Advertisement

Reproducibility of 4D cardiac computed tomography feature tracking myocardial strain and comparison against speckle-tracking echocardiography in patients with severe aortic stenosis

Open AccessPublished:January 27, 2022DOI:https://doi.org/10.1016/j.jcct.2022.01.003

      Abstract

      Background

      Myocardial strain is an established parameter for the assessment of cardiac function and routinely derived from speckle tracking echocardiography (STE). Novel post-processing tools allow deformation imaging also by 4D cardiac computed tomography angiography (CCT). This retrospective study aims to analyze the reproducibility of CCT strain and compare it to that of STE.

      Methods

      Left (LV) and right ventricular (RV), and left atrial (LA) ejection fraction (EF), dimensions, global longitudinal (GLS), circumferential (GCS) and radial strain (GRS) were determined by STE and CCT feature tracking in consecutive patients with severe aortic stenosis evaluated for transcatheter aortic valve implantation.

      Results

      106 patients (mean age 79.9 ​± ​7.8, 44.3% females) underwent CCT at a median of 3 days (IQR 0–28 days) after STE. In CCT, strain measures showed good to excellent reproducibility (intra- and inter-reader intraclass correlation coefficient ≥0.75) consistently in the LV, RV and LA. In STE, only LV GLS and LA GLS yielded good reproducibility, whereas LV GCS and LV GRS showed moderate, and RV GLS and free wall longitudinal strain (FWLS) poor reproducibility. Agreement between CCT and STE was strong for LV GLS only, while other strain features displayed moderate (LV GCS, LA GLS) or weak (LV GRS, RV GLS and FWLS) inter-modality correlation.

      Conclusion

      LV, RV and LA CCT strain assessments were highly reproducible. While a strong agreement to STE was found for LV GLS, inter-modality correlation was moderate or weak for LV GCS, LV GRS, and RV and LA longitudinal strain, possibly related to poor reproducibility of STE measurements.

      Graphical abstract

      Keywords

      Abbreviations

      CCT
      4-dimensional cardiac computed tomography
      CMR
      cardiac magnet resonance
      CV
      chamber view
      EDA/EDV
      end diastolic area/volume
      EF
      ejection fraction
      ESV
      end systolic volume
      FAC
      fractional area change
      FT
      feature tracking
      FWLS
      free wall longitudinal strain
      GCS
      global circumferential strain
      GLS
      global longitudinal strain
      GRS
      global radial strain
      ICC
      intraclass correlation coefficient
      LA
      left atrium
      LV/RV
      left/right ventricle
      SAX
      short axis
      STE
      speckle tracking echocardiography
      TAVI
      transcatheter aortic valve implantation

      1. Introduction

      Endo- and myocardial strain are established determinants in the evaluation of myocardial contractility,
      • Voigt J.U.
      • Pedrizzetti G.
      • Lysyansky P.
      • et al.
      Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging.
      describing the degree of deformation of cardiac walls within the cardiac cycle. Deformation of a cardiac cavity can occur along its long axis (global longitudinal strain – GLS), its perimeter (global circumferential strain – GCS), or by thickening of its wall (global radial strain – GRS).
      • Geyer H.
      • Caracciolo G.
      • Abe H.
      • et al.
      Assessment of myocardial mechanics using speckle tracking echocardiography: fundamentals and clinical applications.
      A large body of evidence supports the high diagnostic and prognostic value of left ventricular (LV) GLS,
      • Kalam K.
      • Otahal P.
      • Marwick T.H.
      Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction.
      while results for LV GCS and GRS are inconclusive.
      • Shen M.T.
      • Yang Z.G.
      • Diao K.Y.
      • et al.
      Left ventricular involvement in arrhythmogenic right ventricular dysplasia/cardiomyopathy predicts adverse clinical outcomes: a cardiovascular magnetic resonance feature tracking study.
      • Florescu M.
      • Magda L.S.
      • Enescu O.A.
      • Jinga D.
      • Vinereanu D.
      Early detection of epirubicin-induced cardiotoxicity in patients with breast cancer.
      • Cho G.Y.
      • Marwick T.H.
      • Kim H.S.
      • Kim M.K.
      • Hong K.S.
      • Oh D.J.
      Global 2-dimensional strain as a new prognosticator in patients with heart failure.
      GLS mainly mirrors the contractility of subendocardial longitudinal fibers that often display impaired function in the early phase of myocardial injury, when LV ejection fraction (LVEF) can still be maintained.
      • Amzulescu M.S.
      • De Craene M.
      • Langet H.
      • et al.
      Myocardial strain imaging: review of general principles, validation, and sources of discrepancies.
      LV GLS provides incremental value beyond LVEF in the prediction of adverse events in patients with chronic heart failure,
      • Sengeløv M.
      • Jørgensen P.G.
      • Jensen J.S.
      • et al.
      Global longitudinal strain is a superior predictor of all-cause mortality in heart failure with reduced ejection fraction.
      ,
      • Park J.J.
      • Park J.B.
      • Park J.H.
      • Cho G.Y.
      Global longitudinal strain to predict mortality in patients with acute heart failure.
      ischemic heart disease,
      • Lai W.
      • Jie H.
      • Jian-Xun D.
      • et al.
      Impact of concomitant impairments of the left and right ventricular myocardial strain on the prognoses of patients with ST-elevation myocardial infarction.
      myocarditis,
      • Fischer K.
      • Obrist S.J.
      • Erne S.A.
      • et al.
      Feature tracking myocardial strain incrementally improves prognostication in myocarditis beyond traditional CMR imaging features.
      valvular heart disease,
      • Schmid J.
      • Kamml C.
      • Zweiker D.
      • et al.
      Cardiac magnetic resonance imaging right ventricular longitudinal strain predicts mortality in patients undergoing TAVI.
      • Canessa M.
      • Thamman R.
      • Americo C.
      • Soca G.
      • Dayan V.
      Global longitudinal strain predicts survival and left ventricular function after mitral valve surgery: a meta-analysis.
      • Alashi A.
      • Khullar T.
      • Mentias A.
      • et al.
      Long-term outcomes after aortic valve surgery in patients with asymptomatic chronic aortic regurgitation and preserved LVEF: impact of baseline and follow-up global longitudinal strain.
      • Ng A.C.T.
      • Prihadi E.A.
      • Antoni M.L.
      • et al.
      Left ventricular global longitudinal strain is predictive of all-cause mortality independent of aortic stenosis severity and ejection fraction.
      and several other cardiac diseases.
      • Shen M.T.
      • Yang Z.G.
      • Diao K.Y.
      • et al.
      Left ventricular involvement in arrhythmogenic right ventricular dysplasia/cardiomyopathy predicts adverse clinical outcomes: a cardiovascular magnetic resonance feature tracking study.
      ,
      • Rowin E.J.
      • Maron B.J.
      • Wells S.
      • et al.
      Usefulness of global longitudinal strain to predict heart failure progression in patients with nonobstructive hypertrophic cardiomyopathy.
      • Lee Chuy K.
      • Drill E.
      • Yang J.C.
      • et al.
      Incremental value of global longitudinal strain for predicting survival in patients with advanced AL amyloidosis.
      • Dons M.
      • Jensen J.S.
      • Olsen F.J.
      • et al.
      Global longitudinal strain corrected by RR-interval is a superior echocardiographic predictor of outcome in patients with atrial fibrillation.
      Speckle-tracking echocardiography (STE) uses unique interference patterns (“speckles”) that are followed across the cardiac cycle to obtain the degree of deformation and derive strain. Speckles result from the naturally occurring inhomogeneous reflection of acoustic signals in the myocardium due to the different orientation of muscle fibers and are displayed in varying grayscales in B-mode echocardiography. More recently, the introduction of novel post-processing tools enable feature tracking (FT) based strain evaluation also by 4D cardiac computed tomography angiography (CCT). FT is an algorithm-based approach, identifying and tracking characteristic features of endo- and epicardial borders.
      • Hor K.N.
      • Baumann R.
      • Pedrizzetti G.
      • et al.
      Magnetic resonance derived myocardial strain assessment using feature tracking.
      In addition to image characteristics (brightness, greyscale), also anatomic landmarks are taken into account to assess strain.
      • Kempny A.
      • Fernández-Jiménez R.
      • Orwat S.
      • et al.
      Quantification of biventricular myocardial function using cardiac magnetic resonance feature tracking, endocardial border delineation and echocardiographic speckle tracking in patients with repaired tetralogy of fallot and healthy controls.
      Broad evidence that supports the clinical value of this promising tool derived from CCT in a real-world setting is widely lacking. Although CCT involves ionized radiation, it might be beneficial in selected patients and can overcome limitations of STE such as poor acoustic window and foreshortening of the ventricles due to suboptimal probe position dictated by the patient's anatomy. Moreover, 4D CCT acquisition protocols are recommended standard in several clinical settings, such as planning of transcathether aortic valve implantation (TAVI)
      • Otto C.M.
      • Nishimura R.A.
      • Bonow R.O.
      • et al.
      2020 ACC/AHA guideline for the management of patients with valvular heart disease: a report of the American college of cardiology/American heart association joint committee on clinical practice guidelines.
      and extracting all available information from these clinically indicated CCT examinations is warranted. The present study aims to determine the reproducibility of CCT strain compared to STE in the left (LV) and right ventricle (RV), and the left atrium (LA).

