Content uploaded by Deqiang Xiao
Author content
All content in this area was uploaded by Deqiang Xiao on Sep 10, 2020
Content may be subject to copyright.
In vivo comparison of two navigation systems
for abdominal percutaneous needle intervention
Deqiang Xiao,
1,2
Yong Li,
3
Huoling Luo,
1,2
Yanfang Zhang,
3
Xuejun Guo,
4
Huimin Zheng,
1
Qingmao Hu,
1,2
Fucang Jia
1,2
1
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, No. 1068, Xueyuan Avenue, Xili Nanshan, Shenzhen,
China
2
Shenzhen College of Advanced Technology, University of Chinese Academy of Sciences, No. 1068, Xueyuan Avenue, Xili
Nanshan, Shenzhen, China
3
Department of Interventional Radiology, Shenzhen People’s Hospital, No. 1017, Dongmen North Rd., Luohu, Shenzhen, China
4
Department of Radiology, Peking University Shenzhen Hospital, No. 1120, Lianhua Rd, Futian, Shenzhen, China
Abstract
Purpose: To compare the accuracy of a Kinect-Optical
navigation system with an electromagnetic (EM) navi-
gation system for percutaneous liver needle intervention.
Materials and methods: Five beagles with nine artificial
tumors were used for validation. The Veran IG4 EM
navigation system and a custom-made Kinect-Optical
navigation system were used. Needle insertions into each
tumor were conducted with these two guidance methods.
The target positioning error (TPE) and the time cost of
the puncture procedures were evaluated.
Results: A total of 18 needle insertions were performed to
evaluate the navigation accuracy of the two guidance
approaches. The targeting error was 6.78 ±3.22 mm
and 8.72 ±3.5 mm for the Kinect-Optical navigation
system and the EM navigation system, respectively.
There is no statistically significant difference in the TPE
between the Kinect-Optical navigation system and the
EM navigation system (p= 0.229). The processing time
with the Kinect-Optical system (10 min) is similar to that
of the Veran IG4 system (12 min).
Conclusions: The accuracy of the Kinect-Optical naviga-
tion system is comparable to that of the EM navigation
system.
Key words: Liver—Needle intervention—
Optical navigation—Kinect camera—Electromagnetic
navigation
Percutaneous ablation is a common treatment for small
liver tumors whose diameters are less than 3 cm [1]. Be-
cause inserting the ablation needle precisely into a
planned tumor position is very difficult, ultrasound (US)
and computed tomography (CT) are used to guide the
percutaneous procedure [2]. US offers real-time two-di-
mensional (2D) image guidance, but 2D US images are
not easy to interpret. CT fluoroscopy also offers real-
time 2D image guidance, but with ionizing radiation. In
addition, preoperative CT guidance provides three-di-
mensional (3D) image information [2], from which an
interventional radiologist designs a surgical plan and
inserts the needle based on the planned path. Because
there is a lack of the direct correspondence between in-
tra- and preoperative positions, the radiologist must
mentally map the needle on the planned path, often
leading to lower targeting accuracy or multiple insertion
trials [3]. Surgical navigation is widely used for instru-
ment placement in neurosurgery and orthopedics [4,5],
and similar strategies can be used in percutaneous liver
needle insertion [6,7].
Navigation systems allow for the precise tracking on
location and orientation of a needle during insertion. The
most popular system used is electromagnetic (EM) nav-
igation [3,6,8–12] because it can track a needle by
attaching a sensor near the needle tip. Another com-
monly used system is optical navigation [7,13–18]. The
tracking volume of optical navigation is larger, and the
location accuracy is higher relative to EM navigation;
however, the light from the optical tracker cannot be
occluded. In addition to tracking techniques, the system
accuracy also depends on physical-to-image registration
that calculates the mapping from a physical space into an
image coordinate system.
