The surgical treatment of voice disorders has been largely developed through a trial-and-error approach. A newly conceived procedure is tried on a few patients and is considered successful if those patients have good outcomes. A more rational approach to surgical design and planning would reduce undesirable outcomes and shorten the time to develop successful surgical techniques. In principle, computational voice simulation can help surgeons predict how changes to vocal fold anatomy lead to changes in the voice. However, current simulation approaches are limited. Most generate only a single voice output based on a fixed vocal fold anatomy. What would be more useful to surgeons is a way to identify the vocal fold anatomy required to produce a desired voice outcome. The objective of this proposal for translational research is to develop a computational tool for surgeons to predict the vocal fold anatomy required to meet a particular vocal demand. The proposed Phonosurgery Optimizer-Simulator (PHONOS) software performs several hundred rounds of voice simulations, with each round producing vocal outputs closer to the desired vocal target than the previous round. In the end, the surgeon will have a set of vocal fold anatomies that produce voice that satisfies the patient's vocal priorities. This will (1) enable a surgeon to compare different surgical options to determine which better meets the patient's vocal priorities and (2) inspire the design of new phonosurgical procedures based on the suggested vocal fold anatomies.
Three specific aims are proposed.
In Aim 1, a Vocal Priority Questionnaire will be developed and validated. This instrument will ask the patients to rate the importance of pitch, loudness, vocal endurance, and vocal clarity in their daily activities. The vocal attributes considered to be more important could then be more heavily favored or weighted during computation.
In Aim 2, the NCVS Voice Simulator, which forms the core of PHONOS, will be adapted for use by surgeons. A series of graphical user interfaces will replace command line inputs to allow intuitive manipulation of vocal fold morphologic parameters in three dimensions.
In Aim 3, the simulator will be integrated with an optimization engine to enable efficient searches for vocal fold anatomies that produce the target vocal outcome. The optimization scheme makes it possible to preferentially weigh vocal attributes considered to be more important by the patient. The software thus developed will enable surgeons to evaluate the relative merits of existing surgical options, influence surgical decision making based on patient preference, and stimulate the design of novel surgeries for vocal folds. This work will shift the paradigm of phonosurgery from a focus on restoring normal vocal fold anatomy to improving vocal function based on patient priorities, thereby transforming decision making in the surgical treatment of voice disorders.
The proposed research to develop computer simulation for surgeons to determine the best treatment for a voice problem is relevant to public health because 30% of the population will experience a voice problem at some point in their lifetimes, causing significant burden on communication, quality of life and productivity. Enhancing surgical design and planning with the proposed software is an example of translating the advances in basic voice science to positive impact on human health, a core mission of the NIH and an explicit objective of the National Institute on Deafness and other Communication Disorders.
| Smith, Simeon L; Titze, Ingo R (2018) Vocal fold contact patterns based on normal modes of vibration. J Biomech 73:177-184 |
