Polycystic Ovary Syndrome (PCOS), a common hormone disorder affecting women between the ages of 15 and 45, is notoriously challenging to diagnose due to its overlapping symptoms with other conditions. However, a recent study conducted by the National Institutes of Health (NIH) reveals that artificial intelligence (AI) and machine learning (ML) can play a pivotal role in effectively detecting and diagnosing PCOS.
PCOS is a disorder characterized by ovaries that do not function properly and are often accompanied by elevated testosterone levels. Its symptoms include irregular periods, acne, excessive facial hair, and hair loss from the head. Additionally, women with PCOS are at an increased risk of developing type 2 diabetes, along with various other health complications, such as sleep disorders, psychological issues, cardiovascular problems, uterine cancer, and infertility.
Dr. Janet Hall, a senior investigator and endocrinologist at the National Institute of Environmental Health Sciences (NIEHS), and a co-author of the study, emphasized the significance of this research: “Given the large burden of under- and mis-diagnosed PCOS in the community and its potentially serious outcomes, we wanted to identify the utility of AI/ML in the identification of patients that may be at risk for PCOS.” She further stated, “The effectiveness of AI and machine learning in detecting PCOS was even more impressive than we had thought.”
The research conducted by the NIH involved a systematic review of published scientific studies that utilized AI/ML to analyze data for the purpose of diagnosing and classifying PCOS. The findings were promising, demonstrating that AI/ML-based programs were successful in detecting PCOS, which often eludes conventional diagnostic methods.
One of the primary challenges in diagnosing PCOS lies in its symptoms, which can overlap with those of other disorders. Dr. Skand Shekhar, senior author of the study and an assistant research physician and endocrinologist at the NIEHS, highlighted this issue: “PCOS can be challenging to diagnose given its overlap with other conditions.” Dr. Shekhar suggested that incorporating AI/ML into electronic health records and clinical settings could significantly improve the diagnosis and care of women with PCOS.
Diagnosing PCOS typically relies on standardized criteria that encompass clinical features (e.g., acne, excess hair growth, and irregular periods), laboratory findings (e.g., high blood testosterone), and radiological results (e.g., ovarian ultrasound). However, these features can co-occur with other disorders such as obesity, diabetes, and cardiometabolic conditions, making PCOS often unrecognized or misdiagnosed.
AI encompasses computer-based systems designed to mimic human intelligence and assist in decision-making, while ML, a subset of AI, focuses on learning from past data to inform future decisions. The ability of AI to process vast amounts of diverse data, including electronic health records, positions it as an ideal tool for diagnosing challenging conditions like PCOS.
The NIH researchers conducted a comprehensive review of peer-reviewed studies published in the past 25 years, from 1997 to 2022, that utilized AI/ML to detect PCOS. With the assistance of an experienced NIH librarian, they identified 135 potentially eligible studies, ultimately including 31 in their analysis. All these studies were observational and evaluated the use of AI/ML technologies in diagnosing PCOS, with roughly half of them incorporating ultrasound images. The average age of the study participants was 29.
Among the 10 studies that employed standardized diagnostic criteria for PCOS diagnosis, the accuracy of detection ranged from an impressive 80% to 90%. Dr. Shekhar emphasized this notable result, stating, “Across a range of diagnostic and classification modalities, there was an extremely high performance of AI/ML in detecting PCOS, which is the most important takeaway of our study.”
The study authors emphasized that AI/ML-based programs have the potential to significantly improve the early identification of PCOS in women, leading to cost savings and a reduced burden on both patients and the healthcare system. To ensure the successful integration of AI/ML for chronic health conditions like PCOS, follow-up studies with robust validation and testing practices will be essential.
The groundbreaking research conducted by the NIH underscores the transformative potential of AI and ML in the detection and diagnosis of PCOS, a condition that has long posed diagnostic challenges. With the continued development and validation of AI/ML-based tools, healthcare providers may soon have more effective means to identify and manage PCOS, ultimately improving the lives of countless women affected by this hormone disorder.
