**4.1. The clinical problem**

phantoms were then imaged with our hybrid scanner, viewed in the 3D visualization software, and manually segmented. The volume of spherical structures were calculated and compared

Testing the 3D hybrid scanner with the 3D thread phantom showed that mean error in the measured values of the distances in the X, Y and Z directions were 3.6%, 2.5% and 5.7% respectively. A one-sample t-test was performed to compare the measured distance values with the known distance value of 1cm, showed there was no statistical significant difference

Validation of volume measurements using the hybrid scanner were carried out by imaging a

agar phantom. The measurements were performed at two different depth settings on the ultrasound machine (10 and 15 cm). The mean errors of the volume measurement were 5.7% and 4.4% for the 10cm and 15cm depth settings respectively, demonstrating that the hybrid

*In-vivo experiments:* After obtaining institutional research board (IRB) approvals, we investigat‐ edtheuseofthescannerinthermalablationtreatmentofprimaryhepatictumors.Figure7shows a 3D US image acquired during the microwave ablation procedure of a primary (hepatocellu‐ lar)tumor.Itshowsapplicationofthehybridmodeinacquiringvolumeslargeenoughtoinclude

**Figure 7.** 3D ultrasound image of a primary (hepatocellular) tumor with two microwave applicators in place. The ap‐ plicators and tumor have been segmented and displayed in 3D allowing the interventional radiologist to examine the placement accuracy of the applicators in the tumor. In addition, the ablation zone has also been superimposed.

embedded in a block of tissue mimicking

between the measured values and expected values between the threads.

scanner can be used to make sufficiently accurate volumetric measurements.

both the ablated tumor region as well as all ablation needles in two different views.

tissue mimicking agar sphere with a volume of 10 cm3

with the expected values.

12 Advancements and Breakthroughs in Ultrasound Imaging

*3.2.3. Validation results*

Prostate Cancer (PCa) is the most commonly diagnosed malignancy in men, and is found at autopsy in 30% of men at age 50, 40% at age 60, and almost 90% at age 90 [66, 67]. Worldwide, it is the second leading cause of death due to cancer in men, accounting for between 2.1% and 15.2% of all cancer deaths [68, 69]. Symptoms PCa are generally absent until extensive local growth or metastases develop. When diagnosed at an early stage, the disease is curable [70, 71], and even at later stages treatment can be effective [72]; however, once the tumor has extended beyond the prostate, the risk of metastases and locally aggressive cancer increases. Clearly, early diagnosis, accurate staging of prostate cancer, and appropriate therapies are critical to the patient's well-being.

In managing patients with possible PCa, the challenges facing physicians are to: (a) diagnose clinically relevant cancers at a curable stage; (b) stage the disease accurately; (c) apply appro‐ priate therapy accurately to optimize destruction of cancer cells while preserving normal tissues and function; and (d) follow patients to assess side effects and therapy effectiveness. This section focuses on improving early PCa diagnosis and staging with the use of 3D ultrasound-guided prostate biopsy.

Since not all cancers are palpable by digital rectal exam (DRE), PCa diagnosis is established by histological examination of prostate tissue obtained most commonly by trans-rectal ultrasound (TRUS)-guided biopsy. Prostate needle biopsy is the only definitive diagnostic modality capable of confirming malignancy, and is now always performed with TRUS guidance.

Since many small tumors are not detected by TRUS or DRE, biopsy samples are obtained from predetermined regions of the prostate known to have a high probability of harboring cancer. These are typically in the peripheral zone (PZ), which harbors 80% of all PCs and a higher proportion of clinically significant ones, and close to the capsule, as most cancers are thought to start within 5*mm* of the prostate capsule. Most centers are now taking 8-12 cores or more as part of their routine assessment [73-76].

TRUS biopsies are now performed with a thin, 18-gauge needle mounted on a spring-loaded gun connected to the TRUS probe, forcing the needle to stay in the imaging plane. Each core is separately identified as to the prostate region from which it was drawn, so that the pathol‐ ogist can report the extent and grade of the cancer within each region.

