| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
© 2004 European Association of Cardio-Thoracic Surgery
MR imaging-based port placement planning for totally endoscopic coronary artery bypass grafting
a Division of General Radiology, Department of Radiology, University Hospital Graz, Auenbruggerplatz 9/P, A-8036 Graz, Austria
* Corresponding author. Tel.: +43-316-385-4618; fax: +43-316-385-3231 Received May 8, 2003; received in revised form January 11, 2004; accepted January 29, 2004
An easy applicable method for pre-operative port position planning for totally endoscopic coronary artery bypass (TECAB) grafting based on magnetic resonance (MR) coronary angiography and image post-processing is introduced and analyzed. For this, combined left main (LM) and left anterior descending (LAD) coronary arteries of 21 subjects (14 patients, 7 healthy volunteers with similar habitus) were investigated in MR by means of transversally orientated, three-dimensional (3D), fat-saturated, breath-hold true fast imaging with steady state precession sequences with real-time navigator-based slice following. For the healthy volunteers, the vertical dimension of the total 3D slab was enlarged to enable TECAB planning. Optimal endoscopic port positions were determined via image analysis and geometric methods. 13.8±2.1 cm mean continuously visible length of combined LM and LAD coronary arteries (no statistical difference between patients and healthy volunteers) allowed visualizing typical regions for suturing of the anastomosis in all 21 cases. The mean horizontal distance of the optimal endoscopic port position from the center of the sternum was 7.0±1.3 cm. In conclusion, MR imaging-based port position planning is feasible. Variability in the determined port positions indicates the necessity of adaption of port positioning even for subjects with similar habitus.
Key Words: Coronary vessels, bypass graft; Heart, surgery; Heart, magnetic resonance; Magnetic resonance, image processing
The traditional approach to coronary artery bypass grafting is through a median sternotomy, which provides optimal access to all cardiac structures. To decrease incisions, robotically and endoscopically assisted techniques of minimally invasive coronary artery surgery at the beating heart were developed. The ultimative goal of these techniques is to perform anastomosis entirely endoscopically [1–8]. The feasibility of such surgery was shown in particular for internal thoracic artery to left anterior descending (LAD) artery anastomosis [2–8]: only three 1-cm incisions are necessary to insert the endoscope and the instruments via ports. However, the conversion rate of totally endoscopic coronary artery bypass (TECAB) grafting to standard procedures is high. The port placement is regarded as an issue of central importance [2–6], because it determines the identification of the target vessel, the working space as well as angles of the endoscope and the instruments with respect to each other and to the best-suited place for anastomosis at the coronary artery. Although the placement of incisions is adapted to the habitus of the patient [5,6], it is typically performed without any prior knowledge of the position of the coronary artery with respect to the body surface. It was previously demonstrated by employing various different approaches, that magnetic resonance (MR) imaging is able to visualize even the middle and distal parts of the coronary arteries [9–13]. The generic property of the coronary MR images of showing the surrounding tissue of chest and abdomen as well as the gentle nature of the investigation suggest using an adapted MR coronary angiography protocol also for pre-operative TECAB planning. The purpose of the present study is to present an easy applicable method to plan port placement for TECAB based on a native three-dimensional (3D) true fast imaging with steady state precession (TrueFisp) protocol and image post-processing. The capabilities of the employed sequence to visualize left main (LM) and LAD coronary arteries were investigated to show the imaging-related feasibility of the procedure. The fictive endoscopic port position for subjects with similar habitus was analyzed to study its variability.
2.1. Subjects and MR imaging Fourteen patients (7 men and 7 women, aged 58±15 years) with suspected coronary heart disease as well as 7 healthy volunteers (all men, aged 30±4 years) with no history of heart or lung disease and similar habitus (with body mass index, BMI 23.1±1.5 kg/m2 and circumference of the chest of 100±7 cm) underwent native, breath-hold MR imaging of the coronary arteries in a 1.5T field strength scanner (Magnetom Sonata, Siemens, Erlangen, Germany). Investigations were performed with subjects in supine position, ECG-gated, native and under breath-hold in inspiration. Informed consent was obtained from all subjects before investigation. For the visualization of the coronary arteries and the chest a 3D, fat-saturated, segmented and transversally orientated TrueFisp sequence with real-time navigator-based slice following was employed. Geometrical parameters of the employed standard TrueFisp sequence were 10 slices per slab, rectangular field of view of 350–380x197–289 mm and resolution of 0.7x1.1–1.2x2.0 mm. Repetition time was chosen between 218.5 and 294.8 ms to give breath-hold periods of approximately 25 s. Whereas for patients only the region containing coronary arteries was covered (because MR coronary angiography was part of a more complete investigation of the heart), transversal slices from 1 to 3 cm above the aortic arch to the lower end of the lobes of the lung were acquired for healthy volunteers to enable TECAB planning. 2.2. Image analysis The acquired 3D slabs were combined to a large 3D volume data set. For measurements within multiplanar reconstructions as well as for 3D reconstructions of the volume data set, a 3D post-processing workstation (Aquarius workstation, TeraRecon, San Mateo, CA) was used.Three-dimensional coordinates of points with distances of 2–5 mm along the combined LM and LAD coronary arteries were extracted in multiplanar reconstruction. Linear interpolation was used for the determination of the continuous visible lengths of the coronary arteries and for the visualization of the variability of the left coronary artery courses.
