Methods:
All patients who underwent isolated SAVR with a tissue valve between January 2012 and January 2020 were included in this study (n= 154). Patients were excluded if they had prior cardiac surgery before the SAVR or did not have adequate MDCT images for analysis. (Figure 1) Clinical information was collected from the patient chart and surgical reports. MDCT images were analyzed by two experienced readers. Each reader was randomly assigned half of the scans to interpret. CT readers were blinded to patient data/operative reports.
Surgical Aortic Valve Replacement
All operations were performed using cardiopulmonary bypass with mild systemic hypothermia (30-34 °C). The majority of procedures were performed with a mini-sternotomy or right thoracotomy, though there were ~20% with full median sternotomy. Intraoperative assessment of the aortic annulus size was conducted after resection of the aortic valve cusps and complete debridement of the calcifications. The sizing obturator was then inserted into the annulus and the largest size valve possible was chosen. The replica end of the sizer was then used to ensure correct valve size. No patients underwent aortic root surgery, coronary artery bypass grafting (CABG) or coronary re-implantation (isolated SAVR only). Valve type was left up to the discretion of the operator. The majority of valves implanted were Edwards Magna Ease.
Annulus/Sinus/Root Measurements
To detect the annulus we obtained two measurements during an end-systolic phase. The annulus was defined as the virtual basal plane of attachment for the aortic valve leaflets. Measurements of the sinus of Valsalva (SOV), sinotubular junction (STJ) and coronary heights were obtained during diastolic phases.
Valve Sizing for MDCT measurements and Assessment for Future ViV Feasibility
We assumed that the sizer (which corresponds to the outer diameter of the valve frame) cannot be forced into an annulus smaller than the valve sizer (i.e.- an average diameter of 24.5 would not be able to accommodate a 25mm valve, so it would be sized for a 23mm valve). Based on the MDCT, we used the average annulus diameter to predict surgical valve sizing as follows;
Average diameter of 19.0 to 20.9mm= Valve size 19mm
Average diameter of 21 to 22.9mm= Valve size 21mm
Average diameter of 23 to 24.9= Valve size 23mm
Average diameter of 25 to 26.9= Valve size 25mm
Average diameter of 27 to 28.9= Valve size 27mm
Averaged diameter >29mm= Valve size 29mm.
We then implanted a virtual valve with the same dimensions as a Carpentier-Edwards Magna Ease at the level of the annulus. We used this valve since it is the most common valve used at our institution. Subsequently, we were able to calculate a virtual valve to coronary ostial distance (VTC) and determine if the patient would be at risk for sinus sequestration, sealing of the aortic root with the virtual valve. We considered patients high risk for future ViV if VTC was <4mm and/or the virtual struts contacted the STJ without room for blood flow around. (6) (FIGURE 2, FIGURE 3)
MDCT Measurements versus Intraoperative Measurements
An important part of virtual valve placement is ensuring that we are placing a proper sized surgical valve. If our measurements on MDCT cannot predict actual valve implanted, then virtual valve sizing may be misguiding. Furthermore, the ability to properly predict SAVR size may be important for procedural planning (i.e.- in patients with small annuli, they might require root enlargement to reduce the risk of patient prosthesis mismatch (PPM)). We therefore evaluated how often we predicted the correct SAVR size implanted. We also assessed the correlation strength between annular diameter, area and perimeter compared to actual valve size implanted.
Statistical Analysis
For patient characteristics, continuous variables were reported as mean with standard deviation (SD) and compared between two groups using a 2-sample independent t tests or Mann–Whitney U test (non-uniform data). A Pearson coefficient of determination (R2) for all diameters, area and perimeter acquired by MDCT imaging was used for comparison with the intraoperative valve chosen. R2values were considered very weak if <0.3, weak if 0.3-0.5, moderate if 0.5-0.7, strong if >0.7. (Moore D.S et al) P values ≤0.05 were considered statistically significant. All statistical analyses were performed with the use of SAS software, version 9.2 (SAS institute, Cary, NC).