Predicting Swallowing Problems During Head and Neck Cancer Radiotherapy Using Ultrasound of the Throat Structures A Study at Hopital Forcilles

What this Trial Is About

Radiotherapy for head and neck cancers increases the risk of swallowing disorders, which can negatively impact nutrition, quality of life, and overall health, leading to higher hospitalization and mortality rates. Early intervention is important and requires precise assessment tools to predict swallowing difficulties, identify affected structures, and guide targeted rehabilitation. Current clinical exams lack precision, and existing complementary methods like videofluoroscopy or fiberoptic endoscopic evaluation are invasive, expose patients to radiation, require many resources, and have unknown prognostic value. Ultrasound imaging offers a non-invasive way to evaluate the morphology and movement of swallowing-related structures such as the tongue, hyoid bone, and suprahyoid muscles. This study plans to develop and assess an ultrasound-based predictive model for swallowing disorders during radiotherapy for head and neck cancer. Ultrasound exams will be done before treatment, and after 7 and 14 days, with blinded operators performing the assessments. Weekly clinical and swallowing tests will also be done by speech therapists, with videofluoroscopy if swallowing disorders are suspected. Participants will include 124 outpatients undergoing radiotherapy over 18 months. Researchers will collect demographic, clinical, ultrasound, and videofluoroscopy data, and monitor patients monthly. They will evaluate how well ultrasound markers predict swallowing disorders and undernutrition, assess measurement reliability, and relate ultrasound findings to quality of life. The primary outcome is the ability of ultrasound to predict swallowing disorders on the first day of radiotherapy. Data analysis includes Cox models and logistic regression to develop predictive scores and assess accuracy, sensitivity, and specificity.

Radiation-Related Dysphagia Development Prediction Using a Two-Step Ultrasonographic Model (R-2D-2)