The attachment sites of the valve leaflets and corresponding papillary muscle attachments were identified and marked on the image.

Those marks were confirmed on two-dimensional (2D)-multi-planar reformatting (MPR) image of the long axis of the heart corresponding to the 3D image. The angle of the cross-section was fine-tuned so that the two marks appear simultaneously in the 2D-MPR image.

Measurements were taken for 11 phases (0%-50%) for cA and cP, respectively. Three measurements were taken per phase in all cases, and the mean value was used as the chordae tendineae length (CL). The maximum CL in the 11 phases for each case was defined as cA (max) and cP (max), and comparisons were made to evaluate the correlation with the body surface area (BSA). The rate of change of CL at each time phase concerning early systole (phase 0%) was graphed to compare the behavior of cA and cP during systole.

All data were analyzed with IBM SPSS Statistics version 29 (IBM, Armonk, NY, United States) and presented as mean ± SD. The homoscedasticity of the mean values of each group was confirmed, and an unpaired t-test was used to compare them. Pearson’s product-moment correlation coefficient was used to analyze the correlation between CL and BSA. Intra-rater reliability was determined by calculating intraclass correlation coefficients (ICC) [ICC (1, 1) and ICC (1, 3)]. The significance level for all statistical tests was < 5%.

The mitral anterior leaflet chordae tendineae could be measured in all cases. In most cases, the cA and cP chordae tendineae could be measured visually, and the leaflets and papillary muscle attachments could be easily identified, even for smaller chordae tendineae that could not be visualized on CT. Individual differences occurred in the number of chordae tendineae originating from the papillary muscles. The number originating from the anterior leaflet of the APM was 3.18 ± 0.60, and that from the anterior leaflet of the PPM was 2.45 ± 0.52, with a significantly higher number originating from the APM. The PPM of Case 10 was smaller than those of the other cases, being displaced toward the apex.

CL exhibited large variations between individuals, with mean cA (max) and cP (max) being 20.2 mm ± 1.95 mm and 23.5 mm ± 4.06 mm, respectively. cP was significantly longer than cA (max). The correlation coefficients (r) with BSA were 0.60 and 0.78 for cA (max) and cP (max), respectively, showing significant correlations (Table 2).

CL revealed constant variation during systole for both cA and cP, with a maximum 1.16-fold elongation in cA and 1.23-fold elongation in cP from early to mid-systole. For cP, CL reached a plateau at 15% and remained elongated until end-systole. Conversely, for cA, after peaking at 15%, CL shortened slightly and then peaked again as end-systole approached (Figure 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Variability of the mitral valve primary chordae tendineae during systole. The primary chordae tendineae from the posterior papillary muscle to the A2-3 region (cP) plateaued at 15%. The primary chordae tendineae from the anterior papillary muscle to the A1-2 region (cA) peaked at 15%, then shortened slightly and again peaked toward end-systole. 0-40%: Systole, 45-100%: Diastole.

We were able to measure the length of the mitral anterior leaflet primary chordae tendineae using 4D-CT. The chordae tendineae are small tissues that move at high speeds, making accurate evaluation challenging. Recently, however, several studies have attempted to measure the length of mitral chordae tendineae with transthoracic echocardiography (TTE), transesophageal echocardiography (TEE), and cardiac CT. TTE measurements have been used for a relatively overtime owing to their simplicity. However, considering that it is challenging to obtain a cross-section that adequately evaluates the 3-dimensional length of the mitral chordae tendineae, this method is somewhat lacking in accuracy[8]. Recently, evaluation by TEE has become widely used, while improvements in techniques such as transgastric echocardiography that produce fewer artifacts have been reported[9]. Although TEE is at present considered the standard method for evaluating CL, this method has some drawbacks, such as the mental and physical stress associated with the examination and the skill needed to perform the technique[10,11]. Despite these challenges, with the advent of multi-slice CT, evaluation using cardiac CT has garnered attention. With helical CT, the mitral valve structure can be obtained as volume data; following this, MPR and 3D image can be created to reconstruct optimal cross-sections for chordae tendineae assessment. This allows for direct and accurate measurement of CL, and for all chordae tendineae to be assessed in a single imaging session[12]. However, the characteristics of 3D measurements make them prone to error, while creating suitable cross-sections for 2D-MPR image measurements is time-consuming. Therefore, in the present study, we combined both methods to obtain simple, reproducible, and accurate results. We identified the points to be measured in 3D image, and then accurately measured them in 2D-MPR image. In most instances, we could evaluate the anterior leaflet primary chordae tendineae directly, although in some cases they were blurred from timing artifacts.

