The past decade has witnessed a significant evolution in the management of small renal masses. As the incidence of renal cancer and incidental findings has increased, so have our options for assessing and treating them. Options currently available for solid enhancing renal tumors include initial biopsy (Bx), Active surveillance (AS), and surgical extirpation (SE) – such as radical nephrectomy (RN) or partial nephrectomy (PN) and probe-based thermal ablation (TA). Nonthermal ablative technologies such as focused high-intensity beam radiotherapy (Cyberknife) and high-voltage bipolar electrical current also known as irreversible electroporation (NanoKnife) have limited applications to support their usage at this time. High-Intensity Focused Ultrasound (HiFU), another form of thermal ablation, has yet to find a place in renal TA.

Among the most commonly utilized modalities are freezing (CRY) and heating with radiofrequency wavelength energy (RFA). This chapter will focus on needle-based thermal ablation as it pertains to percutaneous renal biopsy and RFA, although the principles can be applied to any needle-based therapy. It will include discussions on patient selection, commercially available probes, timing of renal biopsy, technique (step-by-step approach), and results. Newer imaging modalities such as multislice computerized tomography (CT), three-dimensional renderings, flat panel detector fluoroscopic imaging, image fusion techniques, robotic needle placement, and CT/US fusion are being investigated. Results of both short-term and long-term series will conclude this chapter.

Indications

The incidence and detection of small renal masses (SRMs) has been increasing. It is believed that widespread use of abdominal cross-sectional imaging has been instrumental in the discovery of incidental renal neoplasms. Most of these SRMs are classified as renal cell carcinomas (RCCs) [1, 2]. TA management has become more widely accepted, especially for patients with clinical stage T1a (<4 cm) solid renal masses. Although initially only offered for patients who were not surgically fit to undergo more extensive extirpative treatment (RN or PN), long-term data make it attractive as an alternative option for select, compliant patients. Several papers have recently reported mid- to long-term oncologic results after TA suggesting improved oncologic outcomes. Olweny et al. reported comparative 5-year oncologic outcomes for RFA versus PN in patients with clinical T1a RCC [3]. They found comparable cancer-specific survival (97.2 % versus 100 % [p = 0.31]) overall survival (97.2 % versus 100 % [p = 0.31]), and local recurrence-free survival (91.7 % versus 94.6 % [p = 0.96]) between the two groups. Psutka et al. reported on long-term oncologic outcomes for 185 patients with T1 RCC with a median follow-up of 6.43 years [4]. While 13 % were retreated for recurrent disease, overall disease-free survival (DFS) was 88.6 % (92.3 % for T1a and 76.2 for T1b). DFS was impacted primarily by tumor stage on multivariate analysis. Zagoria et al. published nearly identical results in a much smaller series of only 24 patients with a median follow-up of 5.0 years [5]. Retreatment rates were greater for tumors larger than 4 cm thus leading to the suggestion that tumors greater than 4 cm should be approached cautiously. One of the limitations of performing RFA under CT guidance has been the determination of end of treatment. Since heat is responsible for cellular death, and we know that time at hyperthermia (over body temperature of 37 °C) will determine minimal temperature to be achieved, it has been our contention that real-time temperature monitoring is essential to successful treatment. For that reason we have been utilizing peripherally placed, needle-guided, fiber-optic temperature probes at 5 mm from the edge of the renal tumor at the time of the ablation to determine when the “edge” of the tumor reaches >60 °C [6]. With the addition of this “extra step,” we feel that RFA can be more effective and can be used for deep and central tumors as well as during salvage procedures to ensure a measurable quantifiable end point [7, 8].

