Diagnostic performance of 320 cardiac MDCT angiography in assessment of PDA either isolated or associated with duct dependent congenital heart disease

Patent ductus arteriosus (PDA) is one of most common congenital heart defects, it's a unique vascular structure that provides direct communication between pulmonary and systemic circulation. MDCT angiography is a good imaging modality for evaluation of the PDAs and detection of their exact morphological type; course and diameters, which is important before percutaneous closure or stenting procedure of the PDA, also for selection of closure hardware. The aim of this study was to assess the role of MDCT angiography in qualitative and quantitative evaluation of PDA and associated cardiac and\or extracardiac anomalies. Echocardiography detected PDA in 28\30 cases while cardiac MDCT detected PDA in all studied 30 cases confirmed by cardiac catheterization and/or operation. MDCT angiography had sensitivity 100% and specificity 100% for PDA detection. PDA originated from aortic isthmus in 15 cases, inferior surface of aortic arch in 11 cases and innominate artery in 4 cases. The most common morphological type of PDA was type A (cone\46.67%) followed by type C (tubular\23.3%), type D (complex\10%), type E (elongated\13.33%) and type B (window\6.67%). The spearman correlation coefficient test demonstrated poor correlation between size of aortic end and MPA (P = 0.75), and between size of pulmonary end and diameter of MPA (P = 0.99) and also demonstrated fair correlation between length of PDA and MPA (P = 0.018). PDA was isolated in 4\30 cases and associated with cardiac and\or extra cardiac anomalies in 26\30 cases included; ASD (n = 18), VSD (n = 16), pulmonary atresia (n = 7), transposition of great arteries (n = 5), teratology of Fallot (n = 4), aortic coarctation (n = 4), persistent truncus arteriosus (n = 3), tricuspid atresia (n = 3), anomalous of pulmonary venous return (n = 3), hypoplastic segment of aorta (n = 2), Ebstein's anomaly (n = 1), bicuspid aortic arch (n = 1) and left hypoplastic heart syndrome (n = 1). Cardiac MDCT angiography was superior to Echocardiography in detection, quantitative and qualitative evaluation of PDA either isolated or associated with congenital cardiac and\or extracardiac anomalies and was superior to Echocardiography in detection of associated extracardiac anomalies rather than associated intra cardiac anomalies.


Introduction
Congenital heart disease (CHD) is considered a major cardiac problem in pediatrics. Patent ductus arteriosus (PDA) represents 5-10% of all congenital malformations and considered one of the most common congenital heart defects in premature infants [1,2].
Ductus arteriosus is a vascular communication between pulmonary and systemic circulations. The ductus arteriosus during fetal life diverts blood towards the descending aorta and placenta away from the fluid filled lungs, constriction of the ductus arteriosus and obliteration of its lumen after birth separate the pulmonary and systemic circulations [3].
Failure of closure of ductus arteriosus in some infants lead to persistent communication between aorta and pulmonary arteries, left to right extra cardiac shunting occur which depends on size of PDA and pulmonary vascular resistance. Congestive heart failure (CHF) may occur in medium and large shunts due to increased pulmonary blood flow and volume overload of the left heart [4].
PDA may present as isolated or associated with other cardiac and extra cardiac anomalies like tricuspid atresia, pulmonary atresia, teratology of Fallot, in these cases PDA is important for oxygenation and post-natal ductus constriction lead to severe hypoxia and cyanosis. Also, it may be associated with severe systemic blood flow restriction like aortic coarctation, aortic stenosis, interrupted aortic arch or left hypo plastic syndrome [5].
Cardiac imaging plays an important role in the diagnosis, management and follow up after surgical procedures [6]. Echocardiography is always the first line study of choice for neonates with congenital heart disease, it provides immediate high-resolution anatomical and physiological information in addition to its safety, speed, noninvasiveness, and easy availability. The main disadvantages were it is operator dependent and the image quality can be degraded in uncooperative children or by poor acoustic window [7,8].
MDCT and Cardiac Magnetic Resonance Imaging (cMRI) provide valuable noninvasive imaging tools, they are helpful in assessing the complex cardiovascular morphology especially the extracardiac association as well as pulmonary artery anatomy and aortopulmonary collateral vessels [9]. Noninvasive imaging of PDA remains a challenge. MDCT angiography enables excellent qualitative and quantitative information of PDAs [10].
Cardiac Magnetic Resonance Imaging (cMRI) can provide both functional and anatomical information. Its use is limited in uncooperative or severely ill pediatric patients, and it is contraindicated in patients with pacemakers. Also, it takes longer time than CT and may require general anesthesia especially in patients less than 5 years old. In addition, the images available do not have the necessary spatial resolution to assess small anatomical structure [11].
Conventional angiography acts as the gold standard cardiac imaging tool. However, it is an invasive method which may cause death in up to 1% of neonates with complex CHD [12].
Multidetector CT enables a detailed evaluation of PDA morphology and size, potential complications such as thrombosis, aneurysms and calcifications which were useful in the diagnosis of PDA, planning of PDA percutaneous closure. MDCT also enables quantitative evaluation of great vessels morphometry including potential pulmonary artery hypertension and Eisenmenger syndrome [13].
Multidetector CT is easy and achievable now with sub mSv doses with short acquisition time minimizing the need for general anesthesia in pediatric imaging. It is helpful when dealing with clinically unstable children especially from intensive care setting [14]. Multiplanar and three-dimensional (3D) images reformatted from multi-slice spiral CT data can demonstrate normal and pathologic cardiovascular structures in patients with congenital heart disease [11].
The aim of this study was to assess the role of cardiac MDCT angiography in qualitative and quantitative evaluation of PDA and associated cardiac and\or extracardiac anomalies.

