Use of transcutaneous laryngeal ultrasound (TLUSG) in vocal fold examination

Abstract

Recognising recurrent laryngeal nerve (RLN) and thus vocal fold (VF) palsy is important before and after thyroid surgery. Pre-operatively, presence of vocal fold palsy (VFP) suggests local advance malignancy or pre-existing damage to RLNs due to previous operation. Surgeons could then formulae the better surgical plan and take extra-caution in preserving contra-lateral RLN to avoid bilateral VFP and permanent tracheostomy. Post-operatively, early diagnosis of VFP allows voice and swallowing rehabilitation to minimise the risk of silent aspiration. Though flexible laryngoscope allows direct visual assessment of VFs, it is invasive and causes discomfort to patients. Transcutaneous Laryngeal Ultrasound (TLUSG) has been reported as a potential alternative in evaluating VFs movement. This thesis aims to study the feasibility and technique of TLUSG. It also evaluates its clinical application in terms of learning curve and outcome of inaccurate assessment. A prospective assessor-blind study was conducted to evaluate the feasibility and diagnostic accuracy in TLUSG. 204 patients’ VFs were first evaluated by TLUSG and followed by validation laryngoscopy. TLUSG can assess 96.1% pre-operative and 94.6% post-operative VFs. Sensitivity and accuracy in detecting a VFP was 93.3% and 97.4% respectively. To determine which sonographic landmark and what patient should do during TLUSG, two prospective studies were conducted. It was found that identifying all three sonographic landmarks, namely false cord, true cord and arytenoid, was not mandatory. Despite similar accuracy, sonographic landmarks false cord (92.7%) or arytenoid (89.8%) have a higher visualisation rate than true cord (36.7%, p < 0.05%). False cord or arytenoid were the sonographic landmarks to be identified during TLUSG. Similarly, performing all three manoeuvres, namely passive respiration, active phonation and Valsalva manoeuvres, were not necessary. In assessing VF, passive manoeuvre is recommended as it has the highest power in differentiating normal mobile VF from VFP. To order to determine the learning curve, 8 surgical residents without TLUSG experience were recruited. After a short formal training, each of them had performed 80 consecutive TLUSG. Their performance was compared with an experienced TLUSG assessor and analysed with Cumulative Sum (CUSUM) chart. Overall, they could assessed 95.1% VFs with sensitivity and accuracy of 82.5% and 93.6% respectively. By CUSUM analysis, surgeons could master skill in TLUSG after seven examinations and assess VF function consistently and accurately after 40 TLUSG. Despite prior experience, 4-6% of the patients’ VF cannot be assessed by TLUSG and up to 6% of inaccurate results were found. Older age (odds ratio [OR] = 1.06), male gender (OR = 13.66), taller in height (OR = 1.10) and shorter distance from cricoid cartilage to incision (OR = 0.66) were independent factors for “unassessable” VFs, whereas older age (OR = 1.03, p= 0.040) was the only factor for incorrect assessment. Compared to TLUSG-confirmed VFP, TLUSG-missed VFP recovered earlier (69 vs. 125 days, p < 0.001) and fewer patients suffered from permanent VFP (0.0% vs. 34.5%, p = 0.027). In conclusion, TLUSG is feasible to act as a non-invasive and easy-tolearn screening tool to select patients for laryngoscopic examination.