RFA Principles and Equipment

The majority of cases thus far have been performed with monopolar expandable needle probes (Starburst, Angiodynamics, Latham, NY; LeVeen, Boston Scientific, Natick, MA) or internally perfused single needles (Cool-Tip, Covidien, Boulder, CO). Radio waves (in the vicinity of 400 MHz) are delivered via the needle probe(s) to the target tissue. The electrical current causes ions in the tissues to move, thus leading to frictional heating. It is the heat thus generated that effectively destroys cells by several mechanisms – protein denaturation, DNA/RNA unraveling, vascular congestion, and ischemic damage. Heating >60 °C causes irreversible cell damage, whereas heating >70 °C will lead to cell death and tissue coagulation. Hemostasis, therefore, is easily achievable utilizing RFA. When target temperature exceeds 100 °C, then vaporization occurs and limits further spread of heat by passive conduction due to the insulating effects of carbonized (“charcoaled”) tissue. A more detailed description is beyond the scope of this chapter but is available [9].

Step by Step

One of the most important aspects of uro-interventional oncology is precise positioning of needle probes, sensors, and/or biopsy needles. Device placement may be performed either with ultrasound, CT guidance, or both using fusion modalities. A renal biopsy can comfortably and safely be performed with a minimal amount of analgesia. This can be achieved utilizing local anesthetic alone or in combination with moderate intravenous sedation. Needle access is dependent upon location of the mass within the kidney and the position of surrounding structures. Many interventionists prefer a posterior approach with the patient placed in the lateral decubitus position with the lesion side dependent. This position may have some stabilizing effect on the kidney within the retroperitoneal space and aids in limiting the effect of breathing motion on the kidney and can limit bowel and lung interposition.

When performing a TA, however, it may be preferable to utilize a general endotracheal anesthetic. This will allow for optimal airway control and patient safety when prone and assist with targeting by utilizing controlled breathing which can eliminate organ movement [10].

Once under general anesthesia patients are typically placed in the prone position, on chest rolls, to assist with ventilation and both arms tucked by the side (Fig. 15.1). Occasionally the arms may need to be raised above the head if the body habitus will not pass easily through the aperture of the CT gantry. One must be careful, however, to avoid brachial plexus nerve injuries should shoulder hyperextension be needed. It is helpful to place an indwelling Foley catheter.

Fig. 15.1
figure 1

Patient on standard CT gantry, under general anesthesia lying prone on chest rolls with arms tucked. Note limited working space even in thin patients such as this (Arrow points towards head of patient)

Full size image

The classic approach involves the interventional radiologist using a hand-guided needle placement based upon his/her understanding of the local target anatomy after creating a mental picture of the optimal placement. This requires a great deal of skill and experience. The best analogy is one of a gunnery sergeant trying to hit a faraway target by firing mortars and making fine adjustments to hit the target by honing in on the target. Initially one marks the entrance site on the patient’s skin surface with use of the CT software grid or a skin surface grid with embedded wires. Then, under maximal sterile barrier, an anesthesia needle is inserted at the site into the subcutaneous tissue and repeat CT imaging is performed confirming the position of the selected access site. A skin incision is made and the subcutaneous tissue may be dissected with a hemostat probe. We typically incrementally advance a 16 g blunt cannula and intermittently perform CT imaging to the periphery of the lesion. Intravenous contrast is sometimes administered to better localize the mass.

Once the position of the blunt cannula is confirmed with CT imaging, the blunt stylet is withdrawn from the cannula, and a combination of fine needle aspiration of the lesion with a 22 g Chiba needle and core needle sampling with an 18 gauge spring-loaded biopsy needle is performed. FNA (20–25 gauge needles) or core biopsy (16–18 gauge needles) may be used, but many interventionalists prefer use of a coaxial needle system allowing both FNA and core samples to be obtained at the same setting. This combination may result in a higher diagnostic yield although many physicians have abandoned the FNA. While leaving the blunt cannula in position as a guide, the thermal probe is inserted in tandem alongside the cannula and repeat CT imaging is obtained confirming adequate position of the probe. Ablation parameters and end points of treatment are device specific. Further discussion is beyond the scope of this chapter. See reference for additional information [9] (Fig. 15.2).

Fig. 15.2
figure 2

(a) Image demonstrates 2 RF ablation probes in place with peripheral fiber-optic temperature probes along side. Video monitor displays right kidney with needle probes into the tumor mass and the temperature probe. Close-up shot of the monitor (b)

Full size image

Advancements to Augment CT-Based Imaging and Guidance

Several CT-based imaging and guidance modalities have been developed and are commercially available but not widely utilized in clinical practice to date. These include conventional CT, CT fluoroscopy, and cone beam CT which can then be fused with other modalities such as positron emission tomography (PET) or real-time imaging modalities like electromagnetic navigation or ultrasound (Table 15.1).