For the analysis of port positions for healthy volunteers, a fictive point of anastomosis
Although the above-described procedure allows specifying port positions with arbitrary incident angles
Analysis can also be performed the other way around: giving a point 2.3. Statistical analysis All data are given as mean±standard deviation (SD). For comparisons of means of continuous variables for male and female patients as well as healthy volunteers analysis of variance was used, where P-values <0.05 were interpreted as statistical significant.
LM and LAD could be visualized in all studies (Fig. 2) and their mean continuous visible lengths were 13.0±2.2 (range 10.3–16.0 cm), 14.1±2.3 (range 11.6–16.8 cm) and 14.4±1.9 cm (range 11.1–17.1 cm) for female patients, male patients and healthy volunteers, respectively. Differences were not statistically significant and the overall mean continuous visible length was 13.8±2.1 cm. Fig. 3 visualizes the variability of the left coronary artery courses for all subjects.
Mean length of the combined LM and LAD coronary arteries at the fictive point of anastomosis was 9.4±0.7 cm. For the case of endoscopic angle of view of  the mean length from the port position to the fictive point of anastomosis was 5.8±0.9 cm. Vertically, in six of the seven healthy volunteers, the port positions were located in the fourth intercostal space, in one the point laid above the fourth rib. Horizontally, the mean distance from the center of the sternum was 7.0±1.3 cm. Results for various assumed port positions in the fifth intercostal space close to the anterior axillary line are summarized in Table 1.
The continuous visible lengths of the combined LM and LAD coronary arteries for patients and healthy volunteers allowed visualizing places of anastomosis, typically located in the middle part of the LAD, in all cases. Usually, there is no problem to depict the total left part of the chest in the images, because the acquisition of transversal slabs minimizes the image extension in anterior–posterior direction. Consequently, TECAB planning is feasible, if the vertical dimensions of the total 3D slab are large enough. The fictive points of anastomosis at healthy volunteers' LAD coronary arteries were chosen at corresponding positions to minimize variability. In practice, the point of anastomosis should be planned according to the localization of the stenosis and the degree to which LAD areas are covered by fat and muscle. Stenoses can be localized, at least via comparison with angiographic results, and MR imaging can detect free regions of the coronary arteries, where additional scans without fat saturation could be advantageous. But even for subjects with similar body mass index, similar habitus and correspondingly chosen points of anastomosis, there exist 1.3 cm SD of horizontal positions around the mean optimal endoscopic port position as well as 8° SD of the endoscopic angle of view starting from standard endoscopic port positions. Fig. 3 suggests that these deviations arise essentially from the variability of their coronary arteries' courses. Assuming similar variability for all port positions (see also Ref. [14]), it is understandable, that TECAB operations without image-based port positioning can lead to sub-optimal working angles of endoscope and instruments, even if patient group-specific port placements are known [5,6]. An important issue, at least for the absolute correctness of the determined port positions and incident angles for the operation, is the question to which extent patient's ‘anatomical’ state during the TECAB operation is reflected in the MR images. Factors, which have to be taken into consideration, are elevation of patient's left shoulder, best-suited cardiac phase as well as right lung ventilation and CO2 insufflation (see Refs. [1–8] for surgical approaches). It poses no principal problem to mimic a specific small elevation of patient's left shoulder during MR imaging. The situation for best-suited reproduction of heart's and lung's status is more difficult, because it directly correlates with parameters connected with image quality. It is well known [15] that image acquisition during heart's mid-diastolic phase is optimal for MR coronary angiography. Moreover, breath holding for periods of 25 s is possible usually only in maximal inspiration and decreasing this time increases the repetition time, which in turn leads to blurred images of the coronaries due to cardiac motion. Whereas the mid-diastolic cardiac phase seems to be adequate for TECAB planning, the inspiration depth corresponding optimally with CO2 insufflation at different pressures [1–8] has to be investigated. The replacement of the applied breath-hold TrueFisp protocol by a similar free breathing protocol, however, should allow acquiring 3D slabs at any inspiration level enabling the same image post-processing described above. In summary, imaging capabilities of the applied MR coronary angiography protocol together with the presented computational algorithm make pre-operative TECAB planning possible. Its potential to decrease the high conversion rate from TECAB procedures to more invasive procedures will be analyzed in a next step.
The manuscript was presented at the European Congress of Radiology 2003, March 7–11, Vienna, Austria. doi:10.1016/j.icvts.2004.01.015
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
Top
Abstract
1. Introduction
was assumed to be at the end of the middle part of the LAD, which is the worst imaging situation that can occur, because suturing of the anastomosis at distal parts of the LAD is unusual. The determination of port positions leading to an incident of the endoscope (or an instrument) at 
and 
via differentiation. The normal vector 
to the heart in 
and 
parallel, respectively, normal to the course of the coronary artery within the tangential plane: 
with angles 
and 
with respect to 
is determined by 
in the chest from which the point 
and the body surface. 
in the present work their determination was restricted to the case 
which corresponds to the optimal endoscopic incident angle for the anastomosis 