Correspondence to: Yanfang Zhang; email: szjieru0755@126.com;
Fucang Jia; email: fc.jia@siat.ac.cn
ª
Springer Science+Business Media New York 2017
Abdominal
Radiology
Abdom Radiol (2017)
DOI: 10.1007/s00261-017-1083-x
The registration in an intervention navigation system
is typically point-to-point registration, and it requires
markers to be attached to patients. For EM navigation,
skin fiducials [3,9,12] are often used, where the corre-
sponding positions from the patient and the image space
are marked for registration. To improve targeting accu-
racy, internal fiducials [6,8,10,11], which are obtained
by inserting EM sensors into a patient’s body near tu-
mors, are used in combination with skin fiducials.
However, the technique is not appropriate for clinical
routines due to the invasiveness of the procedure. Re-
cently, EM navigation systems integrated with intraop-
erative US imaging have been gradually applied in
clinical practice [19]; internal anatomical markers were
extracted (manually or automatically) from the tracked
US images and were matched with corresponding pre-
operative points for the registration [20,21]. For optical
navigation, skin fiducials [14–18] and internal fiducials
[7] were both studied and utilized for registration. Al-
though the marker-based registration is simple and
effective, the process of marker attachment or localiza-
tion can be inconvenient or traumatic.
In our previous work, a Kinect-Optical navigation
system based on markerless registration was proposed for
percutaneous needle intervention [22]. In this study, we
aim to compare the accuracy of the Kinect-Optical system
and a commercial EM navigation system. Five beagles
containing a total of nine artificial tumors were used for
validation. The needle insertions were performed under
the guidance of the commercial system (Veran IG4, Veran
Medical Technologies, MO, USA) [23] and the Kinect-
Optical navigation system. Finally, the targeting error
was evaluated in postoperative CT images, and the total
puncture procedure time was recorded.
Materials and methods
System design
The workflow of the EM navigation system consists of
planning, registration, tracking, and visualization. Dur-
ing the planning phase, the tumor centroid and an entry
point marked on the abdominal surface are used to
generate a trajectory in the preoperative CT images. Two
groups of positions marked by the EM sensors are
automatically localized in the image space and physical
space for registration. After registration, a needle at-
tached to an EM sensor is tracked and mapped into the
image space through the registration transform. Two CT
slices and the needle tip position are visualized to guide
the insertion procedure, including entry point searching,
insertion angle adjusting, and needle advancing. Real-
time registration was implemented to compensate for
respiratory motion in the Veran IG4 system. The
homogeneous respiratory state with preoperative CT
images can be captured when the time-course curve of
the registration error reaches a minimum.
In the Kinect-Optical system, the workflow includes
four steps that are similar to those in EM navigation.
First, the liver tumor is extracted in the preoperative CT
volumes; the tumor centroid and an entry point on the
abdominal surface are selected for insertion trajectory
planning. Then, two abdominal surfaces are extracted
from the preoperative CT images and the intraoperative
RGB and depth (RGB-D) images. The registration is
performed based on a surface-matching algorithm [22].
After registration, the instrument is positioned by the
optical tracking system and mapped into the preopera-
tive CT space. Finally, a needle is inserted into the
planned tumor centroid according to the visualization
module (Fig. 1). The graphic processing unit (GPU)
acceleration was used to realize the real-time registration
for respiratory motion compensation.
In the Kinect-Optical system, a Polaris Spectra passive
optical tracking system (Northern Digital Inc., Waterloo,
ON, Canada) was applied to locate the instruments. A
second-generation Microsoft Kinect camera (Kinect V2
for Windows, http://www.microsoft.com/en-us/kinectfor
windows) was utilized to acquire a real-time 3D point
cloud of the abdominal surface for registration. A
17-gauge ablation puncture needle (Cool-tip
TM
RF abla-
tion single electrode kit, Covidien, Medtronic, Dublin,
Ireland) attached to a dynamic reference frame (DRF) was
calibrated before tracking [24]. The software runs on a
Lenovo
Ò
Thinkpad W540 laptop computer (2.7 GHz
quad-core processor, 8 GB RAM).
