Preoperative planning of left-sided valve surgery with 3D computed tomography reconstruction models: sternotomy or a minimally invasive approach?

Abstract

OBJECTIVES

With the emergence of a new concept aimed at individualization of patient care, the focus will shift from whether a minimally invasive procedure is better than conventional treatment, to the question of which patients will benefit most from which technique? The superiority of minimally invasive valve surgery (MIVS) has not yet been proved. We believe that through better patient selection advantages of this technique can become more pronounced. In our current study, we evaluate the feasibility of 3D computed tomography (CT) imaging reconstruction in the preoperative planning of patients referred for MIVS.

METHODS

We retrospectively analysed all consecutive patients who were referred for minimally invasive mitral valve surgery (MIMVS) and minimally invasive aortic valve replacement (MIAVR) to a single surgeon in a tertiary referral centre for MIVS between March 2014 and 2015. Prospective preoperative planning was done for all patients and was based on evaluations by a multidisciplinary heart-team, an echocardiography, conventional CT images and 3D CT reconstruction models.

RESULTS

A total of 39 patients were included in our study; 16 for mitral valve surgery (MVS) and 23 patients for aortic valve replacement (AVR). Eleven patients (69%) within the MVS group underwent MIMVS. Five patients (31%) underwent conventional MVS. Findings leading to exclusion for MIMVS were a tortuous or slender femoro-iliac tract, calcification of the aortic bifurcation, aortic elongation and pericardial calcifications. Furthermore, 2 patients had a change of operative strategy based on preoperative planning. Seventeen (74%) patients in the AVR group underwent MIAVR. Six patients (26%) underwent conventional AVR. Indications for conventional AVR instead of MIAVR were an elongated ascending aorta, ascending aortic calcification and ascending aortic dilatation. One patient (6%) in the MIAVR group was converted to a sternotomy due to excessive intraoperative bleeding. Two mortalities were reported during conventional MVS. There were no mortalities reported in the MIMVS, MIAVR or conventional AVR group.

CONCLUSIONS

Preoperative planning of minimally invasive left-sided valve surgery with 3D CT reconstruction models is a useful and feasible method to determine operative strategy and exclude patients ineligible for a minimally invasive approach, thus potentially preventing complications.

INTRODUCTION

Minimally invasive aortic valve replacement (MIAVR) and minimally invasive mitral valve surgery (MIMVS) have proved to be safe alternatives for conventional surgery in terms of mortality and complications over the past two decades. Several studies have shown shorter hospital stay, reduced blood loss, reduced ventilation time and reduced postoperative pain in favour of minimally invasive valve surgery (MIVS), with comparable complication rates. These studies, however, did show prolonged cardiopulmonary bypass (CPB) and clamping times [1–6] in MIVS, which potentially lead to worse neurological outcomes [7].

Consensus has not yet been reached regarding absolute contraindications for MIMVS due to both differences in surgical and perfusion techniques, as well as different approaches used by surgeons and centres [8]. It has therefore been difficult to test potential superiority of the minimally invasive technique over conventional surgery.

However, with the emergence of a new concept of individualization of patient care, the focus will shift from whether a minimally invasive procedure is better than conventional treatment, to the question of which patients will benefit most from which technique? Therefore, in our tertiary university referral centre for MIVS, a dedicated multidisciplinary heart-team, consisting of cardiothoracic surgeons, interventional cardiologists and echocardiography specialized cardiologists, evaluates patient suitability for MIVS. The standard workup for MIVS to assess vascular eligibility includes a computed tomography (CT) and computed tomography angiography (CTA). In addition to the standard workup, we use 3D reconstruction models to allow the surgeon to have an optimal preoperative surgical view of the operating field in order to foresee any potential difficulties or complications. The 3D reconstruction enables the surgeon to determine both the ideal intercostal space (ICS) for incision in relation to the diaphragm, as well as the ideal cannulation site for CPB.

We believe that through this preoperative planning method for MIVS advantages of this minimally invasive technique can be more pronounced, leading to less conversions and other perioperative complications. Eventually, with this knowledge, superiority of MIVS could potentially be proved in prospective studies in the future.

