Researchers use AI to assess patients’ vocals after surgery on the larynx

“For some, the voice changes only slightly, while for others, it can be a life-changing situation. Imagine calling someone on the phone, emergency services, police, etc. – and the one you’re calling does not understand anything. Or even not hear you – as the phone’s noise removal system will cut it out,” said Dr. Rytis Maskeliunas, professor with the Department of Multimedia Engineering, Faculty of Informatics and chief researcher at Kaunas University of Technology Faculty of Informatics.

Often, these types of surgeries require the removal of one or both vocal cords, and sometimes, the entire larynx, to remove a malignant laryngeal tumor. This causes a patient to use the part of the laryngeal structure left after the operation, what the researchers call a “substitute voice.” The team also noted that the quality of the voice and speech declines after the operation, making it more difficult for the patient to communicate.

Automatization of post-operation

The researchers from Kaunas University of Technology Faculty of Informatics (KTU IF) and Lithuanian University of Health Sciences (LSMU) published their findings on using artificial intelligence (AI) to assist in voice analysis in the journal Applied Sciences.

The research team created an artificial intelligence-based algorithm to assess the “substitute voice” of a patient after the excision of the larynx due to cancer. The team was able to determine if a patient had voice pathologies with the assistance of AI signal processing. To automate the post-operative process and analysis, the researchers used AI.

Dr. Maskeliūnas mentioned the uniqueness of the study. “Previous studies have never successfully used artificial intelligence as an expert assistant in voice analysis,” he stated. Due to the innovation, the process after surgery is now automated. This allows for the researchers to install automated screening solutions, allowing medical professionals to follow the course of the disease and diagnose pathologies.

New index to evaluate and assess disease in the larynx post operation

In the study, the researchers announced their discovery, known as the Acoustic Substitution Voice Index (ASVI). The index will allow them to evaluate each patient’s “substitute” voice after oncological laryngitis, or cancer in the larynx. This is done by using acoustic parameters of voice signal combined with artificial intelligence techniques to assess pathologies of the vocal cords.

“Until now, in medical practice, there was no suitable method to objectively and quantitatively evaluate the “substitute” voice. Our proposed ASVI algorithm allows the evaluation process to be done automatically in a concise time,” said Virgilijus Ulozas, professor at the Lithuanian University of Health Sciences (LSMU).

The future goals of using AI for voice analysis

In the future, the team wants this novel discovery to allow patients to perform their own voice tests at home without the need of a medical facility, if there are notable changes in the voiced, allowing patients to notice and evaluate any vocal changes early.  The first clinical trials of the Acoustic Substitution Voice Index have been shown positive results, and researchers hope to one day further develop the ASVI into a mobile app available to specialists and patients, where individuals can upload an audio recording, allowing changes in the voice to be quickly analyzed. This would let the AI system analyze any changes in voice and decide on a treatment option quickly and accurately.