      2. Methods

      2.1 Study cohort

      Between October 2019 and March 2021, consecutive patients with severe aortic stenosis evaluated for TAVI at Bern University Hospital, Switzerland were enrolled in an institutional registry, which is part of the SwissTAVI registry (NCT01368250). Inclusion criteria for the present study were the conduction of a complete CCT scan (0–100% of RR interval with 5% increments reconstructions and 20 phases) and transthoracic echocardiography at our institution prior TAVI. Patients unable to provide written consent were excluded. The study was approved by the local ethics committee and was conducted in accordance with the Declaration of Helsinki.

      2.2 Image acquisition and measurements

      Clinically indicated retrospectively ECG-gated CCT imaging was performed using a dual-source 128-row multisclice CT (Somaton Definition Flash; Siemens Healthcare, Erlangen, Germany) as previously described.
      • Okuno T.
      • Corpataux N.
      • Spano G.
      • et al.
      True-severe stenosis in paradoxical low-flow low-gradient aortic stenosis: outcomes after transcatheter aortic valve replacement.
      Scan parameters were as following: reference tube voltage was set to 100–120 kv and reference tube-current-time product 300 mAsref according to the body weight; rotation time 0.28 ​s; slice collimation 128 ​× ​0.6 ​mm; pitch value 0.17 for spiral acquisition 0–100% of the RR interval. Automatic current modulation (CareDose4D) was used for raw data acquisition. Each patient received an intravenous injection of 40–120 ​ml of contrast medium at a flow rate of 4–5 ​ml/s depending on body-weight. Image acquisition was performed during an inspiratory breath-hold in a cranio-caudal direction and images were reconstructed in 1 ​mm increment using an I30f kernel (SAFIRE, strength 3). CCT datasets in 5% increments were reconstructed throughout the entire cardiac cycle, resulting in 20 reconstructions per scan (0–100%). STE was performed during clinical routine on machines of different vendors by investigators with varying experience following a standardized protocol
      • Mitchell C.
      • Rahko P.S.
      • Blauwet L.A.
      • et al.
      Guidelines for performing a comprehensive transthoracic echocardiographic examination in adults: recommendations from the American society of echocardiography.
      that, among others, included a short axis stack (SAX) of the LV, overview as well as LV- and RV focused depictions in the apical 4-chamber view (CV), LA-focused apical 2-CV and LV-focused apical 2- and 3-CV.
      Both CCT and STE images underwent standardized analysis on dedicated workstations by investigators blinded to the clinical information and findings of the other imaging modality (i.e. authors HG for CCT, JZ for STE, and BB for reproducibility analysis). Echocardiographic images were analyzed by speckle-tracking using TOMTEC Arena 2D cardiac performance analysis, (TOMTEC Imaging Systems GmBH, Unterschleissheim, Germany), whereas Medis Suite v.3.0, (Medis Medical Imaging, Leiden, The Netherlands) was used for segmentation (Medis Suite 3D Viewer) and strain-analysis (Medis Suite QStrain) of CCT images. Both applications use semi-automatic tools to trace endo- and epicardial borders in each view in endsystole and enddiastole. Endo- and epicardial tracings were manually checked for accuracy in each frame throughout the entire cardiac cycle and manually revised by the investigator if required. Prior to analysis, readers rated the quality of the images they considered most suitable for strain analysis on a scale from 1 to 4.1 – poor (e.g. severe arrhythmia or breathing artefact), 2 – moderate (e.g. challenging blood pool-endocardium delineation in all frames), 3 – good (e.g. lack of optimal epicardium soft-tissue delineation in parts in selected frames), 4 – excellent (no limitations). LV GLS was averaged from apical 2-, 3- and 4-CV after tracing of endocardial borders with exclusion of papillary muscles (Fig. 1). For CCT images, the reader previously reformatted 2-, 3- and 4-CV and a SAX stack. LV GCS and GRS were assessed in 3 SAX slices (at the height of the mitral valve, the papillary muscles and the apex). RV and LA strain and dimensions were determined monoplane in the 4- and 2-CV, respectively. For the LA contours the pulmonary veins and the left atrial appendage were excluded (Fig. 1). LA GLS (equivalent to LA reservoir strain) was defined as the peak of the LA strain curve compared to ventricular enddiastole as baseline. LV dimensions and LVEF were calculated by Simpson's biplane method in both STE and CCT.
      Fig. 1
      Fig. 1Global longitudinal strain (GLS) in endsystolic frames of a patient with normal (patient 1) and highly impaired (patient 2) left- and right ventricular and left atrial function. Positive and negative strain values are denoted in red and blue color, respectively. Dark and bright colors represents a low and a high degree of deformation, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
      STE and CCT datasets of 15 randomly selected patients underwent additional analysis, (including multiplanar reconstruction for CCT) by the same and an additional reader blinded to previous findings in order to assess reproducibility.

