|
|
||||||||
J Thorac Cardiovasc Surg 2006;132:264-269
© 2006 The American Association for Thoracic Surgery
Surgery for Congenital Heart Disease |
a Department of Radiology and Diagnostic Imaging, University of Alberta, Edmonton, Alberta, Canada
b Department of Surgery, University of Alberta, Edmonton, Alberta, Canada
c Department of Dentistry, University of Alberta, Edmonton, Alberta, Canada
d Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Alberta, Canada
e COMPRU, Misericordia Community Hospital, Caritas Health Group, Edmonton, Alberta, Canada
Received for publication September 23, 2005; revisions received December 20, 2005; accepted for publication February 3, 2006. * Address for reprints: Elizabeth Ngan, MD, MSc, 2A2.41 Walter C. Mackenzie Health Sciences Centre, 8440-112 St, Edmonton, Alberta, Canada T6G 2B7 (Email: ema2{at}ualberta.ca).
| Abstract |
|---|
|
|
|---|
METHODS: In 6 patients with pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries, anatomic models of the pulmonary vasculature were rapid prototyped from computed tomographic angiographic data. The surgeons used the models for preoperative and intraoperative planning. The models' accuracy and utility were assessed with a postoperative questionnaire completed by the surgeons. An independent cardiac radiologist also assessed each model for accuracy of major aortopulmonary collateral artery origin, course, and caliber relative to conventional angiography.
RESULTS: Of the major aortopulmonary collateral arteries identified during surgery and conventional angiography, 96% and 93%, respectively, were accurately represented by the models. The surgeons found the models to be very useful in visualizing the vascular anatomy.
CONCLUSION: This study presents the novel vascular application of rapid prototyping to pediatric congenital heart disease. Anatomic models are an intuitive means of communicating complex imaging data, such as the pulmonary vascular tree, which can be referenced intraoperatively.
| Introduction |
|---|
|
|
|---|
The goal of corrective surgery is to repair the intracardiac defects, establish right ventricular to pulmonary arterial continuity, and connect the MAPCAs to the pulmonary arterial system.
4-7
This process may require a single operation (single-stage unifocalization and complete intracardiac repair) or multiple operations (sequential unifocalization with subsequent intracardiac repair).
4-7
One critical component of this surgery is the identification and isolation of the MAPCAs, a process made difficult by the patient-specific variations in anatomy. Adequate unifocalization of the MAPCAs is the key to successful surgery.
In more complex cases, the preoperative workup includes conventional angiography, computed tomographic (CT) angiography, or magnetic resonance (MR) angiography to map out the pulmonary vascular supply.
8-12
The precise 3-dimensional (3D) relationship of MAPCAs to the aorta and native pulmonary arteries is significant and should be rendered into a useful, intuitively graspable format that is easily referenced intraoperatively. Rapid prototyping provides a potential solution to this problem.
Rapid prototyping is the fabrication of solid models from computer-generated virtual 3D surface models. The virtual model is first broken down into thin slices or layers. A rapid prototyping machine then builds the solid model layer upon layer, resulting in a physical replica of the virtual model.
13
In medical applications of this technology, the virtual models are typically constructed from CT or MR imaging data sets.
The most extensive medical applications of rapid prototyping technology are in head and neck, oral, and craniomaxillofacial surgery.
14-16
Models of the bony anatomy of the head and neck are used to help guide surgery and reconstruction. Relatively limited numbers of vascular applications of this technology have also been developed. These vascular applications include models of heart valves,
17-20
aortic aneurysms,
21
intracranial aneurysms,
22
carotid arteries,
23
and embryonic hearts.
24,25
| Methods |
|---|
|
|
|---|
Imaging Studies
All patients underwent both conventional angiography and CT angiography of the aorta and pulmonary arterial vasculature. Because of the complexity of this type of congenital anomaly, CT angiography was considered part of the standard preoperative evaluation. No additional CT angiography was performed for the purpose of this study. The CT angiograms were obtained with 0.625-mm or 1.25-mm thickness slices with 0.625-mm reconstruction and a slice overlap of 50% on a 16 slice General Electric Light Speed CT scanner (GE Medical Systems, Milwaukee, Wis). The patients were under general anesthesia, and positive-pressure breath-hold technique was used during image acquisition. Intravenous contrast was administered at a dose of 2 mL/kg over 10 to 15 seconds. The patients were 4 months 6 days to 2 years 4 months old at the time of CT angiography, with 5 of the 6 patients 6 months old or younger. The CT was performed at least 2 weeks before the operation, which allowed time for the model to be built.
