To investigate the robustness and reliability of clinical long-axis CMR imaging for assessment of AVJ excursion and recoil by applying a semi-automatic tracking system.Six human subjects (47±14 yr) in the absence of any left ventricular (LV) and/or right ventricular (RV) dysfunction were enrolled and underwent CMR scan. The protocol was approved by the SingHealth centralized institutional review board, and informed consent was obtained from each participant. All subjects were imaged on a 3T scanner (Ingenia, Philips Healthcare, Best, Netherlands) using steady-state free precession (SSFP) cine gradient echo sequences. Standard survey images were acquired followed by a two-, three-, and four-chamber views, as well as 5 short-axis slices parallel to the mitral valve ring covering the left ventricle from base to apex. With these views as guidance, 18 radial slices were acquired with 10° angular equidistance in the LV long axis, defined from the apex to the centre of the mitral valve orifice (Fig. 1(a)). Typical imaging parameters were as follows: field of view 300 x 300 mm, voxel size 1.04 x 1.04 x 8.00 mm, 40 frames/cardiac cycle, SENSE factor 2, breath-hold time 7s. An in-house developed program ¹ was applied to semi-automatically track the AVJ deformation as a function of time in all 18 planes (Fig. 1(b)). Briefly, a small rectangle (called mask) was manually drawn in the end-diastole frame containing the AVJ point of interest, which was automatically tracked throughout the cardiac cycle using the template matching algorithm ². Four clinically useful data were then extracted: Sm, positive peak systolic velocity; Em, early diastolic velocity; Am, late diastolic velocity during atrial contraction; and maximal displacement. Time to peak systolic and diastolic velocities as well as maximal displacement were also measured. The CMR-derived motion parameters were averaged over all, and every two, three and four AVJ points. In addition, slices 4, 10 and 16 corresponding to routine CMR long-axis views (Fig. 1(b)) were selected and used to generate results based on an average of 6 AVJ points (Fig. 2(a) and (b)).Our proposed tracking system successfully extracted the motion parameters for all the AVJs of all subjects enrolled in the study. Figure 2(a) and (b) showed the computed AVJ velocity and displacement curves of one 39-yr-old male healthy volunteer in slices 4, 10 and 16. The extracted values of 6-point mean Sm, Em, Am and maximal displacement were 8.7, 13.3, 7.6 cm/s and 15.3 mm, respectively. Figure 2(c) presented the peak AVJ velocities and displacements as a function of annular segment along the AVJ. Figure 3 and Table 1 compared the motion parameters based on different number of AVJ points. These results demonstrate, first, the CMR-derived mean AVJ motion parameters were not affected by the number of uniformly selected AVJ points along the annulus. Second, the 6-point mean results based on routinely acquired long-axis CMR views agree well with those averaged over all the 36 AVJ points.As important clinical indicator with both structural and functional information, AVJ motion parameters can be successfully derived by applying a previously introduced semi-automatic tracking and post-processing system based on multiple radially rotational long-axis CMR. Routinely acquired two-, three-, and four-chamber CMR views seem to be robust and reliable enough without additional imaging planes. Potential added clinical value by exploiting full long-axis cine sets needs to be further investigated.