Since prostate volume sampled by the biopsy is small, and PCa is often multi-focal, involving only a small volume of the prostate in the early stages of the disease [77, 78], the probability for obtaining a sample of the tumour on biopsy is small. Thus, a negative biopsy may be, in fact, false, and the patient may be harbouring cancer at an early and curable stage. Various reports have shown that the false negative rate ranges from 10% to 25% [73, 74]. Since cancer is still present in 1/10 to 1/4 of patients with a negative first biopsy, the current biopsy procedure is still suboptimal [74, 79]. Clearly, an improved procedure with improved planning and recording of biopsy locations is necessary to resolve these issues.

Due to the increasing number of younger men with early and potentially curable PCa under‐ going repeated prostate biopsy, it is therefore vital not to re-biopsy the same area if the original biopsy was negative, and it is particularly vital to re-biopsy the same area if a possible abnormal area was detected on first biopsy as ASAP [80]. Thus, the locations of the cores obtained from the prostate must be known accurately to help guide the physician during the repeat biopsy [81, 82], to help in correlating any imaging evidence of the disease, and to provide improved planning for subsequent therapy.

modeling and segmentation algorithms, and system validation using a multi-modal US/CT

3D Ultrasound Imaging in Image-Guided Intervention

http://dx.doi.org/10.5772/55230

15

Our approach involves the use of a device composed of two mechanisms shown as a schematic in Figure 8. The system is composed of an articulated multi-jointed stabilizer and a transducer

**Figure 8.** A schematic diagram of the mechanical tracker, which supports the TRUS transducer and attached cradle. This configuration constrains the TRUS probe motion to three degrees-of-freedom and one degree of translation along the axis of the probe. The system is mounted at the base of a stabilizer while the linkage allows the TRUS trans‐ ducer to be manually manipulated about a remote center of motion (RCM), which is at the center of the ultrasound

The end-firing TRUS transducer with the biopsy needle guide in place is mounted to the mechanical tracking mechanism in a manner where the US probe is free to rotate around its longitudinal axis (Fig. 8). The tracking assembly is attached to a stabilizer, which is mounted on a free-standing cart. Thus, the physician can manipulate the tracking mechanism freely, insert the transducer through the anus, and rotate the transducer in order to acquire a 3D image of the prostate. The tracking linkage contains angle-sensing encoders mounted to each joint in order to transmit to the computer the angles between the arms. This arrangement allows the computer to determine the relative position of the transducer as it is being manipulated. Since the biopsy gun is mounted onto the transducer and its position relative to the transducer is

The mechanical tracking device is a spherical linkage assembly, in which the axis of the joints converge to a common point on the remote center of motion (RCM). The RCM design mini‐ mizes targeting errors within the prostate. As the TRUS transducer is constrained through a stationary point, the physician's movements are replicated at a scaled down rate (minified through the RCM), minimizing changes in morphology and dislocation of the prostate. In addition, the RCM enables a precision equivalent to that of robotic assisted machines. Thus,

prostate phantom.

tracking mechanism.

transducer tip.

calibrated, the needle location can be calculated.

#### **4.2. Multi-modality directed prostate biopsy**

A variety of imaging techniques and molecular imaging probes are being investigated to improve early detection of PCa. Different magnetic resonance imaging (MRI) techniques have been evaluated using body and endo-rectal coils, contrast enhancement, and different pulse sequences [83-85] resulting in disease detection sensitivity and specificity of 80-88% and 75-95%, respectively [84, 86, 87]. Positron emission tomography (PET) (combined with CT or MRI) is used to detect early disease, with the newer PET imaging probes proving to be the more promising [88-90]. Although progress has been made with improved PET and MRI techniques, they do not yet have ideal specificity or sufficient accuracy to assess the grade of the cancer; thus a biopsy of suspicious lesions on MRI or PET is required to provide a definitive diagnosis and grade of the disease. Systems have been developed to perform biopsies in the MRI suite; however, the cost of the equipment and prolonged use of the MRI is extremely expensive and likely prohibitive given the large number of patients requiring biopsy. Un‐ fortunately, conventional 2D TRUS guidance of the biopsy procedure limits the physician's ability to target locations identified as suspicious on other modalities.

As we currently do not have a highly sensitive and specific imaging test for local staging of PCa, there is a growing belief that the optimal method to guide prostate biopsy will involve not just one, but a combination of imaging modalities. 3D TRUS imaging combined with functional or molecular imaging from another imaging modality such as radiopharmaceutical imaging (PET, SPECT), or magnetic resonance imaging (MRS, MRI) may provide the best approach for guiding prostate biopsy.