      2.3 Statistical analysis

      Statistical analysis was conducted with IBM SPSS Statistics 25 (IBM Corp., Armon, New York, USA). Results are presented as mean ​± ​standard deviation or as frequencies and percentage if appropriate. Non-normally distributed variables were provided as median and interquartile range (IQR). Reproducibility of measures was evaluated by intraclass correlation coefficient (ICC), determined with a two-way random effects model for intra- and interreader variability in 15 patients. Intraclass correlation was interpreted according to Koo et al. (ICC ≥0.9 – excellent, ICC ​= ​0.75–0.89 – good, ICC ​= ​0.5–0.74 – moderate, ICC<0.5 – poor).
      • Koo T.K.
      • Li M.Y.
      A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
      Agreement between CCT and STE was visualized in Bland-Altman plots with determination of the mean bias between modalities. Upper and lower limits of agreement (LOA) were defined as mean bias ​± ​1.96 x standard deviation. Bivariate Pearson correlation was used to describe correlation between CCT and STE measurements and interpreted according to Schober et al. (r ​≥ ​0.9 – very strong correlation, r ​= ​0.7–0.89 – strong, 0.4–0.7 –moderate, r ​= ​0.1–0.39 – weak).
      • Schober P.
      • Boer C.
      • Schwarte L.A.
      Correlation coefficients: appropriate use and interpretation.
      Mean relative disagreement was calculated as ratio between the mean of both measurements (CCT and STE) and the differences between both measurements. Mean disagreement across groups stratified by image quality and time delay between STE and CCT were compared by unpaired t-tests or by Mann-Whitney U test.

      3. Results

      Among 310 consecutive patients with severe aortic stenosis referred for the evaluation of TAVI, 106 patients were included in the present study (Fig. 2) of whom 91 (86%) underwent TAVI after CCT and STE imaging. Another n ​= ​7 patients were referred to surgical aortic valve replacement and n ​= ​8 patients were not deemed eligible for TAVI. Mean age was 79.9 ​± ​7.8 years and 59 (56%) were male (Table 1). Median time delay between CCT and STE was +3 days (IQR 0–28 days). For the complete CCT including CT angiography of the thorax and potential TAVI vascular access sites, patients received 85.5 ​± ​14.3 ​ml (median 90 ​ml) contrast agent and the median total dose-length product was 957 (IQR 680 to 1179) mGycm. Image quality (IQ) was rated as excellent in 66 (62.3%) of patients (34% good, 2.8% moderate and 0.9% poor) in CCT and in 10 (9.4%) patients (38.7% good, 36.8% moderate, 15.1% poor) in STE. STE apical views were available in 104 (98%), SAX in 88 (83%), RV in 94 (88%) and LA in 104 (98%) of patients. STE loops were assessed with 51.3 ​± ​7.9 frames per second and a mean heart rate of 73.2 ​± ​13.4 bpm. Mean heart rate in CCT was 69.8 ​± ​13.7 bpm, resulting in a mean temporal resolution of 23.6 ​± ​4.9 frames per second. No patient was excluded due to poor image quality.
      Fig. 2
      Fig. 2Study consort flow chart. Abbreviations: CCT – 4D cardiac computed tomography, TAVI – transcatheter aortic valve implantation.
      Table 1Patient characteristics.
      Total (n ​= ​106)
      Age (years)79.9 ​± ​7.8
      Gender (female)47 (44%)
      Body-Surface-Area (m2)1.87 ​± ​0.25
      Body-Mass-index (kg/m2)27.5 ​± ​5.9
      NYHA class II or III100 (94%)
      STS PROM score4.8 ​± ​4.46
      History of CAD58 (54.7%)
      History of prior myocardial infarction13 (12.3%)
      Arterial hypertension95 (89.6%)
      Atrial fibrillation27 (25.5%)
      Diabetes mellitus24 (22.6%)
      Permanent pacemaker8 (7.5%)
      Implantable Cardioverter Defibrillator4 (3.8%)
      LV mass (g)224.7 ​± ​72
      LV mass index (g/m2)120.5 ​± ​36.8
      LVEDV (ml)114 ​± ​54
      LVEF (%)51.6 ​± ​13.6
      SV (ml)55.1 ​± ​19.2
      SV index29.6 ​± ​9.7
      AVA (cm2)0.87 ​± ​0.3
      AV mean gradient (mmHg)31.7 ​± ​15.7
      AV peak gradient (mmHg)52.3 ​± ​24
      LV end-diastolic pressure (mmHg)14.4 ​± ​12.3
      Hemoglobin (g/l)12.2 ​± ​1.9
      Creatinine (μmol/l)122.8 ​± ​62.2
      Abbreviations: AVA – aortic valve area, CAD – coronary artery disease, EF – ejection fraction, LV – Left ventricular, NYHA – New York Heart Association, STS PROM-predicted risk of mortality according to the Society of Thoracic Surgeons Risk Score, SV - stroke volume.

