Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Sep 24;45(5):e70032. doi: 10.1111/cpf.70032

Reproducibility of diastolic function parameters in repeated ECG‐gated SPECT myocardial perfusion imaging and echocardiography

Aaro Krohns 1, Tomi P Laitinen 2,3, Tiina M Laitinen 2, Saara Sillanmäki 2,3,✉
PMCID: PMC12459082  PMID: 40990226

Abstract

Objective

To study the reproducibility of diastolic function parameters from myocardial perfusion imaging (MPI) using ECG‐gated single photon emission computed tomography (SPECT), and echocardiography in repeated imaging.

Methods

We studied the reproducibility of MPI diastolic function parameters peak‐filling rate (PFR) and time to peak filling (TTPF) as well as echocardiographic parameters E/A, E/e', and E‐wave deceleration time (DT). The study population consisted of 21 patients who underwent routine MPI with repeated rest acquisition and echocardiography. In a test–retest setting, appropriate diastolic measures were successfully obtained using SPECT in 20 patients, with E/A measured in 18, E/e' in 17, and DT in 16 patients.

Results

We found that PFR was well reproducible with the intraclass correlation coefficient (ICC) 0.887, and coefficient of variation (CV%) 10.5. However, TTPF was poorly reproducible (ICC 0.012, and CV% 17.5). E/A was highly reproducible (ICC 0.963, and CV% 12.5). Decent reproducibility was observed for E/e' (ICC 0.809, and CV% 18.6) and DT (ICC 0.833, and CV% 12.0). We further studied the correlation between these parameters. PFR (EDV/s) had negative correlation with DT (r = −0.538, p = 0.039) and E/A had positive with E/é (r = 0.689, p = 0.002). No other significant correlations were observed (p > 0.05 for all). We also examined how parameters classified patients as normal or abnormal regarding the diastolic function. E/A and E/e´, as well as E/A and PFR showed similar classifications in 88% of cases, with a Kappa value of 0.433, p = 0.074 for both.

Conclusions

PFR obtained from repeated SPECT studies, along with the E/A ratio, is highly reproducible.

Keywords: echocardiography, diastolic function, myocardial perfusion imaging, peak‐filling rate, reproducibility, single‐photon emission computed tomography

1. INTRODUCTION

Diastolic dysfunction is defined by the heart's inability to relax and fill the ventricle during diastole (Jeong and Dudley, 2015; Maack and Bhm, 2011). Heart failure (HF) with preserved ejection fraction (HFpEF) is a condition where the occurrence of diastolic dysfunction characterizes HF without a reduced ejection fraction (EF) (Jeong and Dudley, 2015). Approximately half of all diagnosed HF cases are HFpEF (Bhatia et al., 2006; Lam et al., 2011; Owan et al., 2006). The most important risk factors for diastolic dysfunction include hypertension, left ventricular (LV) hypertrophy, coronary artery disease (CAD), and aging (Nagueh et al., 2009; Xu et al., 2021). Asymptomatic mild LV diastolic dysfunction is found in 21% of the population, and asymptomatic moderate or severe diastolic dysfunction is found in 7% of the population (Jeong and Dudley, 2015). HFpEF typically occurs before the reduction of EF (Rosenberg and Manning, 2012).

Diastolic dysfunction is traditionally measured with echocardiography. The E/A is the ratio of peak velocity blood flow from LV relaxation in early diastole (the E wave) to peak velocity flow in late diastole caused by atrial contraction (the A wave). Normally, the E phase typically accounts for approximately two‐thirds of ventricular filling, with maximal inflow velocity occurring during this phase. The atrial contraction phase is relatively minor in filling under normal conditions. Therefore, a reduced E/A can indicate impaired relaxation (Mizunobu et al., 2013). E/A decreases with age (Carvalho et al., 2013) while E/e' increases (D'andrea et al., 2018). However, sometimes the E/A ratio can be normal (pseudonormalization) even though the underlying pathophysiology of diastolic dysfunction is present. The E/e' ratio is the ratio of the peak early filling velocity (E) to the peak velocity of the mitral valve annulus (e'), providing an estimate of LV filling pressures and helping assess diastolic function (Hillis et al., 2004; Ommen et al., 2000).

The diastolic function can also be measured from myocardial perfusion imaging (MPI) using the ECG‐gated single‐photon emission computed tomography (SPECT) (Mizunobu et al., 2013). The study is predominantly used to evaluate myocardial ischemia and scarring. MPI SPECT is accurate in assessing diastolic function, particularly in patients with normal EF and diastolic dysfunction (Mizunobu et al., 2013).

Transthoracic echocardiography has good reproducibility and is the standard noninvasive method of evaluating diastolic function (Bahrami et al., 2021; Frikha et al., 2015). Previous research using radionuclide angiography suggests that the diastolic parameter peak filling rate (PFR) has good and time to peak filling (TTPF) has poor reproducibility (Muntinga et al., 1997). Yet, to our knowledge, no studies have reported on the test–retest reproducibility of ECG‐gated MPI in measuring diastolic function parameters. Adequate reproducibility of the measured parameters is crucial for their reliable application in research and clinical settings. For this reason, we aimed to study the reproducibility of MPI SPECT parameters and compare their performance with echocardiography.

