Minimal invasive aortic valve replacement: associations of radiological assessments with procedure complexity

Surgical valve replacement is standard treatment of severe aortic stenosis (AS). Minimally invasive aortic valve replacement (mini-AVR) is an alternative to conventional open-heart surgery to treat severe AS. Mini-AVR aims to minimize the degree of surgical intrusiveness by operating through a smaller incision (also known as ‘minimal access AVR’). Nevertheless, this procedure can still be considered an invasive surgical procedure. Because the term “Mini-AVR” has achieved common acceptance, we opted to use this as well in the remainder of this manuscript. Compared to conventional aortic valve replacement (CAVR), mini-AVR has shown to reduce transfusion incidences, postoperative pain, ICU stay, hospitalization, renal failure, and wound infection [1‐3]. On the other hand, the limited exposure to the heart during mini-AVR makes myocardial protection and de-airing more challenging. Recent randomized controlled trials have shown that mini-sternotomy did not result in shorter hospital stay, faster recovery, improved survival, or less transfusion of blood compared to CAVR [4, 5]. In addition, mini-AVR has also been associated with increased cardiopulmonary bypass (CPB) time, aortic cross-clamp (AoX) time, overall operating time, and higher costs [6]. Several studies have demonstrated that prolonged CPB time and AoX time is associated with post-operative morbidity and mortality in both low- and high-risk cardiac patients [7‐10]. An important advancement of mini-AVR procedures was the introduction of sutureless valves. Multiple studies have demonstrated that the use of sutureless valves reduces CPB and AoX time and are therefore associated with less post-operative complications [11‐14].

To further improve mini-AVR procedures, preoperative imaging can be used to support surgical planning to decrease procedure complexity. 3D reconstructed Computer Tomography (CT)-images are commonly used for this goal, either for visual evaluation or for measuring distances between the incision and aortic valve [15‐19].

In a recent study, it was found that the access angle, which is the angle between the aortic root and the incision at the annulus, was significantly associated with annulus access difficulty, CPB time, and AoX time in patients treated through a mini-sternotomy [20]. Other radiological characteristics that have been suggested to be associated with procedure complexity are the diameter of the aortic valve annulus, the distance from the ascending aorta to the sternum, and the extent of calcifications in the aortic valve and the ascending aorta [21, 22]. The reported association between access angle and procedural complexity was based on a relatively small patient population, which limited reliable statistical analysis for additional CT-based measures.

The primary purpose of this study is to determine whether CT-image based measurements are associated with mini-AVR surgical times. Additionally, we evaluate whether this hypothesized association of CT-image based parameters with the outcomes differ between sutured and sutureless valve replacement surgeries.

One hundred patients that underwent mini-AVR and had a pre-operative CT scan were included in this study. The median age was 71 [IQR: 63, 75] years and 34 patients were female. Sixteen patients presented New York Health Association (NYHA) functional class III or IV symptoms. The median predicted operative risk by means of logistic EuroSCORE I was 4.3 [IQR: 2.9, 6.3]. Ten patients had mild left ventricular dysfunction, while the rest had normal left ventricular function. Eighty-three patients had severe aortic stenosis and 30 patients had a bicuspid aortic valve. Baseline characteristics, cardiovascular comorbidities and hemodynamic data are listed in Table 1.

Fifty-five patients only underwent a CT and forty-five patients underwent both a CT and CTA. Ten scans were excluded for the access angle and access distance measurements: in these scan acquisitions the manubriosternal joint was not scanned, therefore it could not be used as a landmark for the incision location of our measurements. In total, 90 access distance and angle measurements were performed. For all patients the calcium volume was determined. Mean access angle was 42 (SD: 11) °, mean access distance was 86 (SD: 13) mm and the mean calcium volume was 689 (SD: 509) mm³. Forty-five annulus measurements were performed. The CT measurements are reported in Table 2.

CPB time and AoX time were 88 (SD: 24) and 63 (SD: 21) minutes, respectively. Ninety-nine patients had an isolated aortic valve replacement, with one mini-AVR procedure that was converted to a full sternotomy because the annulus could not be fully exposed to effectively perform the procedure. There was one patient that had a non-isolated aortic valve replacement, in which the patient required additional repair of the ascending aorta. Intra-operative data is summarized in Table 3.

In the univariable regression models, calcium volume, all annulus measures, sutureless valves, and age were significantly associated with CPB and AoX time (p < 0.05). Access angle and access distance were not associated with CPB and AoX time. Increasing levels of calcium volume and annulus measures showed an association with an increase in the CPB and AoX time, whereas advancing age and sutureless valves showed an association with a decrease in CPB and AoX time. After adjustment for confounding variables age and sutureless valves, only calcium volume was significantly associated with CPB (adjusted β-coefficient 0.002, 95%-CI (0.005, 0.019), p-value = 0.002) and AoX time (adjusted β-coefficient 0.010, 95%-CI (0.004, 0.016), p-value = 0.002) (Table 4).

There was no significant interaction found between suture technique and any of the annulus dimensions. Both suture technique and calcium were significant in the CPB and AoX models (p < 0.001), but their interaction coefficients were not (p = 0.24 for the CPB model and p = 0.15 for the AoX model).