Animal preparation
The animal trials were approved by the institutional review
board for animal care and research. Five healthy beagles
(approximately 10 kg weight) were approved for use in the
study. They were anesthetized, supinely fixed upon the CT
scan bed, endotracheally intubated, and monitored by a
professional veterinarian throughout the experiments.
Each beagle was implanted with two artificial liver tumors.
The artificial tumors were generated by percutaneously
injecting 2 mL of lipiodol into the liver parenchyma using
an 18-gauge sheath needle. All critical structures within the
liver were avoided during the injection. For each beagle,
two artificial tumors were implanted in the left and right
liver lobes. The animals underwent CT scanning to con-
firm the tumor location and size after implantation. The
experiments were immediately conducted after the suc-
cessful tumor implantation for each animal. When the
experiments were completed, the five anesthetized beagles
were euthanized with an intravenous injection of 1 mL/kg
potassium chloride.
Experimental procedure
For each tumor, needle insertion was performed twice
with the Veran IG4 EM navigation system and the Ki-
D. Xiao et al.: In vivo comparison of two navigation systems
nect-Optical navigation system. The experiments were
conducted in a random order of the two systems for each
beagle. The needle was inserted by a medical graduate
who had practiced more than 100 times on a phantom
using the Veran IG4 and Kinect-Optical systems. The
details of the experimental procedure with these two
methods are as follows:
The procedure for Veran IG4 system-guided needle
insertion (Fig. 2A) includes the following steps:
(1) Scout scan CT acquisition. Scout scan CT (Siemens
SOMATOM Definition AS 20-Slice, Erlangen, Ger-
many) was performed to locate the liver tumors in
the beagles.
(2) EM sensor attachment. Six EM sensors were pasted
on the abdominal surface near the liver tumor before
CT scan to produce two groups of corresponding
positions of fiducials for registration.
(3) Preoperative CT acquisition. The beagle’s abdomen
was scanned at the end of the expiratory state
to acquire preoperative CT images (0.5 mm 9
0.5 mm 91 mm), and the expiratory breath-hold
was obtained by ventilation interruption at the end-
expiration phase. The scanned area contained all the
EM sensors.
(4) Trajectory planning. In the preoperative CT images,
the tumor centroid was selected as the target of the
planned trajectory; all critical structures and ribs
were avoided in the planned path. To reduce the
difficulties of needle alignment, the trajectory was set
on the plane of the transverse slice as much as pos-
sible.
(5) System setting. The position of the tracker was
adjusted when the instrument or sensor could not
be located. The EM sensors were tracked and mat-
ched with the corresponding points in preoperative
CT images for physical and image space registra-
tion. The registration error was computed in real
time.
(6) Needle insertion. The tracked needle (20-gauge,
Chiba
TM
biopsy needle attached with an EM sensor
on the tip) was mapped into the CT space after
registration. Under a real-time visualization guide,
the operator searched for the entry point, adjusted
the insertion angle, and advanced the needle to the
Fig. 1. A screenshot of the visualization interface of the
Kinect-Optical system. A 2D CT slice (left) and 3D abdomi-
nal surfaces (right) are synchronously displayed; the green
segment denotes the planned trajectory, and the red seg-
ment denotes the real needle insertion progress during
navigation.
D. Xiao et al.: In vivo comparison of two navigation systems
planned target. The insertion was performed at the
end of expiration when the registration error curve
reached its minimum.
(7) Postoperative CT acquisition. To evaluate the tar-
geting error of the needle insertion, postoperative CT
images were also acquired. The imaging settings for
postoperative CT acquisition remained the same as
that of the preoperative scan.
With the Kinect-Optical system guidance (Fig. 2B),
the experimental procedure consists of scout scan CT
acquisition, preoperative CT acquisition, trajectory
planning, system setting, needle insertion, and postop-
erative CT acquisition. The steps different from the
Veran IG4 navigation system are described below.