#### **4.3. 3D TRUS-guided prostate biopsy system**

Since ultrasound imaging is the clinical standard for image-guided biopsy of the prostate, we have developed a 3D TRUS-based navigation system that provides a reproducible record of the 3D locations of the biopsy targets throughout the procedure and allows fusion with MR images with identified lesions for targeting.

The system we have developed is a mechanical 3D biopsy system that maintains the procedural workflow, minimizing costs and physician retraining. This mechanical system has 4 degreesof-freedom (DOF) and has an adaptable cradle that supports commercially available end-firing TRUS transducers used for prostate biopsy [16]. It also allows real time tracking and recording of the 3D position and orientation of the biopsy needle as the physician manipulates the TRUS transducer. The following describes the components of the system, including hardware, modeling and segmentation algorithms, and system validation using a multi-modal US/CT prostate phantom.

Due to the increasing number of younger men with early and potentially curable PCa under‐ going repeated prostate biopsy, it is therefore vital not to re-biopsy the same area if the original biopsy was negative, and it is particularly vital to re-biopsy the same area if a possible abnormal area was detected on first biopsy as ASAP [80]. Thus, the locations of the cores obtained from the prostate must be known accurately to help guide the physician during the repeat biopsy [81, 82], to help in correlating any imaging evidence of the disease, and to provide improved

A variety of imaging techniques and molecular imaging probes are being investigated to improve early detection of PCa. Different magnetic resonance imaging (MRI) techniques have been evaluated using body and endo-rectal coils, contrast enhancement, and different pulse sequences [83-85] resulting in disease detection sensitivity and specificity of 80-88% and 75-95%, respectively [84, 86, 87]. Positron emission tomography (PET) (combined with CT or MRI) is used to detect early disease, with the newer PET imaging probes proving to be the more promising [88-90]. Although progress has been made with improved PET and MRI techniques, they do not yet have ideal specificity or sufficient accuracy to assess the grade of the cancer; thus a biopsy of suspicious lesions on MRI or PET is required to provide a definitive diagnosis and grade of the disease. Systems have been developed to perform biopsies in the MRI suite; however, the cost of the equipment and prolonged use of the MRI is extremely expensive and likely prohibitive given the large number of patients requiring biopsy. Un‐ fortunately, conventional 2D TRUS guidance of the biopsy procedure limits the physician's

As we currently do not have a highly sensitive and specific imaging test for local staging of PCa, there is a growing belief that the optimal method to guide prostate biopsy will involve not just one, but a combination of imaging modalities. 3D TRUS imaging combined with functional or molecular imaging from another imaging modality such as radiopharmaceutical imaging (PET, SPECT), or magnetic resonance imaging (MRS, MRI) may provide the best

Since ultrasound imaging is the clinical standard for image-guided biopsy of the prostate, we have developed a 3D TRUS-based navigation system that provides a reproducible record of the 3D locations of the biopsy targets throughout the procedure and allows fusion with MR

The system we have developed is a mechanical 3D biopsy system that maintains the procedural workflow, minimizing costs and physician retraining. This mechanical system has 4 degreesof-freedom (DOF) and has an adaptable cradle that supports commercially available end-firing TRUS transducers used for prostate biopsy [16]. It also allows real time tracking and recording of the 3D position and orientation of the biopsy needle as the physician manipulates the TRUS transducer. The following describes the components of the system, including hardware,

ability to target locations identified as suspicious on other modalities.

planning for subsequent therapy.

approach for guiding prostate biopsy.

**4.3. 3D TRUS-guided prostate biopsy system**

images with identified lesions for targeting.

**4.2. Multi-modality directed prostate biopsy**

14 Advancements and Breakthroughs in Ultrasound Imaging

Our approach involves the use of a device composed of two mechanisms shown as a schematic in Figure 8. The system is composed of an articulated multi-jointed stabilizer and a transducer tracking mechanism.

**Figure 8.** A schematic diagram of the mechanical tracker, which supports the TRUS transducer and attached cradle. This configuration constrains the TRUS probe motion to three degrees-of-freedom and one degree of translation along the axis of the probe. The system is mounted at the base of a stabilizer while the linkage allows the TRUS trans‐ ducer to be manually manipulated about a remote center of motion (RCM), which is at the center of the ultrasound transducer tip.