The aim of our current study is to evaluate the feasibility and usefulness of 3D CT reconstruction in preoperative planning of patients undergoing left-sided valve surgery and to graphically present the different considerations, operative strategies and contraindications for either MIVS or conventional surgery in our tertiary university referral centre.

MATERIALS AND METHODS

Institutional Review Board and Ethics Committee approval was obtained before performing our retrospective research analysis.

Population

We retrospectively analysed all consecutive patients who were referred for MIMVS and MIAVR between March 2014 and 2015 to a single surgeon in a single tertiary university referral centre for MIVS.

Patients who underwent aortic valve replacement (AVR) or MVS in combination with atrial fibrillation ablation (AF-ablation)—surgery, concomitant MVS/AVR, coronary artery bypass grafting (CABG) or planned ascending aorta repair surgery were excluded. Patients who underwent a reoperation were excluded as well. All patients underwent prospective multidisciplinary preoperative planning.

Computed tomography

To assess the eligibility for minimally invasive surgery, patients underwent a preoperative retrospective electrocardiography-gated multidetector computed tomography (Somatom Definition Flash, Siemens, Forchheim, Germany) of the heart and the aortic root up to the aortic arch. Furthermore, a high-pitch CTA of the peripheral arteries was performed after intravenous administration of 75 cc of Iopromide 300 mg I/ml (Ultravist 300, Bayer, Berlin, Germany).

The 2D images were evaluated and measurements, like vascular dimensions, were being realized by a cardiothoracic radiologist.

3D reconstruction

Using a 3D workstation (C-station, PS-medtech, Amsterdam, Netherlands), 2D CT data were used to acquire a 3D reconstruction model. Within 3D medical imaging various acquisition technologies exist (e.g. CT, magnetic resonance imaging, echography). A common factor amongst all these technologies is that the final acquired dataset is often represented as a stack of images that contain samples taken from a discrete spatial location within a scanned object. Each sample within these images represents a measurement made by the acquisition technology. Depending on the modality and its underlying technology, the value of the samples represent different physical quantities. For instance, with CT imaging the samples will represent the densities of the scanned object. The higher the sample value, the higher is the density in the scanned object.

Vesalius3D (PS-medtech, Amsterdam, Netherlands) is a software program used in combination with the C-station that reconstructs this stack of images into a 3D volume by means of a visualization volumetric light transport mode. In this model, light rays are considered to travel through the volume towards an imaging plane. At every point along the light ray, the radiant energy of the ray is modified through the absorption and emittance of energy by the volume (based on the information contained in the samples of the acquired dataset). Every light ray hitting the imaging plane counts as a sample towards a visual reconstruction of the volume, subsequently resulting in a tangible 3D model allowing the surgeon to assess the patient's anatomy with an overall surgical view.

Even in urgent cases, it is possible to produce and assess a 3D reconstruction. It takes about 1 min to load the CT images into the workstation when the station is connected to the institution's radiology department. Production of the 3D reconstruction takes about 30 s.

Preoperative planning

All patients were preoperatively discussed in a dedicated multidisciplinary heart-team consisting of cardiothoracic surgeons, interventional cardiologists and echocardiography specialized cardiologists.

All patients received a coronary angiogram or a coronary computed tomography angiography, in order to detect potential coronary artery disease.

For MVS, the mechanism of mitral regurgitation (MR) was clarified through conventional transthoracic echocardiography (TTE) and 3D transoesophageal echocardiography (TEE). If surgery was indicated and the patient was in a relatively good condition, they would be considered for either conventional MVS or MIMVS. For less thriving patients, transcatheter mitral valve plasty (MVP) (MitraClip) placement could offer an alternative treatment to surgery. In MIMVS workup echocardiography is also used to screen for aortic regurgitation (AR). AR greater than Grade 2 is a contraindication for MIMVS because of antegrade cardioplegia leakage into the left ventricle and coronary malperfusion [9]. With use of CTA imaging, we screen for ascending aortic calcification or dilatation and peripheral artery disease. Transthoracic aortic clamp placement can cause aortic wall damage in patients with severe calcification of the ascending aorta [10] and is therefore considered a contraindication for MIMVS in our centre. Dilatation of more than 40 mm is a contraindication for MIMVS because of aortic wall weakness. Peripheral artery disease and calcification of the peripheral vessels increase the risk of local aortic perforation, retrograde dissection and potential provocation of thrombo-embolic events during retrograde perfusion in MIMVS [11]. A diameter of less than 0.6 cm of the femoral artery is considered to be too narrow for peripheral cannulation because a 19 Fr cannula needs to fit. Furthermore, CT imaging is used to screen for potential chest deformities which lead to difficulties in approach during MIMVS.