(1) Preoperative CT acquisition. The beagle was covered
with a sterile drape before the CT scan; the operative
region remained exposed in a rectangular shape
above the liver. The scanned area contained the en-
tire operative region, and the same preoperative CT
imaging settings used in the EM navigation were
applied.
(2) System setting. The main steps are tracker adjust-
ment and registration. Based on two extracted point
clouds of the pre- and intraoperative abdominal
surfaces, real-time registration was performed using
a 2D-shape-based surface-matching algorithm.
Evaluation
To evaluate the puncture accuracy under the two dif-
ferent guidance methods, the target positioning error
(TPE) was calculated. The TPE was defined as the dis-
tance from the needle tip to the planned tumor centroid
in the postoperative CT images. Furthermore, the com-
ponent of the TPE (path deflection error) in the per-
pendicular direction to the needle and the component of
the TPE (depth deflection error) along the needle direc-
tion were calculated [25]. The puncture time was also
recorded.
Several statistical method analyses were conducted on
the three error measurements. The mean value, the
standard deviation (SD), the median value, the minimum
value, and the maximum value were calculated for the
TPE, the path deflection error, and the depth deflection
error. Paired ttest was performed on the targeting errors
to compare the puncture accuracy between the two
navigation systems.
Results
A total of 18 needle insertions were conducted on five
beagles using two guidance methods. Of the ten im-
planted tumors, two in each beagle, only nine were tar-
geted; one tumor was not included because of failure.
Fig. 2. ANeedle insertion was conducted under the guidance of the Veran IG4 system, and Bthe Kinect-Optical system was
used to guide needle insertion.
D. Xiao et al.: In vivo comparison of two navigation systems
The mean tumor diameter was 4.7 ±2.3 mm
(2.3–8.8 mm) and the average tumor depth was
71.6 ±14.0 mm (52.7–89.6 mm). Table 1shows the
TPE, the path deflection error, and the depth deflection
error for the two guidance approaches. The processing
times for the puncture procedures are listed in Table 2.
The position of the needle tip with respect to the planned
tumor in the postoperative CT images is shown in Fig. 3.
The paired ttest on the TPE was conducted using
IBM SPSS Statistics V22. The TPE with the Kinect-
Optical system was not significantly different from
that with the Veran IG4 (t
8
= 1.304, p= 0.229); the
Table 1. Error statistics (mm) of TPE, path deflection error, and depth deflection error for the EM navigation, and the Kinect-Optical navigation
Mean SD Median Min Max
EM navigation
TPE 8.72 3.50 9.25 3.11 15.68
Path deflection 6.32 3.46 6.60 1.23 11.84
Depth deflection 4.85 3.79 6.10 0.38 10.27
Kinect-Optical navigation
TPE 6.78 3.22 6.34 2.56 13.86
Path deflection 4.70 2.88 4.72 1.17 9.41
Depth deflection 3.84 3.53 2.64 0.04 10.18
Table 2. Time cost (min) for each step of the puncture procedure with the two guidance methods
Marker pasting CT scan Trajectory planning System setting Needle insertion Total time
EM navigation 1 1 3 2 5 12
Kinect-Optical navigation – 1 3 4 2 10
Fig. 3. Needle tip position relative to the targeted tumor in
postoperative CT images after insertion guided by the two
methods. The top row shows the CT transverse slices. The
second row denotes the corresponding 3D visualization
after volume rendering. A1 and A2: planned targeted tu-
mor; B1 and B2: EM navigated insertion with the Veran IG4
system; C1 and C2: the Kinect-Optical system navigated
insertion.
D. Xiao et al.: In vivo comparison of two navigation systems
differences in the path deflection and depth deflection
errors between them were also not statistically significant
(p= 0.276, 0.606). The statistical power for the paired
ttest performed on the TPE was 0.21 [26].