Study population
This prospective study was conducted on 30 cases suspected to have PDA by clinical examination or echocardiography referred from Pediatric cardiology and Cardio-thoracic departments to Radio-diagnosis and medical imaging department between March 2020 to April 2021 for assessment of PDA and its associated cardiac and\or extra cardiac anomalies.
Patients unwilling or lost for follow up, patients with allergy to the iodinated contrast material, and impaired renal function (creatinine level > 1.5 mg/dl) were excluded from study.
Privacy and confidentiality of all patient's data were guaranteed and their coded number for every patient's file that include all investigations.
All patients submitted to the following Study planning • All the patients had Echocardiographic reports of suspected or diagnosed PDA and underwent MDCT angiography of the heart and great vessels to confirm the diagnosis or to answer specific anatomic question raised by inconclusive echocardiography findings before planning the adequate management. • Our gold standard in study was cardiac catheter or surgery.
• Review of Echocardiographic findings and consultation with the referring physician was attempted prior to the study to discuss the clinical background of the case. • Checking renal functions to exclude patients with impaired renal function. • Proper history taking from the parents.

Preprocedural preparations
• Reassurance of the parents by description of the procedure to them. • Placing of intravenous cannula (20-to 24- Older cases (n = 3) were responding satisfactorily to verbal reassurance to be able to completely suspend respiration.

Image acquisition
• Patients were scanned using 320-row multidetector CT scanner (Aquilion One, Toshiba Medical Systems, Otawara, Japan). • The patient lay supine on the CT table, proper positioning of the patient on table was done to ensure the heart lies in the iso-center of the gantry for spatial resolution optimization. • Application of ECG electrodes to chest wall after skin preparation with alcohol. • Insertion of the intravenous (IV) line and injection testing with saline was done to ensure good IV access with no extravasation. • Thereafter, obtaining of a scanogram (frontal & lateral views) where the scan ranges from neck root entailing proximal common carotid and subclavian arteries down to the level of portal vein inferiorly. • Non-ionic, non-diluted contrast material (Ultravist 300, Schering AG, Germany or Omnipaque 300, Nycomed, Amersham) was injected in 28 and 2 patients respectively according to body weight with maximum dose was 2 ml /kg through the peripherally inserted IV cannula using dual syringe mechanical power injector (Stellant D, Medrad, Indianola, PA, USA) with flow rate 1-1.5 ml/sec increased to 3 ml/ sec in older children. • IV saline chaser injection of 1 ml/kg was given immediately after the contrast injection to improve the contrast homogenicity and opacification. • Manual Bolus tracking was applied after 10-15 s from contrast material injection timing (for upper limb venous line) and after 20 s (for lower limb venous line), the scan is initiated after opacification of both ventricles. • All scans were performed in the cranio-caudal direction, with CT parameters adapted to the patients' weight. The patients were scanned using a singlephase retrospective ECG gated CTA volume scan with a rotation time of 0.35 s and a tube voltage of 80 kV increased to 100 kV in 3 older children. The quality of images was reviewed before end of examination. • The patient was kept under observation for 15-30 min till recovery of sedation.
Image reconstruction and post processing Full volumes were reconstructed in 0.5 mm-thickness slice. Post-processing of multi-detector CT (MDCT) scans was performed by dedicated remote workstation (Vitrea Fx, Vital Images, USA).