Model Construction
The models of the CT data sets were created with Mimics 8.11 software (Materialise, Leuven, Belgium) and a rapid prototyping machine. In Mimics, a virtual 3D model was created of the pulmonary arterial vasculature and aorta by a radiology resident and a pediatric cardiac radiologist. Smoothing of the model was performed with Magics 8.01 software (Materialise) or haptic modeling software and hardware (FreeForm Modeling System, version 7.0 and PHANTOM Desktop Haptic Device; SensAble Technologies, Woburn, Mass). This virtual model was converted into a solid acrylic or plastic anatomic model with a rapid prototyping machine, either the Stratasys Prodigy Plus (Stratasys Inc, Eden Prairie, Minn) or the InVision si2 3-D printer (3D Systems, Valencia, Calif). The image manipulation and rapid prototyping were carried out in the Medical Modeling Research Laboratory, COMPRU, Caritas Health Group, Edmonton, Alberta, Canada. The models were sterilized and used within the operative field at the time of surgery.
To study the accuracy of this construction technique, several test objects containing both flat and cylindric surfaces were imaged and then modeled in the same manner as the patient data. The objects and the models were then measured with digital calipers (Absolute Digimatic calipers; Mitutoyo, MTI Canada Ltd, Mississauga, Ontario, Canada).
Model Evaluation
The assessment of the models was 2-fold. First, after each operation, the surgeons completed a short questionnaire to rate the model's overall usefulness and accuracy. The surgeons rated utility and accuracy on 3-point scales (very useful, somewhat useful, or not useful, and very accurate, minor inaccuracies, or major inaccuracies, respectively). MAPCAs identified at the time of surgery were also quantified. Each MAPCA on the models was also rated on a 3-point scale (accurate, minor inaccuracy, or inaccurate).
Second, an independent pediatric cardiac radiologist compared the models with the angiograms. The cardiac radiologist compared the number of MAPCAs on the model with the number demonstrated on the corresponding angiogram. The accuracy of the origin, course, and caliber of each MAPCA on the model was rated on a 3-point scale (accurate, minor inaccuracy, or major inaccuracy). Minor inaccuracy was defined as acceptable accuracy with a minimal and clinically irrelevant discrepancy between the model and angiogram. Major inaccuracy was defined as unacceptable accuracy, potentially misleading to the surgeon and not reflecting the angiographic findings. The models were therefore compared to the absolute standard, surgery, and to the imaging standard, conventional angiography.
The funding organizations, Caritas Health Group and the Government of Canada through Western Economic Diversification Canada, had no role in the study design; in data collection, analysis, and interpretation; in writing this report; or in the decision to publish the results.
| Results |
|---|
|
|
|---|
|
|
|
|
|
A total of 31 MAPCAs were identified on the models or on the angiograms. Figure 5 summarizes the findings of the independent pediatric cardiac radiologist who compared each of the models with the angiograms, rating the origin, course, and caliber of each model MAPCA for accuracy. Of the MAPCA origins, 93% were rated as having acceptable accuracy, and 97% of the MAPCAs received an acceptable accuracy rating for course and caliber. The MAPCA in case 4 not identifiable on CT was rated as having unacceptable accuracy in all categories. The MAPCA in case 3 that was not identified as a MAPCA but rather included on the model as a branch of the right pulmonary artery received an inaccurate origin rating. Only 2 MAPCAs were rated as having minor inaccuracies. In both cases, the inaccuracy was in the caliber of a stenosis. Although it was not part of the questionnaire completed by the cardiac radiologist, he also noted several cases in which the models were superior to the angiograms. The most obvious example of this is the MAPCA that was not identified on the angiogram but was accurately included on the model, matching the operative findings (case 2).
|
| Discussion |
|---|
|
|
|---|
Technical Considerations
Models of soft tissue structures are more difficult to construct than models of bony structures because of the much smaller variation in Hounsfield units, which is significant. Obviously, contrast administration greatly helps in modeling vascular structures. In this particular application, careful attention to anatomy is required to separate the pulmonary arteries from the pulmonary veins. As a result, virtual model construction cannot be automated and should be performed by a clinical specialist. One technical limitation of the models is their fragility, particularly at stenoses. As a result, the models must be handled with care. The techniques presented in this article represent a starting point in optimizing the model construction of congenital heart disease.
Clinical Implications
Of MAPCAs identified during surgery and conventional angiography, 96% and 93%, respectively, were accurately represented by the models. These values are in line with previous accuracy reports of CT angiography and MR angiography of this condition.