      3.1 Reproducibility

      In CCT, all measurements showed good (ICC >0.75–0.9) to excellent (ICC >0.9) intra- and interreader variability (Table 2 and Graphic Abstract). All CCT derived LV apical measurements (LV EDV, LV EF, LV GLS), and LA EDV as well as LA EF showed both excellent intra- and interreader variability. In STE, only LA ESV yielded excellent reproducibility, while LV apical measurements (LV EDV, LV EF, LV GLS) and LA GLS were still well reproducible. However, short axis (LV GCS and LV GRS) and RV measurements were of moderate or poor robustness. RV GLS and RV free wall longitudinal strain (FWLS) were not significantly correlated across readers in STE.
      Table 2Reproducibility (intraclass correlation coefficient (ICC) and 95% CI) of CCT and STE measurements.
      Intraclass correlation coefficients were interpreted according to Koo et al.
      • Koo T.K.
      • Li M.Y.
      A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
      Abbreviations: CCT – Cardiac computed tomography, EDV/A – end diastolic volume/area, EF – ejection fraction, ESV/A – end systolic volume/area, FAC – fractional area change, FWLS – free wall longitudinal strain, GCS – global circumferential strain, GLS – global longitudinal strain, GRS – global radial strain, ICC – intraclass correlation coefficients, STE – speckle tracking echocardiography.

      3.2 Agreement of CCT and STE

      We observed relevant discrepancies between most CCT and STE measurements (for absolute values see Supplemental Table S1). Only LV apical measurements (LV EDV, LV EF, and LV GLS) showed good agreement and strong correlation, but also considerable limits of agreement (LOA) (Fig. 3). Mean LV GLS was −13.2 ​± ​5.3% in CCT and −16.0 ​± ​5.4% in STE, resulting in a mean inter-modality bias of 2.79% (LOA 9.1 to −3.6%). LVEF was 48.8 ​± ​17% in CCT and 51.6 ​± ​13.6% in STE (mean bias −2.88%; LOA -25.2 to 19.5%). CCT slightly underestimated both variables (less negative values for LV GLS, lower values for LVEF), while LV EDV was consistently overestimated (mean bias ​+ ​68.7 ​ml (LOA -11.5 to 149.9 ​ml) in CCT compared to STE. SAX measurements (LV GRS and LV GCS) showed wide LOAs and moderate (GCS) or poor (GRS) correlation between both modalities. Consistent with the poor reproducibility in STE, STE RV measurements did not correlate to them of CCT, except for moderate correlation of RV EDA (Fig. 4). LA GLS, as well as LA ESV and LA EF moderately correlated between modalities but exhibited wide LOAs (Fig. 5).
      Fig. 3
      Fig. 3Agreement and correlation of CCT and STE measurements in the LV. Correlation coefficients were interpreted according to Schober et al.
      • Schober P.
      • Boer C.
      • Schwarte L.A.
      Correlation coefficients: appropriate use and interpretation.
      Abbreviations: CCT – 4D cardiac computed tomography, EDV – end diastolic volume, EF – ejection fraction, GCS – global circumferential strain, GLS – global longitudinal strain, GRS – global radial strain, LV – left ventricle, SD – standard deviation, STE – speckle tracking echocardiography.
      Fig. 4
      Fig. 4Agreement and correlation of CCT and STE measurements in the RV. Correlation coefficients were interpreted according to Schober et al.
      • Schober P.
      • Boer C.
      • Schwarte L.A.
      Correlation coefficients: appropriate use and interpretation.
      Abbreviations: CCT – 4D cardiac computed tomography, EDA – end diastolic area, FAC – fractional area change, FWLS – free wall longitudinal strain, GLS – global longitudinal strain, RV – right ventricle, SD – standard deviation, STE – speckle tracking echocardiography.
      Fig. 5
      Fig. 5Agreement and correlation of CCT and STE measurements in the LA. Correlation coefficients were interpreted according to Schober et al.
      • Schober P.
      • Boer C.
      • Schwarte L.A.
      Correlation coefficients: appropriate use and interpretation.
      Abbreviations: CCT – 4D cardiac computed tomography, EDV – end diastolic volume, EF – ejection fraction, GLS – global longitudinal strain, LA – left atrium, SD – standard deviation, STE – speckle tracking echocardiography.

      3.3 Impact of image quality

      To investigate whether poor agreement was related to image quality (IQ), we determined the mean relative deviation (“mean relative error”) from the average value of both modalities to allow for a comparison of measurements with different units (Fig. 6). Lowest mean relative deviation was found for LV EF (20.5%) and LV GLS (27.1%) and highest for LV GRS (84.8%) and RV GLS (71.2%). Excluding patients with moderate or poor image quality in STE was associated with a significant reduction in the mean relative deviation of LV GLS (p ​= ​0.003), LV GRS (p ​= ​0.01), RV EDA (p ​< ​0.001), RV FWLS (p ​= ​0.026), LA ESV (p ​= ​0.001), and LA EF (p ​= ​0.008) (Supplemental Table S1). No improvement was seen after exclusion of patients with poor or moderate IQ (n ​= ​4) in CCT and a time delay between STE and CCT >1 day (n ​= ​70).
      Fig. 6
      Fig. 6Spiderplotted relative mean deviation between CCT and STE measurements in different subgroups. Abbreviations: CCT – 4D cardiac computed tomography, EDV/A – end diastolic volume/area, EF – ejection fraction, ESV/A – end systolic volume/area, FAC – fractional area change, FWLS – free wall longitudinal strain, GCS – global circumferential strain, GLS – global longitudinal strain, GRS – global radial strain.

