamic contrast-enhanced MR imaging and 1H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imag 2004;20: Kurhanewicz J, Vigneron DB, Hricak H, Narayan P, Carroll P, Nelson SJ. Three-dimensional H-1 MR spectroscopic imaging of the in situ human prostate with high ( cm3) spatial resolution. Radiology 1996;198: Scheidler J, Hricak H, Vigneron DB, et al., Prostate cancer: Localization with three-dimensional proton MR spectroscopic imaging-clinicopathologic study. Radiology 1999;213: Mulkern RV, Bowers JL. Calculating spectral modulations of AB systems during PRESS acquisitions. Magn Reson Med 1993;30: Mulkern RV, Bowers JL. Density matrix calculations of AB spectral modulations: Quantum mechanics meets in vivo spectroscopy. Conc Magn Reson 1994;6:

19 34. Mulkern RV, Bowers JL, Peled S, Kraft RA, Williamson DS. Citrate signal enhancement with a homonuclear J-refocusing modification to double-echo PRESS sequences. Magn Reson Med 1996;36: Brown TR, Kincaid BM, Ugurbil K. NMR chemical shift imaging in three dimensions. Proc. Natl. Acad. Sci. 179: ; Maudsley, A.A.; Hilal, S.K.; Perman, W.H.; Simon, H.E. Spatially resolved high-resolution spectroscopy by "four-dimensional" NMR. J. Magn. Reson. 51: ; Vigneron D, Males R, Hricak H, Noworolski S, Carrol PR, Kurhanewicz J. Prostate cancer: Correlation of 3D MRSI metabolite levels with histologic grade. Proceedings of RSNA, Chicago, Zakian KL, Sircar K, Hricak H, Chen H-N, Shukla-Dave A, Eberhardt S, Muruganandham M, Ebora L, Kattan MW, Reuter VE, Scardini PT, Koutcher JA. Correlation of proton MR spectroscopic imaging with Gleason score based on step-section pathologic analysis after radical prostatectomy. Radiology 2005;234: Choi H, Zhou XJ, Ma J. Use of perfluorocarbon (PFC) in magnetic resonance spectroscopy (MRS) of prostate: A method to improve the linewidth and quality of spectra. Book of Abstracts, RSNA 2004; Noworolski SM, Henry RG, Vigneron DB, Kurhanewicz J. Dynamic contrast-enhanced MRI in normal and abnormal prostate tissues as defined by biopsy, MRI and 3D MRSI. Magn Reson Med 2005;53: Ito H, Kamoi K, Yokoyama K, Yamada K, Nishimura T. Visualization of prostate cancer using dynamic contrast-enhanced MRI: comparison with transrectal power Doppler ultrasound. Brit J Radiol 2003;76: Preziosi P, Orlacchio A, Di Giambattista G, Di Renzi P, Bortolotti L, Fabiano A, Cruciani E, Pasqualetti P. Enhancement patterns of prostate cancer in dynamic MRI. Euro Radiol 2003;13: Padhani AR, Gapinski CJ, Macvicar DA, Parker GJ, Suckling J, Revell PB, Leach MO, Dearnaley DP, Husband JE. Dynamic contrast enhanced MRI of prostate cancer: correlation with morphology and tumor stage, histological grade and PSA. Clin Radiol 2000;55: Turnbull LW, Buckley DL, Turnbull LS, Liney GP, Knowles AJ. Differentiation of prostatic carcinoma and benign prostatic hyperplasia: correlation between dynamic Gd-DTPA enhanced MR imaging and histopathology. J Magn Reson Imag 1999;9: Koroshetz WJ, Gonzalez G. Diffusion-weighted MRI: An ECG for "Brain Attack". Ann Neurol 1997;41: Lutsep HL, Albers GW, DeCrespigny A, Kamar GN, Marks MP, Moseley ME. Clinical utility of diffusion-weighted magnetic resonance imaging in the assessment of ischemic stroke. Ann Neurol 1997;41: Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a timedependent field gradient. J Chem Phys 1965;42: Turner R, Le Bihan D, Chesnick AS. Echo-planar imaging of diffusion and perfusion. Magn Reson Med 1991;19: Hosseinzadeh K, Schwarz SD. Endo-rectal diffusion weighted imaging in prostate cancer to differentiate malignant and benign peripheral zone tissue. J Magn Reson Imag 2004; Issa B. In vivo measurement of the apparent diffusion coefficient in normal and malignant prostatic tissues using echo-planar imaging. J Magn Reson Imag 2002;16: Gibbs P, Tozer DJ, Liney GP, Turnbull LW. Comparison of quantitative T2 mapping and diffusion-weighted imaging in the normal and pathologic prostate. Magn Reson Med 2001;46:

