Lunit has demonstrated that artificial intelligence (AI) can improve the detection rate of breast cancer, with its study published in Nature Communications.
On the 10th, Lunit announced that the results of a large-scale, multi-center prospective study using its AI-powered mammography analysis solution, Lunit INSIGHT MMG, had been published in the international journal Nature Communications.
This study is a large-scale prospective investigation conducted in a "single reading" environment, where a single radiologist interprets mammograms, as seen in countries like South Korea and the United States. The research aimed to validate the effectiveness of AI in such settings.
Lunit demonstrated that AI enhances diagnostic accuracy without increasing unnecessary follow-up examinations.
Led by Professor Jang Yoon-woo from the Department of Radiology at Soonchunhyang University Seoul Hospital and Professor Ryu Jung-kyu from Kyung Hee University Gangdong Hospital, the study was conducted from February 2021 to December 2022. It involved 24,543 women aged 40 and older who underwent national health screening at six major university hospitals in South Korea: Soonchunhyang University Seoul Hospital, Kyung Hee University Gangdong Hospital, Nowon Eulji University Hospital, Konkuk University Hospital, Inje University Busan Paik Hospital, and CHA University Bundang CHA Hospital.
The research team compared mammogram interpretations by breast radiologists alone with those assisted by AI, following up with patients for one year to analyze the results.
The findings showed that AI-assisted readings increased the breast cancer detection rate (CDR) to 5.7 per 1,000 cases, a 13.8% improvement compared to 5.0 per 1,000 cases without AI.
Furthermore, an additional simulation study revealed that when general radiologists utilized AI, the breast cancer detection rate improved from 3.9 per 1,000 cases to 4.9 per 1,000 cases, marking a 26.4% increase.
AI was particularly effective in detecting small tumors (less than 20mm) and early-stage breast cancers without lymph node metastasis.
Professor Jang Yoon-woo, the study's lead author, stated, “This study confirms that AI can improve breast cancer detection rates in a ‘single reading’ environment. It benefits not only experienced breast radiologists but also general radiologists with less experience in mammogram interpretation. AI has the potential to contribute to the early detection of breast cancer.”