Artificial Intelligence is Capable of Identifying Signs of Cancer in a Person's Voice Prior to Medical Determination
In a groundbreaking development, a new project named 'Bridge2AI-Voice' is transforming the landscape of biomedical research within the US National Institute of Health's 'Bridge to Artificial Intelligence' (Bridge2AI) consortium. This nationwide endeavour aims to apply AI to complex biomedical challenges, with a particular focus on early detection of laryngeal cancer.
The study, led by Dr. Phillip Jenkins, a postdoctoral fellow in clinical informatics at Oregon Health & Science University, has identified several key acoustic features that could potentially revolutionize the way we screen for laryngeal cancer.
By analysing voice recordings, AI models can detect subtle acoustic changes caused by vocal fold lesions, including early-stage carcinoma. This non-invasive method could lead to earlier diagnosis and better patient outcomes, as it enables screening and monitoring before clinical symptoms or invasive procedures like endoscopy and biopsy are required.
The most useful features for distinguishing pathological voices from healthy ones include the harmonic-to-noise ratio (HNR), variations in pitch (fundamental frequency), amplitude variation, tone clarity, and noise levels. In particular, the harmonic-to-noise ratio, especially its variability, has shown promise as a voice-based marker for early detection and monitoring of vocal fold lesions.
However, more research is needed with larger, more diverse populations to refine these features and validate their clinical utility. Particularly, researchers need to gather more data, particularly on women, to move towards an AI tool that recognizes vocal fold lesions.
If refined and validated, a voice-based health tool could potentially enter pilot testing in the next couple of years. This development could make early screening for laryngeal cancer vastly easier, potentially catching it after a few seconds of speaking rather than months of hoarseness.
It's important to note that currently, nasal endoscopy or biopsy are the gold standards for diagnosing laryngeal cancer, but they are invasive and only recommended after symptoms appear or a clear abnormality is found. The idea of detecting cancer through voice recordings is part of a growing movement in medicine to find biomarkers in everyday signals.
The study has been published in the journal Frontiers in Digital Health, providing a significant step forward in the fight against laryngeal cancer. Lingering hoarseness can be an early warning sign of laryngeal cancer, and this AI-powered approach could offer a scalable, simple screening tool that could triage patients for further diagnostic procedures.
The prognosis for laryngeal cancer treatment is variable, with a 35% to 78% survival rate over five years. Early detection and treatment could significantly improve these figures, making this research a promising development in the field of oncology.
- This groundbreaking project, Bridge2AI-Voice, leverages artificial intelligence to address complex medical-conditions like early detection of laryngeal cancer within the science field.
- In the realm of health-and-wellness, research has identified key acoustic features such as the harmonic-to-noise ratio, pitch variations, amplitude variation, tone clarity, and noise levels that could revolutionize the screening process for laryngeal cancer.
- If refined and validated, this voice-based health tool, stemming from the intersection of science, medicine, and technology, could enter pilot testing and potentially offer a simpler, scalable screening method for laryngeal cancer.
- Given the variable prognosis for laryngeal cancer treatment with a 35% to 78% survival rate over five years, the potential for early detection and treatment, through research like Bridge2AI-Voice, can significantly improve these figures, marking a promising development in the field of oncology.