20 52. Chan I, Wells III W, Mulkern RV, Haker SJ, Zhang J, Zou KH, Maier SE, Tempany CMC. Detection of prostate cancer by integration of line-scan diffusion, T2-mapping and T2-weighted MR imaging; a multi-channel statistical classifier. Med Phys 2003;30: Gudbjartsson H, Maier SE, Mulkern RV, Morocz IA, Patz S, Jolesz FA. Line scan diffusion imaging. Magn Reson Med 1996;36: Kooy HM, Cormack RA, Mathiowitz G, Tempany C, D'Amico AV: A software system for interventional magnetic resonance image-guided prostate brachytherapy. Comput Aided Surg 2000; 5(6): Cline HE, Hynynen K, Hardy CJ, et al: MR temperature mapping of focused ultrasound surgery. Magn Reson Med 1994; 31(6): Jolesz FA, Hynynen K: Magnetic resonance image-guided focused ultrasound surgery. Cancer J 2002; 8(Suppl 1):S100-S Hynynen K, Darkazanli A, Damianou CA, Unger E, Schenck JF: Tissue thermometry during ultrasound exposure. Eur Urol 1993; 23(Suppl 1): Jolesz FA, Nabavi A, Kikinis R: Integration of interventional MRI with computer-assisted surgery. J Magn Reson Imaging 2001; 13(1): Fry WJ, Barnard JW, Fry FJ, Krunins RF, Brennan JF: Ultrasonic lesions in mammalian central nervous system, Science. 1955; 122: Hynynen K, Pomeroy O, Smith DN, et al: MR imaging-guided focused ultrasound surgery of fibroadenomas in the breast: a feasibility study, Radiology 2001; 219(1): Huber PE, Jenne JW, Rastert R, et al: A new noninvasive approach in breast cancer therapy using magnetic resonance imaging-guided focused ultrasound surgery. Cancer Res 2001; 61(23): Tempany CM, Stewart EA, McDannold N, et al: MR imaging-guided focused ultrasound surgery of uterine leiomyomas: a feasibility study. Radiology 2003; 226(3): The American Joint Committee on Cancer (AJCC) Cancer Staging Manual, Sixth Edition Hodge KK, McNeal JE, Terris MK, Stamey TA. Random systematic versus directed ultrasound guided transrectal core biopsies of the prostate. Journal of Urology 142:71, Presti JC Jr, Chang JJ, Bhargava V, Shinohara K. The optimal systematic prostate biopsy scheme should include 8 rather than 6 biopsy cores: results of a prospective clinical trial. Journal of Urology 163: , Terris MK., Wallen EM, et al. (1997). "Comparison of mid-lobe versus lateral systematic sextant biopsies in the detection of prostate cancer." Urol Int 59(4): Roehl KA, Antenor JA, et al. (2002). "Robustness of free prostate specific antigen measurements to reduce unnecessary biopsies in the 2.6 to 4.0 ng./ml. range." J Urol 168(3): Norberg M, Egevad L, et al. (1997). "The sextant protocol for ultrasound-guided core biopsies of the prostate underestimates the presence of cancer." Urology 50(4): Rabbani R, Stroumbakis N, Kava BR, Incidence and clinical significance of false negative sextant prostate biopsies. J Urol 1998; 159: Wefer AE., Hricak H, et al. (2000). "Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology." J Urol 164(2): Terris MK., McNeal JE, et al. (1993). "Efficacy of transrectal ultrasound-guided seminal vesicle biopsies in the detection of seminal vesicle invasion by prostate cancer." J Urol 149(5): Keetch DW, McMurtry JM, et al. (1996). "Prostate specific antigen density versus prostate specific antigen slope as predictors of prostate cancer in men with initially negative prostatic biopsies." J Urol 156(2 Pt 1): Epstein JI, Walsh PC, Sauvagerot J et al. Use of repeat sextant and transition zone biopsies for assessing extent of prostate cancer. J Urol 158;