Image interpretation
Images were interpreted guided by the anatomical and segmental/sequential approach as follows:

Statistical analysis
Statistical analysis of the present study was conducted by SPSS V.20. Qualitative data was presented using number and percentage. Quantitative data presented as mean and standard deviation (SD). For categorical variables, Chisquare test was used for analysis. The level of significance was adopted at P < 0.05.
All the patients had Echocardiographic reports of suspected or diagnosed PDA and underwent MDCT angiography of the heart and great vessels to confirm the diagnosis. We compared cardiac MDCT findings with the data collected by cardiac catheterization and/or operation (gold standard).
Echocardiography detected PDA in 28 cases (93.3%) while MDCT angiography detected PDA in all studied 30 cases confirmed by cardiac catheterization and/or operation. MDCT angiography had sensitivity 100% and specificity 100% for PDA detection with the advantage of giving precise anatomical details and volume rendered images which are very helpful before surgery for accurate surgery planning.

Qualitative and quantitative assessment of PDA by cardiac MDCT
Twenty six cases had left sided PDA and 4 cases had right sided PDA. PDA originated from aortic isthmus in 15 cases (Fig. 1), inferior surface of aortic arch in 11 cases (Fig. 2, 3) and innominate artery in 4 cases (Fig. 4). The morphological type of PDA was evaluated according to Krichenko et al. classification [15], it was found that the most common type of PDA was type A (cone) detected in 14 (46.67%) cases followed by type C (tubular) detected in 7 (23.3%) cases (Fig. 2,3), type D (complex) detected in 3 (10%) cases, type E (elongated) detected in 4 (13.33%) cases ( Fig. 1) and type B (window) detected in 2 (6.67%) cases (Fig. 5). According to tortuosity of PDA, it was straight (type I) in 15 cases, has one turn (type II) in 9 cases, and has multiple turns and tortuous (type III) in 6 cases.
The size of aortic end of PDA, pulmonary end of PDA, length of PDA, diameter of main pulmonary artery, diameter of aortic isthmus and descending aorta were calculated and compared according to the type of PDA as listed in Table 1.
The spearman correlation coefficient test demonstrated poor correlation between size of aortic end and MPA (P = 0.75), and between size of pulmonary end and diameter of MPA (P = 0.99). Fair correlation between length of PDA and MPA (P = 0.018). Also demonstrated poor correlation between size of aortic end and diameter of RPA (P = 0.57), between size of pulmonary end and diameter of RPA (P = 0.4), and between length of PDA and RPA (P = 0.12). The spearman correlation coefficient test demonstrated poor correlation between size of aortic end and diameter of LPA (P = 0.3), and between size of pulmonary end and diameter of LPA (P = 0.07). Fair correlation between length of PDA and LPA (P = 0.027), the above findings shown in Table 2 and Figs. 6, 7.

Intra-cardiac anomalies associated with PDA
The intracardiac defects encountered within the study were classified into atrial and ventricular septal defects. The most common ventricular septal defect (VSD) was perimembranous VSD (8 cases, 26.6%). The most common atrial septal defect (ASD) was ostium secundum ASD (12 cases, 40%).
Cardiac MDCT missed the diagnosis of ASD in 1 case and VSD in 1 case, hard to interpret ASD in 1 case and VSD in 2 cases due to cardiac motion during the examination and accurately diagnosed other types of ASD and VSD. Chi square test revealed non statistically significant difference between Echo and MDCT angiography in detection of ASD (P = 0. 95) and VSD (P = 0.90) as listed in Table 4.