Discussion
The navigation errors were comparable between the
Veran IG4 and the Kinect-Optical systems. The accuracy
of the Veran IG4 system was 8.72 mm in this experiment.
Banovac et al. [6] developed a similar EM navigation
method for abdominal needle insertion, and the system
accuracy in two porcine models was 8.3 ±3.7 mm,
which is similar to our results. The accuracy of the Ki-
nect-Optical system was 6.78 mm, which is a consider-
ably improvement over similar work (19.4 mm)
conducted by Seitel et al. [27], who developed an optical
navigation system based on the first version of the Kinect
camera. The accuracies of these two systems might not be
high enough for guiding a biopsy on small tumors with
diameter of less than 12 mm, but the targeting error can
be reduced by performing the needle insertions twice; the
deflection of the first insertion is used to refine the sec-
ond pass by adjusting the original trajectory [28]. The
statistical power for the paired ttest was smaller than
0.8, which indicates that the sample size is not large en-
ough, but it was limited by the IRB agreement.
Different registration techniques were applied in the
two navigation systems. Marker-based registration used
in the Veran IG4 system can be simple but most likely
leads to low accuracy when the attached fiducials shift
after preoperative imaging, although shifts are not
common during clinical routines. The distance from
fiducials to the target and their distribution also influence
the registration error [17]. To paste fiducials on
abdominal skin would occlude operative areas, and an
optimal entry point for insertion might be lost due to
fiducial occlusion. Moreover, sterilizing the EM sensor is
necessary for clinical application. Within the Kinect-
Optimal system, all the aforementioned issues were
avoided due to the markerless registration applied; thus,
the precision of Kinect-Optical navigation is higher than
that of the EM method. Furthermore, the markerless
registration also simplified the workflow of procedure.
Both navigation systems can compensate for respiratory
motion based on real-time registration. The Veran IG4
system captures real-time intraoperative fiducial posi-
tions for registration. In the Kinect-Optical system, the
abdominal surface acting as the intraoperative data was
extracted in real time, and the feasibility of respiratory
motion tracking via the Kinect camera has been studied
[29,30].
The puncture procedure time was estimated for the
two guidance approaches. Due to the markerless regis-
tration, the marker attachment step was avoided; time
was saved during the surgical preparation for the
Kinect-Optical system. Compared to the Veran IG4
system, a surface-based matching algorithm was used in
the Kinect-Optical system; it takes more time to obtain
point clouds for registration during the system setting.
Because of the 3D visualization scheme applied in the
Kinect-Optical system, the operator could easily align
the virtual needle on the planning path; therefore, the
time for needle insertion is reduced relative to that
needed for the Veran IG4 system. Considering the total
procedure time, the time cost of all steps in Kinect-
Optical system is not significantly different from that of
the Veran system.
Lipiodol was utilized for constructing the artificial
tumors because lipiodol can be visualized in CT images
and because the shape of lipiodol retention is relatively
stable during the puncture procedure [31]. In the relevant
works [6,7], artificial tumors were generated by mixing
liquid agar into CT contrast medium [32], which could
result in the similar lesion imaging to ours. The volume
of lipiodol was set as 2 mL to construct a small tumor
(diameter is less than 10 mm); the margin of the small
tumor is clear, then the tumor centroid can be accurately
located. In trajectory planning and targeting error cal-
culation, the tumor centroids were manually extracted by
an experienced radiologist from pre- and postoperative
CT images.
The respiration compensation applied in this study is
similar to breath holding. The breath-hold technique is
always used under local or general anesthesia. For gen-
eral anesthesia, the patient is unconscious and cannot
follow the clinician’s instructions to hold their breath;
breath-hold can be realized by ventilation interruption,
but this would make the patient uncomfortable. To apply
the breath-hold technique under sedation or general
anesthesia without harm, we acquired the preoperative
CT images at the end of expiration and inserted the
needle during the end of expiration after real-time reg-
istration. Because there is a relatively long pause at the
end of each expiration and respiration is the most
reproducible during this phase [33], the clinician may
watch the registration error curve displayed in real time,
and when the registration error is approaching the min-
imum at the end of expiration, the clinician can insert the
needle at this time to avoid introducing a large error.