In previous reports on autopsies, mitral anterior leaflet CL was 15 mm-20 mm[13] and 17.5 mm ± 0.25 mm[14]. While this study had similar findings, the CLs measured in our study were slightly longer than those measured in previous autopsies. Moreover, CLs measured by the previous studies using TEE tended to be longer than those obtained through autopsies, i.e., 21.3 mm ± 2.8 mm[9] and 19.3 mm ± 0.50 mm[11]. The lengths obtained by autopsies were shorter than those obtained by live measurements because the chordae tendineae in the dead bodies lacked proper tension. In the present study, cP (max) was longer than cA (max) and APM had more chordae tendineae origins than PPM. A similar tendency was observed in a study of pig hearts, with cA 16.04 mm ± 2.16 mm and cP 20.56 mm ± 3.35 mm (Sengda), suggesting that it is possible to accurately measure cA and cP with 4D-CT[14]. The ICC (1, 1) and ICC (1, 3) in our study were 0.963 and 0.987 for cA and 0.970 and 0.990 for cP, respectively, suggesting very high reproducibility of this method.

cA (max) and cP (max) exhibited large individual differences in the present study, and a significant correlation was found between BSA and CL. It is known that left ventricular end-diastolic diameter correlates with BSA in normal hearts; however, larger bodies and hearts may have longer CL[15,16]. Although there have been no reports on differences in CL based on body size, it may be necessary to take body size into account, especially when making preoperative predictions. It could be beneficial to investigate the relationship between heart size and CL by comparing various echocardiographic parameters to CL.

Variation in CL was observed during systole in the present study. Conventionally, the chordae tendineae were thought to be rigid, non-elastic tissues composed mainly of elastin and collagen; however, recent studies have indicated that they have some degree of elasticity[17-19]. In the present study, a maximum 1.16-fold elongation was observed in cA from early to mid-systole and 1.23-fold elongation in cP. This indicates the importance of timing (15%-40%) in CL measurement, because measuring CL without specifying the timing on cardiac CT may underestimate it. In the present study, cP exhibited a PV loop-like variation with increasing intracardiac pressure, whereas cA exhibited a slight shortening of CL once it peaked, and then another peak toward end-systole. This may distort the pressure applied to the mitral anterior leaflet during systole. In vitro studies have suggested that in normal hearts, pressure on the mitral valve is uniform, but the displacement of the papillary muscle may result in non-uniform pressure[20]. The left ventricle is also known to undergo torsional movement during systole, which may have some effect on the displacement of the papillary muscle or the length of chordae tendineae[21]. Moreover, a recent analysis found that intracardiac blood flow is not linear but spiral, suggesting the pressure applied to the mitral valve in the body is not uniform[22]. While we were unable to determine the cause of this, dynamic assessments and blood flow analyses of the papillary muscles and left ventricular wall should be performed going forward.

As described above, while 4D-CT was seen as useful for CT measurements, the risks associated with radiation exposure and the use of contrast media in CT imaging should be taken into consideration when using this technique. That said, this imaging method does not require any special treatment or technique and can be performed simultaneously with coronary artery CT. If an institution routinely performs coronary artery CT as a preoperative examination, this method can be simultaneously implemented with no increase in radiation exposure or other burdens on patients. In the present study, we focused on two of the more important and thicker chordae tendineae; however, we believe that relatively thin chordae tendineae close to the commissure could also be evaluated if suitable cross-sections were prepared. Variations in the shape of the papillary muscles were also seen in the cases in this study. In case 10, a smaller PPM than that in the other cases was noticed and resulted in a very long cP. It is very important to take these individual differences into account when performing mitral valvuloplasty, and CT is an excellent tool for evaluating the entire mitral valve apparatus, including the papillary muscles. Using this study as a stepping stone, we would like to pursue further research by evaluating patients with pathological conditions to help make 4D-CT a useful preoperative examination for mitral valvuloplasty.

The limitations of this study include a small sample size, a narrow range of participants restricted to older individuals, and the absence of direct measurement of chordae tendineae. The chordae tendineae behavior in older individuals may differ from that in younger patients.