2. METHODS

2.1. Study Population

This study was approved by the Research Ethics Committee of Kuopio University Hospital, and all patients provided written informed consent. Twenty‐one randomly selected patients (13 women, eight men, mean age 66 years) who had undergone a standard MPI at Kuopio University Hospital were enrolled. The most common indication for MPI was the suspicion of CAD underlying the patient's symptoms (65% of the study cases). Twelve patients had hypertension, and nine had dyslipidemia. CAD was present in five patients, including two who had previously undergone coronary artery bypass grafting. Eight patients had type 2 diabetes, and one had type 1 diabetes. Asthma was reported in three patients, while two had a diagnosis of sleep apnea. One patient had hypertrophic cardiomyopathy, and another had severe renal disease. More detailed information on the study population is described in Table 1 and the previous study (Sillanmäki et al., 2016).

Table 1.

Study Population Characteristics.

Number (N) of study subjects Population mean (SD) Range Median
Age (years) N = 21 65.7 ± 9.7 43.1–79.9 67.3
Weight (kg) N = 21 84 ± 20 49–122 79
Height (cm) N = 21 166 ± 10 152–183 168
BMI (kg/m2) N = 21 30.6 ± 5.7 20.7–45.4 28.8
EDV (ml) N = 20 84 ± 37 39–164 80
ESV (ml) N = 20 27 ± 27 5–110 25
EF (%) N = 20 73 ± 15 31–92 73
PFR (EDV/s) N = 20 2.39 ± 0.72 0.98–3.94 2.42
PFR (ml/s) N = 20 223 ± 78 122–238 210
TTPF (ms) N = 20 171.3 ± 21.8 128.0–233.5 171.8
E/A N = 18 1.17 ± 0.75 0.60–3.95 0.93
E/e' N = 17 9.9 ± 4.0 4.4–19.3 8.9
DT N = 16 241 ± 67 160–384 234
Open in a new tab

Abbreviations: A, atrial diastolic filling; BMI, body mass index; DT, deceleration time of E wave; e’, peak velocity of mitral annulus; E, early diastolic filling; EDV, end‐diastolic volume; EDV/s, end‐diastolic volumes per second; EF, ejection fraction; ESV, end‐systolic volume; PFR, peak filling rate; SD, standard deviation; TTPF, time to peak filling.

2.2. Myocardial Perfusion Imaging

The patients underwent a 1‐day MPI protocol for stress and rest imaging. Patients underwent a pharmacological stress protocol to induce an increase in coronary blood flow (Verberne et al., 2015) using Regadenoson 400 µg (Rapiscan Pharma Solutions EU Ltd) administered intravenously, followed by 300 MBq Tc‐99m‐tetrofosmin radiotracer injection while riding a bicycle ergometer at 20‐60 W for 4 min. After 30 min from the injection, the patients were imaged with SPECT. For the rest study, patients received 700 MBq Tc‐99m‐tetrofosmin. 15 min before administering the second dose, patients received 5 mg of isosorbide nitrate p.o. to ensure the myocardial viability was not underestimated (Verberne et al., 2015). In the test‐rest setting, the first rest MPI SPECT was acquired 45 min after the radiotracer injection. After the first rest MPI SPECT acquisition, the patients stood up for a few minutes, and the second rest MPI SPECT was carried out right after to study the reproducibility (Sillanmäki et al., 2016).

A Siemens Symbia SPECT camera (Siemens Medical Solutions USA, Hoffman Estates, Illinois, USA) was used for image acquisition. The SPECT data were acquired with 16 cardiac frames and 25 beats per angle on average. If the patient's weight was over 100 kg, the acquisition time was extended to 30 beats per angle. The rest MPI sets were reconstructed iteratively with HybridRecon‐Cardiology (Hermes Medical Solution AB, Stockholm, Sweden) by an experienced physicist using side‐by‐side and blinded techniques. The imaging and reconstruction methods are described in more detail in our previous study (Sillanmäki et al., 2016).

LV function was analyzed retrospectively from MPI rest data using the QGS 2017 program (Cedars‐Sinai Medical Center, Los Angeles, California, USA). At the beginning of the rest MPI data analysis, the myocardium outlining was checked and modified manually, if necessary. In addition, manual masking was used if extracardiac activity hindered myocardial outlining. EF, end‐diastolic volume (EDV), and end‐systolic volume (ESV) were calculated from the volumes determined by the endocardial surface at different time points (Germano et al., 2007). The LV filling rate/time curve was computed from the first derivative of the volume/time curve. PFR was defined as the greatest filling rate in early diastole, corresponding to the peak value of the first derivative of the diastolic portion of the time–activity curve. The filling was considered in absolute terms (ml/s) and in relation to end‐diastolic volume (EDV/s). TTPF was assessed according to the interval between the end‐systole and the PFR.