      4. Discussion

      In the present study, we compared myocardial strain in CCT against STE in patients with severe aortic stenosis. The findings can be summarized as follows: a) 4D CCT provides images of high quality that allow assessment not only of LV and LA, but also of RV feature tracking strain with high reproducibility; b) using STE, only LV GLS and LA GLS were reproducible and c) agreement of CCT to STE was strong for LV GLS only, but can be increased if STE images of poor quality are excluded.
      Lower reproducibility of STE derived LV GCS and GRS in comparison to GLS was already reported by other studies.
      • Sugimoto T.
      • Dulgheru R.
      • Bernard A.
      • et al.
      Echocardiographic reference ranges for normal left ventricular 2D strain: results from the EACVI NORRE study.
      • Chan J.
      • Shiino K.
      • Obonyo N.G.
      • et al.
      Left ventricular global strain analysis by two-dimensional speckle-tracking echocardiography: the learning curve.
      • Buss S.J.
      • Schulz F.
      • Mereles D.
      • et al.
      Quantitative analysis of left ventricular strain using cardiac computed tomography.
      In addition to poor image quality in the short axis view – particularly in the apical segments, also lower lateral resolution of echocardiography can contribute to poor reproducibility.
      • Pedrizzetti G.
      • Claus P.
      • Kilner P.J.
      • Nagel E.
      Principles of cardiovascular magnetic resonance feature tracking and echocardiographic speckle tracking for informed clinical use.
      2D STE requires the speckles to move in parallel direction to the ultrasound lines on that specific image plane which enables a better tracking in the apical views compared to the SAX.
      • Langeland S.
      • Wouters P.F.
      • Claus P.
      • et al.
      Experimental assessment of a new research tool for the estimation of two-dimensional myocardial strain.
      Moreover, the echocardiographic acquisition of an appropriate SAX and the post-processing for GCS and GRS assessments requires higher expertise, if compared to GLS derived from the long axis views.
      • Chan J.
      • Shiino K.
      • Obonyo N.G.
      • et al.
      Left ventricular global strain analysis by two-dimensional speckle-tracking echocardiography: the learning curve.
      Nevertheless, other studies reported higher reproducibility of STE LV strain acquisition compared to the results in our cohort.
      • Sugimoto T.
      • Dulgheru R.
      • Bernard A.
      • et al.
      Echocardiographic reference ranges for normal left ventricular 2D strain: results from the EACVI NORRE study.
      ,
      • Luis S.A.
      • Yamada A.
      • Khandheria B.K.
      • et al.
      Use of three-dimensional speckle-tracking echocardiography for quantitative assessment of global left ventricular function: a comparative study to three-dimensional echocardiography.
      ,
      • Obokata M.
      • Nagata Y.
      • Wu V.C.
      • et al.
      Direct comparison of cardiac magnetic resonance feature tracking and 2D/3D echocardiography speckle tracking for evaluation of global left ventricular strain.
      In contrast to our analysis, these studies systematically excluded patients with poor image quality (7–25% of patients), which might contribute to this inconsistency.
      STE RV and LA strain were less reproducible than STE LV strain in our study. Both, RV and LA strain were derived from a single slice, whereas LV strains were averaged from 3 SAX (base, papillary muscle, apex) or 3 apical slices (2-, 3- and 4CV), respectively. Hence, RV and LA acquisitions are more susceptible to foreshortening and imaging artefacts. The assessment of RV GLS requires complete depiction of the RV from base to the apex and clear delineation of the endocardium from the blood pool. Both can be complicated by the limited echocardiographic accessibility of the RV due to its proximity to the sternum and the complex right ventricular geometry with thin and trabeculated walls that are difficult to delineate from soft tissue.
      • Venkatachalam S.
      • Wu G.
      • Ahmad M.
      Echocardiographic assessment of the right ventricle in the current era: application in clinical practice.
      The significantly lower volume estimates (RV EDA) in STE compared to CCT also suggest that full visualization of the RV cavity was not consistently achieved in STE, potentially contributing to poor reproducibility. None of these limitations exists in CCT and the present study demonstrated that RV strain can be assessed with high reproducibility by CCT FT.
      Most strain features had a relevant inter-modality bias in our study. Not surprisingly, agreement between STE and CCT was highest in parameters with good reproducibility in both modalities (dimensional LV and LA measurements, LVEF and LV- and LA GLS) and limited for LV SAX- (LV GCS and GRS) and RV measurements. Correlation between modalities improved after exclusion of patients with poor or moderate IQ in STE. Accuracy of STE strain is highly dependent from the acoustic window and IQ, and similar findings were reported from a study comparing CMR and STE.
      • Obokata M.
      • Nagata Y.
      • Wu V.C.
      • et al.
      Direct comparison of cardiac magnetic resonance feature tracking and 2D/3D echocardiography speckle tracking for evaluation of global left ventricular strain.
      CCT consistently overestimated volumetric variables (LV EDV, RV EDA and LA ESV) and underestimated (resulting in less negative values) LV- and LA GLS compared to STE. Previous studies, comparing both modalities also reported an underestimation of LV- and LA GLS in CCT but described a higher agreement between modalities compared to our findings.
      • Buss S.J.
      • Schulz F.
      • Mereles D.
      • et al.
      Quantitative analysis of left ventricular strain using cardiac computed tomography.
      ,
      • Ammon F.
      • Bittner D.
      • Hell M.
      • et al.
      CT-derived left ventricular global strain: a head-to-head comparison with speckle tracking echocardiography.
      • Szilveszter B.
      • Nagy A.I.
      • Vattay B.
      • et al.
      Left ventricular and atrial strain imaging with cardiac computed tomography: validation against echocardiography.
      • Marwan M.
      • Ammon F.
      • Bittner D.
      • et al.
      CT-derived left ventricular global strain in aortic valve stenosis patients: a comparative analysis pre and post transcatheter aortic valve implantation.
      • Gegenava T.
      • van der Bijl P.
      • Hirasawa K.
      • et al.
      Feature tracking computed tomography-derived left ventricular global longitudinal strain in patients with aortic stenosis: a comparative analysis with echocardiographic measurements.
      However, these studies systematically excluded a large number of patients with poor IQ in STE (up to about 50% of patients) and, besides Buss et al.,
      • Buss S.J.
      • Schulz F.
      • Mereles D.
      • et al.
      Quantitative analysis of left ventricular strain using cardiac computed tomography.
      increments of every 10% of the cardiac cycle were rendered, resulting in only 10 reconstructions per cardiac cycle. We hypothesize that, similar to CMR,
      • Claus P.
      • Omar A.M.S.
      • Pedrizzetti G.
      • Sengupta P.P.
      • Nagel E.
      Tissue tracking technology for assessing cardiac mechanics: principles, normal values, and clinical applications.
      ,
      • Miskinyte E.
      • Bucius P.
      • Erley J.
      • et al.
      Assessment of global longitudinal and circumferential strain using computed tomography feature tracking: intra-individual comparison with CMR feature tracking and myocardial tagging in patients with severe aortic stenosis.
      also the lower temporal resolution of CCT compared to STE contributes to inter-modality bias, since maximal deformation (peak strain) is less likely to be captured in sequences with a lower number of frames. Also in our study, the number of frames varied between CCT (20 reconstructions (5%) in each RR-interval) and STE (mean 43.4 ​± ​9.8 frames in each RR-interval).
      These results demonstrate that strain values derived from STE and CCT are not interchangeable and separate cut-off values have to be considered to implement CCT strain to clinical practice. Despite a systematic underestimation, clinical interpretation (but not cut-off values) might be similar for CCT longitudinal strain and two studies demonstrated the value of CCT LV GLS in the prediction of mortality in patients undergoing TAVI.
      • Gegenava T.
      • van der Bijl P.
      • Vollema E.M.
      • et al.
      Prognostic influence of feature tracking multidetector row computed tomography-derived left ventricular global longitudinal strain in patients with aortic stenosis treated with transcatheter aortic valve implantation.
      ,
      • Fukui M.
      • Xu J.
      • Thoma F.
      • et al.
      Baseline global longitudinal strain by computed tomography is associated with post transcatheter aortic valve replacement outcomes.
      Another study comparing LV strain assessed by CCT, CMR tagging and STE reported higher correlations of CMR to CCT versus CMR to STE, pointing to a high validity of CCT strain.
      • Tavakoli V.
      • Sahba N.
      Cardiac motion and strain detection using 4D CT images: comparison with tagged MRI, and echocardiography.
      However, despite its limitations, STE will remain the standard modality to determine strain, but CCT can be a valuable adjunct in patients with clinically indicated CCT or impaired acoustic window in STE. Whereas this study could proof the reproducibility of CCT strain acquisition in the LV, the LA, and for the first time also in the RV, the diagnostic accuracy of RV strain has to be validated against CMR strain imaging in future trials. In addition to LV strain, the assessment of RV and LA might offer the potential to more adequately predict outcomes and tailor management in patients that undergo interventions in which these variables might be interest, such as transcatheter interventions on the mitral- and tricuspid valve.