Although the performance of this method is not as good
as that of spontaneous breath holding, it could still lar-
gely reduce the respiratory motion; this was proven by
Yaniv et al. [10].
Image-guided percutaneous procedures have become
a common clinical approach for the treatment of hepatic
tumors; different imaging techniques can be applied
during the procedure. CT is always used for path plan-
ning and post-procedural assessment. US is the most
widely utilized technique in percutaneous ablations be-
cause it can monitor the insertion visually in real time
and does not cause ionizing radiation [19]. Magnetic
D. Xiao et al.: In vivo comparison of two navigation systems
resonance imaging (MRI) is also an option when the
tumor is not well visualized in CT or US. In this study,
the Kinect camera exhibited considerable potential as an
intraoperative imaging modality in clinical applications
for percutaneous insertion guidance.
Navigation tools combined with different imaging
modalities (e.g., US, CT, and MRI) provide a large
potential for increasing clinical success in percutaneous
biopsies and ablations. Mauri et al. studied the image
fusion between the US and the pre-acquired second
imaging modality (e.g., CT and MRI) and combined that
technique with the EM navigation to guide thermal
ablation; they found that the treatment success rate can
be improved with their approach for the liver tumors
undetectable with US [19]. In this study, we integrated
CT and a Kinect sensor with the optical navigation,
which not only eliminated the procedures of marker
attachment and localization but also obtained a com-
parable accuracy to a conventional EM system. Other
advanced imaging techniques such as contrast-enhanced
US (CEUS) and positron-emission tomography/com-
puted tomography (PET/CT) would be promising if
combined with conventional navigation tools because
CEUS can target small residual viable tumor tissue
immediately after ablation [34,35], and PET/CT allows
for the targeting of the vital part of a lesion with short
acquisitions [36].
There are several limitations in this study. The animal
sample size and the number of needle insertions are
small; therefore, more animal trials are needed for fur-
ther validation. All trajectories were set in a plane
(transverse slice) in the preoperative CT images, and out-
of-plane trajectories sometimes are preferred in clinical
routine; thus, various trajectory designs for each tumor
should be considered in further validations. Only one
operator was responsible for insertions in this work; the
insertion of needles by more operators with different
experience levels would be more meaningful.
Conclusion
We validated two navigation methods on animal models
for percutaneous needle insertions guided by a com-
mercial EM navigation system (Veran IG4) or a Kinect-
Optical system. The experimental results indicate that
both the accuracy and the processing time are compa-
rable between the two systems. In conclusion, the Kinect-
Optical navigation system can be an alternative for
guiding percutaneous liver needle interventions.
Acknowledgements. We thank Dr. Gregory C Sharp for language
improvement.
Compliance with ethical standards
Funding This study was supported by the following grants: National
Key Research and Development Program (No. 2016YFC0106502),
NSFC-GD Union Foundation (No. U1401254), NSFC-Shenzhen Un-
ion Foundation (No. U1613221), Guangdong Provincial Key Science
and Technology Program (No. 2015B020214005), Guangdong Science
and Technology Program (No. 2013B010404028), Shenzhen Key
Science and Technology Development Program (No. JCYJ20140509
174140681, JCCJY20140415162543027, JCYJ20150630114942306, and
CXZZ20150430161339354).
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical approval All applicable institutional and/or national guidelines
for the care and use of animals were followed.
Informed consent Statement of informed consent was not applicable
since the manuscript does not contain any patient data.