2.3. Echocardiography

Standard transthoracic two‐dimensional echocardiography was performed for all patients. Repeated high‐quality measurements of E/A and E/e' were lacking for a couple of patients. For this reason, the reproducibility analysis was performed on E/A in 18 patients and on E/e' in 17 patients. Deceleration time (DT) in ms was measured from the peak of the E‐wave down its deceleration slope to the point where the trace meets the zero‐velocity baseline from 16 patients (in one case, DT was not analyzable). Echocardiography was performed using the Philips Epiq 7 Ultrasound System and processed using Qlab (version 10.1; Philips Healthcare, Andover, Massachusetts, USA). The ultrasound scanner was equipped with an X5–1 3D adult echo transducer (1–5 MHz). Image acquisition was performed in the lateral decubitus position. Echocardiography studies were repeated after a 5 min break after the first acquisition. The patients were asked to stand up or sit between the two sessions, depending on their clinical condition, and move their upper extremities. Subsequently, they were repositioned to the lateral decubitus position (Germano et al., 2007).

2.4. Statistical Analyses

Reproducibility and interobserver variability of PFR, TTPF (both analysed from rest MPI), E/A, E/e´ and DT were calculated using the intraclass correlation coefficient (ICC), concordance correlation coefficient (CCC), and coefficient of variation (CV%). ICC was calculated according to Koo et al. (Koo and Li, 2016), CCC according to Lin (Lin, 1989), and CV% according to Glüer et al. (Glüer et al., 1995). The absolute values of the differences between the two parameter measurements were calculated. The Shapiro‐Wilk W‐test was used to determine the normality of the parameters and possible explanatory variables, with p > 0.05 assumed as a normal distribution. PFR, TTPF, and E/e' were normally distributed in their original form, while E/A and DT were normally distributed after logarithmic correction. Pearson's and Spearman's correlations were used to study the correlations in normally distributed and non‐normally distributed parameters, respectively. T‐test and Mann‐Whitney's U‐test were used to evaluate the impact of gender on the reproducibility of the studied parameters. The first and second acquisition measurements were viewed at separate time points to increase the objectivity and independence of the results.

The prevalence of diastolic dysfunction was further studied. The normal range for E/A (0.6–1.6) was selected to match the mean age of the study population (Carvalho et al., 2013). Similarly, the normal upper limit for E/e´ (15.2) matched the mean age of the study population (D'andrea et al., 2018). The normal threshold for PFR was used >1.7 EDV/s and for TTPF < 208 ms (Verberne et al., 2015). Cohen's Kappa was used to quantify the level of agreement between two methods of classifying diastolic dysfunction. Kappa values were interpreted based on the following ranges: 0–0.2 indicates slight agreement, 0.2–0.4 reflects fair agreement, 0.4–0.6 signifies moderate agreement, 0.6–0.8 represents substantial agreement, and 0.8–1.0 denotes almost perfect agreement (Landis and Koch (1977)). Data analyses were performed with IBM SPSS Statistics (ver. 29.0.1.0). Data are presented as mean ± SD, or median and interquartile range if the variable was not normally distributed (Figure 1).

Figure 1.

Figure 1

Open in a new tab

View of the Single‐Photon Emission Computed Tomography and echocardiography output. An example of ECG‐gated myocardial perfusion imaging and echocardiography results. (A) Image shows myocardial perfusion of the left ventricle during the end‐diastole (ED) and end‐systole (ES) phases. The top three images are from the apex (top image) to the base of the heart. Following these are horizontal long‐axis (HLA) and vertical long‐axis (VLA) views of the left ventricle. On the right, the volume curve (red) and its first derivative (black) illustrate heart filling. The data is presented as EDV/s (end‐diastolic volume per second). The imaging uses 16 ECG‐gated SPECT, dividing the cardiac cycle into 16‐time bins on the X‐axis. (B) and (C) images are gained with ultrasound. The E‐wave represents early filling due to ventricular relaxation, and the A‐wave represents the atrial filling phase. The mitral annular velocity is recorded with tissue Doppler and represents the myocardial movement (e') during diastole.

3. RESULTS

The study population characteristics are shown in Table 1. One patient had atrial fibrillation (AF), two patients had left bundle branch block, one patient had right bundle branch block, and two patients had right bundle branch block with left anterior fascicular block. The average QRS duration was 112 ± 39 ms.

We found the PFR obtained from SPECT studies to be well reproducible: ICC 0.887 (95% CI 0.738‐0.953), CCC 0.88, and CV% 10.5. The TTPF from SPECT studies was poorly reproducible (Table 2). The E/A values calculated from the echocardiography measurements were well reproducible: ICC 0.963 (95% CI 0.869‐0.987) (p < 0.001), CCC 0.96, and CV% 12.5. The reproducibility of E/e' and DT was nearly as good (Table 2). For MPI diastolic measurements, the calculated difference in repeated studies did not correlate statistically significantly with age, body mass index (BMI), EF, or systolic blood pressure. For E/A, we found that the calculated difference in repeated studies correlated statistically significantly with EF (r = 0.502, p = 0.04). Sex did not have a statistically significant effect on the reproducibility of the studied parameters (data not shown).