The end-firing TRUS transducer with the biopsy needle guide in place is mounted to the mechanical tracking mechanism in a manner where the US probe is free to rotate around its longitudinal axis (Fig. 8). The tracking assembly is attached to a stabilizer, which is mounted on a free-standing cart. Thus, the physician can manipulate the tracking mechanism freely, insert the transducer through the anus, and rotate the transducer in order to acquire a 3D image of the prostate. The tracking linkage contains angle-sensing encoders mounted to each joint in order to transmit to the computer the angles between the arms. This arrangement allows the computer to determine the relative position of the transducer as it is being manipulated. Since the biopsy gun is mounted onto the transducer and its position relative to the transducer is calibrated, the needle location can be calculated.

The mechanical tracking device is a spherical linkage assembly, in which the axis of the joints converge to a common point on the remote center of motion (RCM). The RCM design mini‐ mizes targeting errors within the prostate. As the TRUS transducer is constrained through a stationary point, the physician's movements are replicated at a scaled down rate (minified through the RCM), minimizing changes in morphology and dislocation of the prostate. In addition, the RCM enables a precision equivalent to that of robotic assisted machines. Thus, the system improves the physician's ability to accurately biopsy a point of interest within the patient's prostate.

#### **4.4. Prostate biopsy procedure**

To perform a 3D US-guided prostate biopsy, the end-firing US transducer is mounted onto the tracking assembly such that the tip of the probe is initially set to the RCM point of the tracker linkage. The physician inserts the TRUS transducer into the patient's rectum and aligns the prostate to the center of the 2D TRUS image. A 3D image of the prostate is then acquired by rotating the transducer 180 degrees about its longitudinal axis (Fig. 1b) [91]. A graphical model of the prostate is then generated by a semi-automatic 3D segmentation algorithm [61, 92-94]. After the prostate model has been constructed, the physician can then manipulate the 3D image on the computer screen and select locations to biopsy. After all of the biopsy targets have been selected, the system then displays the 3D needle guidance interface (Fig. 9), which facilitates the systematic targeting of each biopsy location previously selected. Other images or infor‐ mation (*e.g*., MRI or PET/CT images), if available, are registered to the 3D TRUS image and displayed as an overlay on the computer screen (Fig. 10).

**Figure 10.** Registered 3D TRUS and MRI images of the same patient showing delineated suspicious lesions identified in the MR images (right panel). The MR images were then registered with the 3D TRUS images (left panel) and the delineated two regions (red and green) superimposed on the 3D TRUS images. These regions were then targeted with

3D Ultrasound Imaging in Image-Guided Intervention

http://dx.doi.org/10.5772/55230

17

As the physician manually manipulates the TRUS transducer, the 3D location and orientation of the transducer and needle trajectory are tracked in real-time throughout the procedure on the computer screen. Figure 9 illustrates the biopsy interface, which is composed of 4 windows: the live 2D TRUS video stream, the 3D TRUS image, and two 3D model views. The 2D TRUS window displays the real-time 2D TRUS image from the US machine. The 3D TRUS window contains a 2D slice of the 3D static model in real-time to reflect the expected orientation and position of the TRUS probe. This correspondence allows the physician to compare the 3D image with the real-time 2D image to determine if the prostate has moved or deformed to a prohibitive extent. After each biopsy, the biopsy location is recorded in 3D from the tracker orientation, and the system is ready for the next biopsy. After the needle is withdrawn, a 3D image may

Clinical studies are being performed at a number of centers to evaluate the clinical impact of fusion of MRI to intra-biopsy 3D TRUS for 3D US-guided targeted biopsy of suspicious MRI lesions on prostate cancer detection and grading. At the London Health Sciences Centre in London, Canada, prostate MR imaging was performed on 31 patients with clinical suspicion for prostate cancer in advance of their 3D TRUS-guided biopsy. T2, diffusion-weighted and dynamic-contrast enhanced MR sequences were collected in a 3T MRI system with an endorectal RF coil. All suspicious lesions in the MR images were then identified and delineated on the images, which were then registered to the 3D TRUS image obtained during the biopsy procedure (see Figure 10). Using the 3D TRUS-guided biopsy system, prostate biopsy cores were targeted toward each suspicious delineated MRI lesion, which were displayed on the 3D TRUS image. A standard 12-core set of random biopsies was also performed on each patient

be obtained to determine if there is any movement or swelling of the prostate.

