Is speckle tracking strongly dependent on the experience level of the evaluator?

Authors

JAKUBÍK Juraj SVAČINOVÁ Jana HRUŠKOVÁ Jana PODROUŽKOVÁ Helena ŠTÍPALOVÁ Tatiana VINCIGUERRA Manlio

Year of publication 2018
Type Conference abstract
Citation
Description Background: Speckle-tracking echocardiography (STE) is a promising diagnostic method of myocardial deformation assessment. It quantifies global contractile function of the left ventricle (LV) via longitudinal (GLS) and circumferential (GCS) strain. It seems that the quality of input data could affect the results of STE analysis. Moreover, the impact of the evaluator's level of experience on final results is still unknown. Purpose: The aim of the study was to elucidate the role of the evaluator's level of experience on validity of STE analysis. Moreover, we tried to find out whether there is a relation between the quality of the input data and the difference in STE results of various evaluators. Method: The visual data of 119 subjects in STE quality were collected from 3 independent sonographers. Each dataset involved GLS of 4-chamber (4CH), 2-chamber (2CH) and 3-chamber (APLAX) apical LV projections and GCS of LV in parasternal short axis projection on the level of papillary muscles (SAX). STE analysis of the same dataset was performed in parallel by 2 evaluators on different experience level (A - highly experienced: >2000 cases, >200 STE analyses; B – intermediate: >100 cases, instructed to STE without previous analyses). The image quality of datasets was assessed and marked in 1–4 scale by highly experienced external sonographer (1 - excellent, 2 - good, 3 - one segment unacceptable, 4 - unacceptable image quality). We compared the degree of match of both evaluators in relation to the dataset image quality. Results: The Spearman's correlation between the evaluator A and B in all analyzed projections was significantly positive: r>0.7, p<0.00001 (4CH: r=0.73; 2CH: r=0.75; APLAX: r=0.74; SAX r=0.69). Significant difference between evaluator A and B was presented only in 4CH GLS (A: -18.43%; B: -17.60%; p<0.001) and SAX (A: -20.03%, B: -20.63%, p<0.01). The mean absolute errors (MAE) in relevant projections correspond to the following values in %; 4CH: 1.05 (0.53 - 2.18), 2CH: 1.27 (0.62 - 2.13), APLAX: 1.1 (0.43 - 1.73), SAX: 1.73 (0.67 - 2.95). The presence of a difference between both evaluators did not correlate with the datasets image quality. Conclusions: One of the five projections (SAX) showed the constant underestimation of strain values by the less experienced evaluator (MAE <2%). On the other hand, the experienced evaluator had a tendency to overestimate strain values in apical 4CH projection (MAE <2%). In conclusion, the results of complex GLS and GCS STE analyses were not strongly affected by the experience level of the evaluator. Moreover, STE results were not significantly affected even if we used various image quality datasets as an input of the analysis.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.