Table 2.

Reproducibility of Parameters Acquired from SPECT and Echocardiography Studies.

Sarake1 Study 1 Mean ± SD Study 2 Mean ± SD ICC (95% CI) CCC CV% Difference between studies Mean ± SD (Abs.±SD) 95% limits of agreement for differences (Abs.)
PFR (EDV/s) (N = 20) 2.33 ± 0.77 2.44 ± 0.71 0.887* (0.738–0.953) 0.88 10.5 0.11 ± 0.34 (0.27 ± 0.23) −0.57 to 0.80 (−0.19 to 0.73)
PFR (ml/s) (N = 20) 214 ± 78.7 232 ± 81.7 0.852* (0.655–940) 0.85 13.6 17.6 ± 41.4 (30.3 ± 32.7) −65.2 to 100.4 (−35.1 to 95.7)
TTPF (ms) (N = 20) 168.9 ± 32.2 173.7 ± 29.2 0.012** (−0.447–0.452) 0.01 17.5 4.8 ± 43.2 (34.3 ± 25.5) −81.6 to 91.2 (−16.7 to 85.3)
E/A (N = 18) 1.10 ± 0.76 1.20 ± 0.74 0.963* (0.869–0.987) 0.96 12.5 0.11 ± 0.17 (0.13 ± 1.61) −0.24 to 0.47 (−0.19 to 0.46)
E/e'(N = 17) 9.7 ± 3.9 10.1 ± 4.5 0.809* (0.552–0.926) 0.80 18.6 0.4 ± 2.6 (1.86 ± 1.86) −4.9 to 5.7 (−1.86 to 5.60)
DT (ms) (N = 16) 235 ± 65 246 ± 75 0.833* (0.595–0.938) 0.82 12.0 11.1 ± 40.5 (29.9 ± 28.5) −69.9 to 92.1 (−27.1 to 87.0)
Open in a new tab

Abbreviations: DT, deceleration time of E wave; EDV/s, end‐diastolic volumes per second; ml/s, milliliters per second; PFR, peak filling rate; SPECT, single‐photon emission tomography; TTPF, time to peak filling; ms, milliseconds; E/A, ratio of peak blood flow velocity in early (E) and atrial (A) phases in echocardiography; E/e’, ratio of peak blood flow in early filling phase (E) and peak mitral annulus velocity (e´); SD, standard deviation; ICC, intraclass correlation coefficient; CI, confidence interval; CCC, concordance correlation coefficient; CV%, coefficient of variation; Abs, when mean value and standard deviation is calculated using absolute values of the difference between Study 1 and Study 2; 95% limits of agreement (Mean ± 2 SD) were calculated using both non‐absolute and absolute (in brackets) values of the difference between Study 1 and Study 2 results.

*

for p < 0.001 and

**

for p = 0.480.

The correlation between parameters estimating LV diastolic function was measured. PFR (EDV/s) had significant negative correlation with DT (r = −0.538, p = 0.039) and E/A had a significant positive correlation with E/é (r = 0.689, p = 0.002). No other significant correlations were observed among diastolic function parameters, including E/A, E/é, DT, PFR (EDV/s), and TTPF (r = −0.069‐0.205, p > 0.05 for all). To reduce methodological disparity, we recalculated the PFR as an absolute volumetric value in ml/s. After removing one clear outlier (Figure 3), the correlation between the echocardiographic E/A ratio and PFR (ml/s) was r = 0.463, p = 0.071; between E/é ratio and PFR (ml/s) was r = −0.121, p = 0.656, and between DT and PFR (ml/s) was r = −0.361, p = 0.186.

Figure 3.

Figure 3

Open in a new tab

Correlation between peak filling rate (PFR) expressed in ml/s and (a) the E/A ratio, (b) the E/e′ ratio, and (c) the deceleration time (DT). One outlier (indicated by a red star) was excluded from the analysis.

We further examined how these parameters classified patients as normal or abnormal regarding diastolic function (Table 3). E/A and E/e´, as well as E/A and PFR, showed similar classifications in 88% of cases, with a Kappa Value of 0.433, p = 0.074 for both (Table 3).

Table 3.

Classification of Parameters as Normal or Abnormal and Agreement between Methods.

Parameter Cut‐off for Normality Normal (n) Abnormal (n) Methods Similarly classified Kappa Value p‐value
E/A <0.6 and >1.6 16 2 E/A versus E/e´ 88% 0.433 0.074
E/e' >15.2 15 2 E/A versus PFR 88% 0.433 0.074
PFR <1.7 (EDV/s) 18 2 E/A versus TTPF 82% −0.085 0.707
TTPF >208 ms 19 1 E/e’ versus PFR 75% −0.143 0.568
E/e’ versus TTPF 81% −0.091 0.696
PFR versus TTPF 85% −0.071 0.732
Open in a new tab

E/A, the ratio of early (E) to late (A) diastolic filling velocities; E/e’, the ratio of early diastolic mitral inflow velocity (E) to early diastolic mitral annular velocity (e'); PFR, peak filling rate; TTPF, time to peak filling; EDV/s: end‐diastolic volume per second.

