Introduction

Three-Dimensional (3D) printing of the heart has become a widely used diagnostic modality for the treatment of congenital heart disease in recent times [1-4]. Since structural malformations in complex congenital heart diseases show heterogeneous features, it is principal to understand patient-specific pathophysiology of the anomalies to establish an optimal treatment strategy [5]. Comprehensive assessment of internal cardiac structures is important especially in Double Outlet Right Ventricle (DORV). Because DORV is manifested as extreme anatomical variability, selection of optimal therapeutic strategy can also be challenging, with potential options ranging from simple intracardiac rerouting to single ventricular Palliation (SVP) [6]. Although several anatomical risk factors precluding anatomical repair recruiting the Left Ventricle (LV) as a systemic ventricle (i.e. non-committed Ventricular Septal Defect [VSD], severe Mitral Valve [MV] straddling, left ventricular hypoplasia, multiple VSDs…) have been suggested [7], determination of surgical strategy for each patient is not always straightforward and is more frequently dependent on personal experiences of individual cardiologists and cardiac surgeons [8,9]. Despite the promising early outcomes after SVP, LVR has remained as a preferable approach, given the long-term consequences of SVP. We hypothesized that the candidacy for LVR in complex DORV can be extended with the aid of Three-Dimensionally Printed Cardiac model (3DM).

Materials and Methods

Patients

From January 2008 to June 2018, 294 consecutive patients were diagnosed as having DORV in our institution. Ninety patients with concomitant atrioventricular septal defect, discordant atrioventricular connection (i.e. congenitally corrected transposition of great arteries, double inlet right or left ventricle, criss-cross heart), or atresia of one atrioventricular valve were excluded. In remaining 204 patients, 102 patients who did not have any LVR precluding complex features, such as remote VSD, both great arteries completely arise from the right ventricle (200% DORV) [10], restrictive VSD, huge or multiple VSDs, LV outflow tract obstruction, ventricular hypoplasia, atrioventricular valve straddling, and tricuspid valve interposition in potential baffle course were excluded, and two patients who died before the definite determination of surgical strategy were also excluded. Consequently, 100 patients with DORV who were allocated to one of the two treatment strategies during immediate postnatal period constituted the study cohort (Figure 1A).

Determination of treatment strategy at initial presentation

Initial treatment plan was established by thorough discussion among the cardiologists, radiologists, and surgeons based on the patient’s echocardiographic findings, catheterization data, magnetic resonance imaging, Computed Tomography (CT) images, and 3DMs since 2015. With these diagnostic modalities, patients were allocated for initial treatment plan either to LVR arm (n = 50) or to SVP arm (n = 50). There were 15 patients with strategic crossover from SVP to LVR in the course of staged palliation, and surgical strategy was finalized as SVP in 35 patients and LVR in 65 patients (Figure 1B).

Operative Procedures

Fifty patients in the SVP arm underwent initial palliation based on the hemodynamic characteristics, principally on the degree of pulmonary blood flow (PBF). Twenty-one patients with excessive pulmonary blood flow were initially managed either with pulmonary artery banding (n = 20) or Norwood-type repair (n = 1). Nineteen patients with insufficient pulmonary blood flow were managed with systemic-pulmonary shunt (n = 16; right modified Blalock-Taussig shunt in 13, left modified BlalockTaussig shunt in 3), stent insertion in the ductus arteriosus (n = 2), or surgical pulmonary valvotomy (n = 1). No initial palliation was performed for 8 patients whose pulmonary blood flow was deemed appropriate. In the remaining 2 patients with remote VSD, there was a strategic crossover towards LVR in the operating theater because intra-cardiac baffling appeared to be technically feasible. As a second stage palliation, Bidirectional Glenn Operation (BDG) was performed in 29 patients and 1 patient is awaiting BDG. One patient with left atrial isomerism and inferior vena cava interruption underwent Kawashima operation as a definitive palliation. Seven patients had strategic crossover towards LVR and 10 patients died prior to BDG. Among the 29 patients with BDG, 14 patients completed entire palliative operation with extracardiac conduit Fontan operation and 5 patients are awaiting Fontan operation. Six patients had strategic crossover towards LVR prior to Fontan operation. There were 2 early mortalities after BDG and 2 patients were lost to follow-up.