11,12
The accuracy of the models is limited by the data set on which they are based, CT angiography. Two MAPCAs found either during surgery or by conventional angiography were not included on the models. In both cases, when the CT data were retrospectively assessed, it was either impossible or difficult to identify the missed MAPCAs. Contributing factors to these MAPCAs being missed were degraded images from table motion artifact and probable occlusion of a small vessel in the interval between the conventional angiogram and the CT angiogram. However, 1 MAPCA identified on the CT study and so included on the model was missed at angiography. This demonstrates the complementary nature of CT angiography with rapid prototyping and conventional angiography in identifying as many MAPCAs as possible before surgery. Because of the need for maximum accuracy of imaging, we believe that CT angiography with rapid prototyping provides useful supplemental information in addition to conventional angiography. In this study, all MAPCAs were identified either on the CT or the angiogram before surgery.
However, simple identification of vessels is not adequate. This study extends the utility of CT angiography by presenting the data in a more useful format, a solid anatomic model showing the course, caliber, and origin of each MAPCA. Solid models communicate information in both visual and tactile formats; however, it is difficult to quantitatively evaluate the conveyance of information, particularly tactile information. The surgeons' overall impression of the ability of the anatomic models to communicate the necessary spatial information was quantified as usefulness in this study. The surgeons rated all the models as very useful for preoperative and intraoperative planning, allowing rapid isolation of MAPCAs during the surgery. One surgeon's impression was that the models decreased operating room time, although this study did not measure operating room time as an outcome variable.
The surgeons had available to them a virtual model on a workstation preoperatively and the solid model both preoperatively and intraoperatively. We view the models as an addition to rather than a replacement for a conventional workstation display. On the basis of this limited experience, the 2-dimensional workstation screens did not convey the information as intuitively as the solid models.
Study Limitations
The number of patients is one of the limiting factors in this study. In a larger study, variables such as operating room time and patient outcome could be examined. Because of the small number of patients, accounting for anatomic variations and comorbidities that affect prognosis would be difficult. Because pulmonary atresia with VSD and MAPCAs is a relatively rare condition, time is required to assess a larger number of patients. The creation of virtual models was relatively labor intensive and required expert knowledge of the pulmonary vasculature. Automated virtual model construction was not possible with the current software used (Mimics).
| Conclusion |
|---|
|
|
|---|
| Footnotes |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. Sodian, D. Schmauss, C. Schmitz, A. Bigdeli, S. Haeberle, M. Schmoeckel, M. Markert, T. Lueth, F. Freudenthal, B. Reichart, et al. 3-dimensional printing of models to create custom-made devices for coil embolization of an anastomotic leak after aortic arch replacement. Ann. Thorac. Surg., September 1, 2009; 88(3): 974 - 978. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Riesenkampff, U. Rietdorf, I. Wolf, B. Schnackenburg, P. Ewert, M. Huebler, V. Alexi-Meskishvili, R. H. Anderson, N. Engel, H.-P. Meinzer, et al. The practical clinical value of three-dimensional models of complex congenitally malformed hearts J. Thorac. Cardiovasc. Surg., September 1, 2009; 138(3): 571 - 580. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Sodian, D. Schmauss, M. Markert, S. Weber, K. Nikolaou, S. Haeberle, F. Vogt, C. Vicol, T. Lueth, B. Reichart, et al. Three-Dimensional Printing Creates Models for Surgical Planning of Aortic Valve Replacement After Previous Coronary Bypass Grafting Ann. Thorac. Surg., June 1, 2008; 85(6): 2105 - 2108. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. S. Kim, A. R. Hansgen, O. Wink, R. A. Quaife, and J. D. Carroll Rapid Prototyping: A New Tool in Understanding and Treating Structural Heart Disease Circulation, May 6, 2008; 117(18): 2388 - 2394. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Jacobs, R. Grunert, F. W. Mohr, and V. Falk 3D-Imaging of cardiac structures using 3D heart models for planning in heart surgery: a preliminary study Interactive CardioVascular and Thoracic Surgery, February 1, 2008; 7(1): 6 - 9. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Sodian, S. Weber, M. Markert, D. Rassoulian, I. Kaczmarek, T. C. Lueth, B. Reichart, and S. Daebritz Stereolithographic Models for Surgical Planning in Congenital Heart Surgery Ann. Thorac. Surg., May 1, 2007; 83(5): 1854 - 1857. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |