Artificial intelligence-derived left ventricular strain in echocardiography in patients treated with chemotherapy.

Текст сохранён. Часть иллюстраций пока доступна только в первоисточнике; загрузка будет проверена повторно.

Introduction

Development of cancer treatment substantially improves prognosis of patients with cancer over the decades [ 1 , 2 ]. However, some patients may face the development of cardiac therapy-related cardiac dysfunction during chemotherapy, which limits their quality of life or worsens patient prognosis [ 3 – 5 ]. To mitigate this anticipated risk and prevent interruption of cancer treatment, early identification of myocardial dysfunction has been recommended [ 6 ].

Global longitudinal strain (GLS) is a clinically feasible echocardiographic parameter for detecting subclinical myocardial dysfunction earlier than declines in left ventricular ejection fraction (LVEF) [ 7 , 8 ]. Recent guidelines recommend using GLS for all patients before initiating cardiotoxic cancer treatment, aiming to stratify cancer therapy-related cardiovascular disease risk and identify significant changes during treatment [ 6 ]. However, GLS has not yet been fully adopted in clinical practice due to limited availability, inter-vendor variability, and the need for specialized expertise [ 9 ].

Artificial intelligence (AI) allows for learning from recorded images and performing automated image analysis for strain echocardiography [ 10 ]. Several studies have suggested that AI-based GLS reduced analysis time and the required skill, overcoming hurdles in GLS implementation [ 11 , 12 ]. This method may achieve accuracy similar to manual measurements [ 13 ]. However, by semi-automatically initiating the contour before tracking, AI algorithms might reduce observer variability compared to conventional tracking [ 10 ]. Furthermore, importantly, studying the reliability of AI-based GLS versus conventional GLS holds significant clinical relevance, especially for patients undergoing chemotherapy, who often show GLS decline while maintaining preserved LVEF [ 7 , 8 ].

Our aims are to investigate the correlation between AI-based GLS and conventional GLS, and whether the agreement may be influenced by cardiac structures or cardiovascular risk. Furthermore, we assess whether AI-based GLS may reduce inter-observer variability between differing expertise levels compared with conventional GLS.

Methods

Patient population

The present study is a retrospective cohort study including 117 patients treated with chemotherapy who underwent strain echocardiography at Tokyo Medical University Hachioji Medical Center between October 2018 and September 2020. A segment was considered missing if partly outside the image sector or if the myocardium was indistinguishable from surrounding structures due to artifacts. Those with poor image quality, defined as ≥ 3 missing segments of the 6 individual myocardial segments of the 4-chamber view, were excluded. Totally, the remaining 94 patients were included in the current analysis.

Information on medical history, laboratory results, and echocardiography were retrospectively collected. The study was approved by the local ethics committee (T2023-0031).

Echocardiography

Echocardiographic examinations were performed by experienced examiners at the Tokyo Medical University Hachioji Medical Center using standard commercially available ultrasound equipment (Vivid E95, GE Healthcare) with a 2.5 MHz phased-array transducer and reviewed on an ECHOPAC workstation (View Pal6.14). The measurement of LVEF and left atrial (LA) volume was obtained from the modified Simpson’s rule [ 14 ].

Diastolic function was assessed from the mitral inflow pattern by pulsed-wave Doppler. Mitral annular early diastolic velocity (e′) was assessed at the septal and lateral sites of the mitral annulus using tissue Doppler imaging. E/A ratio, e′ mean and E/e′ mean ratio were calculated [ 15 ].

LV hypertrophy by indexed LV mass (LVMi) > 115 g/m² in men or > 95 g/m² in women, LA volume enlargement by indexed LA volume (LAVi) > 34 mL/m², and abnormal e′ by a septal e′ <7 cm/sec or lateral e′ <10 cm/Sects. [ 15 , 16 ].

Conventional GLS

Conventional GLS was measured using speckle-tracking analyses using the semi-automatic analysis method implemented in commercially available software (View Pal6.14). The end-diastolic (ED) was defined by the automatic ECG trigger algorithm of the analysis software. After using the ECG trigger algorithm, the timing of end-systole was manually corrected in some cases using the aortic valve closure signal obtained by pulsed-wave Doppler in the left ventricular outflow tract or by continuous-wave Doppler through the aortic valve. The observer manually corrected the region of interest (ROI) by visual assessment of the endocardial and epicardial borders. GLS was measured from the 4-chamber view, based on the maximum absolute value of strain derived from strain curves as recommended [ 9 ].

AI-based GLS, Caas qardia [ 17 ]

AI-based GLS was obtained offline using a vendor-neutral speckle-tracking software platform (Caas Qardia 1.1, Pie Medical Imaging, Maastricht, NL). Core lab validation of the software platform was presented earlier [ 17 ]. In addition, the software version used in the present study contained a novel AI-based endocardial contour detection algorithm. The workflow consists of two analysis steps (Fig. 1 ). First, an AI algorithm automatically segmented the endocardial border at the ED reference frame using a U-Net convolutional neural network. The neural network is trained specifically for LV endocardium segmentation using standardized echocardiographic views (e.g., apical 4 chamber view). By default, the software proposes two ED frames which coincide with the first and second R-wave times of the acquired image run as determined by the ultrasound machine software. These ED frames were adjusted by the analyst if necessary. Both the ED endocardial contour and ED timings were corrected in case these were deemed to be determined incorrectly by the algorithm (Fig. 1 ).

Fig. 1 Schema of the LV longitudinal strain workflow. ROI, region of interest; t cycle . cardiac cycle duration; GLS, global longitudinal strain; HR, heart rate; FR, frame rate

The second analysis step uses the AI-based endocardial contour at the first ED phase, and the two ED phase timings as input to a speckle-tracking algorithm. By means of cross-correlation the current location of speckles is tracked for each image frame of the heart cycle at several endocardial ROI. GLS is computed based on the total myocardial line length and presented as time-strain graphs and numerical values [ 9 ].

Tracking quality was inspected visually. In case of inaccurate tracking the endocardial ROI was manually adjusted by the analyst, and tracking was reprocessed.

Statistical analysis

Categorical variables are expressed as frequencies (percentages) and continuous variables as means ± standard deviation or median (25th and 75th percentiles) according to the distribution of the variables. The agreements between AI-based GLS and measured GLS were assessed by Bland-Altman plots with corresponding 95% limits of agreement. Inter-observer and intra-observer agreements of estimated-GLS were assessed with the intraclass correlation coefficient (ICC).

All exams for AI-based GLS were assessed by a physician (AK) with no experience with strain echocardiography (“GLS beginner”), blinded to demographic and clinical parameters. After a very short training (< 30 min of theory and practice), Inter-observer reproducibility (with an experienced cardiologist, RT) were tested with a separate complete GLS analysis of randomly selected 30 patients. A GLS beginner was considered as a physician who had never assessed GLS tracking, while an experienced echocardiographer had certification as a national echocardiographic specialist.

All analyses were performed using R version 3.4.0 (R Development Core Team, Vienna, Austria). A two-sided p-value < 0.05 was considered statistically significant. No imputation was performed.

Results

Baseline patient characteristics

Among patients included in the current analysis, mean age was 69 ± 12 years, 73% were female, median body mass index was 22 ± 4 kg/m², 43% had hypertension, 14% had diabetes mellitus, 52% were diagnosed with breast cancer, 77% were treated with anthracycline drug, mean systolic blood pressure (BP) was 131 ± 22 mmHg, and median b-type natriuretic peptide (BNP) was 20 (10 to 78) pg/ml. Table 1 .

Table 1 Study Population ( N = 94) Mean ± SD / n (%) Age, years 69 ± 12 Female, N (%) 69 (73.4%) Body mass index, kg/m² 22 ± 4 Body mass index ≥ 25 kg/m 2 , N (%) 23 (24.5%) Medical history, N (%) Hypertension 40 (42.6%) Diabetes 13 (13.8%) Heart failure 2 (2.1%) Myocardial infarction 4 (4.3%) Cancer type Breast 49 (52.1%) Lymphoma 21 (22.3%) Gastric/colorectal 6 (6.4%) Others* 18 (19.1%) Anti-cancer treatments, N (%) Surgical intervention 49 (52.1%) Radiotherapeutic intervention 13 (13.8%) Chemotherapy, N (%) Anthracycline chemotherapy 72 (76.6%) HER2-targeted therapy 21 (22.3%) Other targeted therapy 36 (38.3%) Systolic BP, mmHg 131 ± 22 Diastolic BP, mmHg 71 ± 14 Heart rate, mg/dl 74 ± 15 Laboratory data Creatinine, mg/dl 0.8 ± 0.5 Hemoglobin, g/dl 11.7 ± 2.0 BNP, pg/ml 20 (10 to 78) Echocardiogram LV end-diastolic diameter, mm 44 ± 5 LV end-systolic diameter, mm 29 ± 5 LV ejection fraction, % 63 ± 8 LV mass index, g/m² 83 ± 22 LV hypertrophy, N (%) 13 (13.8%) LA volume index, ml/m² 28 ± 9 LA volume index > 34 ml/m², N (%) 21 (22.3%) Septal e′, cm/sec 6.8 ± 2.0 Lateral e′, cm/sec 8.7 ± 2.7 Septal e′<7 cm/sec or lateral e‘<10 cm/sec, N (%) 63 (70.0%) AI-based GLS, % -19.0 ± 3.7 Conventional GLS, % -15.9 ± 3.4 Values are expressed as median (25th to 75th percentile) or n (%) HER2, human epidermal growth factor receptor-2; BP, blood pressure; BNP, B-type natriuretic peptide; LV, left ventricular; LA, left atrial; GLS, global longitudinal strain *Others included, urologic, hepatic, lung, hematologic (i.e., multiple myeloma and myelodysplastic syndromes) cancers

For echocardiographic parameters, the mean LVEF was 63 ± 8%, 14% of patients had a LVH, 22% had LA enlargement, and 70% had increased LV stiffness. Table 1 .

Correlation between AI-based GLS and conventional GLS

Mean GLS in the entire population was − 19.0 ± 3.7% and − 15.9 ± 3.4% for the AI method and the conventional method, respectively. There was a significant correlation between the methods (rho = 0.74; p < 0.01, Fig. 2 ). The Bland-Altman analysis of between method differences showed a bias of -3.1% with limits of agreement ± 5.0%. Figure 2 .

Fig. 2 Correlation plots and Bland-Altman plots between AI-based and conventional GLS. GLS, global longitudinal strain

Agreements between AI-based and conventional GLS across different cardiac structure, function and cardiovascular diseases/risks

The agreement between AI-based and conventional GLS across different categories of abnormal cardiac structure and function (i.e., LV systolic function, hypertrophy, and regional wall motion abnormalities, and diastolic function) and presence/absence of cardiovascular disease and/or risk factors (i.e., hypertension, diabetes, heart failure, myocardial infarction) are presented in Table 2 . Overall, the agreements between AI-based and measured GLS were not modified by these categories (p-for-interaction > 0.10). Table 2 .

Table 2 Agreement between AI-based GLS and conventional GLS across cardiac structure, function and cardiovascular diseases/risk factors Number Average diff Limits of agreement Pearson rho P -for-interaction LVEF ≤ 60% 30 -2.12 -7.62 to 3.39 0.74 0.74 > 60% 64 -3.55 -8.14 to 1.04 0.62 LV hypertrophy Absence 81 -3.29 -8.43 to 1.86 0.67 0.40 Presence 13 -1.89 -5.63 to 1.84 0.91 LAVi ≤ 30 ml/m 2 56 -3.02 -8.22 to 2.18 0.68 0.38 > 30 ml/m 2 38 -3.21 -8.08 to 1.67 0.80 Septal e′<7 cm/sec or lateral e‘<10 cm/sec Absence 27 -3.38 -8.12 to 1.36 0.76 0.32 Presence 63 -3.08 -8.27 to 2.11 0.62 LV asynergy Absence 79 -3.26 -7.94 to 1.42 0.61 0.55 Presence 15 -2.22 -8.82 to 4.38 0.72 Cardiovascular disease and/or risk factors* Absence 43 -2.61 -7.49 to 2.27 0.64 0.53 Presence 51 -3.67 -8.73 to 1.39 0.79 LVEF, left ventricular ejection fraction; LV, left ventricular; LAVi, left atrial volume index *Cardiovascular diseases and/or risk factors included hypertension, diabetes, myocardial infarction and heart failure

Agreements between GLS beginner and experienced echocardiographer using AI-based and conventional GLS

Inter-observer ICC coefficients between GLS beginner and experienced echocardiographer showed numerically higher reproducibility in the AI method than the conventional method (inter-observer agreement [95%CI] = 0.81 [0.64 to 0.90] in the AI method vs. 0.62 [0.34 to 0.80] in the conventional method). The Bland-Altman analysis of between examiner differences showed a bias of 0.91% with limits of agreement ± 3.5% in AI method, and a bias of -5.07% with limits of agreement ± 5.74% in conventional method (Fig. 3 ).

Fig. 3 Agreements between GLS beginner and experienced echocardiographer using AI-based and conventional GLS. AI, artificial intelligence; GLS, global longitudinal strain

Discussion

In the present study, our results showed that, in patients eligible for GLS-based assessments, AI-based GLS was moderately correlated with conventional GLS. The agreements were consistent across abnormalities of cardiac structure and function (i.e., LV systolic function, LV hypertrophy, LA enlargement and diastolic dysfunction), and cardiovascular diseases/risks. Additionally, the agreement between GLS beginner and an experienced echocardiographer was numerically better in the AI-based method than the conventional method. These findings support the utility of AI-based GLS and enhance its application among patients undergoing for chemotherapies in clinical practice.

The incorporation of AI into cardiovascular image analysis software is ongoing [ 18 ]. Its fundamental applications include the assessment of cardiac function by LVEF and GLS [ 10 , 18 ]. There are roughly two AI approaches for echocardiography-based GLS calculation. The first approach, such as used in the present study, utilizes deep learning to perform endocardial or full cardiac wall segmentation. This segmentation subsequently serves as input for a motion calculation algorithm, (e.g., by speckle tracking or optical flow methods), estimating the displacement and translation of the cardiac wall throughout the cardiac cycle [ 19 , 20 ]. The second approach utilizes deep learning for both the segmentation and motion tracking steps [ 21 ].

To date, performance and validation data of these methods often comprise of head-to-head comparison between the AI-based GLS method and conventional-based method. Herein, correlation coefficients between the methods ranged from moderate to excellent ranging between 0.54 and 0.93, respectively [ 19 – 22 ]. For example, in a community-based cohort including 561 patients with a risk of developing heart failure, the agreement between manual and fully automated GLS showed fairness (rho = 0.69), and its agreement was improved with semi-automated GLS (rho = 0.84). However, the limited expertise in performing strain echocardiography may constrain the generalizability of these findings [ 22 ]. With this regard, our results showed numerically better reproducibility of AI-based GLS method than conventional GLS method. These findings may be explained by the fact that our AI approach can result in a hybrid approach combining AI for segmentation as initialization step for subsequent strain calculation, subsequently reducing the inter observer variability, and suggested this AI tool could be more user-friendly, particularly for unexperienced physicians.

Patients with abnormalities of LV structure and function were more likely to have lower value of GLS, compared with those not [ 23 , 24 ]. Along with technical issues (i.e., tracing process, LV segment identification), different cardiac morphology and/or cardiovascular diseases may impact the agreement between AI-based GLS and conventional GLS [ 20 ]. Nonetheless, our results showed that the agreements between two GLS approaches were consistent across different abnormal cardiac structure, function and cardiovascular diseases/risks. These results are reassuring and may boost the applicability of our AI-based GLS approach.

Previously, the EACVI/ASE Strain Standardization Taskforce assessed the inter-vendor variability of nine conventional-based GLS method [ 25 ]. Future investigations by the Taskforce could reveal the performance and robustness characteristics of the different AI approaches for calculating GLS, i.e., based fully on deep learning, or a hybrid approach combining AI for segmentation as initialization step, and subsequent using the widely adopted speckle tracking or optical flow methods. Importantly, the merits of our AI-based GLS lie in assessing GLS values within pre-existing images through web-based software (eliminating the need for desktop installation) and in achieving better reproducibility of AI-based GLS across both inexperienced and experienced echocardiographers, compared to conventional GLS. Patients who are candidates for cancer treatment are recommended to undergo GLS measurements at least every three months during the treatment period [ 26 , 27 ]. Therefore, our AI-based echocardiogram may hold potential for increasing the relevance of GLS assessments for patients treated with chemotherapies in clinical practice.

Limitations

The results should be interpreted in light of the following limitations. This retrospective study has a small sample size, which may limit the generalizability of our findings and necessitate further statistical power to confirm them. AI-based GLS value was determined through assessment with good image quality, < 2 missing segments of the 6 individual myocardial segments in the 4-chamber view. All echocardiographic images were captured by experienced echocardiographic examiners; hence, any differences of generating the images due to varying levels of expertise were not accounted for in this analysis. Nevertheless, the utilization of an AI-based methodology might offer a solution to the challenge of precisely delineating the endocardium. The timing of performing echocardiography, particularly the duration after chemotherapy may impact patient cardiac function, however, we lacked a consistent duration period between chemotherapies and echocardiography assessment. Additionally, the longer training period for GLS beginners may influence our results. An adequately powered prospective study following structured training should be warranted.

Although GLS value may vary with vendors and examiners [ 25 , 28 ], in the current analysis, GLS values, which experienced examiners measured using the conventional method, were considered as a standardized value. The influence of observer experience level on performance and the concordance of AI-based GLS values with those from different imaging modalities, such as cardiac magnetic resonance imaging, are sparsely reported and warrant further investigation [ 29 ].

Conclusions

In patients treated with chemotherapy. AI-based GLS was moderately correlated with conventional GLS. The agreement of AI-based GLS values measured by beginner and experienced echocardiographers was numerically better than that of conventional GLS.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Acknowledgements

We thank Maarten Heusinkveld and Peter Huntjens for providing software of artificial intelligence-based strain echocardiography.

Author contributions

Drafting: A.K. Y.I., and M.K.Analysis: A.K. and M.K.Data collection: A.K. and R.T.Critical review of the manuscript: R.T., S.Y., T.K., K.S., N.T.Supervision: N.T.

Funding

Open access funding provided by Tokyo Medical University JSPRINTA.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

References

1. Miller KD Siegel RL Lin CC Cancer treatment and survivorship statistics, 2016 Cancer J Clin 2016 66 271 289 10.3322/caac.21349 27253694 Miller KD, Siegel RL, Lin CC et al (2016) Cancer treatment and survivorship statistics, 2016. Cancer J Clin 66:271–289 10.3322/caac.21349 27253694

2. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J et al (1990) Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. European journal of cancer (Oxford, England: 2013;49:1374 – 403 10.1016/j.ejca.2012.12.027 23485231

3. Ewer MS Ewer SM Cardiotoxicity of anticancer treatments Nat Rev Cardiol 2015 12 547 558 10.1038/nrcardio.2015.65 25962976 Ewer MS, Ewer SM (2015) Cardiotoxicity of anticancer treatments. Nat Rev Cardiol 12:547–558 25962976 10.1038/nrcardio.2015.65

4. Zamorano JL Lancellotti P Rodriguez Muñoz D 2016 ESC position paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: the Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC) Eur Heart J 2016 37 2768 2801 10.1093/eurheartj/ehw211 27567406 Zamorano JL, Lancellotti P, Rodriguez Muñoz D et al (2016) 2016 ESC position paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: the Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC). Eur Heart J 37:2768–2801 27567406 10.1093/eurheartj/ehw211

5. Koene RJ Prizment AE Blaes A Konety SH Shared Risk factors in Cardiovascular Disease and Cancer Circulation 2016 133 1104 1114 10.1161/CIRCULATIONAHA.115.020406 26976915 PMC4800750 Koene RJ, Prizment AE, Blaes A, Konety SH (2016) Shared Risk factors in Cardiovascular Disease and Cancer. Circulation 133:1104–1114 26976915 10.1161/CIRCULATIONAHA.115.020406 PMC4800750

6. Lyon AR López-Fernández T Couch LS 2022 ESC guidelines on cardio-oncology developed in collaboration with the European Hematology Association (EHA), the European Society for Therapeutic Radiology and Oncology (ESTRO) and the International Cardio-Oncology Society (IC-OS) Eur Heart J 2022 43 4229 4361 10.1093/eurheartj/ehac244 36017568 Lyon AR, López-Fernández T, Couch LS et al (2022) 2022 ESC guidelines on cardio-oncology developed in collaboration with the European Hematology Association (EHA), the European Society for Therapeutic Radiology and Oncology (ESTRO) and the International Cardio-Oncology Society (IC-OS). Eur Heart J 43:4229–4361 36017568 10.1093/eurheartj/ehac244

7. Santoro C Arpino G Esposito R 2D and 3D strain for detection of subclinical anthracycline cardiotoxicity in breast cancer patients: a balance with feasibility Eur Heart J Cardiovasc Imaging 2017 18 930 936 10.1093/ehjci/jex033 28379383 Santoro C, Arpino G, Esposito R et al (2017) 2D and 3D strain for detection of subclinical anthracycline cardiotoxicity in breast cancer patients: a balance with feasibility. Eur Heart J Cardiovasc Imaging 18:930–936 28379383 10.1093/ehjci/jex033

8. Thavendiranathan P Poulin F Lim KD Plana JC Woo A Marwick TH Use of myocardial strain imaging by echocardiography for the early detection of cardiotoxicity in patients during and after cancer chemotherapy: a systematic review J Am Coll Cardiol 2014 63 2751 2768 10.1016/j.jacc.2014.01.073 24703918 Thavendiranathan P, Poulin F, Lim KD, Plana JC, Woo A, Marwick TH (2014) Use of myocardial strain imaging by echocardiography for the early detection of cardiotoxicity in patients during and after cancer chemotherapy: a systematic review. J Am Coll Cardiol 63:2751–2768 24703918 10.1016/j.jacc.2014.01.073

9. Voigt JU Pedrizzetti G Lysyansky P Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging Eur Heart J Cardiovasc Imaging 2015 16 1 11 10.1093/ehjci/jeu184 25525063 Voigt JU, Pedrizzetti G, Lysyansky P et al (2015) Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. Eur Heart J Cardiovasc Imaging 16:1–11 25525063 10.1093/ehjci/jeu184

10. Nolan MT Thavendiranathan P Automated quantification in Echocardiography JACC Cardiovasc Imaging 2019 12 1073 1092 10.1016/j.jcmg.2018.11.038 31171260 Nolan MT, Thavendiranathan P (2019) Automated quantification in Echocardiography. JACC Cardiovasc Imaging 12:1073–1092 31171260 10.1016/j.jcmg.2018.11.038

11. Belghitia H Brette S Lafitte S Automated function imaging: a new operator-independent strain method for assessing left ventricular function Arch Cardiovasc Dis 2008 101 163 169 10.1016/S1875-2136(08)71798-4 18477943 Belghitia H, Brette S, Lafitte S et al (2008) Automated function imaging: a new operator-independent strain method for assessing left ventricular function. Arch Cardiovasc Dis 101:163–169 18477943 10.1016/s1875-2136(08)71798-4

12. Medvedofsky D Kebed K Laffin L Reproducibility and experience dependence of echocardiographic indices of left ventricular function: side-by-side comparison of global longitudinal strain and ejection fraction Echocardiography 2017 34 365 370 10.1111/echo.13446 28185312 Medvedofsky D, Kebed K, Laffin L et al (2017) Reproducibility and experience dependence of echocardiographic indices of left ventricular function: side-by-side comparison of global longitudinal strain and ejection fraction. Echocardiography 34:365–370 28185312 10.1111/echo.13446

13. Ingul CB Torp H Aase SA Berg S Stoylen A Slordahl SA Automated analysis of strain rate and strain: feasibility and clinical implications J Am Soc Echocardiogr 2005 18 411 418 10.1016/j.echo.2005.01.032 15891750 Ingul CB, Torp H, Aase SA, Berg S, Stoylen A, Slordahl SA (2005) Automated analysis of strain rate and strain: feasibility and clinical implications. J Am Soc Echocardiogr 18:411–418 15891750 10.1016/j.echo.2005.01.032