Table 15.1 CT-based imaging and guidance modalities for renal ablation
Full size table

All of the techniques focus on visualizing a safe linear trajectory from a selected entry point on the skin surface to the target. From there, either manually or robotically, the needle is placed on the selected entry point on the skin, oriented to the correct angle, and inserted along the trajectory until the target is reached [11, 12].

Few centers have embraced these new technologies. Possible barriers include a high cost of entry as the equipment can be expensive to obtain, the need for additional training, and the lack of long-term data on the efficacy of such procedures. Additionally, operator comfort with present techniques is most likely a factor. However, it has been shown these initial entry costs may be recouped through long-term cost savings of CT-guided procedures which are estimated at saving between $3,625 and $5,155 per procedure [13–15].

Additionally, the large variety of available systems and a lack of understanding on how they compare to one another makes it difficult to choose a particular modality. The following review of many of the available technologies may prove helpful.

Cone Beam CT

Digital fluoroscopy systems can provide high-quality 3D reconstruction of cannulated structures but is unable to image soft tissues well [16]. Cone beam CT utilizes a rotating C arm with a large flat panel detector to capture over 200 X-ray images from varying angles in a process known as rotational angiography. This scan takes only 6 s and allows easy access to the patient. These images are then combined in what is termed cone beam reconstruction to create 3D CT-like volumes (Fig. 15.3). The resulting images can then be scrolled through similar to CT in order to plan appropriate needle pathways for interventional procedures. Three commercially available systems applying this technique are the DynaCT (Artis Zeego, Siemens Medical Solutions, Erlangen, Germany), Innovact (GE Healthcare, Schenectady, New York), and XperCT (Phillips Healthcare, Amsterdam, Netherlands). The images can be coupled with Artis Zeego iGuide technology (Siemens Medical Solutions, Erlangen, Germany) which assists in planning and mapping the trajectory. The tumor is first marked on coronal, axial, and sagittal planes. The skin entry point and trajectory are then chosen and the software displays the desired needle path which is then projected live on the patient with laser crosshairs (Fig. 15.4). Laser navigation systems like these have been shown to greatly improve target point accuracy. For instance, Moser et al. were able to decrease their target point error from a mean 3.5 mm freehand to 2.0 mm using a laser navigation system in spinal injection procedures [17]. Real-time fluoroscopy can then be used to compare the needle path with the planned trajectory. Manual needle placements allow for tactile feedback for the surgeon.

Fig. 15.3
figure 3

(a) Sagittal, coronal, and axial views obtained with the cone beam CT. The angles of view can be altered on the computer to coordinate with “in-line” viewing “down the barrel” of the needle. These images are also utilized for 3D reconstructions (b)

Full size image
Fig. 15.4
figure 4

(a) Demonstration of fingertip control of needle placement using laser crosshairs (Artis Zeego, Siemens). (b) Note the open space in which to work utilizing cone beam Fluoroscopic technique

Full size image

Tovar-Arriaga et al. additionally incorporated the DLR/KUKA Light Weight Robot III with the cone beam CT system. This robotic arm integrates the imaging and guidance technology and orients the needle itself while still utilizing manual insertion [12]. They measured their accuracy in a phantom model utilizing an Artis Zeego imaging system for error visualization. They determined the compilation of errors from the robot calibration, camera, and image construction to result in an average of 1.2 ± 0.4 mm from tip of the needle to the target distances [12].

This technique has been used by several centers in the ablation of small renal tumors with results showing recurrence-free rates >90 % [18, 19]. Additionally, it has successfully been implemented as a biopsy technique with accuracies as high as 95 % [20].