4. DISCUSSION

We found that PFR is well reproducible when assessed using the MPI SPECT, whereas TTPF showed poor reproducibility. The E/A was well reproducible and consistent with previous research findings (Bahrami et al., 2021; Frikha et al., 2015). The reproducibility of PFR and DT was almost the same level as E/A. We further studied the correlation between the diastolic ultrasound and SPECT parameters. PFR (EDV/s) correlated negatively with DT (r = −0.538, p = 0.039), and when expressed relative to EDV (PFR ml/s), its correlation with E/A nearly approached statistical significance (r = 0.463, p = 0.071). We also found that the PRF and E/A methods classified most cases (88%) similarly as normal or abnormal diastolic function. Based on the good ICC and CCC values, PFR could potentially serve as a useful method for research purposes at the population level. Figure 2

Figure 2.

Figure 2

Open in a new tab

Correlation between the two measurements of each diastolic parameter used to assess reproducibility. The graphs illustrate the correlations between the two measurements of each diastolic parameter, with normal ranges marked in light green. (a) shows the ratio of peak early diastolic filling velocity (E) to atrial diastolic filling velocity (A) (E/A) in two repeated measurements. Similarly, (b) displays the ratio of peak early diastolic filling velocity (E) to the peak velocity of the mitral annulus during early diastole (e') (E/e') in two measurements. The peak filling rate (PFR) measurements are presented in (c), while (d) presents the results of the time to peak filling (TTPF) measurements.

In this study, we found that CV% for PFR was 10.5%, while for TTPF, it was 17.5%. Part of the variability we observed arises from differences in imaging itself (test–retest variability), and some from intra‐observer variability. The impact of intraobserver variability (analysis‐reanalysis) has been studied previously with Tc‐99m sestamibi MPI SPECT in a few studies. These studies reported CV% values for TTPF ranging from 1% to 5%, and for PFR from 4% to 7% (Nakae et al., 2007; Yamamoto et al., 2007, 2008). This suggests that most of the variability detected can be explained by the test–retest part of the study. Both physiological and technical factors can also affect the reproducibility. We observed that if the filling curve becomes more flattened, the point at which TTPF is defined can shift significantly between measurements. This is not seen with PFR. In one case, software allocated both PFR and TTPF to the second filling phase, rather than positioning them in the first filling phase. As the software automatically detects these points and does not allow manual adjustment, this error cannot be corrected. These fluctuations in the filling curve can be caused by temporary alterations in heart rate, respiration, or blood pressure (Clements et al., 2000; Harada et al., 1995; Rajiah et al., 2023). These potential influences should be acknowledged when evaluating TTPF measurements.

We further studied the correlation between different parameters evaluating diastolic function. We found a good correlation among ultrasound parameters. There was also negative correlation with PFR (EDV/s) correlated negatively with DT (p = 0.039), while PFR (ml/s) showed a nearly significant (p = 0.071) positive correlation with E/A. The inverse relationship between PFR and DT is aligned with physiological expectations, as a higher early filling rate shortens transmitral flow deceleration time. The trend toward a positive association between PFR and E/A suggests shared physiological determinants, including LV compliance, relaxation, and left atrial–ventricular pressure gradients. Clinically, these findings suggest that PFR may help detect early alterations in LV filling dynamics. However, the correlation is neither exact nor uniform across all diastolic parameters. Several factors may explain this difference between methodologies. First, cardiac ultrasound, diastolic function parameters are considered to provide complementary rather than comprehensive information, with no single parameter fully capturing overall diastolic performance. Consequently, correlations between individual variables are not consistently strong. Secondly, the two modalities interrogate related but not identical physiological constructs, because Doppler measures instantaneous transmitral flow velocities, whereas SPECT calculates peak filling rate from count‐based time–volume curves averaged across several cycles. Third, echocardiography and SPECT were performed up to an hour apart, and changes in heart rate, rhythm, blood pressure, or patient posture during this interval could have influenced preload and, consequently, diastolic filling. Moreover, beat‐to‐beat fluctuations captured by Doppler are largely averaged out in SPECT, which derives its indices from multiple cardiac cycles. Finally, the study included only small sample of participants, so the analysis lacked power to detect higher correlations, and a single outlier (that was excluded from analysis) had a disproportionate effect on the coefficients (Figure 3). These observations indicate that SPECT‐derived diastolic indices should be considered complementary to, rather than interchangeable with, Doppler echocardiography and underscore the need for larger, contemporaneously acquired datasets to clarify the true extent of inter‐modality agreement. The methods classified cases as having normal or abnormal diastolic function with reasonably good agreement. One to two of our study subjects, depending on the technique used, were categorized as having abnormal diastolic dysfunction. The agreement analysis suggests that, although there is some concordance among the methods for classifying diastolic function, the level of agreement falls just below statistical significance.