References
1. McGahan JP, Dodd GD III (2001) Radiofrequency ablation of the
liver: current status. Am J Roentgenol 176(1):3–16
2. Goldberg SN, Grassi CJ, Cardella JF, et al. (2009) Image-guided
tumor ablation: standardization of terminology and reporting cri-
teria. J Vasc Interv Radiol 20(7):S377–S390
3. Kru
¨cker J, Xu S, Glossop N, et al. (2007) Electromagnetic tracking
for thermal ablation and biopsy guidance: clinical evaluation of
spatial accuracy. J Vasc Interv Radiol 18(9):1141–1150
4. Peters TM (2006) Image-guidance for surgical procedures. Phys
Med Biol 51(14):R505
5. Yaniv Z, Cleary K (2006) Image-guided procedures: a review.
Technical Report CAIMR TR-2006-3, Georgetown University,
Washington, DC, USA
6. Banovac F, Tang J, Xu S, et al. (2005) Precision targeting of liver
lesions using a novel electromagnetic navigation device in physio-
logic phantom and swine. Med Phys 32(8):2698–2705
7. Maier-Hein L, Tekbas A, Seitel A, et al. (2008) In vivo accuracy
assessment of a needle-based navigation system for CT-guided
radiofrequency ablation of the liver. Med Phys 35(12):5385–5396
8. Wood BJ, Zhang H, Durrani A, et al. (2005) Navigation with
electromagnetic tracking for interventional radiology procedures: a
feasibility study. J Vasc Interv Radiol 16(4):493–505
9. Banovac F, Wilson E, Zhang H, Cleary K (2006) Needle biopsy of
anatomically unfavorable liver lesions with an electromagnetic
navigation assist device in a computed tomography environment.
J Vasc Interv Radiol 17(10):1671–1675
10. Yaniv Z, Cheng P, Wilson E, et al. (2010) Needle-based interven-
tions with the image-guided surgery toolkit (IGSTK): from phan-
toms to clinical trials. IEEE Trans Biomed Eng 57(4):922–933
11. Kru
¨cker J, Xu S, Venkatesan A, et al. (2011) Clinical utility of real-
time fusion guidance for biopsy and ablation. J Vasc Interv Radiol
22(4):515–524
12. Appelbaum L, Solbiati L, Sosna J, et al. (2013) Evaluation of an
electromagnetic image-fusion navigation system for biopsy of small
lesions: assessment of accuracy in an in vivo swine model. Acad
Radiol 20(2):209–217
13. Wacker FK, Vogt S, Khamene A, et al. (2006) An augmented
reality system for MR image–guided needle biopsy: initial results in
a swine model. Radiology 238(2):497–504
14. Nicolau S, Pennec X, Soler L, et al. (2009) An augmented reality
system for liver thermal ablation: design and evaluation on clinical
cases. Med Image Anal 13(3):494–506
15. Oliveira-Santos T, Peterhans M, Hofmann S, Weber S (2011)
Passive single marker tracking for organ motion and deformation
detection in open liver surgery. In: Information processing in
computer-assisted interventions. Springer, Berlin, pp 156–167
16. von Jako CR, Zuk Y, Zur O, Gilboa P (2013) A novel accurate
minioptical tracking system for percutaneous needle placement.
IEEE Trans Biomed Eng 60(8):2222–2225
17. Lin Q, Yang R, Cai K, et al. (2015) Strategy for accurate liver
intervention by an optical tracking system. Biomed Opt Express
6(9):3287–3302
18. Bale R, Schullian P, Schmuth M, et al. (2016) Stereotactic
radiofrequency ablation for metastatic melanoma to the liver.