      5. Limitations

      Limitations are based on the nature of this retrospective observational study. STE and CCT were not consistently performed at the same day, which might contribute to an underestimation of the correlation of strain measures between modalities. However, in comparison to other cardiac disease like acute myocardial infarction, myocarditis, or takotsubo cardiomyopathy, severe aortic stenosis can be considered to be a mostly stable disease with slow progression and changes in cardiac contractility over time. Changes in ventricular loading conditions over time can also impact strain values,
      • Yotti R.
      • Bermejo J.
      • Benito Y.
      • et al.
      Validation of noninvasive indices of global systolic function in patients with normal and abnormal loading conditions: a simultaneous echocardiography pressure-volume catheterization study.
      but in our analysis we found no difference depending on the time delay between CCT and STE. A relevant number of patients was excluded from analysis due to lacking or incomplete CCT or STE, which might result in selection bias. Calculation methods, workstations and vendors for imaging- and strain-acquisition varied between modalities. CCT strain was determined by feature tracking while echocardiography was analyzed by speckle tracking and we cannot exclude a systematic bias resulting from varying algorithm results specific for the used applications. Neither CCT-, nor STE image acquisition were dedicatedly performed for strain analysis but in clinical routine following standardized protocols. Instructed investigators and experts in echocardiography might provide images of better quality that are more feasible for strain analysis. Hence, this study might underestimate reproducibility of both modalities and particularly of echocardiographic assessments.

      6. Conclusion

      4D CCT feature tracking represents a reproducible method to assess left- and right ventricular and left atrial myocardial strain. Whereas agreement between modalities was strong for left ventricular global longitudinal strain, only moderate to poor correlation was depicted in right ventricular and left atrial strain, the latter possibly related to limited reproducibility in echocardiography.

      Funding

      None.

      Declaration of competing interest

      No specific funding was obtained for this work. The study was approved by the local ethics committee and was conducted in accordance with the Declaration of Helsinki. All patients provided written informed consent. The paper is not under consideration elsewhere, nor has the paper's contents been previously published. All authors have significantly contributed to the study and have the following disclosures: Dr. Okuno receives speaker fees from Abbott. Dr. Huber has received research grants from the Swiss National Science Foundation, the Helmut-Hartweg Foundation and the Foundation to Fight against Cancer, all for work outside the submitted study. Dr. Stortecky is the recipient of research grants to the institution from Edwards Lifesciences, Medtronic, Abbott Vascular and Boston Scientific, is a consultant for BTG and Teleflex and has received speaker fees from Boston Scientific. Dr. Windecker has received research grants to his institution from Abbott, Amgen, Boston, Biotronik, and St. Jude Medical, he has received no speaker fee. Dr. Pilgrim has received research grants to his institution from Edwards Lifesciences, Symetis, and Biotronik; has received speaker fees from Boston Scientific; and has received reimbursement for travel expenses from St. Jude Medical. Dr. Gräni received research funding from Swiss National Science Foundation and Innosuisse outside of the submitted work. Further Dr. Gräni received travel fees from Amgen and Bayer. All other authors report no conflicts.

