Evaluation of an unbalanced right dominant atrioventricular canal with a single 4D flow dataset Authors: Stephen Ream, Laura Schoeneberg, Wilson King Published: 2026 ABSTRACT Unbalanced atrioventricular (AV) canal anatomy presents a clinical challenge for the pediatric cardiology and cardiothoracic surgical team regarding the pre-operative assessment and ultimately, whether the optimal surgery should consist of a univentricular or biventricular repair. We present such a case of a patient with an unbalanced, right dominant, AV canal where a novel approach to preoperative imaging utilizing a relatively brief cardiac MRI consisting of a single 4D flow dataset was used to accurately assess ventricular volumes and AV valve sizes. This pre-operative imaging assessment was instrumental in the decision-making process for a successful single ventricle palliation for this patient. Case report A 7-day old term male born at 38w5d was diagnosed prenatally with right dominant unbalanced atrioventricular (AV) canal and coarctation of the aorta. A postnatal echocardiogram confirmed the diagnosis and demonstrated borderline left heart structures including the mitral valve and left ventricle. The patient was started on prostaglandin to maintain systemic perfusion in the setting of known coarctation. The cardiology and surgical teams were faced with the challenge of determining which surgical option was most appropriate to offer the patient and family. The key clinical question was whether the left ventricle was adequate for biventricular repair, or if not, whether the patient should undergo a single ventricle repair. The initial images obtained by transthoracic echocardiography were insufficient to accurately characterize the left ventricular volume given the morphology and were unable to determine if the left ventricle was “apex forming”. Additionally, AV valve-focused echocardiographic indices of biventricular repair potential were felt to be borderline with an AV valve index of 0.3 and an LV inflow index of 0.63. Cardiac magnetic resonance (CMR) imaging was thus obtained to better evaluate left ventricular volume and Qp: Qs ratio to determine biventricular candidacy. Cardiac MRI was performed utilizing a GE Discovery 3 T 750 W scanner and a 16 channel GEM large flex coil. The patient was intubated and sedated with general anesthesia. Ferumoxytol 3 mg/kg was administered over 15 min prior to scan. A single 4D flow pulse sequence was performed at a resolution of 1.2 × 1 × 1 mm reconstructed to 0.6 × 0. 86 × 0.86 mm with 20 phases reconstructed to 25 phases (FA 15°, BW 62.5 MHz, TR 5 ms TE 2.5 ms). The overall scan time was 23 min and 50 s. Ventricular volumes and flow data were obtained from the 4D flow sequence as seen in Figs. 1 and 2. Myocardial borders and vessels were sharp throughout the cardiac cycle, increasing our confidence in the dataset. The left ventricle was significantly hypoplastic, with the left ventricular end-diastolic volume measured at 18 mL/m2, and an LVEDV/RVEDV ratio of 0.15:1. The left AV valve was hypoplastic with a 4-chamber measurement of 7 mm with z-score of -2.4. The left AV valve to total AV valve cross-sectional area 0.23:1 and left AV valve to right AV valve inflow ratio by 4D flow of 0.3:1. 4D flow also confirmed significant pulmonary overcirculation with a Qp: Qs of 5:1 by branch pulmonary arteries to aortic flow comparison and 7:1 comparing pulmonary vein flow to caval flow. The inlet ventricular septal defect was moderate in size, measuring 4.6 mm in the 4-chamber view. A 3D printed model (Fig. 3.) was also constructed to help illustrate the ventricular size discrepancy and 3D relationships of the anatomy. Fig. 14D Flow CMR. (A) Inverse reconstruction apical 4 chamber angulation, demonstrating non-apex forming hypoplastic left ventricle. (B) 4D Flow demonstrating disproportionate inflow, with LV receiving only 23% of total AV inflow. (C) 4D Flow in diastole re-demonstrating the discrepant ventricular inflow. (D) Apical 4 chamber 4D Cine showing hypoplastic, non-apex forming LV and (E) a similar plane in systole showing the primum and secundum ASDs via 4D flow. (F) AV valve view en face with 4D flow demonstrating the unbalanced nature of the AV valves with significantly more inflow to RV than LV 4D Flow CMR. (A) Inverse reconstruction apical 4 chamber angulation, demonstrating non-apex forming hypoplastic left ventricle. (B) 4D Flow demonstrating disproportionate inflow, with LV receiving only 23% of total AV inflow. (C) 4D Flow in diastole re-demonstrating the discrepant ventricular inflow. (D) Apical 4 chamber 4D Cine showing hypoplastic, non-apex forming LV and (E) a similar plane in systole showing the primum and secundum ASDs via 4D flow. (F) AV valve view en face with 4D flow demonstrating the unbalanced nature of the AV valves with significantly more inflow to RV than LV Fig. 2(A) 3D reconstruction of 4D flow sequence demonstrating the hypoplastic aortic arch in the foreground with patent ductal arch in the background. (B) same image with 4D Flow (A) 3D reconstruction of 4D flow sequence demonstrating the hypoplastic aortic arch in the foreground with patent ductal arch in the background. (B) same image with 4D Flow Fig. 33D printed model of the patient’s anatomy emphasizing the small comparative size of the left ventricle and illustrating the 3D relationships of the anatomy in general 3D printed model of the patient’s anatomy emphasizing the small comparative size of the left ventricle and illustrating the 3D relationships of the anatomy in general Discussion The initial surgical options considered by the team included: (1) an aortic arch reconstruction/repair along with pulmonary artery banding to restrict pulmonary blood flow in the setting of AV canal; (2) a Stage 1 Norwood operation with Sano shunt; (3) “Hybrid procedure” with stenting of the patent ductus arteriosus (PDA) in the cardiac catheterization lab as well as surgical pulmonary artery banding simultaneously. Depending on the initial surgical intervention, subsequent surgical options included: (1) Continuing down a univentricular pathway with a Comprehensive Stage II (if initial operation was a Hybrid procedure) (2) Glenn operation (if initial operation was a Norwood) followed by an eventual Fontan operation. Alternatively, if a biventricular repair was later thought to be feasible, the original repair could be taken down as applicable and the patient might be able to proceed with complete AV canal repair. Current data demonstrates that in older pediatric patients with AV canal, left ventricular volumes as small as 20 mL/m2 may be able to safely undergo eventual biventricular repair [1]. Cardiac MRI precisely quantified the left ventricular volume in our patient at 18 mL/m2, below this threshold. CMR was also able to accurately quantify the left AV valve size, and the Qp: Qs ratio. The information determined unfavorable candidacy for biventricular repair. The VSD was also smaller than suggested by echocardiogram, making arch repair and PA band less attractive. Based on MRI and echo data, the patient underwent a Stage I Norwood/Sano repair and did well. This case illustrates how a single 4D flow dataset can be used to accurately obtain structure, function, and flow information in a clinically impactful manner, all in a single acquisition in less than 25 min.