HEALTH

How Your Voice Could Help Spot Throat Cancer

Portland, Oregon, USAThu Aug 14 2025

Artificial intelligence might soon be able to detect throat cancer just by listening to a person's voice. This is according to a recent study that focused on men with larynx cancer. The larynx, also known as the voice box, is where the vocal cords are located. When cancer affects this area, it can change the way a person sounds.

Key Findings

Researchers found that men with larynx cancer have unique differences in their voices. These differences are caused by abnormal growths in the vocal cords. The study used over 12,500 voice recordings from people across North America. The AI analyzed these recordings and found specific patterns that could indicate cancer.

Unique Voice Patterns

One key discovery was that men with larynx cancer had a different harmonic-to-noise ratio in their voices. This ratio measures the amount of noise in a person's speech. The AI also noticed differences in the pitch of their voices. These findings suggest that voice analysis could be a simple and non-invasive way to detect throat cancer early.

Importance of Early Detection

Early detection is crucial because it can significantly improve survival rates. The study found that the five-year survival rate for throat cancer can be as high as 78% if caught early. However, it drops to 35% if detected late. This makes early diagnosis a critical factor in treating the disease.

Future Research

The study focused on men, but researchers hope to expand their findings to include women. They plan to gather more data and test the AI in clinical settings. This could lead to the development of voice-based tools that can help detect throat cancer in its early stages.

Next Steps

While the current findings are promising, more research is needed. The AI needs to be trained with a larger dataset and tested thoroughly. If successful, this technology could revolutionize how throat cancer is detected and treated.

questions

    How does the AI differentiate between benign vocal cord lesions and malignant ones, and what is the success rate of this differentiation?
    If the AI can detect cancer by listening to voices, will it start giving unsolicited medical advice during phone calls?
    How accurate is the AI in detecting laryngeal cancer in women, and what steps are being taken to improve this accuracy?

actions