The main limitation in our study is a relatively small study population, and larger study is needed to confirm these findings. Also, with small sample size, it is not possible to form subgroup analyses, in which reproducibility could be studied in different patient groups, or factors related to reproducibility could be more precisely supported. Additionally, some patients had conduction abnormalities, and one had AF, which could potentially influence the results. However, this reflects real‐world conditions in patient populations. The influence of age on diastolic parameters also presents a challenge when evaluating diastolic function. We accounted for this by using age‐dependent normal values. Yet, this factor does not affect the reproducibility of the test–retest results, which was the primary focus of our study. Most of our patients also had other underlying medical conditions, such as hypertension, dyslipidemia, type II diabetes mellitus, or CAD. As these conditions are associated with higher prevalences of HFpEF (Abudureyimu et al., 2022), the presence of these factors should be considered when evaluating the results of this study. However, we think this is a strength of our research since our study population corresponds well with the patients usually studied using MPI SPECT. Due to the significant radiation exposure associated with MPI SPECT, it was not feasible to include healthy individuals as controls in the study. Therefore, we relied on cutoff values established in previous research (Verberne et al., 2015).

MPI SPECT is a well‐established clinical tool, particularly valuable for evaluating CAD and myocardial scarring. Additionally, LV volumes and systolic function are typically assessed when interpreting MPI findings (Mizunobu et al., 2013). Our findings indicate that MPI SPECT diastolic measurement (PFR) demonstrates good reproducibility, and it might assist in identifying patients with diastolic dysfunction.

5. CONCLUSION

We conclude that PFR acquired from repeated MPI SPECT studies is well reproducible, and the reproducibility is almost the same level as that of the widely used echocardiography parameter E/A. This is clinically relevant since SPECT might be used to identify diastolic dysfunction, especially in patients with normal EF. This information could help guide better patient care, yet further investigation in larger population studies is needed.

AUTHOR CONTRIBUTIONS

Aaro Krohns (A.K.) Gathering of acquisition data, Statistical analyses, Planning, and writing of the original draft. Saara Sillanmäki (S.S.) Patient recruiting, echo measurements, Statistical analyses, writing ‐ review & editing, Interpretation of results, Investigation, Funding. Tomi Laitinen (T.P.L.): Study design, Statistical analyses, Writing ‐ review & editing, Investigation, Supervision, Funding. Tiina Laitinen (T.M.L): Reconstructions of MPI SPECT studies, review & editing. All authors read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ACKNOWLEDGEMENTS

This study was financially supported by the Finnish Foundation for Cardiovascular Research (10.11.2020 for T.P.L.). S.S. has achieved funding from the Research Committee of the Kuopio University Hospital Catchment Area for State Research Funding (5063586). Open access publishing facilitated by Ita‐Suomen yliopisto, as part of the Wiley ‐ FinELib agreement.

Krohns, A. , Laitinen, T.P. , Laitinen, T.M. & Sillanmäki, S. (2025) Reproducibility of diastolic function parameters in repeated ECG‐gated SPECT myocardial perfusion imaging and echocardiography. Clinical Physiology and Functional Imaging, 45, e70032. 10.1111/cpf.70032