For AVR, risk factors, gradation of aortic stenosis by TTE and comorbidities were assessed. If intervention was indicated, the patient's physical condition and comorbidities were evaluated. When patients were in stable condition and eligible, they were to be considered for surgery. Patients in less thriving conditions would be considered for a transcatheter solution (transcatheter aortic valve intervention). In workup for MIAVR, TEE and CTA imaging is used to assess the ascending aorta. Dilatation of the aorta of more than 45 mm is considered to be a contraindication for MIAVR because dilatation results in a weaker aortic wall, which potentially increases intraoperative complication risk [12].

As in MIMVS, severe calcification of the ascending aorta is a contraindication for MIAVR for the same reason [10]. The position of the aortic root is assessed as well. When the root is located too caudally, an upper hemisternotomy up to ICS 3 will not be sufficient to visualize the surgical field properly. Finally, the length of the ascending aorta is reviewed. Patients with aortic elongation have a more fragile aorta and are therefore more at risk for intraoperative complications.

Surgical technique, minimally invasive mitral valve surgery

When performing MIMVS, a right-sided minithoracotomy of 5 cm is made in the fourth or fifth ICS. The right lung is deflated to optimize cardiac exposure. Visualization is accomplished by a 5 mm 30° endoscope that is placed in the anterior axillary line in the same ICS through a trocar. CPB is established by cannulation of the right femoral artery and vein under TEE guidance. An aortic needle is placed in the proximal ascending aorta for antegrade cardioplegia delivery, while cross-clamping is achieved through insertion of a transthoracic aortic clamp through a small incision in the third ICS. The mitral valve is approached through the so-called Waterston's groove, and mitral valve pathology is assessed by use of nerve hooks. Additionally, the left ventricle is filled with water to assess location and mechanism of regurgitation. The valve is either replaced (MVR) or repaired (MVP). We prefer to initially repair the valve. Several techniques exist to repair MR; resection, neochordal placement, cleft closure and annuloplasty. After MVP or MVR is performed, TEE is used to determine repair success and postoperative valvular gradient and presence of potential systolic anterior motion.

Surgical technique, minimally invasive aortic valve replacement

Our current approach for MIAVR is through an upper hemisternotomy until the third ICS. An alternative route for MIAVR, which is used less frequently in the literature, is through a minithoracotomy in the second ICS. The arterial component of CPB is established centrally in the ascending aorta. The venous component of CPB is achieved percutaneously through femoral vein cannulation. Cardioplegia is delivered directly, antegradely in the aortic root or selectively in patients with AR. The aortic valve is approached through a hockey-stick incision in the proximal, ventral portion of the ascending aorta. The aortic valve and its pathology are assessed and the valve is excised. Either a mechanical or biological prosthesis is then fixed in the aortic annulus. The ascending aorta is closed with a Blalock suture.

Statistical analysis

Statistical analysis was performed using the SPSS statistical software package 21.0 for Windows (SPSS, Inc., Chicago, IL, USA).

Normality of continuous variables was tested with the Shapiro–Wilk test. Continuous normally distributed data were presented as mean ± standard deviation. Continuous non-normally distributed data were presented as median and range. Frequencies were presented as absolute numbers and relative percentages.

