How AI detects the hidden diseases doctors aren’t looking for

AI-powered opportunistic screening technology is enabling healthcare providers to detect undiagnosed diseases by analyzing routine CT scans and other medical imaging for conditions beyond their original purpose.

The breakthrough case: Will Studholme’s emergency room visit for food poisoning led to an unexpected osteoporosis diagnosis after AI technology analyzed his routine abdominal CT scan and identified a collapsed vertebra.

• The 58-year-old male patient would likely have gone undiagnosed without the AI screening, as he didn’t fit the typical osteoporosis patient profile
• The early detection enabled doctors to begin preventative treatment with annual drug infusions to improve bone density

How the technology works: Artificial intelligence algorithms are trained on thousands of tagged medical scans to identify early indicators of various diseases that human radiologists might miss during routine examinations.

• The AI systematically analyzes imaging data that often goes unused, looking for signs of conditions like osteoporosis, heart disease, fatty liver disease, and diabetes
• The technology is designed to complement human expertise, with radiologists reviewing AI findings before they’re reported to doctors
• Training data must include diverse ethnic groups to ensure the technology works effectively across populations

Implementation and results: Oxford NHS hospitals have been at the forefront of implementing this technology, particularly for osteoporosis screening.

• Since officially adopting Nanox.AI’s osteoporosis screening tool in 2020, Oxford hospitals have seen up to a six-fold increase in identifying patients with vertebral fractures
• The success has led to expanded trials at hospitals in Cambridge, Cardiff, Nottingham and Southampton
• The technology is being developed by only a handful of companies, with Nanox.AI offering products for screening osteoporosis, heart disease, and fatty liver disease

Healthcare system impacts: While the technology shows promise for early disease detection, it presents challenges for healthcare system resources.

• The increased identification of potential conditions requires additional confirmatory testing and follow-up care
• Healthcare providers must adapt their systems to handle the increased patient load, such as Oxford’s nurse-delivered fracture prevention service
• Despite initial resource strain, early detection and prevention could lead to long-term cost savings, particularly for conditions like osteoporosis where complications often result in hospitalization

Future implications: The emerging field of AI-powered opportunistic screening represents a shift toward more proactive healthcare, though careful consideration of healthcare system capacity and resource allocation will be crucial for successful implementation.

• Healthcare providers will need to balance the benefits of early detection against the strain on resources and potential for unnecessary testing
• The technology could help reduce healthcare disparities by screening all patients regardless of demographic profiles
• Success will depend on developing efficient pathways for handling increased diagnosis rates and ensuring appropriate follow-up care