Turkish scientists detect early-stage lung cancer through voice analysis

A team of Turkish researchers has pioneered a remarkable method to identify early-stage lung cancer by analyzing the hidden, subtle nuances of human speech, harnessing the power of artificial intelligence.

A research member, Dr. Yusuf Kahya, from Ankara University's Faculty of Medicine’s Thoracic Surgery Department, announced that their study achieved an accuracy rate of approximately 90 percent, daily Hürriyet reported.

Speaking at the Turkish Thoracic Society Congress held in the southern city of Antalya, Kahya explained that by analyzing vocal features imperceptible to the human ear through artificial intelligence, they were able to detect early signs of lung cancer.

The groundbreaking research was conducted under the leadership of Dr. Haydar Ankışhan from Ankara University, with contributions from doctors Haluk Ulucanlar, İslam Aktürk, Kübra Alphan Kavak, Ulaş Bağcı and Bülent Yenigün, along with Kahya.

The study involved multiple parameters and compared the speech of 50 healthy individuals with 50 lung cancer patients.

Kahya provided further details about the study conducted by a large research team, underlining the significance of early detection of lung cancer and highlighting the growing role of AI in medicine.

“Lung cancer remains the leading cause of cancer-related deaths worldwide. The main challenge continues to be late diagnosis,” he said. Many studies in the field of lung cancer focus on early detection and numerous methods are being developed for this purpose, he added, noting that AI has become widely used in this direction.

“In collaboration with researchers from Ankara University and Health Sciences University, we developed an AI-based program that analyzes everyday speech to detect early-stage lung cancer. Our program successfully distinguished between healthy individuals and those with lung cancer with a high degree of accuracy."

Kahya pointed out that the analysis of speech features, similar to how a person's voice changes when they have a cold or the flu, can reveal health conditions that are not detectable by the human ear.

He noted that while the findings of their small-scale study are promising as a potential new diagnostic method, the technology is not yet ready for clinical use and requires further development.

Although 90 percent is a significant accuracy rate, the method needs to be validated with larger groups. If a particular success rate is attained, the new technology may eventually be employed as a long-term diagnostic technique, according to Kahya.