While allowing for fast and accurate soft tissue imaging, the cone beam CT imaging systems are still hindered by respiratory-dependent organ motion. To minimize the error, both the imaging and the procedure must be done at identical points in respiration which is usually held at end expiration. Additionally, while some real-time information of needle placement can be gathered through fluoroscopy, the majority of the imaging is performed intermittently.

Electromagnetic Navigation

Electromagnetic navigation (EMN) has been shown in phantom models to increase accuracy and reduce procedure time when combined with CT fluoroscopy guidance [21]. These systems consist of a field generator producing an alternating electromagnetic field which induces a voltage in small coils which have been placed into the needles. The resulting voltage is measured and used to calculate the current position and the orientation of the coil. Passive fiducial markers placed on the patient’s skin allow for fusion of prior CT imaging with real-time electromagnetic navigation. Holzknecht et al. implemented the Ultraguide (Tirat Hacarmel, Israel) electromagnetic tracking system for 50 image-guided biopsies and found the deviation between tip and target to be 2.2 ± 2.1 mm [22]. However, these results do not necessarily translate to improved outcomes for patients. In comparing CT-guided percutaneous lung biopsy alone to that with an electromagnetic navigation system in 60 patients, Grand et al. using the ig4 EMN system (Veran Medical Inc., St Louis, MO) found that adding electromagnetic navigation gave no statistical improvement in operative time, radiation dose, number of needle repositions, or diagnostic yield [23]. Additionally, the necessary hardware for these systems, including instruments with the appropriate coils, can be expensive.

Ultrasound CT Fusion

Several systems have been developed allowing for real-time imaging and feedback during percutaneous procedures [24]. One such system described by Venkatesan et al. combines electromagnetic device tracking and computed tomography (CT) with ultrasound (US) and FDG positron emission tomography (PET) imaging fusion [25]. This system fuses a prior PET/CT done on average 2 weeks before admission, with a navigation CT scan done with the patient in position and a sequence of tracked intraoperative US images using the PercuNav (Philips, Eindhoven, Netherlands) software. Utilizing the information from PET, a physician can better localize the target. Passive fiducial markers are placed near the entry site and are coregistered to the markers and radiopaque grid in the preprocedural CT. Intraoperatively, an electromagnetic field generator is placed about 30 cm from the sterile field. The electromagnetic tracking space is then registered to the navigational images by pointing the tracked needle to each fiducial at the interrupted point in ventilation and averaging the signal observed. During needle insertion they could then display PET/CT images, the fused real-time US images, and the electromagnetic-tracked needle location and trajectory. On 14 patients in which the system was used by Venkatesan et al., a verification CT scan showed a basic tracking error of 5.85 ± 4.48 mm. In performing this technique for 36 biopsies, 31 (86 %) of them were diagnostic. Additionally, one patient received hepatic RFA without complication of short-term recurrence at 56 days [25]. Krucher et al. used a similar system on 12 patients undergoing kidney tumor RFA ablation with an average tip to target error of 3.4 ± 2.1 mm [26].

Another US-based system was explored by Hung et al. It exclusively used real-time 2D US with a multiplanar Global Positioning System (MyLab, 70XVGm Biosound Esaote, Indianapolis, IN) which can track the magnetic sensors mounted on the ultrasound and treatment probes. First 2D US images are compiled into a 3D volume which is used as a planning image then overlaid as axial, sagittal, and coronal views onto real-times US images. The location of the probe based on GPS data is then overlaid onto the fused real-time and planning US images. Using this system Hung at al. was able to ablate 32 virtual tumors marked by gold fiducial markers in the renal parenchyma in 16 canine kidneys with a target to tip distance averaging 1.8 mm. They found the learning curve for the system to be quite steep with their initial experiences having much larger errors (6.3 mm) [27].

Other systems have been developed that utilize real-time ultrasound in order to track a moving target. Hong et al. describe a CT/US fusion system in which they use image feedback from visual sensors to control a robotic needle holder (termed servoing) [28]. In this case, their robot automated alignment of a needle using a rigidly placed ultrasound fixed to the base of a 7° of freedom manipulator. By putting the ultrasound in the same plane as the needle, they could ensure direct visualization of the location of the needle tip. Then, by extracting the image based on the change in density at the tumor edge, the instrument will move to compensate for the motion of organs secondary to respiration. In a phantom study, they were able to follow an oscillating moving target at 10 Hz feedback rate with an error of 1.7 mm with most of this error arising from soft tissue needle deflection [28].