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

REFERENCES

  1. Abudureyimu, M. , Luo, X. , Wang, X. , Sowers, J.R. , Wang, W. , Ge, J. et al. (2022) Heart Failure with Preserved Ejection Fraction (HFpEF) in Type 2 Diabetes Mellitus: from Pathophysiology to Therapeutics. Journal of Molecular Cell Biology, 14(5), mjac028. Available from: 10.1093/jmcb/mjac028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bahrami, H.S.Z. , Pedersen, F.H.G. , Myhr, K.A. , Møgelvang, R. & Hassager, C. (2021) Feasibility, Repeatability, and Reproducibility of Contemporary Diastolic Parameters and Classification. The International Journal of Cardiovascular Imaging, 37(3), 931–944. Available from: 10.1007/s10554-020-02069-z [DOI] [PubMed] [Google Scholar]
  3. Bhatia, R.S. , Tu, J.V. , Lee, D.S. , Austin, P.C. , Fang, J. , Haouzi, A. et al. (2006) Outcome of Heart Failure with Preserved Ejection Fraction in a Population‐Based Study. New England Journal of Medicine, 355(3), 260–269. Available from: 10.1056/nejmoa051530 [DOI] [PubMed] [Google Scholar]
  4. Carvalho, J.C. , Farand, P. , Do, H.D. , Brochu, M.C. , Bonenfant, F. & Lepage, S. (2013) Effect of Age and Sex on Echocardiographic Left Ventricular Diastolic Function Parameters in Patients with Preserved Ejection Fraction and Normal Valvular Function. Cardiology Journal, 20(5), 513–518. Available from: 10.5603/CJ.2013.0137 [DOI] [PubMed] [Google Scholar]
  5. Clements, I.P. , OLSON, L.J. , SCANLON, P.D. , GERTZ, M.A. & MULLANY, C.J. (2000) The Effect of Respiration on Left Ventricular Diastolic Filling as Assessed by Radionuclide Ventriculography. Nuclear Medicine Communications, 21, 55–63. Available from: 10.1097/00006231-200001000-00010 [DOI] [PubMed] [Google Scholar]
  6. D'andrea, A. , Vriz, O. , Ferrara, F. , Cocchia, R. , Conte, M. , M.D.iM.aio et al. (2018) Reference Ranges and Physiologic Variations of Left E/e' Ratio in Healthy Adults: Clinical and Echocardiographic Correlates. Journal of Cardiovascular Echography, 28(2), 101–108. Available from: 10.4103/jcecho.jcecho_57_17 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Frikha, Z. , Girerd, N. , Huttin, O. , Courand, P.Y. , Bozec, E. , Olivier, A. et al. (2015) Reproducibility in Echocardiographic Assessment of Diastolic Function in a Population Based Study (The STANISLAS Cohort Study). PLoS One, 10(4), e0122336. Available from: 10.1371/journal.pone.0122336 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Germano, G. et al. (2007) Quantitation in Gated Perfusion SPECT Imaging: The Cedars‐Sinai Approach. Journal of Nuclear Cardiology, 14(4), 433–454. Available from: 10.1016/j.nuclcard.2007.06.008 [DOI] [PubMed] [Google Scholar]
  9. Glüer, C.C. , Blake, G. , Lu, Y. , Blunt1, B.A. , Jergas1, M. & Genant1, H.K. (1995) Accurate Assessment of Precision Errors: How to Measure the Reproducibility of Bone Densitometry Techniques. Osteoporosis International, 5(4), 262–270. Available from: 10.1007/BF01774016 [DOI] [PubMed] [Google Scholar]
  10. Harada, K. , Takahashi, Y. , Shiota, T. , Suzuki, T. , Tamura, M. , Ito, T. et al. (1995) Effect of Heart Rate on Left Ventricular Diastolic Filling Patterns Assessed by Doppler Echocardiography in Normal Infants. American Journal of Cardiology, 76(8), 634–636. Available from: 10.1016/S0002-9149(99)80175-1 [DOI] [PubMed] [Google Scholar]
  11. Hillis, G.S. , Møller, J.E. , Pellikka, P.A. , Gersh, B.J. , Wright, R.S. , Ommen, S.R. et al. (2004) Noninvasive Estimation of Left Ventricular Filling Pressure by E/e′ Is a Powerful Predictor of Survival after Acute Myocardial Infarction. Journal of the American College of Cardiology, 43(3), 360–367. Available from: 10.1016/j.jacc.2003.07.044 [DOI] [PubMed] [Google Scholar]
  12. Jeong, E.M. & Dudley Jr., S.C. (2015) Diastolic Dysfunction: Potential New Diagnostics and Therapies'. Circulation Journal, 79(3), 470–477. Available from: 10.1253/circj.CJ-15-0064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Koo, T.K. & Li, M.Y. (2016) A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. Journal of Chiropractic Medicine, 15(2), 155–163. Available from: 10.1016/j.jcm.2016.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lam, C.S.P. , Donal, E. , Kraigher‐Krainer, E. & Vasan, R.S. (2011) Epidemiology and Clinical Course of Heart Failure With Preserved Ejection Fraction. European Journal of Heart Failure, 13, 18–28. Available from: 10.1093/eurjhf/hfq121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Landis, J.R. & Koch, G.G. (1977) The Measurement of Observer Agreement for Categorical Data. Biometrics, 33(1), 159. [PubMed] [Google Scholar]
  16. Lin, L.I.‐K. (1989) A Concordance Correlation Coefficient to Evaluate Reproducibility. Biometrics, 45(1), 255. Available from: 10.2307/2532051 [DOI] [PubMed] [Google Scholar]