This study shows that annulus dimensions and calcium volume are associated with CPB and AoX time, however, after adjusting for age and suture technique, the association between annulus dimensions and procedure times lost statistical significance. With increasing size in annulus dimensions and calcium volumes, CPB and AoX time increases. The use of sutureless valves and increasing age showed an association with a decrease in surgical time. In contrast to previous findings, in our population, access angle and distance were not associated with the procedure complexity and outcome.

In the study of Ellatar et al. [20], access angle was significantly associated with AoX time. We did not find such an association of access angle with neither AoX time nor CPB time. Our population is an extension of that same population, which could suggest that this previous finding was due to the limited experience in the procedure. In another study of Gilmanov et al. [2], access angle and distance were suggested to be associated with outcome parameters, which also contradicts with our findings. In their study, access angle and access distance were defined as the distance from ascending aorta to sternum and the angle between the sternum midline and position of the ascending aorta at the level of the main pulmonary artery. The discrepancy between these findings and the present study could be due to the differences in treatment selection. In our study, only patients that were treated through mini-sternotomy were included in this study, whereas the study of Gilmanov included mini-sternotomy and mini-thoracotomy procedures. Another reason why access angle and distance might not be associated with outcome parameters, is that the surgeons who performed the surgery have increased in skill over the years, and thus have become less dependent on the aortic anatomy. In Martella et al. [24], preoperative CT scans were used to plan the incision for right thoracotomies. Their results showed that if the incision location is perpendicular to the plane of the aortic valve, the surgeon has a better view on the exposed valve. It was shown that for this procedure, anterior and medially positioned aortas are more challenging because the angle towards the valve becomes more difficult, especially for right thoracotomies. Our study indicates that procedure complexity decreases with age. This may be explained by the fact that the elasticity of the aortic annulus and ascending aorta diminishes with advancing age. In contrast to our study, a previous research showed that increased annulus size was associated with a shorter procedural time, which was explained by the fact that the increased size in aortic annulus diameter and the elongation of the aorta gives the surgeon more working space [25]. This relation was not found in our study.

Our findings suggest that with each increment 100 mm³ of calcium surgical procedure time increases approximately by 1 min. So, a severely calcified aortic valve with an approximate volume of 500 mm³ would increase the CPB and AoX time by 5 min, which is a considerable amount of time since the average of our CPB and AoX time was 88 and 63 min respectively.

Surgical times are reduced by using sutureless valves, which has been confirmed by multiple studies [25‐28]. This suggests that the choice of sutureless valves over sutured valves should always to be considered to decrease AoX time, in order to diminish the chance of post-operative morbidity and mortality [8, 10, 29]. In our study, the use of sutureless valves showed an association with a decrease of AoX time approximately by 26 min on average, which is a significant amount on an average of 63 min. Our study also suggests that there is no interaction between the annular dimensions and the choice of suture technique, indicating that the relation of annulus measures and calcium volume on surgical time is similar for both techniques.

Mini-AVR and CAVR are not the only choice of intervention for AS. The patient can also be treated through the less invasive transcatheter aortic valve implantation (TAVI). TAVI is favoured for high-risk and intermediate-risk patients who clinically are frail and old and have an increased risk for surgery. Mini-AVR and CAVR are favoured for low-risk and intermediate-risk patients, who might have endocarditis and might require additional interventions like revascularization of the coronaries. Surgery is considered when the aortic valve annulus is out of range for TAVI, the aortic root morphology is unfavourable for TAVI, and when the morphology of the valve (bicuspid valves, degree of calcification) is unfavourable for TAVI [30]. For the therapeutic choice between TAVI, mini-AVR and CAVR, a dedicated heart team assesses each patient based on previous cardiologic history and baseline characteristics, calculate the risk of surgery, evaluate the feasibility of TAVI or surgery, and local experience. TAVI is associated with increased pacemaker implantation, vascular complications and paravalvular leakage [31]. For mini-AVR and CAVR these complications are less common. However bleeding complications, acute kidney injury and new-onset atrial fibrillation happen more frequent when compared with TAVI [32]. When an institute is able to perform both surgical as percutaneous valve replacement, the heart team can evaluate technical suitability and risk-benefit ratio and decide the best course of action.

Our study suffers from a number of limitations: Only a single rater performed the CT measurements. The measures used to assess procedure complexity (CPB and AoX time) is only a derivative of the complexity and may under- or overestimate the surgical complexity. Although various steps of the measurements have been automated, which should reduce interobserver variation, the robustness of these measures and their association with the outcome parameters could be validated by more raters. The surgeries have been performed by multiple surgeons, which might have influenced the variability in surgical time. The image data was not uniform over the whole dataset. Non-contrast CT, CTA and multi-phase CTA were used based on what was available, because it is not a standard procedure to perform a CT before mini-AVR surgery. The number of bicuspid aortic valves was larger in the sutured group of patients, which might indicate bias to the selection of type of valve.

We have shown that an increase in annulus size and calcium volume is associated with increased aortic cross-clamp and cardiopulmonary bypass time in patients with a mini-AVR. However, after adjusting for age and suture technique, the association between annulus dimensions and procedure times lost statistical significance.

Additionally, the study confirms that the implantation of sutureless valves and advancing age is associated with decreased aortic cross-clamp and cardiopulmonary bypass time. In contrast to previous studies, access angle was not associated with procedure complexity.