Robotic Arm Assistant Platforms

Other robotic assistant platforms have been developed that do not track moving objects but instead work to improve the accuracy of needle insertions. One such system described by Koethe et al. uses the MAXIO robotic assistance platform (Perfint Healthcare, Chennai, India) with 5° of freedom that will place the needle into position. The platform is placed next to the patient and registers with the CT table by a mechanical docking mechanism, optical registration, and tilt sensing (Fig. 15.5). Physicians can plan the target and needle pathway on coronal, axial, and sagittal views which is then displayed as a 3D reconstruction. The software then instructs the operator to move the CT table to a particular z-axis location. The robotic guide arm receives the CT images along with the planned needle pathway, moves to the appropriate position, and the physician manually inserts the needle through the needle guide. Koethe et al. tested this system using a custom-made opaque phantom which showed a shorter mean needle tip to target distance using the robotic assistant platform compared with the freehand technique (6.5 ± 2.5 mm vs. 15.8 ± 9.2 mm) [29].

Fig. 15.5
figure 5

(a) Maxio robotic platform (Courtesy of Perfint Healthcare Corp., Redmond, WA, USA) (b) Close-up of the actual needle holder

Full size image

Another system named iSYS – 1 (Medizintechnik GmbH, Kitzbühel Austria) uses joystick controls to maneuver a needle held by a robotic arm into position and to the correct angle. Physicians can then manually insert the needle towards the target.

These can be compared to the PAKY-RCM (Percutaneous Access of the Kidney and Remote Center of Motion) (Johns Hopkins, Baltimore, MD) developed by Solomon et al. and Patriciu et al. which is a robotic system with 11° of freedom which is mounted to a large frame and attached to the CT table. The robot takes advantage of the laser light that is in every CT scanner in order to achieve section selection to achieve patient registration. The mechanical arm is placed so the tip of the needle is in the skin dermotomy site which is planned by spiral CT imaging. The robot itself will then move the needle to the correct trajectory and under CT fluoroscopy advance the needle using a rolling dowel mechanism to the target location. In phantom studies, they showed a mean tip to target error of 1.66 mm. In 21 percutaneous procedures performed on patients, the robot was able to meet the target adequately 17 (81 %) times with the remaining four requiring joystick fine tuning [30, 31]. This allows for minimal radiation exposure to the clinician but eliminates any tactile feedback by the physician on needle insertion.

Camera Feedback

Camera-based augmented reality systems can also be used to provide real-time feedback during procedures. Nicolau et al. developed a system using passive markers placed on the patient’s abdomen and two calibrated cameras to register 3D CT images to the patient. These cameras provide real-time feedback about patient motion, including respiratory motion. This information allows them to complete a respiratory gating technique, in which guiding information is provided regularly at the point in the respiratory cycle at which they performed preoperative CT images. Using this technique they were able to reduce their errors to 2 mm in an abdominal phantom and below 5 mm in patients [32].

Challenges

There are several challenges faced by all guidance systems. One of the biggest is that of patient motion, particularly respiratory-dependent motion which is difficult to control. Many systems work to image and act at the same points in the respiratory cycle and rely on the target in being in the same location with varying results. Additionally, tissue deformation caused by needle insertion can often change target position. Algorithms have been developed to try to account for this deformation and guide a needle appropriately [33]. Lastly, the accumulation of error from each step of the process from imaging, registration of the patient, equipment positions, and to robotic motion error needs to be minimized to ensure properly reaching the target. Even small errors in the initial targeting angle are compounded over the distance traveled by the needle and can result in large errors in the ablation zone (Fig. 15.6).

Fig. 15.6
figure 6

Small errors in the initial angle at the skin entry can lead to large differences in the attempted and actual ablation zones. As the needle goes deeper into tissues and continues to veer off course, the effect will be exaggerated over these larger distances

Full size image