**4.5. Clinical evaluation of 3D TRUS/MRI-guided biopsy**

and used as an internal control.

the 3D TRUS-guided biopsy system shown in Figs. 8 and 9.

**Figure 9.** The 3D US-guided prostate biopsy system interface is composed of 4 windows: (top left) the 3D TRUS image dynamically sliced to match the real-time TRUS probe 3D orientation, (bottom left) the live 2D TRUS video stream, (right side) and the 3D location of the biopsy core is displayed within the 3D prostate models. The targeting ring in the bottom right window shows all the possible needle paths that intersect the preplanned target by rotating the TRUS about its long axis. This allows the physician to move the TRUS probe to the target (highlighted by the red dot) in the shortest possible distance. The segmented tumor to be targeted is outlined and rendered in red.

**Figure 10.** Registered 3D TRUS and MRI images of the same patient showing delineated suspicious lesions identified in the MR images (right panel). The MR images were then registered with the 3D TRUS images (left panel) and the delineated two regions (red and green) superimposed on the 3D TRUS images. These regions were then targeted with the 3D TRUS-guided biopsy system shown in Figs. 8 and 9.

As the physician manually manipulates the TRUS transducer, the 3D location and orientation of the transducer and needle trajectory are tracked in real-time throughout the procedure on the computer screen. Figure 9 illustrates the biopsy interface, which is composed of 4 windows: the live 2D TRUS video stream, the 3D TRUS image, and two 3D model views. The 2D TRUS window displays the real-time 2D TRUS image from the US machine. The 3D TRUS window contains a 2D slice of the 3D static model in real-time to reflect the expected orientation and position of the TRUS probe. This correspondence allows the physician to compare the 3D image with the real-time 2D image to determine if the prostate has moved or deformed to a prohibitive extent. After each biopsy, the biopsy location is recorded in 3D from the tracker orientation, and the system is ready for the next biopsy. After the needle is withdrawn, a 3D image may be obtained to determine if there is any movement or swelling of the prostate.

## **4.5. Clinical evaluation of 3D TRUS/MRI-guided biopsy**

the system improves the physician's ability to accurately biopsy a point of interest within the

To perform a 3D US-guided prostate biopsy, the end-firing US transducer is mounted onto the tracking assembly such that the tip of the probe is initially set to the RCM point of the tracker linkage. The physician inserts the TRUS transducer into the patient's rectum and aligns the prostate to the center of the 2D TRUS image. A 3D image of the prostate is then acquired by rotating the transducer 180 degrees about its longitudinal axis (Fig. 1b) [91]. A graphical model of the prostate is then generated by a semi-automatic 3D segmentation algorithm [61, 92-94]. After the prostate model has been constructed, the physician can then manipulate the 3D image on the computer screen and select locations to biopsy. After all of the biopsy targets have been selected, the system then displays the 3D needle guidance interface (Fig. 9), which facilitates the systematic targeting of each biopsy location previously selected. Other images or infor‐ mation (*e.g*., MRI or PET/CT images), if available, are registered to the 3D TRUS image and

**Figure 9.** The 3D US-guided prostate biopsy system interface is composed of 4 windows: (top left) the 3D TRUS image dynamically sliced to match the real-time TRUS probe 3D orientation, (bottom left) the live 2D TRUS video stream, (right side) and the 3D location of the biopsy core is displayed within the 3D prostate models. The targeting ring in the bottom right window shows all the possible needle paths that intersect the preplanned target by rotating the TRUS about its long axis. This allows the physician to move the TRUS probe to the target (highlighted by the red dot) in the

shortest possible distance. The segmented tumor to be targeted is outlined and rendered in red.

patient's prostate.

**4.4. Prostate biopsy procedure**

16 Advancements and Breakthroughs in Ultrasound Imaging

displayed as an overlay on the computer screen (Fig. 10).

