Turkish scientists’ AI app detects early-stage lung cancer from voice with 90% accuracy

Turkish researchers have developed an artificial intelligence-supported application capable of detecting lung cancer at an early stage by analyzing a person’s voice.

Led by Dr. Haydar Ankishan from the Stem Cell Institute at Ankara University, and involving scientists from a range of disciplines, the project focuses on the early diagnosis of lung cancer — a disease typically identified only after reaching advanced stages.

Voice structure reveals early warning signs

Speaking to reporters, Dr. Ankishan explained the innovative foundation of their work.
“In our study, we considered the structure of the voice, the anatomical features of the lungs, and the circulatory system,” he said. “We proposed that the voice could reveal vital information about the presence of lung cancer. We pursued this idea.”

Following nearly 18 months of research, the team achieved remarkable results.
“We have found that lung cancer, particularly in its first stage, can be diagnosed early with an accuracy rate of over 90 percent,” Dr. Ankishan stated.

Two-minute voice recordings yield vital clues

Dr. Bulent Mustafa Yenigun, a faculty member at Ankara University’s School of Medicine and part of the research team, emphasized the life-saving potential of early detection.

“The most critical factor in treating lung cancer is identifying it as early as possible,” he said.

The research involved a group of 100 individuals — 50 patients diagnosed with lung cancer and 50 healthy participants. Each subject read from a specially prepared two-minute speaking script, with their voices recorded for analysis.

“All recordings were made and analyzed under identical conditions,” Dr. Yenigun added. “Our application detects subtle changes in the voice and alerts users when there are indications that pathological lung cancer might be present.”