The protocol was approved by the Institutional Review Board of Siriraj Hospital (Si 351/2019). Between 1992 and 2019, 170 patients with d-TGA underwent ASO in the Faculty of Medicine Siriraj Hospital, a tertiary care center in Bangkok, Thailand. Of those who underwent ASO (170 cases), seventeen patients follow up at other hospitals. Fifteen patients were peri-operative dead. Ninety-five patients did not meet institutional inclusion criteria as the majority of them are still under 16 years of age.

Our institutional inclusion criteria are 1) patients in early adolescence, 2) experience any symptomatic chest pain or changes in clinical status, 3) innate (preoperative) abnormal coronary patterns and intend to participate in competitive sport. Therefore, only forty-three patients were included to this study. Therefore, data were reviewed from forty-three patients who had undergone ASO and received CCTA based on the criteria for imaging after ASO.

Preoperative coronary artery patterns were identified and addressed by both preoperative echocardiography and operative note (direct inspection by surgeon).

All clinical data including ECG, echocardiography, exercise stress test (EST), CCTA, cardiovascular risk factors such as hypertension and obesity, medications and reoperation during the follow-up period were collected from the Pediatric Cardiology Division database.

Individual operative notes were reviewed to obtain initial coronary artery patterns visualized during surgery. The Leiden classification was used to describe coronary artery patterns [24]. A usual coronary artery pattern was defined as left anterior descending artery (LAD) and circumflex coronary artery (LCx) arising from sinus 1 and right coronary artery (RCA) from sinus 2 (1LCx-2R) (Fig. 1). Clinical data from the Cardiology clinic database and telephone interviews regarding myocardial ischemia-related symptoms such as chest pain during exertion and relieved by rest, recurrent episodes of typical and atypical angina chest pain without improvement by medical treatment of musculoskeletal and gastrointestinal tract problems, and dyspnea on exertion were obtained and recorded.

All CCTA was performed by using a contrast-enhanced, dual-source, 256 slices CT system (Somatom Definition, Siemens Healthcare, Germany) according to established guidelines [17,25]. Images in post-ASO patients were performed to assess coronary artery courses, neo-aorta and neo-pulmonary vessels, and myocardial function. An ECG-gated CCTA of the entire chest was acquired. Prior to CCTA, laboratory tests and ECG were performed. Patients who had a heart rate >70 beats per min (bpm) usually required pre-medication, such as a single dose oral beta-blocker under the supervision of attending cardiologists to obtain an optimal heart rate during the examination. In patients with a heart rate below 70 bpm, a prospective triggering was applied to reduce the dose of radiation. However, in patients with a heart rate above 70 bpm, retrospective gating or adaptive sequential scan was considered. Acquisition parameters were selected based on the patient’s size to reduce radiation dose. The acquisition was performed after intravenous administration of contrast media with a concentration of iodine 350–370 mg/ml and a flow speed of 1.2–1.8 ml/s (depending on the patient’s body weight and intravenous access size). The acquisition was automatically triggered after reaching the threshold of saturation in the descending aorta. Reconstructions were performed with an interval of 10% of the R-R interval for the phase from 0% to 70% using retrospective gating (diastole) and with the same intervals with prospective triggering in the range automatically selected by the scanner. The images were then reconstructed in the workstation and interpreted by two cardiologists (CV, PWC) and a radiologist (KP) trained in cardiac imaging.

Two cardiologists analyzed coronary artery anatomy. Coronary artery anomalies (CAA) included, 1) significant coronary stenosis (visual grading ≥ 50% luminal narrowing), 2) acute take-off angle (angle between the proximal coronary artery and the aortic wall <45°), 3) interarterial course, and 4) prepulmonic course as described in the literature [25–28].

Analysis was performed using SPSS for Windows, version 22 (released 2013. IBM Corp, Armonk, NY, USA). Mean with standard deviation or median with interquartile range were used to describe continuous data, while percentages were used to describe categorical data. Patients were divided according to the presence of either usual coronary artery pattern (1LCx-2R), or unusual coronary pattern (other than 1LCx-2R pattern). Univariate analyses were initially performed to compare findings between the two groups. Predictive models for significant chest pain symptoms were analyzed using binary logistic regression. All statistical analyses were two-tailed, and a significance level of 0.05 was used to calculate the adjusted odds ratio and 95% confidence interval.

