Noninvasive flow evaluations of coronary artery bypass grafting using dynamic cardiac CT.

We aimed to investigate the correlation of graft flow measurements between transit-time flow measurement (TTFM) during coronary artery bypass grafting (CABG) surgery and dynamic cardiac CT after the surgery.Fourteen patients underwent CABG with TTFM and postoperative dynamic cardiac CT; 11 internal thoracic artery (ITA) grafts and 15 saphenous venous grafts (SVGs) were included for analysis. Pearsons correlation analysis was performed for the comparisons of the TTFM and cardiac dynamic CT flow parameters.TTFM was not significantly correlated with the CT flow of the ITA grafts (r = -0.23, P = .49), but it had a very strong correlation with the CT flow of the SVGs (r = 0.83, P < .01).In patients who underwent CABG surgery, dynamic cardiac CT enabled quantitative evaluation of SVG flow, with good correlation with TTFM.

Introduction

Coronary artery bypass grafting (CABG) has been the standard strategy for revascularization of multivessel disease and left main trunk lesion and is the most common cardiac surgery worldwide.[ In the Synergy between PCI with Taxus and Cardiac Surgery (SYNTAX) trial, Thuijs et al reported that CABG provided a significant survival benefit in patients with triple-vessel disease after a 10-year follow-up period.[ Among patients who underwent CABG between 2013 and 2016, the proportion of triple-vessel CABG was approximately 35%.[ The combination of internal thoracic artery (ITA) grafts and saphenous venous grafts (SVGs) has been commonly used for multivessel CABG.

For the clinical outcomes, graft patency, advancement of the disease in native vessels, and comorbid conditions were the reported major predictors of long-term mortality, particularly in patients aged 65 years or older.[ Of these, long-term graft patency remains one of the most important challenges. Intraoperative quality control of the grafts using transit-time flow measurement (TTFM) has been shown to reduce the rate of adverse events and graft failure,[ and it has been recommended in the European Association for Cardio-Thoracic Surgery and the European Society of Cardiology guidelines for myocardial revascularization.[

Cardiac computed tomography (CT) is widely used for the anatomical evaluation of graft patency after CABG; however, it cannot provide the functional information of the graft. A previous study suggested the usefulness of dynamic cardiac CT for the functional assessment of coronary vessels.[ We hypothesized that cardiac dynamic CT enables not only anatomical but also functional evaluation of the graft in the management of patients undergoing CABG. Therefore, the purpose of our study was to investigate the correlation of the flow measurements between TTFM during surgery and dynamic cardiac CT after surgery.

Materials and methods

Patient population

The study population comprised 16 patients who underwent primary elective CABG with TTFM and postoperative dynamic cardiac CT within 2 weeks after surgery between July 2017 and February 2018. The study was approved by the ethics committee of the Kumamoto University Hospital (approve number, #2341) and informed consent was obtained from all patients. Two patients were excluded due to poor image quality. The remaining 14 patients (11 ITA grafts and 15 SVGs) were included for analysis: sequential and composite grafts were excluded. The demographic and clinical data of the patients are shown in Table 1.

Table 1

Patient characteristics.

Number of patients	14
Age (years)	73.0 ± 6.7
Female/male	3/11
Number of grafts (ITA grafts/SVGs)	11/15
Body weight (kg)	56.9 ± 10.8
Body mass index (kg/m2)	21.9 ± 3.5
Average heart rate (beats/minutes)	62.3 ± 4.5

CABG surgery and TTFM

All SVGs were harvested using an open technique. ITA grafts were harvested in a skeletonized fashion. All procedures were performed through median sternotomy, standard cannulation, and extracorporeal circulation for on-pump procedures or with the use of stabilizers for off-pump procedures (12 off-pump and 2 on-pump). The coronary vessel diameter was measured using a surgical metal probe. Distal and Y-anatomizes were made using 8–0 polypropylene. The proximal anatomizes were made using 6–0 polypropylene during partial aortic clamping.

The TTFM values [i.e., mean graft flow volume (ml/minutes) and pulsatility index] of all grafts were recorded intraoperatively in a standardized manner. In particular, at 5 minutes after the patient was weaned from cardiopulmonary bypass and the hemodynamic condition was assessed as being stable, the TTFM values and respective flow curves were obtained using a dedicated TTFM system (Medistim VeriQ; Medistim USA Inc., Plymouth, MN, USA). The TTFM probe was fitted precisely around the middle portion for the ITA grafts and proximally around for the SVGs.