14. Lang RM Badano LP Mor-Avi V Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging Eur Heart J Cardiovasc Imaging 2015 16 233 270 10.1093/ehjci/jev014 25712077 Lang RM, Badano LP, Mor-Avi V et al (2015) Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging 16:233–270 25712077 10.1093/ehjci/jev014

15. Nagueh SF Smiseth OA Appleton CP Recommendations for the evaluation of left ventricular diastolic function by Echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging J Am Soc Echocardiogr 2016 29 277 314 10.1016/j.echo.2016.01.011 27037982 Nagueh SF, Smiseth OA, Appleton CP et al (2016) Recommendations for the evaluation of left ventricular diastolic function by Echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 29:277–314 27037982 10.1016/j.echo.2016.01.011

16. Lang RM Badano LP Mor-Avi V Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging J Am Soc Echocardiogr 2015 28 1 39e14 10.1016/j.echo.2014.10.003 25559473 Lang RM, Badano LP, Mor-Avi V et al (2015) Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 28:1–39e14 25559473 10.1016/j.echo.2014.10.003

17. Spitzer E Camacho B Mrevlje B Brandendburg HJ Ren CB Echocardiography Core Laboratory Validation of a Novel Vendor-Independent web-based Software for the Assessment of Left Ventricular Global Longitudinal strain J Cardiovasc Imaging 2023 31 135 141 10.4250/jcvi.2022.0130 37488918 PMC10374390 Spitzer E, Camacho B, Mrevlje B, Brandendburg HJ, Ren CB (2023) Echocardiography Core Laboratory Validation of a Novel Vendor-Independent web-based Software for the Assessment of Left Ventricular Global Longitudinal strain. J Cardiovasc Imaging 31:135–141 37488918 10.4250/jcvi.2022.0130 PMC10374390

18. Lin A Pieszko K Park C Artificial intelligence in cardiovascular imaging: enhancing image analysis and risk stratification BJR Open 2023 5 20220021 37396483 10.1259/bjro.20220021 PMC10311632 Lin A, Pieszko K, Park C et al (2023) Artificial intelligence in cardiovascular imaging: enhancing image analysis and risk stratification. BJR Open 5:20220021 37396483 10.1259/bjro.20220021 PMC10311632

19. Myhre PL Hung CL Frost MJ External validation of a deep learning algorithm for automated echocardiographic strain measurements Eur Heart J Digit Health 2024 5 60 68 10.1093/ehjdh/ztad072 38264705 PMC10802824 Myhre PL, Hung CL, Frost MJ et al (2024) External validation of a deep learning algorithm for automated echocardiographic strain measurements. Eur Heart J Digit Health 5:60–68 38264705 10.1093/ehjdh/ztad072 PMC10802824

20. Salte IM Østvik A Smistad E Artificial Intelligence for Automatic Measurement of Left ventricular strain in Echocardiography JACC Cardiovasc Imaging 2021 14 1918 1928 10.1016/j.jcmg.2021.04.018 34147442 Salte IM, Østvik A, Smistad E et al (2021) Artificial Intelligence for Automatic Measurement of Left ventricular strain in Echocardiography. JACC Cardiovasc Imaging 14:1918–1928 34147442 10.1016/j.jcmg.2021.04.018

21. Jiang J Liu B Li YW Hothi SS Clinical service evaluation of the feasibility and reproducibility of novel artificial intelligence based-echocardiographic quantification of global longitudinal strain and left ventricular ejection fraction in trastuzumab-treated patients Front Cardiovasc Med 2023 10 1250311 10.3389/fcvm.2023.1250311 38045908 PMC10693341 Jiang J, Liu B, Li YW, Hothi SS (2023) Clinical service evaluation of the feasibility and reproducibility of novel artificial intelligence based-echocardiographic quantification of global longitudinal strain and left ventricular ejection fraction in trastuzumab-treated patients. Front Cardiovasc Med 10:1250311 38045908 10.3389/fcvm.2023.1250311 PMC10693341

