Polycystic Ovary Syndrome (PCOS) is a global concern for 10% – 20% of women in their reproductive age. It can also contribute to the development of Type 2 Diabetes, Cardiovascular Disease, Endometrial Cancer and various Psychological Disorders.
The diagnostic delay for PCOS is not for lack of diagnostic criteria. It is a consequence of heterogeneous symptom profiles, fragmented care pathways and a healthcare system that has historically underinvested in women’s health conditions. However, Artificial Intelligence is showing considerable improvements and now a large number of investors have shown significant interest.
What are the Rates of PCOS in the UK?
Based on the latest population estimates, there are over 3 million women with PCOS in the UK. The incidence of PCOS diagnoses in UK primary care has risen significantly in recent years, a trend that likely reflects both genuine increases and improving awareness among clinicians.The demographic with the highest estimates of PCOS prevalence is women aged 20-24 and women living in areas of high deprivation.
Can AI and Machine Learning Diagnose PCOS Accurately?
AI-assisted PCOS detection is promising, but limitations exist. A 2023 National Institute of Environmental Health Sciences (NIEHS) study found that AI and Machine Learning could diagnose PCOS with an accuracy of 80% – 90% when standard diagnosis criteria were used.
The researchers found that AI performed ‘extremely’ well at multiple diagnoses and classifications. Additionally, a 2024 study published in Frontiers in Endocrinology reported 85% AUC on average when machine learning was employed to predict PCOS before a diagnosis was provided.
More from Artificial Intelligence
- OpenAI Trained Its AI To Never Talk About Goblins And The Internet Has Questions
- OpenAI Will Now Be Using AWS For Its Models – What Does This Mean For Users?
- AI In The Shadows? Unofficial And Unapproved AI May Already Be Powering Your Business
- What Will Monetisation Look Like For Creators In The AI Era?
- Google Partners With SAP: What Is A Multi-Agent AI?
- Meta Will Be Using AI Chat Controls To Manage Parental Controls – How Will It Work?
- It May Not Be Intentional, But AI Bias Is Real, And It’s Already Distorting Real-World Outcomes: Experts Comment
- How Do NVIDIA Chips Power Alibaba’s AI Car Systems?
Real Clinical Deployment Remains the Challenge
Many AI models do not take into account diversity when training the algorithms on data coming from large projects. The results from many AI models do not reach clinically acceptable levels of reliability, which hinders their deployment. The large bias in data used to build AI models, coupled with methodologies that focus on maximising discrimination metrics like AUC, strongly impact clinical reliability.
Top AI Trends for PCOS Right Now
From electronic health record mining to flag women who have reports of common symptoms of PCOS to AI-powered ultrasounds interpreting known PCOS symptoms quicker:
Electronic Health Record (EHR) Mining
One of the options that has a better chance of successful near-term application is using machine learning to analyse EHR data and potentially identify women with previously diagnosed PCOS. A patient with infrequent menstrual cycles with accompanying androgen markers and a family history of diabetes would likely never be investigated for PCOS in the current system. However, a properly designed AI system would be able to identify this woman and provide justification for PCOS testing.
AI-Powered Ultrasound Interpretation
PCOS) builds fluid filled cysts on the ovaries that can be diagnosed via Ultrasound Imaging. Unfortunately, the precision of Ultrasound Imaging varies from operator to operator, however, AI can help improve the consistency of PCOS diagnosis via ultrasound. Improved ultrasound PCOS imaging may become more accessible to the general population thanks to transfer-learning models, which can train the AI on ultrasound imaging datasets frameworks, in combination with advanced PCOS symptoms imaging.
PCOS Diagnosis via AI-Driven Health Apps
There has been mammoth interest in tracking PCOS related symptoms using AI. Many AI applications can score risk of PCOS and other PCOS-related symptoms based on the self reported menstrual, behavioural and health data (such as mood and weight, and physical and mental changes such as acne, hair loss or hair growth). These applications are not clinical tools, however, can help clinicians make better decisions through tracked data.
Personalised Treatment and Management Support
AI applications extend beyond a PCOS diagnosis. After diagnosis, managing a lifelong complex condition is similarly challenging. AI tools are designed to develop personalised treatment guidance, e.g. who may benefit from lifestyle change vs. who may benefit from pharmacological management, predicting response to ovulation induction and longer-term risk management for metabolic complications.
How Do Japanese Treat PCOS?
The Japanese framework takes a structured, stepwise approach: lifestyle modification and weight reduction are mandated as the first-line intervention for overweight patients, before any pharmacological treatment is considered. This lifestyle-first philosophy aligns with the international evidence base, which consistently finds that even modest weight reduction improves insulin sensitivity, ovulatory function and hormonal balance in overweight women with PCOS.
What the Next Funding Round Looks Like for PCOS Startups
The FemTech market is expanding, but at an unequal pace. In 2024, the FemTech market reached 55% global growth and was the fastest emerging digital health market segment. Analysts predict FemTech will grow to $206 billion by 2033. Just in March 2026, FemTech startups raised $72 million. In order to attract growth stage investments, PCOS startups need to prove distinct clinical differentiation and the most viable approach towards implementation of their offerings than they did in the past.
Recent Highlighted Investments
In July 2025, Solence, an AI and PCOS digital therapeutic, was able to raise €1.6million in seed funding. Solence was founded by Clara Stephenson who, despite having textbook symptoms of PCOS, was undiagnosed for 10 years. LEVY Health, an intelligence amplification firm, was recently nominated for the top FemTech startups to invest in and in 2024, secured $.5 million in seed funding. LEVY Health aims to reduce PCOS diagnosis times from years to weeks with its intelligence amplification software.
What Investors Want to See
The investors backing PCOS and broader women’s health AI tools in 2025 and 2026 are increasingly focused on clinical validation and real-world integration rather than impressive demo-day accuracy metrics. The FemTech funding landscape has matured enough that the conversation has shifted from ‘can AI do this?’ to ‘can AI do this in a way that integrates with NHS or insurance systems, achieves regulatory clearance and actually reaches the women who need it?’
