AI Accurately Diagnoses Rare Hormone Disorder from Hand Images

A new artificial intelligence (AI) system developed by researchers at Kobe University can accurately diagnose acromegaly, a rare and often delayed-diagnosis endocrine disorder, simply by analyzing images of the back of the hand and a clenched fist. This privacy-conscious approach promises to improve referral systems and reduce healthcare disparities.

Understanding Acromegaly

Acromegaly is a rare, progressive condition typically occurring in middle age, characterized by the abnormal growth of hands and feet, changes in facial features, and effects on bone and organ growth. It’s caused by the overproduction of growth hormone. If left untreated, acromegaly can lead to life-threatening complications and reduce life expectancy by approximately 10 years. Kobe University researchers note that diagnosis is often delayed, frequently taking up to a decade due to the slow progression of the disease and its relative rarity.

The AI Diagnostic Tool

Traditionally, diagnosing acromegaly relies on a combination of clinical evaluation, hormone level testing, and imaging scans. Although, researchers at Kobe University have pioneered a new method leveraging the power of AI. The team, led by endocrinologist FUKUOKA Hidenori, focused on analyzing images of the back of the hand and clenched fist, avoiding the privacy concerns associated with facial recognition technology. Eurekalert reports that this approach was chosen because hands are routinely examined during clinical assessments for signs of the condition.

The AI model was trained and validated using over 11,000 images donated by 725 patients across 15 medical facilities in Japan. Published in the Journal of Clinical Endocrinology & Metabolism, the research demonstrates that the AI model achieves high sensitivity and specificity in diagnosing acromegaly, even outperforming experienced endocrinologists when evaluating the same images. Kobe University highlights that this level of accuracy achieved without facial features makes the approach particularly practical for broader disease screening.

Privacy and Future Applications

A key advantage of this AI system is its focus on patient privacy. By utilizing images of the hands, the researchers circumvent the ethical and logistical challenges associated with collecting and analyzing facial data. OHMACHI Yuka, a Kobe University graduate student, explained that the team specifically chose to focus on the hands to address these privacy concerns. Eurekalert

Looking ahead, the Kobe University team plans to expand the model’s capabilities to detect other conditions identifiable through hand images, such as rheumatoid arthritis, anemia, and finger clubbing. They envision this technology as a tool to complement clinical expertise, reduce diagnostic errors, and facilitate earlier intervention. The goal is to create a medical infrastructure that connects suspected cases to specialists and supports non-specialist physicians in regional healthcare settings, thereby reducing healthcare disparities. SciTechDaily

Key Takeaways

• An AI developed by Kobe University can diagnose acromegaly from hand images with high accuracy.
• The system prioritizes patient privacy by avoiding the use of facial photographs.
• Early diagnosis of acromegaly is crucial, as untreated the condition can reduce life expectancy.
• Researchers aim to expand the AI’s capabilities to detect other conditions visible in hand images.

This research, funded by the Hyogo Foundation for Science Technology, involved collaboration with researchers from multiple universities and hospitals across Japan, including Fukuoka University, Hyogo Medical University, and Nagoya University.