Search

Cochlear implant (CI) is the most effective solution for the restoration of hearing function in severe to profound sensorineural hearing loss. The main objective of this PhD thesis is to answer the question: « Does AI-assisted programming of a cochlear implant provide similar or better results than manual programming, with less time investment for the patient and the clinician? ». A first study that was conducted in 4 French-speaking centers in Belgium and in France showed a large variability of CI programming within and between centers. This context supports the use of a tool using artificial intelligence, such as the Fitting to Outcome eXpert® (FOX®). A pilot study on 2 subjects who had limited results with their manually programmed CI showed encouraging benefits. Secondly, FOX® was used on a larger scale in 2 prospective interventional studies. The first study evaluated the benefits of using FOX® in former patients who had been implanted and formerly manually programmed for several years. 89% of the subjects showed better auditory performances (psychoacoustic battery A§E®) with the FOX® map than with the manual map. This significant but not systematic gain is discussed, considering the impact of FOX® as a novelty among other factors. The second study compared the manual and FOX® types of programming in a randomised longitudinal study with a crossover design in newly implanted "naïve" subjects. The crossover design involved each group changing programming type at one year post-activation. The FOX® group achieved better speech test results more rapidly than the manual group and maintained these results at one year post-activation. Due to the crossover, FOX® improved the results of 81% of the patients initially programmed manually. In terms of time gain, the number of programming sessions decreased by 55% over one year in patients initially programmed with FOX® compared to those initially programmed manually. A general discussion and a conclusion nuance the results. AI-assisted programming provides support to clinicians and shows positive results in terms of hearing performance and time gain. Current clinical practice is heading in the direction of developing advanced technological solutions up to remote treatment, which FOX® will enable. L’implant cochléaire (IC) est la solution la plus efficace, en termes de restauration de la fonction auditive, dans le cadre d’une surdité neurosensorielle sévère à profonde. L’objectif principal de cette thèse de doctorat est de répondre à la question : « La programmation d’un implant cochléaire assistée par l’intelligence artificielle donne-t-elle des résultats semblables voire meilleurs que ceux obtenus par la programmation manuelle, avec moins d’investissement en temps pour le patient et le clinicien ? ». Une première étude réalisée auprès de 4 centres francophones en Belgique et en France rend compte de la grande variabilité des programmations d’ICs intra et inter-centres. Ce contexte justifie l’apport d’un outil utilisant l’intelligence artificielle, comme le Fitting to Outcome eXpert® (FOX®). Une étude pilote sur 2 sujets présentant des résultats limités avec leur IC programmé manuellement relève des bénéfices encourageants. Ensuite, FOX® est utilisé à plus grande échelle dans 2 études prospectives interventionnelles. La première évalue son apport chez d’anciens patients implantés et programmés manuellement depuis des années. 89% des sujets présentent de meilleures performances auditives (batterie psychoacoustique A§E®) avec la map FOX® qu’avec la map manuelle. Ce gain significatif mais non systématique se discute, en tenant compte, entre autres, de l’impact du FOX® en tant que nouveauté. La deuxième compare chez de nouveaux sujets implantés « naïfs » les types de programmation, manuelle et FOX®, dans une étude longitudinale randomisée avec un cross-over. Le design cross-over implique que chaque groupe change de type de programmation à un an post-activation. Le groupe FOX® obtient plus rapidement de meilleurs résultats aux tests vocaux que le groupe manuel, résultats qu’il maintient à un an post-activation. Suite au cross-over, FOX® améliore les résultats de 81% patients initialement programmés manuellement. Au niveau du gain de temps, le nombre de sessions de programmation diminue de 55% sur un an chez les patients initialement programmés avec FOX® par rapport à ceux initialement programmés manuellement. Une discussion générale et une conclusion nuancent les résultats. La programmation assistée par l’intelligence artificielle fournit une aide aux cliniciens et montre des résultats positifs en termes de performance auditive et de gain de temps. La pratique clinique actuelle va dans le sens du développement de solutions technologiques de pointe allant jusqu’au traitement à distance, ce que permettra FOX®. (BIFA - Sciences biomédicales et pharmaceutiques) -- UCL, 2022

Background: COVID-19 has heterogeneous manifestations, though one of the most common symptoms is a sudden loss of smell (anosmia or hyposmia). We investigated whether olfactory loss is a reliable predictor of COVID-19., Methods: This preregistered, cross-sectional study used a crowdsourced questionnaire in 23 languages to assess symptoms in individuals self-reporting recent respiratory illness. We quantified changes in chemosensory abilities during the course of the respiratory illness using 0-100 visual analog scales (VAS) for participants reporting a positive (C19+; n=4148) or negative (C19-; n=546) COVID-19 laboratory test outcome. Logistic regression models identified singular and cumulative predictors of COVID-19 status and post-COVID-19 olfactory recovery., Results: Both C19+ and C19- groups exhibited smell loss, but it was significantly larger in C19+ participants (mean±SD, C19+: -82.5±27.2 points; C19-: -59.8±37.7). Smell loss during illness was the best predictor of COVID-19 in both single and cumulative feature models (ROC AUC=0.72), with additional features providing no significant model improvement. VAS ratings of smell loss were more predictive than binary chemosensory yes/no-questions or other cardinal symptoms, such as fever or cough. Olfactory recovery within 40 days was reported for ~50% of participants and was best predicted by time since illness onset., Conclusions: As smell loss is the best predictor of COVID-19, we developed the ODoR-19 tool, a 0-10 scale to screen for recent olfactory loss. Numeric ratings ≤2 indicate high odds of symptomatic COVID-19 (10