Among the 50 patients in the initial LVR arm, 23 patients underwent palliation (pulmonary artery banding in 14, systemicpulmonary shunt in 6 [right modified Blalock-Taussig shunt in 4, left modified Blalock-Taussig shunt in 1, and central shunt in 1], Norwood-type operation in 2, and right ventricle to pulmonary artery conduit insertion in 1). Two patients underwent BDG and 1 patient with inferior vena cava interruption underwent Kawashima operation prior to LVR. Forty-three patients completed LVR and 7 patients are awaiting repair. Among the 15 patients with strategic crossover from the SVP arm, 13 patients completed LVR and 2 patients are awaiting repair. Thus, 56 patients (43 from initial LVR arm, 13 with strategic crossover from initial SVP arm) have completed LVR. One and a half ventricle hemodynamics was established in 7 patients (3 from initial LVR arm, 4 patients with strategic crossover from initial SVP arm after BDG). Among the 56 patients with completed LVR, intraventricular baffling was done toward aortic valve in 45 patients (to the native orthotopic aortic valve in 40, to the posteriorly tranlocated aortic valve upon modified Nikaidoh procedure in 5) and toward pulmonary valve in 11 patients (in combination with arterial switch operation in 10, with modified Yasui operation in 1). Surgical management for the reconstruction of right ventricular outflow tract comprised conventional right ventricular outflow widening with or without pulmonary valvotomy or branch pulmonary artery angioplasty in 23 patients, right ventricle to pulmonary artery conduit insertion in 11 patients, utilization of anteriorly translocated pulmonary artery upon modified Nikaidoh operation in 5 patients. In 17 patients, no surgical intervention was performed for the reconstruction of right ventricular outflow tract.

Three-Dimensional Printing

Entire 3DM manufacturing processes are 1) Acquisition of high quality CT images, 2) 3D reconstruction of blood pool, 3) Segmentation and coloring of reconstructed images, 4) Acquisition of the cavitary data along the surface of reconstructed blood pool, 5) Data processing to STL format, and 6) Actual 3D printing [11]. High quality CT images were obtained by electrocardiogram-gated, breath tracking cardiac CT with contrast media (Somatom Definition Flash, Siemens, Erlangen, Germany). Images of intra-cardiac structures were obtained by thresholding algorithm of contrast enhanced blood pool, and surface of the 3DM was smoothed manually by the attending radiologists. Using this reconstructed intra-cardiac structure, we made two different models: rigid model and soft model. Rigid model is a replica of entire cardiac structure with adjacent great vessels and made for general anatomical comprehension, while soft model is the reconstruction of cavitary surface of the blood pool and made for the understanding of the special relationship between the anatomical structures. This cavitary wall model had 1 mm thickness with soft material for better manipulation and surgical simulation. All models were made in real size of the patients’ heart except for one model from an adult patient which was reduced by 1/4 magnification. Rigid models were made of gypsum-like material (VisiJet PXL Core powder, VisiJet PXL clear binder and color bonds; 3D Systems, Rock Hill, SC, USA) using Projet 460 printer, and soft models were made of rubber-like resin (TangoPlus FullCure resin, Stratasys Ltd., Eden Prairie, MN, USA) using Objet Connex Series printer. CT image reconstruction and 3D printing required about 1 to 2 weeks after CT examination, depending on the size of the models. Entire 3DM manufacturing processes are illustrated in Figure 2.

