Search

Thyroid nodules are usually explored non-invasively by ultrasonography followed by an invasive cytopunction in case of suspected cancer. Echogenicity of the thyroid nodule is a key parameter to differentiate benign and malignant nodules but its visual determination lacks reproducibility. The objective of this r.tudy was to define and evaluate a standardized measurement cf thvroid natures' echogenicity. In this monocentric study, ultrasound data .(2D recorded images, echogenicity and EU-TIRADS (European Thyroid Imaging Reporting And Data System) score determined visually by an experienced physician) and cytological results of cytopunctions performed between June 2016 and January 2019 were retrieved retrospectively. A manual segmentation on the 20 recorded images of the nodules, the healthy thyroid and the cervical muscles was performed to measure the nodules' echogenicity normalized by the healthy thyroid (Nt) and the muscles (Nm). Spearman's correlations were used to compare the parameters and a comparative analysis of diagnostic performance by ROC (Receiving Operator Characteristic) curves by using cytological Bethesda results as reference. 201 patients (224 nodules) with thyroid cytopunctions were included and analyzed. Concerning the Spearman's correlations, the measured parameters were moderately correlated to the EU-TIRADS (maximal p=-0,60, p

Thyroid nodules are usually explored non-invasively by ultrasonography followed by an invasive cytopunction in case of suspected cancer. Echogenicity of the thyroid nodule is a key parameter to differentiate benign and malignant nodules but its visual determination lacks reproducibility. The objective of this r.tudy was to define and evaluate a standardized measurement cf thvroid natures' echogenicity. In this monocentric study, ultrasound data .(2D recorded images, echogenicity and EU-TIRADS (European Thyroid Imaging Reporting And Data System) score determined visually by an experienced physician) and cytological results of cytopunctions performed between June 2016 and January 2019 were retrieved retrospectively. A manual segmentation on the 20 recorded images of the nodules, the healthy thyroid and the cervical muscles was performed to measure the nodules' echogenicity normalized by the healthy thyroid (Nt) and the muscles (Nm). Spearman's correlations were used to compare the parameters and a comparative analysis of diagnostic performance by ROC (Receiving Operator Characteristic) curves by using cytological Bethesda results as reference. 201 patients (224 nodules) with thyroid cytopunctions were included and analyzed. Concerning the Spearman's correlations, the measured parameters were moderately correlated to the EU-TIRADS (maximal p=-0,60, pDans cette étude monocentrique, les données échographiques (images, échogénicité visuelle, score EU-TIRADS par médecin expérimenté) et cytologiques (Bethesda) ont été recueillies de façon rétrospective , entre juin 2016 et janvier 2019. Une segmentation manuelle sur ces images de la thyroïde saine, des muscles cervicaux et des nodules a été réalisée afin de mesurer l'échogénicité nodulaire normalisée par la thyroide saine (Nt) et les muscles (Nm). Un test de corrélation entre ces mesures et le score EU-TIRADS puis une analyse comparative des performances diagnostiques par courbes ROC ont été réalisés en utilisant pour référence les cytologies. 201 patients avec cytoponction étaient inclus (224 nodules) et analysés. Les paramètres étudiés étaient modérément corrélés (valeur max p=-0.60 pour Nt) mais de façon significative avec le score visuel EU-TIRADS (p