Extra-cardiac anomalies associated with PDA
Regarding aortic anomalies (Figs. 3, 5), Echocardiography missed aortic coarctation in 1 case, hypoplastic aortic segment in 3 cases, right sided aortic arch in 3 cases. However, MDCT angiography missed one case of aortic valve stenosis due to motion artifact that interfere with good interpretation of aortic valve anatomy as listed in Tables 5 and 6.
Different aortic arch branching patterns were detected by MDCT angiography; left sided aortic arch with normal branching pattern (n = 18), left sided arch with bovine branching pattern (n = 6), right sided aortic arch with mirror image branching pattern (n = 3), left sided with aberrant right subclavian artery (n = 1) (Fig. 1), and right sided aortic arch with aberrant left subclavian (n = 1).
Regarding coronary arteries anomalies, twenty-two cases had no coronary anomalies, 5 cases were hard to be interpretable due to motion artifact (high heart rate), 2 cases had single coronary artery and 1 case had ectatic left coronary artery with cameral fistula end at right ventricle (Fig. 1).
Chi square test revealed statistically significant difference between Echocardiography and cardiac MDCT in detection of left pulmonary artery and right pulmonary artery anomalies (P = 0.05 for left pulmonary artery and P = 0.043 for right pulmonary artery) as listed in Table 7.
Cardiac MDCT missed diagnosis of one case with pulmonary stenosis and 2 cases with hypoplastic pulmonary artery branches due to cardiac motion which degrade the quality of the image as listed in Table 8.
Regarding pulmonary venous abnormalities, pulmonary veins were normal in 27 cases and venous anomalies were diagnosed in 3 cases; (TAPVR) total anomalous pulmonary venous return detected in 1 (3.33%) case, (PAPVR) partial anomalous pulmonary venous return

Radiation dose
The MDCT scanner automatically estimated the absorbed radiation dose in the form of dose length product (DLP) (radiation dose for a predetermined scanned length). For transforming the DLP to effective radiation dose in milliSievert, it is agreed to multiply the DLP by conversion coefficient factor according to age. Thirty MSCT examinations were done, and the mean absorbed radiation dose was 91.79 ± 88.45 mGy per scan and the mean effective dose was 2.61 ± 1.38 mSv as illustrated in Table 9.