Dynamic cardiac CT data acquisition

CT examinations were performed using a 320-detector row CT scanner (Aquilion One Geneis Edition; Canon Medical Systems, Otawara, Tochigi, Japan). This scanning protocol was modified based on previously reported methods.[ Dynamic cardiac CT was scanned at mid-diastole during 15 to 20 (median, 18) cardiac cycles with prospective electrocardiogram gating. Iodinated contrast medium at 420 mgI/kg (Oiparomin 370; Fuji Pharmaceutical Company, Toyama, Japan) was injected over a fixed duration of 17 seconds, followed by a 30-ml saline flush, which was injected at the same rate as the contrast medium. Image acquisition was determined by the test bolus data, from the arrival time to the ascending aorta to 10 seconds after the peak time. The scan parameters included gantry rotation time of 0.275 seconds, tube voltage of 100 kV, tube current of 100 mA, and field of view of 200 mm. Coronary CT angiography was obtained as a boost scan during the dynamic scanning, with tube voltage of 100 kV and tube current of 400 mA at 8 seconds after the peak time that was obtained by the test bolus data. All CT images were reconstructed using model-based iterative reconstruction (FIRST, Canon Medical Systems). All patients received sublingual nitroglycerin and β-blockers before the CT study.

CT flow measurement

For increasing the precision of the CT flow measurement, the dynamic data sets were interpolated with 3 additional data sets using motion coherence image processing software (PhyZiodynamics; Ziosoft Inc., Tokyo, Japan), which can perform deformable registration to track all the voxels throughout multiple phases and can interpolate images between phases to generate new phases. The spherical volume of interest (VOI) was located on the middle portion of the ITA grafts and on the proximal portion of the SVGs; the VOIs were automatically tracked throughout all phases. The time–density curve of the grafts was generated. The CT flow of the graft vessels was defined as the maximum upslope from the time–density curves.

Visual evaluation for standard coronary CT angiography

From the boost scan data of the dynamic cardiac CT, standard coronary CT angiographic images were generated. Two reviewers consensually performed visual evaluation of the grafts, according to a 4-point grading, as follows: 4 (excellent), the entire graft and anastomosis could be clearly identified; 3 (good), the graft patency could be evaluated but the anastomosis was unclear; 2 (fair), the graft patency was partially obscured; and 1 (poor), the graft could not be identified.

Statistical analysis

Continuous data were expressed as mean ± SD. Pearsons correlation analysis was performed for the pairwise comparisons of the TTFM and cardiac dynamic CT flow parameters. The strength of correlation was described, as follows: perfect correlation (r = 1); very strong correlation (r = 0.8 to 0.99); strong correlation (r = 0.6 to 0.79); moderate correlation (r = 0.4 to 0.59); weak correlation (r = 0.2 to 0.39); and no to very weak correlation (r = 0 to 0.19). Statistical analysis was performed with the computer software (MedCalc version 17.2, MedCalc, Mariakerke, Belgium). A P value of< .05 was considered statistically significant.

Results

For the ITA grafts, the CT flow was not significantly correlated with the mean graft flow volume by TTFM (r = −0.23, P = .49) (Fig. 1a). On the other hand, a very strong correlation was found between the CT flow of the SVGs and the mean graft flow volume by TTFM (r = 0.83, P < .01) (Fig. 1b). The CT flow was not significantly correlated with the pulsatility index for both the ITA grafts (r = 0.11, P = .76) and the SVGs (r = −0.34, P = .21). The mean visual score for the standard coronary CT angiography was 3.7, and there were no patients who had scores of 1 (fair) and 2 (poor). The mean estimated effective radiation dose for dynamic cardiac CT was 4.4 ± 0.9 mSv.

Figure 1

The correlation of graft flow measurements between TTFM and dynamic cardiac CT. (a) The CT flow of the ITA grafts is not significantly correlated with the mean graft flow volume by TTFM (r = −0.23, P = .49). (b) A very strong correlation is found between the CT flow of the SVGs and the mean graft flow volume by TTFM (r = 0.83, P < .01).