Clinical studies are being performed at a number of centers to evaluate the clinical impact of fusion of MRI to intra-biopsy 3D TRUS for 3D US-guided targeted biopsy of suspicious MRI lesions on prostate cancer detection and grading. At the London Health Sciences Centre in London, Canada, prostate MR imaging was performed on 31 patients with clinical suspicion for prostate cancer in advance of their 3D TRUS-guided biopsy. T2, diffusion-weighted and dynamic-contrast enhanced MR sequences were collected in a 3T MRI system with an endorectal RF coil. All suspicious lesions in the MR images were then identified and delineated on the images, which were then registered to the 3D TRUS image obtained during the biopsy procedure (see Figure 10). Using the 3D TRUS-guided biopsy system, prostate biopsy cores were targeted toward each suspicious delineated MRI lesion, which were displayed on the 3D TRUS image. A standard 12-core set of random biopsies was also performed on each patient and used as an internal control.

The results of this study showed that MRI-3D TRUS fusion was successfully performed and the targeted biopsy needle cores had a significantly higher rates of prostate malignancy (30.0%) compared to random, sextant cores (10.0%). In total, prostate cancer was biopsy confirmed in 11 patients; however, only 7 of these patients had abnormal MRI findings (even in retrospective analysis) and were sampled with targeted MRI-3D TRUS fusion. Random sampling detected the remaining four patients. A significantly higher percentage of the targeted biopsy cores (47+/-26%) contained cancer compared to the randomly sampled cores (28+/-26%), and for 3 patients, the MRI-targeted cores detected a higher Gleason cancer grade than the random cores, modifying potential treatment modalities. This study showed that MRI-3D TRUS fusion allows for superior sampling of prostate cancer visible on MRI. This technology may benefit both cancer detection and accurate malignancy grading for appropriate therapeutic management; however, further testing is needed to establish the full utility of this technology.

**Author details**

**References**

graphics. 2000;20(2):559-71.

2008;4(2):180-91.

2009;36(2):411-28.

2004;31(3):539-48.

Aaron Fenster1,2,3, Jeff Bax1,2, Hamid Neshat1,2, Nirmal Kakani3

1 Robarts Research Institute, University of Western Ontario, London, Canada

2 Biomedical Engineering Department, University of Western Ontario, London, Canada

[1] Elliott ST. Volume ultrasound: the next big thing? Br J Radiol. 2008;81(961):8-9.

[2] Downey DB, Fenster A, Williams JC. Clinical utility of three-dimensional US. Radio‐

[3] Boctor EM, Choti MA, Burdette EC, Webster Iii RJ. Three-dimensional ultrasoundguided robotic needle placement: an experimental evaluation. Int J Med Robot.

[4] Hummel J, Figl M, Bax M, Bergmann H, Birkfellner W. 2D/3D registration of endo‐

[5] Carson PL, Fenster A. Anniversary paper: evolution of ultrasound physics and the role of medical physicists and the AAPM and its journal in that evolution. Med Phys.

[6] Wei Z, Wan G, Gardi L, Mills G, Downey D, Fenster A. Robot-assisted 3D-TRUS guided prostate brachytherapy: system integration and validation. Med Phys.

[7] Smith WL, Surry K, Mills G, Downey D, Fenster A. Three-dimensional ultrasoundguided core needle breast biopsy. Ultrasound in Med and Bio. 2001;27(8):1025-34.

[8] Chin JL, Downey DB, Onik G, Fenster A. Three-dimensional prostate ultrasound and

[9] Chin JL, Downey DB, Mulligan M, Fenster A. Three-dimensional transrectal ultra‐ sound guided cryoablation for localized prostate cancer in nonsurgical candidates: a

[10] Smith WL, Fenster A. Optimum Scan Spacing for Three-Dimensional Ultrasound by Speckle Statistics. Ultrasound in Medicine and Biology. 2000;26(4):551-62.

[11] Fenster A, Tong S, Sherebrin S, Downey DB, Rankin RN. Three-dimensional ultra‐

its application to cryosurgery. Tech Urol. 1996;2(4):187-93.

feasibility study and report of early results. J Urol. 1998;159(3):910-4.

sound imaging. SPIE Physics of Medical Imaging. 1995;2432:176-84.

scopic ultrasound to CT volume data. Phys Med Biol. 2008;53(16):4303-16.

3 Department of Medical Imaging, University of Western Ontario, London, Canada

and Cesare Romagnoli3

3D Ultrasound Imaging in Image-Guided Intervention

http://dx.doi.org/10.5772/55230

19