Discussion
PDA could be seen isolated or associated with other congenital cardiac and\or extra cardiac anomalies. A detailed description of the PDA is important before attempting percutaneous closure of the PDA and during selection of closure hardware [16].
Cardiac catheterization was traditionally used to complement Echocardiography and to visualize extracardiac great vessels and remains the gold standard for direct hemodynamic assessment [17]. Cardiac MDCT is an important modality used for the accurate depiction of complex cardiovascular anatomic features before and after surgery [18].
Potential interventional complications (dissection, occlusion, bleeding, etc.) with cardiac catheterization are absent in cardiac MDCT, also it is applicable in the case of bleeding diathesis. Conventional angiography has the disadvantage of being invasive, taking a long time, and requiring anaesthesia in the paediatric population. However, hemodynamic information such as pressure curves and oxygen saturation data cannot be derived from MDCT examination [19].
The presence of clinical symptoms in patients with PDA is related to the magnitude of the shunt through the duct [11], in the present study we found two cases of the studied cases were asymptomatic and the other 28 cases had symptoms.
In the current study, Echocardiography missed detection of two cases of PDA, this came in agreement with Hu et al. [20] and Nie et al. [21], they stated that Echocardiography can only identify relatively large ones, in contrast to study did by Eltatawy et al. [22], they reported that Echocardiography detect PDA in all cases of their study but overestimate one PDA.
Lin et al. [23] reported that PDA originated from inferior surface of aortic arch in all cases. However, in the present study we detected origin of PDA from aortic isthmus in 15 cases, from inferior surface of aortic arch in 11 cases and from innominate artery in 4 cases.
Krupinski et al. [13], documented the occurrence of type A PDA with a wide aortic and narrow pulmonary orifice in the majority of the cases (16 cases\50%) followed by type C in 9 cases (28%), also they reported that types D (with multiple constrictions) and type E (elongated with distal constriction) ducts were marginal and made up only a few percent of patients (2/32) 6%. Also, Morgan et al. [10] reported that most common type of PDA was type A. We reported in this study that type A PDA was the most common PDA type (14/30) represented 46.67% and type B was the least common type of PDA (2/30) represented 6.67%.
We reported larger aortic orifice than pulmonary orifice (4.6 ± 1.8 mm vs. 2.4 ± 1.04) in type A PDA. This came in acceptance with Krupinski et al. [13], they provided quantitative evaluation confirmed the qualitative findings of larger aortic than pulmonary orifices in patients with the type A morphology of PDA (10.2 ± 5.2 mm vs. 5.1 ± 2.7 mm, P = 0.0001).
The spearman correlation coefficient test in this study demonstrated fair correlation between length of PDA and MPA (P = 0.018). Also demonstrated fair correlation between length of PDA and LPA (P = 0.027) which agreed with Krupinski et al. [13], they reported that diameters of PDA correlated with diameters of pulmonary artery and this correlation was strongest between PDA diameter at the narrowest site and diameter of main pulmonary artery.
Eltatawy et al. [22] reported 14 cases with PDA, only two of them were isolated with no intracardiac or another vascular anomaly. We noted in current study four cases were isolated PDA and 26 cases were associated with other cardiac and\or extra cardiac anomalies.  Regarding associated cardiac and extra cardiac anomalies in the present study; 15 cases with duct dependent pulmonary anomalies, 7 cases with duct dependent systemic anomalies and 5 cases with duct dependent mixing circulation anomalies (TGA). However, in the study did by Lin et al. [23], they reported all cases of their study were duct dependent pulmonary circulation.
Enaba et al. [24], reported 15 cases of tetralogy of Fallot (25%), 12 cases of tricuspid atresia (20%), four cases of Ebstein's anomaly (6.5%) and 7 cases of pulmonary atresia and stenosis (11.5%). Also, Seif El-Din et al. [16] reported 4 cases of tetralogy of Fallot associated with PDA.   MDCT angiography was superior to Echocardiography in assessment of extra cardiac aortic anomalies, Echocardiography missed aortic coarctation in one case which was similar to Al-Azzazy et al. [19], they reported that Echocardiography missed two cases of coarctation in their study from 24 cases of aortic coarctation with significant difference between Echocardiography and cardiac MDCT in the detection of aortic coarctation.
Echocardiography in the current study missed detection of two cases of hypoplastic segment of the aorta which came close to Eltatawy et al. [22], they missed detection of two cases of aortic arch hypoplasia.    Demonstration of the origin of coronary arteries may be part of the routine workup for patients with PDA undergo diagnostic evaluation and therapeutic interventions [25].
In the present study we detected coronary anomalies in 3 (10%) cases (2 had single coronary artery arise from right coronary cusp and 1 case had ectatic left coronary artery) which was similar to study did by Dotan et al. [25], they reported (10.8%) anomalous origin of coronary artery associated with PDA.
Abd El-Gaber et al. [26] reported atresia (absence) of the main pulmonary artery (14 cases\33.3%) was the   commonest encountered type of pulmonary artery anomalies associated with PDA. However, we reported main pulmonary artery atresia in 4 (13.33%) out of 18 cases, dilated MPA in 7 (23.33%) cases, MPA hypoplasia in 6 (20%) cases, subpulmonic stenosis in 1 (3.33%) case. Chi square test in the current study revealed non statistically significant difference between both Echocardiography and MDCT angiography in detection of main pulmonary artery anomalies (P = 0.49) but revealed statistically significant difference between both Echocardiography and MDCT angiography in detection of left pulmonary artery and right pulmonary artery anomalies (P = 0.05 for left pulmonary artery and P = 0.043 for right pulmonary artery).This came in agreement with Liu et al. [27], they missed detection of MPA anomalies in 7 cases, 46 cases of RPA and LPA anomalies by Echocardiography due to restricted acoustic windows and operator dependency. Also agreed with Chandrashekhar et al. [8], they also reported failure of Echocardiography to visualize 3 cases of MPA anomalies, 6 cases of RPA anomalies and 2 cases of LPA anomalies due to limited acoustic windows, low spatial resolution and structures are obscured by overlying bone and aerated lung.
The relatively small numbering of the studied cases was our major limitation in this study that not giving full idea about the diagnostic efficacy of cardiac MDCT.

Conclusion
Cardiac MDCT angiography was superior to Echocardiography in detection, quantitative and qualitative evaluation of PDA either isolated or associated with congenital cardiac and\or extracardiac anomalies and was superior to Echocardiography in detection of associated extracardiac anomalies rather than associated intra cardiac anomalies.
We recommended that patients should be carefully selected for CT imaging, and strategies for radiation exposure reduction need to be applied. Increasing the sample size in the future is recommended.