We analyzed 43 cases (67% male) that received coronary CTA after ASO. The median age at CCTA was 15.6 years (interquartile range; IQR, 12–21.3 years). There were 14 (33%) cases of simple d-TGA, 24 (56%) d-TGA with ventricular septal defect, and five (11%) Taussig-Bing anomalies. The median age at surgery was 16.5 days (IQR, 12–58 days) in simple d-TGA, and three months (IQR, 1–6 months) in d-TGA with ventricular septal defect (VSD) and Taussig Bing anomaly patients. Preoperatively, a usual coronary artery pattern was observed in 23 of 43 (53%) patients. During the follow-up period, five cases required percutaneous balloon pulmonary valvuloplasty (PBPV) due to severe pulmonic valve stenosis. Three patients underwent right ventricle to pulmonary valve conduit replacement due to severe subvalvular and valvular pulmonary stenosis from a small annulus. One patient had significant aortic root dilatation and needed aortic root replacement. Two patients underwent coronary artery bypass graft (CABG) surgery. All patients, except the two post-CABG patients, had good left ventricular systolic function (left ventricular ejection function (LVEF) 65.5% ± 7.9%).

The most common types of unusual coronary patterns were 2LCxR and 1L-2CxR (5 patients in each group), each accounting for 11.6% (Fig. 2). Twenty-two (51%) cases were found to have CAA from CCTA. Twelve patients had acquired abnormal coronary patterns whereas, ten patients had CAA (prepulmonic course) related to innate abnormal coronary patterns. Postoperative CAA were more often observed in patients with unusual coronary patterns (90% vs. 17.4% P < 0.001, Table 1). Of those who had CAA, fourteen patients were asymptomatic, and eight patients experienced chest pain or change in functional class.

Characteristics of symptomatic patients are shown in Tables 2 and 3. Two of these patients had significant coronary artery stenosis and the evidence of myocardial ischemia from adenosine stress cardiovascular magnetic resonance (CMR) and underwent CABG surgery. The first patient (#1 in Tables 2 and 3) with a usual coronary pattern experienced dyspnea on exertion and a decline in functional class I to functional class II 24 years after ASO. An exercise stress test showed depression in ECG lead II, III, aVF at less than 7 METs. Adenosine stress CMR showed myocardial ischemia at basal to apical anterior and anteroseptal wall. Coronary CTA showed an interarterial course of LMCA that was significantly compressed by the left pulmonary artery (LPA) and aorta. The LVEF improved from 36% five months after CABG surgery to 63% one year postoperatively. The second patient (#2 in Tables 2 and 3) with an unusual coronary artery pattern (1RL-2Cx) experienced typical angina pain and was admitted to the hospital due to cardiogenic shock one year after ASO. Myocardial perfusion imaging with thallium-201 was performed showing a large area of infarction at the anteroseptal and posteroinferior ventricular wall. The intraoperative finding was 80% stenosis of the coronary ostial at the implantation site that supplied blood to the RCA and LAD. The patient underwent coronary artery bypass graft (CABG) and suffered long-term residual myocardial dysfunction.

Apart from two patients who have myocardial ischemia and underwent CABG, the remaining of symptomatic patients (Tables 2 and 3) did not have the evidence of myocardial ischemia from adenosine stress CMR or EST. Although there was no evidence of myocardial ischemia, six out of eight had CAA from CCTA. All the remaining symptomatic patients did not required surgery.

Among subjects with a single coronary artery and common ostium of the RCA and LAD 60% (6 out of 10 patients) reported multiple episodes of chest pain (all symptomatic patients also had prepulmonic course of coronary artery after ASO), compared with 9% (3 of 33 cases) of patients with other coronary artery patterns (adjusted OR = 14.3%, 95% CI (2.5–82.3); P = 0.003) (Table 4). Of those who have preoperative unusual coronary patterns, prepulmonic course was found in 11 (55%) patients including 1R-2LCx, 1RL-2Cx, 1RLCx, and 2LCxR (Figs. 2 and 3). Six of 11 (54%) patients in this group experienced significant chest pain (crude OR = 11.6%, 95% CI (2.1–62.2), P = 0.001). Interarterial courses were seen in nine patients (20%), and two of those patients reported significant chest pain (22%). Among them, seven patients had an interarterial course of LAD or LMCA, while two patients had an interarterial course of RCA (Fig. 4). Acute take-off angle occurred in three patients (6.9%); each had high take-off (coronary ostia located above the sinotubular junction) [29], but none had significant coronary stenosis (Fig. 5). The mean aortic root z score in patients post-ASO was 4.71 ± 2.4 [30].