Cardiovasc Interv. Radiol 39(8):1128–1135
D. Xiao et al.: In vivo comparison of two navigation systems
19. Mauri G, Cova L, De Beni S, et al. (2015) Real-time US-CT/MRI
image fusion for guidance of thermal ablation of liver tumors
undetectable with US: results in 295 cases. Cardiovasc Interv
Radiol 38(1):143–151
20. Mauri G, De Beni S, Forzoni L, D’Onofrio S, Kolev V, Lagana
`M,
Solbiati L (2014) Virtual navigator automatic registration tech-
nology in abdominal application. In: 2014 36th annual interna-
tional conference of the IEEE engineering in medicine and biology
society, 2014. IEEE, pp 5570–5574
21. Mauri G (2015) Expanding role of virtual navigation and fusion
imaging in percutaneous biopsies and ablation. Abdom Imaging
40(8):3238–3239
22. Xiao D, Luo H, Jia F, et al. (2016) A Kinect camera based navi-
gation system for percutaneous abdominal puncture. Phys Med
Biol 61(15):5687–5705
23. Appelbaum L, Sosna J, Nissenbaum Y, Benshtein A, Goldberg SN
(2011) Electromagnetic navigation system for CT-guided biopsy of
small lesions. Am J Roentgenol 196(5):1194–1200
24. Birkfellner W, Watzinger F, Wanschitz F, Ewers R, Bergmann H
(1998) Calibration of tracking systems in a surgical environment.
IEEE Trans Med Imaging 17(5):737–742
25. Wallach D, Toporek G, Weber S, Bale R, Widmann G (2014)
Comparison of freehand-navigated and aiming device-navigated
targeting of liver lesions. Int J Med Robot 10(1):35–43
26. Hintze J (2011) PASS 11. NCSS, LLC. Kaysville. Utah, USA.
www.ncss.com.
27. Seitel A, Bellemann N, Hafezi M, et al. (2016) Towards markerless
navigation for percutaneous needle insertions. Int J Comput Assist
Radiol Surg 11(1):107–117
28. Toporek G, Engstrand J, Nilsson H, et al. (2015) Intra-interven-
tional image fusion and needle placement verification for percuta-
neous CT-guided interventions. Int J Comput Assist Radiol Surg
10(Suppl 1):S37–S38
29. Xia J, Siochi RA (2012) A real-time respiratory motion monitoring
system using KINECT: proof of concept. Med Phys 39(5):2682–
2685
30. Tahavori F, Alnowami M, Wells K Marker-less respiratory motion
modeling using the Microsoft Kinect for Windows. In: SPIE
Medical imaging, 2014. International Society for Optics and Pho-
tonics, pp 90360 K–90310 K
31. Yue J, Sun X, Cai J, et al. (2012) Lipiodol: a potential direct sur-
rogate for cone-beam computed tomography image guidance in
radiotherapy of liver tumor. Int J Radiat Oncol Biol Phys
82(2):834–841
32. Tsuchida M, Yamato Y, Aoki T, et al. (1999) CT-guided agar
marking for localization of nonpalpable peripheral pulmonary le-
sions. Chest 116(1):139–143
33. Sainani NI, Arellano RS, Shyn PB, et al. (2013) The challenging
image-guided abdominal mass biopsy: established and emerging
techniques ‘if you can see it, you can biopsy it’. Abdom Imaging
38(4):672–696
34. Mauri G, Porazzi E, Cova L, et al. (2014) Intraprocedural contrast-
enhanced ultrasound (CEUS) in liver percutaneous radiofrequency
ablation: clinical impact and health technology assessment. Insights
Imaging 5(2):209–216
35. Mauri G, Cova L, Ierace T, et al. (2016) Treatment of metastatic
lymph nodes in the neck from papillary thyroid carcinoma with
percutaneous laser ablation. Cardiovasc Interv Radiol 39(7):1023–
1030
36. Shyn P, Tatli S, Sahni V, et al. (2014) PET/CT-guided percutaneous
liver mass biopsies and ablations: targeting accuracy of a single 20 s
breath-hold PET acquisition. Clin Radiol 69(4):410–415
D. Xiao et al.: In vivo comparison of two navigation systems
- A preview of this full-text is provided by Springer Nature.
- Learn more
Preview content only
Content available from Abdominal Radiology
This content is subject to copyright. Terms and conditions apply.