Statistical Analysis

Categorical variables are presented as numbers with percentages, and continuous variables are presented as mean with standard deviation or median with Interquartile Range (IQR) or entire range according to the distribution of data. Kolmogorov Smirnov method was used for verification of normal distribution. Kaplan-Meier survival estimate with log-rank test were used for analysis of time to adverse events. Risk factors for final destination to the SVP arm (35/100) and predictors of strategic crossover from the SVP arm to LVR arm (15/50) were identified with logistic regression analysis. Variables for univariable analysis included patient’s demographic data, year of registration, LVR precluding features, associated anomalies, and the use of 3DM. Variables with a p-value less than 0.1 on univariable analysis were used for multivariable analysis. A p-value less than 0.05 was considered statistically significant. All statistical analyses were conducted using SPSS Statistics version 25 (IBM, Armonk, NY, USA) or R software version 3.5.1 (www.r-project.org).

Results

Determinants of final SVP and strategic crossover

Among the 100 patients of entire study cohort, 35 patients eventually remained in the SVP arm after repeated evaluation in each step of the Fontan track. Significant determinants for final SVP strategy were LV hypoplasia (OR 5.16, 95% CI 1.83 to 14.55, p = 0.002), MV straddling (OR 11.94, 95% CI 2.39 to 59.76, p = 0.003), and preoperative use of 3DM turned out to lower the probability to remain in the SVP arm (OR 0.23, 95% CI 0.05 to 0.95, p = 0.042) (Table 1). In the 50 patients with initial SVP arm, there were 15 patients with strategic crossover towards LVR in the course of Fontan track. Tricuspid valve interposition within the potential baffle course (OR 4.66, 95% CI 1.07 to 20.37, p = 0.041) and preoperative use of 3DM (OR 8.99, 95% CI 1.90 to 42.49, p = 0.006) were identified as predictors of strategic crossover from the SVP arm towards LVR arm (Table 2).

Mortalities and reinterventions

There were 5 surgical mortalities and 1 late death after the completion of LVR, and 6 surgical mortalities and 6 inter-stage mortalities (4 after the initial palliation, 2 after the BDG) in the course of SVP (Figure 1). Overall survival since first hospital registration was better in patients with final LVR strategy (n = 65) than those with final SVP strategy (n = 35) (88.8% vs. 64.6% at 5 year; log-rank p = 0.002; Figure 3A). When the survival rates of the patients who completed staged operation either with LVR (n = 56) or SVP (n = 14) were compared, however, there was no significant inter-group difference (88.9% in LVR strategy vs. 100% in SVP strategy at 5 year; log rank p = 0.19; Figure 3B). Eight patients underwent reinterventions after LVR (3 LV outflow tract relief, 3 permanent pacemaker insertion, 1 remnant VSD closure, 1 pulmonary valve replacement, and 1 tricuspid valve repair). Two patients underwent reinterventions after SVP (1 LV outflow tract relief, 1 pericardiostomy). Freedom from reoperation or catheter-based reintervention was not significantly different between the two groups (82.4% in LVR strategy vs. 81.7% in SVP strategy at 5 year; log rank p = 0.91; Figure 3C).

Influence of 3DM

Among the patients who completed LVR (n = 56), median cardiopulmonary bypass time and aortic cross-clamping time in the patients with (n = 13) or without (n = 46) the preoperative use of 3DM were not significantly different (cardiopulmonary bypass time: 170 min vs 176 min, p = 0.82, aortic cross-clamping time: 106 min vs 108 min, p = 0.96). Since the introduction of preoperative use of 3DM in patients with DORV in 2015, 3DMs were made in 26 patients, of whom strategic crossover from initial SVP to LVR took place in 8 patients. Detailed information and strategic changes of the 26 patients with preoperative 3DM are summarized in Table 3 and Figure 4.

Comments

In this study, we found potential benefit of 3DM usage in the determination of surgical strategy favoring LVR in patients with complex DORV. Although MV straddling and LV hypoplasia were identified as LVR precluding factors in this study, absolute contraindications (or indications) for LVR is still controversial [7,8,12]. In practice, LVR precluding factors manifest themselves as a spectrum rather than discrete features (i.e. remoteness of the VSD, degree of LV hypoplasia, extent of the subvalvular apparatus encroachment into the course of intracardiac baffling…). The difficulties in delineating a subset not amenable to LVR may be attributed not only to non-standardized and inconsistent assessment of each morphologic feature but to the differences in individual surgeon’s ability to overcome these obstacles. Therefore, comprehensive and standardized preoperative assessment of intra-cardiac structures is crucial to formulate an optimal surgical strategy for each patient [13,14]. In this regard, utilization of 3DM may be beneficial to reach a consensus on optimal surgical strategy for each patient and to predict potential complications related to the LVR [15-18]. Compared to the 3-dimenstionally reconstructed images on two-dimensional displays [19], 3DM provides real 3-dimensional spatial relationship with tactile sensation [20].