      Appendix A. Supplementary data

      The following is the supplementary data to this article:

      References

        • Voigt J.U.
        • Pedrizzetti G.
        • Lysyansky P.
        • et al.
        Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging.
        J Am Soc Echocardiogr. 2015; 28 (official publication of the American Society of Echocardiography): 183-193
        • Geyer H.
        • Caracciolo G.
        • Abe H.
        • et al.
        Assessment of myocardial mechanics using speckle tracking echocardiography: fundamentals and clinical applications.
        J Am Soc Echocardiogr. 2010; 23 (official publication of the American Society of Echocardiography) (quiz 453-355): 351-369
        • Kalam K.
        • Otahal P.
        • Marwick T.H.
        Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction.
        Heart. 2014; 100: 1673-1680
        • Shen M.T.
        • Yang Z.G.
        • Diao K.Y.
        • et al.
        Left ventricular involvement in arrhythmogenic right ventricular dysplasia/cardiomyopathy predicts adverse clinical outcomes: a cardiovascular magnetic resonance feature tracking study.
        Sci Rep. 2019; 9: 14235
        • Florescu M.
        • Magda L.S.
        • Enescu O.A.
        • Jinga D.
        • Vinereanu D.
        Early detection of epirubicin-induced cardiotoxicity in patients with breast cancer.
        J Am Soc Echocardiogr. 2014; 27 (official publication of the American Society of Echocardiography): 83-92
        • Cho G.Y.
        • Marwick T.H.
        • Kim H.S.
        • Kim M.K.
        • Hong K.S.
        • Oh D.J.
        Global 2-dimensional strain as a new prognosticator in patients with heart failure.
        J Am Coll Cardiol. 2009; 54: 618-624
        • Amzulescu M.S.
        • De Craene M.
        • Langet H.
        • et al.
        Myocardial strain imaging: review of general principles, validation, and sources of discrepancies.
        Eur Heart J Cardiovasc Imaging. 2019; 20: 605-619
        • Sengeløv M.
        • Jørgensen P.G.
        • Jensen J.S.
        • et al.
        Global longitudinal strain is a superior predictor of all-cause mortality in heart failure with reduced ejection fraction.
        JACC. Cardiovasc Imaging. 2015; 8: 1351-1359
        • Park J.J.
        • Park J.B.
        • Park J.H.
        • Cho G.Y.
        Global longitudinal strain to predict mortality in patients with acute heart failure.
        J Am Coll Cardiol. 2018; 71: 1947-1957
        • Lai W.
        • Jie H.
        • Jian-Xun D.
        • et al.
        Impact of concomitant impairments of the left and right ventricular myocardial strain on the prognoses of patients with ST-elevation myocardial infarction.
        Front cardiovasc med. 2021; 8: 659364
        • Fischer K.
        • Obrist S.J.
        • Erne S.A.
        • et al.
        Feature tracking myocardial strain incrementally improves prognostication in myocarditis beyond traditional CMR imaging features.
        JACC. Cardiovasc Imaging. 2020; 13: 1891-1901
        • Schmid J.
        • Kamml C.
        • Zweiker D.
        • et al.
        Cardiac magnetic resonance imaging right ventricular longitudinal strain predicts mortality in patients undergoing TAVI.
        Front cardiovasc med. 2021; 8: 644500
        • Canessa M.
        • Thamman R.
        • Americo C.
        • Soca G.
        • Dayan V.
        Global longitudinal strain predicts survival and left ventricular function after mitral valve surgery: a meta-analysis.
        Semin Thorac Cardiovasc Surg. 2021; 33: 337-342
        • Alashi A.
        • Khullar T.
        • Mentias A.
        • et al.
        Long-term outcomes after aortic valve surgery in patients with asymptomatic chronic aortic regurgitation and preserved LVEF: impact of baseline and follow-up global longitudinal strain.
        JACC. Cardiovasc Imaging. 2020; 13: 12-21
        • Ng A.C.T.
        • Prihadi E.A.
        • Antoni M.L.
        • et al.
        Left ventricular global longitudinal strain is predictive of all-cause mortality independent of aortic stenosis severity and ejection fraction.
        Eur Heart J Cardiovasc Imaging. 2017; 19: 859-867
        • Rowin E.J.
        • Maron B.J.
        • Wells S.
        • et al.
        Usefulness of global longitudinal strain to predict heart failure progression in patients with nonobstructive hypertrophic cardiomyopathy.
        Am J Cardiol. 2021; 151: 86-92
        • Lee Chuy K.
        • Drill E.
        • Yang J.C.
        • et al.
        Incremental value of global longitudinal strain for predicting survival in patients with advanced AL amyloidosis.
        JACC Cardiovasc Oncol. 2020; 2: 223-231
        • Dons M.
        • Jensen J.S.
        • Olsen F.J.
        • et al.
        Global longitudinal strain corrected by RR-interval is a superior echocardiographic predictor of outcome in patients with atrial fibrillation.
        Int J Cardiol. 2018; 263: 42-47
        • Hor K.N.
        • Baumann R.
        • Pedrizzetti G.
        • et al.
        Magnetic resonance derived myocardial strain assessment using feature tracking.
        JoVE. 2011; : 2356
        • Kempny A.
        • Fernández-Jiménez R.
        • Orwat S.
        • et al.
        Quantification of biventricular myocardial function using cardiac magnetic resonance feature tracking, endocardial border delineation and echocardiographic speckle tracking in patients with repaired tetralogy of fallot and healthy controls.
        J Cardiovasc Magn Reson. 2012; 14: 32
        • Otto C.M.
        • Nishimura R.A.
        • Bonow R.O.
        • et al.
        2020 ACC/AHA guideline for the management of patients with valvular heart disease: a report of the American college of cardiology/American heart association joint committee on clinical practice guidelines.
        J Am Coll Cardiol. 2020; 77: 450-500
        • Okuno T.
        • Corpataux N.
        • Spano G.
        • et al.