21 74. Stroumbakis N, Cookson MS, Reuter VE, et al. Clinical significance of repeat sextant biopsies in prostate cancer patients. Urology Suppl 49; Gore JL, Shariat SF, Miles BJ, Kadmon D, Jiang N, Wheeler TM, Slawin KM. Optimal combinations of systematic sextant and laterally directed biopsies for the detection of prostate cancer. Journal of Urology. 165:1554-9, Stewart CS, Leibovich BC, Weaver AL, Lieber MM. Prostate cancer diagnosis using a saturation needle biopsy technique after previous negative sextant biopsies. Journal of Urology 166(1):86-92, Perrotti M, Han KR, et al. (1999). "Prospective evaluation of endorectal magnetic resonance imaging to detect tumor foci in men with prior negative prostastic biopsy: a pilot study." J Urol 162(4): Chodak GW, Thisted RA, Gerber GS, Johansson JE, Adolfsson J, Jones GW, Chisholm GD, Moskovitz B, Livne PM, Warner J. Results of conservative management of clinically localized prostate cancer. New England Journal of Medicine 330: , Stamey TA, Freiha FS, McNeal JE, Redwine EA, Whittemore AS, Schmid HP. Localized prostate cancer. Relationship of tumor volume to clinical significance for treatment of prostate cancer. Cancer 71(3 Suppl): , Hata N, Jinzaki M, Kacher D, et al: MR imaging-guided prostate biopsy with surgical navigation software: device validation and feasibility. Radiology 2001; 220(1): Cormack RA, D'Amico AV, Hata N, et al: Feasibility of transperineal prostate biopsy under interventional MR guidance. Urology 2000; 56(4): D'Amico AV, Tempany CM, Cormack R, et al: Transperineal magnetic resonance image guided prostate biopsy. J Urol 2000; 164(2): Warfield SK, et al. "Capturing Intraoperative Deformations." Medical Image Analysis, to appear. 84. Hirose M, Bharatha A, Hata N, Zou KH, Warfield SK, Cormack R, D'Amico A, Kikinis R, Jolesz F, Tempany C, "Quantitative MR Imaging Assessment of Prostate Gland Deformation before and during MR Imaging-guided Brachytherapy", Acad. Rad., 9, 8, pp Bharatha A, Hirose M, Hata N, Warfield SK, Ferrant M, Zou KH, Suarez-Santana E, Ruiz- Alzola J, D'Amico A, Cormack RA, Kikinis R, Jolesz FA, Tempany CM. ``Evaluation of Three- Dimensional Finite Element-based Deformable Registration of Pre- and Intra-Operative Prostate Imaging.'' Med Phys 28: , Haker S, Warfield SK, Tempany CMC. Landmark-Guided Surface Matching and Volumetric Warping for Improved Prostate Biopsy Targeting and Guidance, MICCAI 2004: 7th International Conference, Saint-Malo, France, September 26-29, 2004, pp Ferrant M, Warfield SK, Nabavi A, Macq B, and Kikinis R, "Registration of 3D Intraoperative MR Images of the Brain Using a Finite Element Biomechanical Model," presented at Third International Conference on Medical Image Computing and Computer-assisted Intervention, Pittsburgh, PA, USA, Shen D, Lao Z, Zeng J, Herskovits EH, Fichtinger G, Davatzikos D, A Statistical Atlas of Prostate Cancer for Optimal Biopsy, Lecture Notes in Computer Science, Volume 2208, Jan 2001, p Shen D, Lao Z, Herskovits EH, Fichtinger G, Davatzikos C, Zeng J. "Statistically Optimized Biopsy Strategy for the Diagnosis of Prostate Cancer," 14th IEEE Symposium on Computer- Based Medical Systems (CMBS'01), 2001 p Peschel RE., Colberg JW "Surgery, brachytherapy, and external-beam radiotherapy for early prostate cancer." Lancet Oncol 4(4): 2000, D'Amico AV, Tempany CM, Schultz D, Cormack RA, Hurwitz M, Beard C, Albert M, Kooy H, Jolesz F, Richie JP. Comparing PSA outcome after radical prostatectomy or magnetic 20