In this study, there was no surgical mortality in 26 patients with 3DM. However, extension of candidacy for LVR in patients with multiple risk factors may lead to higher postoperative mortality and morbidity, including the development of LV outflow tract obstruction and atrioventricular block [12,21]. In the previous studies, use of 3DM has been also known to be advantageous for the surgical training and preoperative simulation [1,22-24]. However, 3DM still has limitations. Because it is made from the still images of CT, 3DM cannot elucidate cardiac movement and function. Changes in intracardiac dimensions and function of the cardiac valves according to the cardiac cycle, regional ventricular wall motion, and exact attachment of fine chordae and subvalvar apparatus can be better visualized by echocardiography [25]. Secondly, direct measurement of cardiac structures from 3DM can be erroneous because imaging data from cardiac CT undergo multiple processing before the reconstruction of 3DM. Last limitation is time and cost [26]. Acquisition of sophisticated radiologic images, conversion of imaging data to a 3D format, and 3D printing process are time-consuming procedures. Therefore, it may not be an appropriate diagnostic modality for the patients who require urgent operation.

Conclusion

LVR can be facilitated with comprehensive preoperative anatomical assessment using the 3DM in patients with complex DORV. However, LVR is still challenging in patients with LV hypoplasia and MV straddling.

References

1. Kim MS, Hansgen AR, Wink O, Quaife RA, Carroll JD (2008) Rapid prototyping: a new tool in understanding and treating structural heart disease. Circulation 117: 2388-2394.
2. Matsumoto JS, Morris JM, Foley TA, Williamson EE, Leng S, et al. (2015) Three-dimensional Physical Modeling: Applications and Experience at Mayo Clinic. Radiographics 35: 1989-2006.
3. Kiraly L (2018) Three-dimensional modelling and three-dimensional printing in pediatric and congenital cardiac surgery. Transl Pediatr 7: 129-138.
4. Anwar S, Singh GK, Varughese J, Nguyen H, Billadello JJ, et al. (2017) 3D Printing in Complex Congenital Heart Disease: Across a Spectrum of Age, Pathology, and Imaging Techniques. JACC Cardiovasc Imaging 10: 953-956.
5. Valverde I, Gomez-Ciriza G, Hussain T, Suarez-Mejias C, Velasco-Forte MN, et al. (2017) Three-dimensional printed models for surgical planning of complex congenital heart defects: an international multicentre study. Eur J Cardiothorac Surg 52: 1139-1148.
6. Ruzmetov M, Rodefeld MD, Turrentine MW, Brown JW (2008) Rational approach to surgical management of complex forms of double outlet right ventricle with modified Fontan operation. Congenital Heart Disease 3: 397-403.
7. Serraf A, Nakamura T, Lacour-Gayet F, Piot D, Bruniaux J, et al. (1996) Surgical approaches for double-outlet right ventricle or transposition of the great arteries associated with straddling atrioventricular valves. J Thorac Cardiovasc Surg 111: 527-535.
8. Barbero-Marcial M, Tanamati C, Atik E, Ebaid M (1999) Intraventricular repair of double-outlet right ventricle with noncommitted ventricular septal defect: advantages of multiple patches. J Thorac Cardiovasc Surg 118: 1056-1067.
9. Koshiyama H, Aoki M, Akiyama S, Nakamura Y, Hagino I, et al. (2014) Successful biventricular repair of double-outlet right ventricle with transposition of the great arteries, pulmonary stenosis, and straddling mitral valve. The Annals of Thoracic Surgery 97: 1431-1433.
