Communication After Laryngectomy and in Chronic Hoarseness: Piloting Current Limitations and the Prospects of AI-based Voice Restauration


Objective:
Individuals who have undergone total laryngectomy or suffer from chronic hoarseness (e.g., due to vocal fold paralysis, benign lesions, or functional dysphonia etc.) often experience severe impairments in verbal communication, leading to frustration, social withdrawal, and reduced quality of life (QoL). Conventional methods of voice rehabilitation, after laryngectomy such as tracheoesophageal prostheses, esophageal speech, and electrolarynx, often lack naturalness of voice sound and intelligibility. This study examines the relationship between QoL, communication ability, and psychosocial factors and evaluates the perceived potential of voice conversion (VC) to improve naturalness and communication confidence.

Methods:
Data collection includes questionnaires on QoL and communicative participation (German translations of namely the Swedish Self Evaluation of Communication Experiences after Laryngectomy (S-SECEL), the University of Washington Quality of Life Questionnaire (UW-QOL), and the Communicative Participation Item Bank (CPIB) and as well as pilot instruments addressing current telecommunication use and expectations regarding VC. Speech recordings are conducted under standardized and interactive conditions to establish a high-quality dataset for training and evaluation of VC models. In a later stage, the trained VC system will be applied to transform pathological or substitute voices into natural-sounding synthetic voices for exploratory listening assessments.

Results:
Preliminary findings indicate marked limitations in QoL, particularly in the speech domain of the UW-QOL. Telecommunication emerged as the least used and most challenging form of interaction, while face-to-face communication in noisy environments was also reported as difficult. Participants frequently expressed frustration with current communication methods and rather low confidence in speech-based interactions. QoL appeared linked to psychosocial factors such as frustration and communicative confidence. Participants anticipated that AI-based VC could improve speech naturalness, intelligibility, and overall QoL.

Conclusion
These preliminary results highlight the central role of psychosocial and communicative dimensions in voice rehabilitation. VC may represent a promising pathway to restore speech naturalness, strengthen communicative confidence, and enhance QoL in individuals with chronic hoarseness or after laryngectomy.

Sarang
Veronika
Vincent
Benedikt
Martin
Philipp
Kim
Kranebitter
Aigner
Mayrhofer
Hagmüller
Aichinger