RESULTS

Mitral valve surgery

Twenty-four patients were referred for evaluation for elective MVS. Three patients required reoperation, 3 patients underwent concomitant AF-ablation surgery, 1 patient received concomitant ascending aortic reconstruction and 1 patient underwent concomitant AVR. Subsequently, 16 patients were included in our study. Fifteen patients (94%) were operated on electively, 1 patient urgently. Eleven patients (69%) underwent MIMVS. Five patients (31%) underwent conventional MVS through a median sternotomy. Of the total number of the MVS group, 14 patients (87%) underwent an MVP. Two patients (13%) underwent MVR for rheumatic mitral valve disease. Baseline characteristics of all patients are mentioned in Table 1.

Table 1:

Baseline characteristics

Patient characteristics	Conventional MVS (n = 5)	MIMVS (n = 11)	Conventional AVR (n = 6)	MIAVR (n = 17)
Age (years) (median, R)	71, (54–79)	62, (36–78)	58, (44–72)	67, (19–83)
Gender (male)	1 (20%)	7 (64%)	5 (83%)	6 (35%)
BMI (mean ± SD)	28.1 ± 4.4	24.7 ± 2.9	26.2 ± 2.6	25.9 ± 3.6
EuroSCORE II (median, R)	1.68, (1.02–2.46)	0.69, (0.50–4.17)	1.40 (0.60–2.33)	1.12, (0.50–2.62)
Elective surgery	5 (100%)	10 (91%)	6 (100%)	16 (94%)
LVF
Normal (LVEF% >50)	5 (100%)	8 (73%)	5 (83%)	16 (94%)
Moderate (LVEF% 30–50)	0	3 (27%)	1 (17%)	1 (6%)
Poor (LVEF <30)	0	0	0	0
COPD	2 (40%)	0	0	2 (12%)
Peripheral artery disease	1 (20%)	0	0	1 (6%)
CVA	0	1 (9%)	0	2 (12%)
Renal failure	0	0	0	0
Hypertension	5 (100%)	11 (100%)	6 (100%)	16 (94%)
Hypercholesterolaemia	5 (100%)	11 (100%)	6 (100%)	16 (94%)
Isolated aortic stenosis	–	–	5 (83%)	16 (94%)
Isolated aortic regurgitation	–	–	0	0
Combined AS/AR	–	–	1 (17%)	1 (6%)
Isolated mitral stenosis	0	0	–	–
Isolated mitral regurgitation	4 (80%)	10 (91%)	–	–
Combined MS/MR	1 (20%)	1 (9%)	–	–
MVP	4 (80%)	10 (91%)	–	–
MVR	1 (20%)	1 (9%)	–	–

MVS: mitral valve surgery; MIMVS: minimally invasive mitral valve surgery; AVR: aortic valve replacement; MIAVR: minimally invasive aortic valve replacement; R: range; BMI: body mass index; EuroSCORE: European System for Cardiac Operative Risk Evaluation; LVF: left ventricular function; LVEF: left ventricular ejection fraction; COPD: chronic obstructive pulmonary disease; CVA: cerebrovascular accident; AS: aortic stenosis; AR: aortic regurgitation; MS: mitral stenosis; MR: mitral regurgitation; MVP: mitral valve plasty; MVR: mitral valve replacement; SD: standard deviation.

Findings, possibly more than 1 per patient, leading to an indication for conventional MVS (Table 2), were: ascending aortic elongation (Fig. 1A), extensive pericardial calcification (Fig. 1B), a slender femoro-iliac tract (Fig. 1C) and calcification of the aortic bifurcation (Fig. 1D). One patient had a history of right-sided thoracic radiotherapy for therapy of breast cancer. Because of expected adhesions, a right anterolateral minithoracotomy was contraindicated.

Table 2:

Contraindications for MIMVS

Contraindication for MIMVS	Number of cases
Calcification of abdominal aorta/aortic bifurcation	3
Iliac tortuosity	2
Slender iliac arteries	1
Elongated ascending aorta	2
Extensive pericardial calcification	1
Previous right-sided thoracic radiotherapy	1

MIMVS: minimally invasive mitral valve surgery.

Figure 1:

Contraindications for MIMVS. (A) Aortic elongation, (B) severe pericardial calcification, (C) slender femoro-iliac tract, (D) abdominal aortic calcification. MIMVS: minimally invasive mitral valve surgery.