10. Bharucha T, Hlavacek AM, Spicer DE, Theocharis P, Anderson RH (2017) How should we diagnose and differentiate hearts with double-outlet right ventricle? Cardiol Young 27: 1-15.
11. Yang DH, Park S-H, Lee K, Kim T, Kim JB, et al. (2018) Applications of Three-Dimensional Printing in Cardiovascular Surgery: A Case-Based Review. Cardiovasc Imaging Asia 2: 166-175.
12. Delius RE, Rademecker MA, de Leval MR, Elliott MJ, Stark J (1996) Is a high-risk biventricular repair always preferable to conversion to a single ventricle repair? The Journal of Thoracic and Cardiovascular Surgery 112: 1561-1569.
13. Pushparajah K, Barlow A, Tran VH, Miller OI, Zidere V, et al. (2013) A systematic three-dimensional echocardiographic approach to assist surgical planning in double outlet right ventricle. Echocardiography 30: 234-238.
14. Zhao L, Zhou S, Fan T, Li B, Liang W, et al. (2018) Three-dimensional printing enhances preparation for repair of double outlet right ventricular surgery. J Card Surg 33: 24-27.
15. Olivieri LJ, Krieger A, Loke Y-H, Nath DS, Kim PCW, et al. (2015) Three-dimensional printing of intracardiac defects from three-dimensional echocardiographic images: feasibility and relative accuracy. J Am Soc Echocardiogr 28: 392-397.
16. Giannopoulos AA, Steigner ML, George E, Barile M, Hunsaker AR, et al. (2016) Cardiothoracic Applications of 3-dimensional Printing. J Thorac Imaging 31: 253-272.
17. Cantinotti M, Valverde I, Kutty S (2017) Three-dimensional printed models in congenital heart disease. Int J Cardiovasc Imaging 33: 137-144.
18. Bhatla P, Mosca RS, Tretter JT (2017) Altering management decisions with gained anatomical insight from a 3D printed model of a complex ventricular septal defect. Cardiol Young 27: 377-380.
19. Farooqi KM, Gonzalez-Lengua C, Shenoy R, Sanz J, Nguyen K (2016) Use of a Three Dimensional Printed Cardiac Model to Assess Suitability for Biventricular Repair. World J Pediatr Congenit Heart Surg 7: 414-416.
20. Garekar S, Bharati A, Chokhandre M, Mali S, Trivedi B, et al. (2016) Clinical Application and Multidisciplinary Assessment of Three Dimensional Printing in Double Outlet Right Ventricle With Remote Ventricular Septal Defect. World J Pediatr Congenit Heart Surg 7: 344-350.
21. Bradley TJ, Karamlou T, Kulik A, Mitrovic B, Vigneswaran T, et al. (2007) Determinants of repair type, reintervention, and mortality in 393 children with double-outlet right ventricle. J Thorac Cardiovasc Surg 134: 967-973.
22. Costello JP, Olivieri LJ, Su L, Krieger A, Alfares F, et al. (2015) Incorporating three-dimensional printing into a simulation-based congenital heart disease and critical care training curriculum for resident physicians. Congenital Heart Disease 10: 185-190.
23. Yoo S-J, Spray T, Austin EH, Yun T-J, van Arsdell GS (2017) Hands-on surgical training of congenital heart surgery using 3-dimensional print models. J Thorac Cardiovasc Surg 153: 1530-1540.
24. Martelli N, Serrano C, van den Brink H, Pineau J, Prognon P, et al. (2016) Advantages and disadvantages of 3-dimensional printing in surgery: A systematic review. Surgery 159: 1485-1500.
25. Farooqi KM, Sengupta PP (2015) Echocardiography and three-dimensional printing: sound ideas to touch a heart. J Am Soc Echocardiogr 28: 398-403.
26. Shiraishi I, Yamagishi M, Hamaoka K, Fukuzawa M, Yagihara T (2010) Simulative operation on congenital heart disease using rubber-like urethane stereolithographic biomodels based on 3D datasets of multislice computed tomography. Eur J Cardiothorac Surg 37: 302-306.