22 resonance imaging-guided partial prostatic irradiation in select patients with clinically localized adenocarcinoma of the prostate. Urology Dec;62(6): D'Amico AV, Tempany CM, Schultz D, Cormack RA, Hurwitz M, Beard C, Albert M, Kooy H, Jolesz F, Richie JP. Comparing PSA outcome after radical prostatectomy or magnetic resonance imaging-guided partial prostatic irradiation in select patients with clinically localized adenocarcinoma of the prostate. Urology Dec;62(6): D'Amico AV, J. W. Moul, et al. (2003). "Surrogate end point for prostate cancer-specific mortality after radical prostatectomy or radiation therapy." J Natl Cancer Inst 95(18): Zelefsky MJ, et al., Comparison of the 5-year outcome and morbidity of three-dimensional conformal radiotherapy versus transperineal permanent iodine-125 implantation for early-stage prostatic cancer. Journal of Clinical Oncology, (2): p Zelefsky MJ., Hollister T, et al. (2000). "Five-year biochemical outcome and toxicity with transperineal CT-planned permanent I-125 prostate implantation for patients with localized prostate cancer." Int J Radiat Oncol Biol Phys 47(5): Blasko JC., Mate T, et al. (2002). "Brachytherapy for carcinoma of the prostate: techniques, patient selection, and clinical outcomes." Semin Radiat Oncol 12(1): Merrick GS, Wallner K, et al. (2001). "Short-term sexual function after prostate brachytherapy." Int J Cancer 96(5): Beyer DC. (2001). "The evolving role of prostate brachytherapy." Cancer Control 8(2): Stokes SH, "Comparison of biochemical disease-free survival of patients with localized carcinoma of the prostate undergoing radical prostatectomy, transperineal ultrasound-guided radioactive seed implantation, or definitive external beam irradiation." Int J Radiat Oncol Biol Phys 47(1): 2000, Lee WR, et al., A prospective quality-of-life study in men with clinically localized prostate carcinoma treated with radical prostatectomy, external beam radiotherapy, or interstitial brachytherapy. Int J Radiat Oncol Biol Phys, (3): p Zelefsky MJ, Yamada Y, et al. (2003). "Improved conformality and decreased toxicity with intraoperative computer-optimized transperineal ultrasound-guided prostate brachytherapy." Int J Radiat Oncol Biol Phys 55(4): King CR, DiPetrillo TA, Wazer DE, Optimal radiotherapy for prostate cancer: predictions for conventional external beam, IMRT, and brachytherapy from radiobiologic models. Int J Radiat Oncol Biol Phys Jan 1;46(1): Talcott, J.A., et al., Long-term treatment related complications of brachytherapy for early prostate cancer: a survey of patients previously treated. J Urol, (2): p Merrick GS., et al., Rectal function following prostate brachytherapy. International Journal of Radiation Oncology, Biology, Physics, (3): p Krumholtz, J. S., J. M. Michalski, et al. (2000). "Health-related quality of life and morbidity in patients receiving brachytherapy for clinically localized prostate cancer." J Endourol 14(4): Alexianu M, Weiss GH. Radical prostatectomy versus brachytherapy for early-stage prostate cancer. J Endourol May;14(4): Chan AJ, Saltzman B. Role of brachytherapy in patients with stage T1c, T2a adenocarcinoma of the prostate. J Endourol May;14(4): D'Amico AV, Cormack R, Kumar S, Tempany CM: Real-time magnetic resonance imagingguided brachytherapy in the treatment of selected patients with clinically localized prostate cancer. J Endourol 2000; 14(4): Thomas MD, Cormack R, Tempany CM, et al: Identifying the predictors of acute urinary retention following magnetic-resonance-guided prostate brachytherapy. Int J Radiat Oncol Biol Phys 2000: 47(4):

23 110. Albert M, Tempany CM, Schultz D, Chen MH, Cormack RA, Kumar S, et al. Late genitourinary and gastrointestinal toxicity after magnetic resonance image-guided prostate brachytherapy with or without neoadjuvant external beam radiation therapy. Cancer 2003;98(5): Clark JA, Talcott JA: Symptom indexes to assess outcomes of treatment for early prostate cancer. Med Care 2001; 39(10): Chinzei K, Kikinis R and Jolesz FA (1999). MR compatibility of mechatronic devices: Design criteria. Medical Image Computing and Computer-Assisted Intervention, Miccai'99, Proceedings. 1679: Chinzei K, Hata N, Jolesz FA and Kikinis R (2000). MR compatible surgical assist robot: System: integration and preliminary feasibility study. Medical Image Computing and Computer- Assisted Intervention - Miccai : Chinzei K and Miller K (2001). "Towards MRI guided surgical manipulator." Med Sci Monit 7(1): DiMaio SP, Pieper S, Chinzei K., Fichtinger G, Tempany C, and Kikinis R. "Robot assisted percutaneous intervention in open-mri." 5th Interventional MRI Symposium, Boston, October 15-16, p