In 2 cases, operative strategy of MIMVS was altered preoperatively based on the 3D reconstruction models. Two patients were cannulated for CPB in the left femoral vessels due to right iliac tortuosity (Fig. 2A). Thoracic anatomy and the level of diaphragm were assessed in all patients and if possible, the minithoracotomy was performed in the fourth ICS (Fig. 2B).

Figure 2:

Operative strategy in MIMVS. (A) Change of cannulation site due to right iliac tortuosity, (B) relation between right hemidiaphragm and thoracic anatomy, choice for minithoracotomy in the fourth intercostal space. MIMVS: minimally invasive mitral valve surgery.

In the conventional MVS group, 2 patients died within 30 days of surgery. Both were already considered high-risk patients prior to surgery. The first patient presented with aortic elongation and expected fragility on 3D reconstruction. Intraoperatively, type A aortic dissection occurred, which was repaired successfully. Postoperatively, a mutation in the SMAD3 gene was found, which causes a syndromic form of aortic dilatation and aortic dissections [13]. The patient eventually died of multiorgan failure due to ischaemia secondary to prolonged cross-clamping time.

The second deceased patient presented with extensive pericardial calcification on preoperative 3D reconstruction and was excluded from MIMVS because of expected intraoperative difficulties. This patient eventually died in the postoperative period due to sustained diffuse epicardial bleeding and low cardiac output.

There were no mortalities reported in the MIMVS group. No patients from the MIMVS group were converted to a sternotomy. Conversion, mortality and other complications are depicted in Table 3.

Table 3:

Complications and mortality rate

Complications and mortality rate	Conventional MVS (n = 5)	MIMVS (n = 11)	Conventional AVR (n = 6)	MIAVR (n = 17)
Conversion	–	0	–	1 (6%)
Aortic dissection	1 (20%)	0	0	0
Postoperative CVA	0	1 (9%)	0	1 (6%)
Rethoracotomy	0	2 (18%)	0	5 (29%)
Tamponade	0	1 (9%)	0	3 (18%)
In-hospital mortality	2 (40%)	0	0	0

MVS: mitral valve surgery; MIMVS: minimally invasive mitral valve surgery; AVR: aortic valve replacement; MIAVR: minimally invasive aortic valve replacement; CVA: cerebrovascular accident.

Aortic valve replacement

Twenty-nine patients were referred for evaluation for MIAVR of which the following patients were excluded due to need for concomitant mitral valve surgery (MVS) (1), CABG (1) and ascending aorta repair (2). Two patients were excluded because of reoperation. Subsequently, 23 patients were included in our study.

One patient was operated on urgently, the others electively. Seventeen patients (74%) underwent a minimally invasive procedure after preoperative evaluation; 6 patients (26%) underwent conventional surgery through a full sternotomy. Patient characteristics are presented in Table 1.

Indications for conventional aortic valve surgery (Table 4), possibly more than 1 per patient, were: severe calcification of the ascending aorta (Fig. 3A); dilatation of the ascending aorta (Fig. 3B) and caudal displacement of the aortic valve (Fig. 3C, Video 1).

Table 4:

Contraindications for MIAVR

Contraindication for MIAVR	Number of cases
Ascending aortic calcification	4
Caudal valve displacement	2
Dilatation/aneurysm of the ascending aorta	1

MIAVR: minimally invasive aortic valve replacement.

Figure 3:

Contraindications for MIAVR. (A) Severe ascending aortic calcification, (B) ascending aortic dilatation, the same patient with caudal aortic valve displacement (C) and calcification of the ascending aorta (D). MIAVR: minimally invasive aortic valve replacement.

One patient within the MIAVR group was converted intraoperatively to a full sternotomy. No mortality within 30 days of surgery was reported in the AVR group. Complications and mortality rate are reported in Table 3.

DISCUSSION

Our current descriptive retrospective feasibility study presents our experience in the preoperative planning of left-sided valve surgery with the use of 3D CT reconstruction models. Our aim was to evaluate the feasibility of this 3D imaging technique in preoperative planning of left-sided valve surgery and to graphically present the different considerations for either MIVS or conventional surgery through a full sternotomy in patients who were referred for MIMVS or MIAVR.