  17. Maack, C. & Böhm, M. (2011) Targeting Mitochondrial Oxidative Stress in Heart Failure. Journal of the American College of Cardiology, 58, 83–86. Available from: 10.1016/j.jacc.2011.01.032 [DOI] [PubMed] [Google Scholar]
  18. Mizunobu, M. , Sakai, J. , Sasao, H. , Murai, H. & Fujiwara, H. (2013) Assessment of Left Ventricular Systolic and Diastolic Function Using ECG‐Gated technetium‐99m Tetrofosmin Myocardial Perfusion SPECT: Comparison With Ultrasound Echocardiography. International heart journal, 54(4), 212–215. Available from: 10.1536/ihj.54.212 [DOI] [PubMed] [Google Scholar]
  19. Muntinga, H.J. , van den Berg, F. , Knol, H.R. , Niemeyer, M.G. , Blanksma, P.K. , Louwes, H. et al. (1997) Normal Values and Reproducibility of Left Ventricular Filling Parameters by Radionuclide Angiography. International Journal of Cardiac Imaging, 13(2), 165–171. Available from: 10.1023/A:1005704415207 [DOI] [PubMed] [Google Scholar]
  20. Nagueh, S.F. , Appleton, C.P. , Gillebert, T.C. , Marino, P.N. , Oh, J.K. , Smiseth, O.A. et al. (2009) Recommendations for the Evaluation of Left Ventricular Diastolic Function by Echocardiography. Journal of the American Society of Echocardiography, 22(2), 107–133. Available from: 10.1016/j.echo.2008.11.023 [DOI] [PubMed] [Google Scholar]
  21. Nakae, I. , Matsuo, S. , Tsutamoto, T. , Matsumoto, T. , Mitsunami, K. & Horie, M. (2007) Assessment of Cardiac Function in Patients with Heart Disease by Quantitative Gated Myocardial Perfusion SPECT. Annals of Nuclear Medicine, 21(6), 315–323. Available from: 10.1007/s12149-007-0032-0 [DOI] [PubMed] [Google Scholar]
  22. Ommen, S.R. , Nishimura, R.A. , Appleton, C.P. , Miller, F.A. , Oh, J.K. , Redfield, M.M. et al. (2000) Clinical Utility of Doppler Echocardiography and Tissue Doppler Imaging in the Estimation of Left Ventricular Filling Pressures: A Comparative Simultaneous Doppler‐Catheterization Study. Circulation, 102(15), 1788–1794. Available from: 10.1161/01.CIR.102.15.1788 [DOI] [PubMed] [Google Scholar]
  23. Owan, T.E. , Hodge, D.O. , Herges, R.M. , Jacobsen, S.J. , Roger, V.L. & Redfield, M.M. (2006) Trends in Prevalence and Outcome of Heart Failure with Preserved Ejection Fraction. New England Journal of Medicine, 355(3), 251–259. Available from: 10.1056/nejmoa052256 [DOI] [PubMed] [Google Scholar]
  24. Rajiah, P.S. , Moore, A. , Broncano, J. , Anand, V. , Kolluri, N. , Shah, D.J. et al. (2023) Diastology with Cardiac MRI: A Practical Guide. Radiographics, 43(9), 1–22. Available from: 10.1148/rg.220144 [DOI] [PubMed] [Google Scholar]
  25. Rosenberg, M.A. & Manning, W.J. (2012) Diastolic Dysfunction and Risk of Atrial Fibrillation: A Mechanistic Appraisal. Circulation, 126(19), 2353–2362. Available from: 10.1161/CIRCULATIONAHA.112.113233 [DOI] [PubMed] [Google Scholar]
  26. Sillanmäki, S. , Koivumäki, T. , Laitinen, T.M. , Hakulinen, M. , Hedman, M. & Laitinen, T. (2016) BMI Influence on the Reproducibility of e CG‐Gated Myocardial Perfusion Imaging Phase Analysis in Comparison with Novel Echocardiographic Dyssynchrony Estimation Methods. Nuclear Medicine Communications, 37(7), 767–774. Available from: 10.1097/MNM.0000000000000508 [DOI] [PubMed] [Google Scholar]
  27. Verberne, H.J. , Acampa, W. , Anagnostopoulos, C. , Ballinger, J. , Bengel, F. , De Bondt, P. et al. (2015) EANM Procedural Guidelines for Radionuclide Myocardial Perfusion Imaging with SPECT and SPECT/CT: 2015 Revision. European Journal of Nuclear Medicine and Molecular Imaging, 42(12), 1929–1940. Available from: 10.1007/s00259-015-3139-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Xu, B. , Liu, L. , Abdu, F.A. , Yin, G. , Mohammed, A.Q. , Xu, S. et al. (2021) Prognostic Value of Diastolic Dysfunction Derived From D‐SPECT in Coronary Artery Disease Patients With Normal Ejection Fraction. Frontiers in Cardiovascular Medicine, 8, 700027. Available from: 10.3389/fcvm.2021.700027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Yamamoto, A. , Takahashi, N. , Ishikawa, M. , Abe, K. , Kobayashi, Y. , Tamai, J. et al. (2008) Relationship between Left Ventricular Function and Wall Motion Synchrony in Heart Failure Assessed by ECG‐Gated Myocardial Perfusion SPECT. Annals of Nuclear Medicine, 22(9), 751–759. Available from: 10.1007/s12149-008-0172-x [DOI] [PubMed] [Google Scholar]
  30. Yamamoto, A. , Takahashi, N. , Munakata, K. , Hosoya, T. , Shiiba, M. , Okuyama, T. et al. (2007) Global and Regional Evaluation of Systolic and Diastolic Left Ventricular Temporal Parameters Using a Novel Program for ECG‐Gated Myocardial Perfusion SPECT ‐ Validation by Comparison with Gated Equilibrium Radionuclide Angiography and Speckle‐Tracking Radial Strain from Echocardiography. Annals of Nuclear Medicine, 21(2), 115–121. Available from: 10.1007/BF03033989 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Articles from Clinical Physiology and Functional Imaging are provided here courtesy of Wiley

RESOURCES