        True-severe stenosis in paradoxical low-flow low-gradient aortic stenosis: outcomes after transcatheter aortic valve replacement.
        Eur Heart J Qual Care Clin Outcomes. 2021; 7: 366-377
        • Mitchell C.
        • Rahko P.S.
        • Blauwet L.A.
        • et al.
        Guidelines for performing a comprehensive transthoracic echocardiographic examination in adults: recommendations from the American society of echocardiography.
        J Am Soc Echocardiogr. 2019; 32 (official publication of the American Society of Echocardiography): 1-64
        • Koo T.K.
        • Li M.Y.
        A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
        J Chiropr Med. 2016; 15: 155-163
        • Schober P.
        • Boer C.
        • Schwarte L.A.
        Correlation coefficients: appropriate use and interpretation.
        Anesth Analg. 2018; 126
        • Sugimoto T.
        • Dulgheru R.
        • Bernard A.
        • et al.
        Echocardiographic reference ranges for normal left ventricular 2D strain: results from the EACVI NORRE study.
        Eur Heart J Cardiovasc Imaging. 2017; 18: 833-840
        • Chan J.
        • Shiino K.
        • Obonyo N.G.
        • et al.
        Left ventricular global strain analysis by two-dimensional speckle-tracking echocardiography: the learning curve.
        J Am Soc Echocardiogr. 2017; 30 (official publication of the American Society of Echocardiography): 1081-1090
        • Buss S.J.
        • Schulz F.
        • Mereles D.
        • et al.
        Quantitative analysis of left ventricular strain using cardiac computed tomography.
        Eur J Radiol. 2014; 83: e123-e130
        • Pedrizzetti G.
        • Claus P.
        • Kilner P.J.
        • Nagel E.
        Principles of cardiovascular magnetic resonance feature tracking and echocardiographic speckle tracking for informed clinical use.
        J Cardiovasc Magn Reson. 2016; 18: 51
        • Langeland S.
        • Wouters P.F.
        • Claus P.
        • et al.
        Experimental assessment of a new research tool for the estimation of two-dimensional myocardial strain.
        Ultrasound Med Biol. 2006; 32: 1509-1513
        • Luis S.A.
        • Yamada A.
        • Khandheria B.K.
        • et al.
        Use of three-dimensional speckle-tracking echocardiography for quantitative assessment of global left ventricular function: a comparative study to three-dimensional echocardiography.
        J Am Soc Echocardiogr. 2014; 27 (official publication of the American Society of Echocardiography): 285-291
        • Obokata M.
        • Nagata Y.
        • Wu V.C.
        • et al.
        Direct comparison of cardiac magnetic resonance feature tracking and 2D/3D echocardiography speckle tracking for evaluation of global left ventricular strain.
        Eur Heart J Cardiovasc Imaging. 2016; 17: 525-532
        • Venkatachalam S.
        • Wu G.
        • Ahmad M.
        Echocardiographic assessment of the right ventricle in the current era: application in clinical practice.
        Echocardiography. 2017; 34: 1930-1947
        • Ammon F.
        • Bittner D.
        • Hell M.
        • et al.
        CT-derived left ventricular global strain: a head-to-head comparison with speckle tracking echocardiography.
        Int J Cardiovasc Imag. 2019; 35: 1701-1707
        • Szilveszter B.
        • Nagy A.I.
        • Vattay B.
        • et al.
        Left ventricular and atrial strain imaging with cardiac computed tomography: validation against echocardiography.
        J Cardiovasc Comput Tomogr. 2020; 14: 363-369
        • Marwan M.
        • Ammon F.
        • Bittner D.
        • et al.
        CT-derived left ventricular global strain in aortic valve stenosis patients: a comparative analysis pre and post transcatheter aortic valve implantation.
        J Cardiovasc Comput Tomogr. 2018; 12: 240-244
        • Gegenava T.
        • van der Bijl P.
        • Hirasawa K.
        • et al.
        Feature tracking computed tomography-derived left ventricular global longitudinal strain in patients with aortic stenosis: a comparative analysis with echocardiographic measurements.
        J Cardiovasc Comput Tomogr. 2020; 14: 240-245
        • Claus P.
        • Omar A.M.S.
        • Pedrizzetti G.
        • Sengupta P.P.
        • Nagel E.
        Tissue tracking technology for assessing cardiac mechanics: principles, normal values, and clinical applications.
        JACC. Cardiovasc Imaging. 2015; 8: 1444-1460
        • Miskinyte E.
        • Bucius P.
        • Erley J.
        • et al.
        Assessment of global longitudinal and circumferential strain using computed tomography feature tracking: intra-individual comparison with CMR feature tracking and myocardial tagging in patients with severe aortic stenosis.
        J Clin Med. 2019; 8: 1423
        • Gegenava T.
        • van der Bijl P.
        • Vollema E.M.
        • et al.
        Prognostic influence of feature tracking multidetector row computed tomography-derived left ventricular global longitudinal strain in patients with aortic stenosis treated with transcatheter aortic valve implantation.
        Am J Cardiol. 2020; 125: 948-955
        • Fukui M.
        • Xu J.
        • Thoma F.
        • et al.
        Baseline global longitudinal strain by computed tomography is associated with post transcatheter aortic valve replacement outcomes.
        J Cardiovasc Comput Tomogr. 2020; 14: 233-239
        • Tavakoli V.
        • Sahba N.
        Cardiac motion and strain detection using 4D CT images: comparison with tagged MRI, and echocardiography.
        Int J Cardiovasc Imag. 2014; 30: 175-184
        • Yotti R.
        • Bermejo J.
        • Benito Y.
        • et al.
        Validation of noninvasive indices of global systolic function in patients with normal and abnormal loading conditions: a simultaneous echocardiography pressure-volume catheterization study.
        Circ Cardiovasc imaging. 2014; 7: 164-172