The concept of preoperative planning can roughly be applied in two different ways. The first is to adapt the surgical technique to each specific patient. Loor and Roselli [14] presented their experience with this method using 3D CT reconstruction whereby they modified the level of the hemisternotomy to the level of the aortic valve of the specific patient. The advantage of this method is that many patients can be included for minimally invasive surgery. The disadvantage is that the surgeon potentially uses various approaches and incisions. This could potentially prolong his learning curve but also results in an overall less-experienced surgical team.

In contrast, in our institution we select patients for our specific surgical technique. For MIAVR, this means an upper hemisternotomy until ICS 3. For MIMVS, this means a right anterolateral minithoracotomy, femorofemoral cannulation for CPB and direct transthoracic aortic clamping. A potential disadvantage of this preoperative planning method would be the exclusion of several patients for minimally invasive surgery. A great advantage, however, is that the surgeon and his surgical team are highly experienced in the specific technique, resulting in fewer complications and shorter operation times [15].

Patient selection for MIVS starts with clinical examination. Patients with thoracic deformities are excluded from MIVS because of potential intraoperative exposure difficulties. CT imaging is used to assess the ascending aorta. Because of the use of an aortic clamp during MIMVS and MIAVR, patients with severe calcification and elongation of the ascending aorta are excluded as well. Echocardiography is used to assess valvular pathology. Severe complex valve disease can also be an indication for conventional surgery. For MIAVR, knowledge of the valvular location is vital. Caudal displacement makes performance of MIAVR through an upper hemisternotomy more difficult or even impossible (Video 1). Finally, peripheral vessel eligibility for MIMVS is assessed with use of CTA. Severe calcification, artheropathy, dilatation, stenosis or tortuosity of the peripheral vessels make the patient less suitable for transcatheter cannulation and retrograde perfusion.

As mentioned above, conventional CT imaging, with or without contrast, has proved to be an indispensable element in all concepts of preoperative planning of MIVS. CT imaging is ideal for vessel-diameter measurement and calcification detection. 2D imaging, however, relies on the acquisition of a limited number of image planes which cannot be altered during review [16]. Therefore, it does not present an overview of the operating field and an assessment of spatial relations is limited. Especially in minimally invasive surgery, where exposure is reduced compared with conventional surgery, preprocedural understanding of the specific anatomy is crucial. 3D reconstruction does allow the surgeon to have such an overview. Although improved outcome, or even non-inferiority, when compared with conventional imaging techniques has not yet been proved due to a lack of randomized data, 3D imaging does seem to give the surgeon the possibility to form a better procedural strategy. An example of preoperative strategy planning in MIMVS is the choice for ideal incision location or ideal cannulation site. In a period where we are still gaining experience in the field of minimally invasive cardiac surgery and operating times are still longer compared with conventional surgery [3, 5], we believe preoperative strategy planning is essential and that 3D reconstruction is a valuable addition to conventional imaging techniques which can be used to realize this.

Limitations

Our study only focuses on and graphically presents the indications, operative strategies and contraindications for the techniques we specifically use in our tertiary referral centre.

We conducted a retrospective study, although preoperative planning was realized prospectively. We do not compare outcomes in order to prove the superiority of preoperative planning with use of 3D CT reconstruction models when compared with conventional CT imaging in preoperative planning of MIVS.

Finally, our study group is relatively small due to a retrospective study period of ∼1 year and the need for patients with indication for solely isolated valve surgery.

CONCLUSION

In our tertiary university referral centre for MIVS, we used 3D CT reconstruction models in addition to conventional CT imaging for preoperative planning of MIVS. Through preoperative planning we were able to exclude patients for MIVS who were not eligible due to anatomical and/or vascular abnormalities. Additionally, preoperative planning allowed us to determine an ideal operative strategy.

We, therefore, conclude that the addition of 3D CT reconstruction models to standard preoperative workup is a feasible method to exclude patients ineligible for MIVS and to determine operative strategy in MIVS. This method shows promising results and justifies the conduction of future comparative, prospective studies.

SUPPLEMENTARY MATERIAL

Video 1