22. Kawakami H Wright L Nolan M Potter EL Yang H Marwick TH Feasibility, reproducibility, and clinical implications of the Novel fully automated Assessment for Global Longitudinal strain J Am Soc Echocardiogr 2021 34 136 145e2 10.1016/j.echo.2020.09.011 33293202 Kawakami H, Wright L, Nolan M, Potter EL, Yang H, Marwick TH (2021) Feasibility, reproducibility, and clinical implications of the Novel fully automated Assessment for Global Longitudinal strain. J Am Soc Echocardiogr 34:136–145e2 33293202 10.1016/j.echo.2020.09.011

23. Haji K Huynh Q Wong C Stewart S Carrington M Marwick TH Improving the characterization of Stage A and B Heart failure by adding global longitudinal strain JACC Cardiovasc Imaging 2022 15 1380 1387 10.1016/j.jcmg.2022.03.007 35926896 Haji K, Huynh Q, Wong C, Stewart S, Carrington M, Marwick TH (2022) Improving the characterization of Stage A and B Heart failure by adding global longitudinal strain. JACC Cardiovasc Imaging 15:1380–1387 35926896 10.1016/j.jcmg.2022.03.007

24. Salas-Pacheco JL Lomelí-Sánchez O Baltazar-González O Soto ME Longitudinal systolic dysfunction in hypertensive cardiomyopathy with normal ejection fraction Echocardiography 2022 39 46 53 10.1111/echo.15267 34904277 Salas-Pacheco JL, Lomelí-Sánchez O, Baltazar-González O, Soto ME (2022) Longitudinal systolic dysfunction in hypertensive cardiomyopathy with normal ejection fraction. Echocardiography 39:46–53 34904277 10.1111/echo.15267

25. Farsalinos KE Daraban AM Ünlü S Thomas JD Badano LP Voigt JU Head-to-Head comparison of global longitudinal strain measurements among nine different vendors: the EACVI/ASE Inter-vendor Comparison Study J Am Soc Echocardiogr 2015 28 1171 1181 10.1016/j.echo.2015.06.011 26209911 Farsalinos KE, Daraban AM, Ünlü S, Thomas JD, Badano LP, Voigt JU (2015) Head-to-Head comparison of global longitudinal strain measurements among nine different vendors: the EACVI/ASE Inter-vendor Comparison Study. J Am Soc Echocardiogr 28:1171–1181e2 26209911 10.1016/j.echo.2015.06.011

26. Kenigsberg B Wellstein A Barac A Left ventricular dysfunction in Cancer Treatment: is it relevant? JACC Heart Fail 2018 6 87 95 10.1016/j.jchf.2017.08.024 29413379 Kenigsberg B, Wellstein A, Barac A (2018) Left ventricular dysfunction in Cancer Treatment: is it relevant? JACC Heart Fail 6:87–95 29413379 10.1016/j.jchf.2017.08.024

27. Čelutkienė J Pudil R López-Fernández T Role of cardiovascular imaging in cancer patients receiving cardiotoxic therapies: a position statement on behalf of the Heart Failure Association (HFA), the European Association of Cardiovascular Imaging (EACVI) and the Cardio-Oncology Council of the European Society of Cardiology (ESC) Eur J Heart Fail 2020 22 1504 1524 10.1002/ejhf.1957 32621569 Čelutkienė J, Pudil R, López-Fernández T et al (2020) Role of cardiovascular imaging in cancer patients receiving cardiotoxic therapies: a position statement on behalf of the Heart Failure Association (HFA), the European Association of Cardiovascular Imaging (EACVI) and the Cardio-Oncology Council of the European Society of Cardiology (ESC). Eur J Heart Fail 22:1504–1524 32621569 10.1002/ejhf.1957

28. Voigt JU Pedrizzetti G Lysyansky P 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 28 183 193 10.1016/j.echo.2014.11.003 25623220 Voigt JU, Pedrizzetti G, Lysyansky P et al (2015) 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 28:183–193 25623220 10.1016/j.echo.2014.11.003

29. Brandt Y, Lubrecht JM, Adriaans BP et al (2024) Quantification of left ventricular myocardial strain: comparison between MRI tagging, MRI feature tracking, and ultrasound speckle tracking. NMR Biomed :e5164 10.1002/nbm.5164 38664924