Validation and Optimization Methods for a Quantitative Characterization of Motion Abnormalities of the Larynx With Echography

The rationale of this study is that Transcutaneous Laryngeal Ultrasonography (TLU) may be a relevant and reliable tool in assessing vocal folds (VF) paralysis secondary to endocrine surgery (thyroid and parathyroid surgery). According to the literature and to our own publications, subjective impression of VF (im)mobility seems sufficient to diagnose a paralysis. However, the investigators think that objective measures, based on a dedicated semi-automatic software, may offer a much powerful implement and needs to be compared to the gold standard technique, i.e. the nasofibroscopy (NF), in a large prospective multicentric cohort.

In a preliminary study, the investigators described TLU anatomical landmarks first in an ex vivo study and secondary in a prospective study based on 50 patients with VF paralysis compared to 50 controls. These three landmarks, the two arytenoids and the anterior insertion of the two vocal folds on the thyroid cartilage, are easy to recognize, and independent of the postoperative delay and the person performing TLU. They define two symmetrical hemi-larynges in healthy subjects. The investigators designed an original software, aiming at offering two quantitative criteria of laryngeal mobility and symmetry, taking into account motion symmetry (symmetry index) and amplitude (mobility index) of the two hemi-larynges. When a subject presented with a unilateral VF paralysis, the mean symmetry index was significantly larger compared to the control group (20.4% ±13.8% versus 6.7% ±4.7%, p<0.0001) and the mean lower mobility index was significantly smaller (2.2% ±17.4% versus 20.5% ±14.8%, p<0.0001). The sensitivity and specificity of this first version of our automatized protocol were 82%.

The aim of the present study is to validate TLU as an alternative to NF in a large independent cohort of patients operated on endocrine surgery. The main goal is to improve the sensitivity and specificity of the objective measures and implement new parameters in the software. By a ready to use tool, the investigators would like to offer the opportunity to any practitioner to assess and quantify VF (im)mobility with their usual US machine.

The subjects will be recruited in the three different clinical centers, highly specialized in endocrine surgery. The aim of 500 subjects included seems feasible thanks to the high activity of these centers. Moreover the latter already have sufficient and qualified staff and technical tools to perform clinical follow up, US, and NF.

The technical acquisition is as follow. The TLU is performed with a 7-14 MHz linear probe, using a portable machine. Subjects are lying on the back, neck slightly extended, with the probe placed over the laryngeal prominence of the thyroid cartilage, including the arytenoids in the axial view. The images are acquired in B-mode, during normal breathing at rest. Typical values for acquisition are: frequency 10 MHz, depth of field of view 4 cm, mechanical index 0.6, gain 45%, dynamic range 85 decibels. A video clip of 10 seconds (30 images per second) is recorded in order to include 5 to 6 complete respiratory cycles. Two images corresponding to physiological abduction and adduction during the same breath cycle are selected. The landmarks, previously described, are selected on the images and tracked by the software.