Discussion

Multivessel CABG involves 2 or 3 grafts between the aorta and coronary artery circulation; arterial (ITA) and venous grafts (SVG) have been widely used as bypass grafts. However, compared with ITA grafts, SVGs were reported to have high failure rates of 10% to 25% in the first 12 to 18 months,[ leading to recurrent angina and myocardial infarction.[ Therefore, follow-up evaluation of the graft flow should be especially important for SVGs. By analyzing the time–density curve of the SVG, our study revealed that dynamic cardiac CT provided functional information of the SVG and its significant correlation (r = 0.87) with TTFM. Furthermore, our results showed that the standard CT angiography that was derived from the dynamic CT data clearly demonstrated anatomical information on graft patency and the anastomosis for arterial and venous conduits. We believe that combined anatomical and functional assessment of the grafts by dynamic cardiac CT has the potential to enable early intervention for grafts with insufficient flow, although further studies should be conducted to verify this.

In this study, the dynamic cardiac CT and TTFM flow evaluations of the ITA grafts were not significantly correlated, suggesting the difficulty and complicated process of using dynamic cardiac CT to assess ITA flow measurement. One of the reasons may be the long course of the arterial grafts from the ascending aorta to the brachiocephalic artery and ITA. Furthermore, although we used wide-coverage CT (16 cm), the entire ITA could not be scanned by the dynamic cardiac CT. On the other hand, the long-term patency of ITA grafts is good, and we considered that the standard visual evaluation of the ITA grafts may be clinically adequate.

Radiation dose reduction is important during dynamic cardiac CT, because the dynamic scan requires repeated radiation exposures of the patients. Therefore, we used a low-kilovoltage scan plus model-based iterative reconstruction techniques.

From the viewpoint of clinical practice, the time required for both TTFM and CT is short (a few minutes to around 15 minutes), and it is considered to be highly practical. TTFM aims to evaluate intraoperative graft failure. On the other hand, dynamic cardiac CT may be useful for assessing postoperative graft failure. The combined use of these modalities may contribute to quality improvement and cost reduction in CABG patient management. Further verification is needed in this regard.

The usefulness of immunofluorescence imaging in the evaluation of intraoperative graft failure has been reported.[ It should be possible to improve the accuracy of intraoperative graft evaluation by combining quantitative functional evaluation by TTFM and qualitative morphological evaluation by immunofluorescence imaging. Furthermore, dynamic cardiac CT enables both qualitative morphological evaluation and quantitative functional evaluation as postoperative graft evaluation, which may improve the quality of patient management after CABG.

There were some limitations in our study. First, wide-coverage CT machine and a dedicated workstation are necessary for dynamic cardiac CT. Moreover, analysis of the dynamic data can be time-consuming, even for experienced radiological technologists or radiologists. Therefore, automatic analysis should be imperative. Second, our study population was small, and there was no patient with graft failure in this study. The relationship of dynamic cardiac CT and graft failure after CABG should be verified in future studies. Third, since sequential and composite grafts were excluded in this study, the validity of dynamic cardiac CT with these grafts is unknown. Lastly, we have not verified the difference in the dynamic cardiac CT findings between off-pump and on-pump CABG. Since the differences between off-pump and on-pump CABG in TTFM have been reported,[ there may be a difference in dynamic cardiac CT as well.

In conclusion, dynamic cardiac CT enabled quantitative evaluation of SVG flow, with good correlation with TTFM, in patients who underwent CABG surgery. Its possible application in clinical practice may be promising.

Author contributions

Conceptualization: Seitaro Oda, Ken Okamoto, Daisuke Utsunomiya.

Data curation: Daisuke Sakabe, Ken Okamoto.

Formal analysis: Seitaro Oda, Yoshinori Funama.

Investigation: Daisuke Sakabe, Yoshinori Funama, Masafumi Kidoh.

Methodology: Seitaro Oda, Yoshinori Funama.

Supervision: Toshihiro Fukui, Osamu Tominaga, Shingo Kato, Osamu Ikeda, Daisuke Utsunomiya.

Visualization: Masafumi Kidoh.

Writing – original draft: Daisuke Sakabe, Seitaro Oda.

Writing – review & editing: Toshihiro Fukui, Osamu Tominaga, Tsuneo Yamashiro, Osamu Ikeda, Daisuke Utsunomiya.