AI advancements may soon allow for early cancer detection through voice analysis, according to researchers' claims.
Artificial Intelligence (AI) is showing promising results in the early detection of throat cancer, particularly laryngeal cancer, by analysing short voice recordings.
Researchers have been piloting voice-based health tools, with one team analysing approximately 12,500 voice recordings from 306 people. The study, published in Frontiers in Digital Health, was led by Phillip Jenkins, a postdoctoral researcher in clinical informatics at Oregon Health and Science University.
The team identified clear differences in harmonic-to-noise ratio and pitch between men with healthy voices, benign lesions, and cancer. AI models trained on over 12,000 voice samples detected variations linked to cancer with encouraging accuracy by analysing voice tone, pitch, and clarity.
Lower harmonic-to-noise ratios indicate vocal fold irregularities, potentially signalling early cancer progression. This approach is non-invasive, faster, and could complement or reduce reliance on invasive diagnostic methods.
However, the current data for women are less conclusive due to smaller sample sizes. No significant patterns were found in women, potentially due to a smaller dataset.
To improve AI model generalizability and accuracy across demographics, the next step is to expand and diversify datasets, particularly increasing female participants. Validating the AI systems through real-world clinical trials is also crucial to assess diagnostic performance and robustness outside research settings.
Addressing ethical sourcing of voice data and ensuring privacy and consent in clinical use is another important consideration. Developing clinically integrated decision support tools that healthcare providers can incorporate in routine check-ups is also a key goal, enabling early screening for laryngeal cancer.
With larger datasets and clinical validation, similar tools for detecting vocal fold lesions might enter pilot testing in the next couple of years. The new AI tool, if further developed, could potentially be better than doctors at diagnosing complicated medical issues, according to Microsoft.
Early detection of laryngeal cancer is crucial because it significantly improves survival rates and treatment outcomes. Smoking, alcohol use, and certain strains of HPV are key risk factors for laryngeal cancer. One of the most common warning signs for laryngeal cancer is hoarseness or changes in the voice that last more than three weeks.
Cancer of the voice box, or larynx, affects more than a million people worldwide and kills roughly 100,000 every year. Survival rates for laryngeal cancer vary from around 35% to 90% depending on how early the disease is diagnosed, according to Cancer Research UK.
The findings could support efforts to find an easier, faster way to diagnose cancerous lesions on the vocal cords. Pilot clinical testing of AI voice analysis tools is anticipated within the next few years, potentially making early, non-invasive throat cancer screening more accessible and affordable worldwide.
[1] Jenkins, P., et al. (2022). Non-Invasive Detection of Laryngeal Cancer Using Harmonic-to-Noise Ratio and Pitch Analysis of Short Voice Recordings. Frontiers in Digital Health.
[3] Microsoft. (2022). AI Shows Promise in Early Detection of Throat Cancer. [online] Available at: https://www.microsoft.com/en-us/research/blog/ai-shows-promise-in-early-detection-of-throat-cancer/
[5] Science Daily. (2022). AI Shows Promise in Early Detection of Throat Cancer. [online] Available at: https://www.sciencedaily.com/releases/2022/03/220317130240.htm
- In the realm of health and wellness, artificial intelligence (AI) is demonstrating potential in the early detection of laryngeal cancer, a type of throat cancer, by analyzing short voice recordings.
- Researchers are exploring the use of voice-based health tools, with studies showing that AI models, trained on over 12,000 voice samples, can detect signs of cancer with remarkable accuracy, by analyzing voice tone, pitch, and clarity.
- To enhance the accuracy and generalizability of AI models across demographics, particularly women, it is crucial to expand and diversify datasets, conduct real-world clinical trials, and address ethical considerations regarding voice data.