24 Figure Captions Figure 1. On the left, T2W FSE axial image of the prostate. On the right, a T1W image at the same spatial location. The arrows indicate a region of the peripheral zone which is hypointense in both the T1 and T2 images. Such a location would be a target for biopsy. Figure 2. On the left, a standard T2W FSE image, with grid showing corresponding voxels used for spectral analysis. On the right, spectra corresponding to the gridded regions. The bottom row of spectra indicate possible cancer. Figure 3. To the left is the standard T2W FSE image corresponding to the spectroscopic slice shown in Figure 2. To the right is the trace diffusion D map calculated from an LSDI DTI acquisition of the same slice. Note the decrease in D from the peripheral zone, extending more to the patients right than left, in the D map. The spectroscopic results contain (Cho + Cr)/Cit ratios consistent with prostate cancer in this same region (see Figure 2) while the standard T2W FSE image does not show particularly suspicious hypointensity in this region. Figure 4. Coronal images displayed in the bore of the 0.5T scanner during MR-guided biopsy. On the left, FGR imaging obtained every 3-4 seconds. The black arrow indicates the biopsy needle artifact. On the right, registered pre-procedure T2W image showing pre-planned target (white arrow). These views are alternated for viewing by the physician. Figure 5. Axial images displayed during the biopsy procedure. On the left, real-time FGR image at 0.5T, with black arrow indicating the biopsy needle artifact. In the middle, registered pre-procedure T2W image at 0.5T showing pre-planned target (arrow). On the right, combined 1.5T T2W and 0.5T T2W image (See Figure 6). These views are alternated for viewing by the physician. Figure 6. Preoperative 1.5T (left), intraoperative 0.5T (middle), and combined (right) T2W MR images. In the combined image, registered preoperative imaging is shown within the prostate, while in the surrounding area the intraoperative image is shown. Note that significant soft tissue deformation has occurred between scans. 23

25 Figure 1. On the left, T2W FSE axial image of the prostate. On the right, a T1W image at the same spatial location. The arrows indicate a region of the peripheral zone which is hypointense in both the T1 and T2 images. Such a location would be a target for biopsy. Figure 2. On the left, a standard T2W FSE image, with grid showing corresponding voxels used for spectral analysis. On the right, spectra corresponding to the gridded regions. The bottom row of spectra indicate possible cancer. 24

26 Figure 3. To the left is the standard T2W FSE image corresponding to the spectroscopic slice shown in Figure 2. To the right is the trace diffusion D map calculated from an LSDI DTI acquisition of the same slice. Note the decrease in D from the peripheral zone, extending more to the patients right than left, in the D map. The spectroscopic results contain (Cho + Cr)/Cit ratios consistent with prostate cancer in this same region (see Figure 2) while the standard T2W FSE image does not show particularly suspicious hypointensity in this region. Figure 4. Coronal images displayed in the bore of the 0.5T scanner during MR-guided biopsy. On the left, FGR imaging obtained every 3-4 seconds. The black arrow indicates the biopsy needle artifact. On the right, registered pre-procedure T2W image showing pre-planned target (white arrow). These views are alternated for viewing by the physician. 25

27 Figure 5. Axial images displayed during the biopsy procedure. On the left, real-time FGR image at 0.5T, with black arrow indicating the biopsy needle artifact. In the middle, registered pre-procedure T2W image at 0.5T showing pre-planned target (arrow). On the right, combined 1.5T T2W and 0.5T T2W image (See Figure 6). These views are alternated for viewing by the physician. Figure 6. Preoperative 1.5T (left), intraoperative 0.5T (middle), and combined (right) T2W MR images. In the combined image, registered preoperative imaging is shown within the prostate, while in the surrounding area the intraoperative image is shown. Note that significant soft tissue deformation has occurred between scans. 26