AI's Impact on Disease Detection

AI Skeptic: "Randomized Clinical Trials for AI are too difficult to implement." Sweden: "Here’s a large-scale RCT with 105,934 participants, testing AI in real-world clinical practice within a national screening program" The MASAI trial, a randomized, controlled, non-inferiority study, tested AI-supported mammography screening against standard double reading in Sweden’s national screening program. Published in The Lancet Digital Health, it provides real-world evidence on AI’s impact in clinical practice. Key results: ✔️ 29% increase in cancer detection (6.4 vs. 5.0 per 1,000 screened participants, p=0.0021) ✔️ 44% reduction in screen-reading workload (61,248 vs. 109,692 total readings) ✔️ No significant rise in false positives (1.5% vs. 1.4%, p=0.92) Importantly, AI did not just detect more cancers—it detected more clinically relevant ones: 🔹 More small, lymph-node negative invasive cancers (270 vs. 217) 🔹 Increased detection of aggressive subtypes, including triple-negative and HER2-positive cancers 🔹 No increase in low-grade ductal carcinoma in situ, reducing concerns about overdiagnosis This trial is a landmark in demonstrating that AI in medicine can and should be tested under the same rigorous standards as new drugs and medical devices. When the stakes are high, clinical evidence—not hype—should drive adoption! Source: https://lnkd.in/d8s5NM9W

For the first time, a machine can look at a routine mammogram and predict a woman's 5-year breast cancer risk quite accurately: no history, no demographics, just the image itself revealing invisible patterns. As someone working inside the pharmaceutical ecosystem, this is interesting and exciting. This is the kind of shift that makes you rethink everything you assumed about early detection. Until now, doctors estimated risk using age, family history, and breast density. Useful signals, but broad, indirect, and often late. We've been predicting population risk far better than individual risk. That is now changing. An image-only AI model can predict 5-year breast cancer risk more accurately than breast density alone. The image reveals patterns the human eye cannot see. This is AI potentially adding healthy years to people's lives. Years without chemotherapy, without surgery, without the fear of late discovery. We're starting to see risk while the body still looks normal. From an industry lens, this changes three fundamentals: •⁠ ⁠Screening shifts from age-based to risk-based. Screening frequency and prevention strategies can now be customized based on actual biological risk, not age brackets. •⁠ ⁠Prevention becomes earlier and more precise, but raises hard questions about false positives, patient anxiety, and long-term follow-up. •⁠ ⁠We face a scale challenge. AI can identify risk at the population level. Healthcare systems must be ready to act without overwhelming clinicians or excluding low-resource settings. The technology is moving faster than our operating models. The real leadership test is no longer whether AI can predict risk. It's whether we can deploy it responsibly, equitably, and at scale without creating gaps in care. The future of oncology will not be defined only by better drugs. It will be defined by how early we dare to see risk and how wisely we choose to act on it. #AI #healthtech

Doctors fear AI will replace them. Instead, it's revealing cancers they couldn't see for 4 more years. The same AI shows them exactly where to look. Think about that. Dr. Cara Antoine—Executive VP at Capgemini—said it perfectly on my Edge of Tomorrow podcast: "AI isn't what we should fear. Staying in the dark is." She's right. While doctors worried about their jobs, AI started catching what they couldn't. Breast cancer with 99% accuracy where radiologists saw nothing. Heart attacks 4 hours before anyone felt chest pain. Emergency rooms cutting wait times by 30%. Traditional Medical Reality: ↳ Radiologists missing 20% of breast cancers ↳ Emergency departments drowning in triage ↳ Doctors spending 70% of time on paperwork ↳ Rural clinics lacking specialist access The AI Revolution: ↳ Cancer spotted years before visible ↳ 30% reduction in ED wait times ↳ Pattern recognition across millions of cases ↳ Clinicians back to actual patient care But here's what stopped me cold: Dr. Antoine talked about people with visual impairments using AI to examine their own eyes. Shop alone for the first time. Walk through spaces they couldn't navigate before. The same tech doctors thought would end their careers is giving independence to people who lost theirs. AI isn't stealing the stethoscope. It's the X-ray vision doctors always wished they had. What changes everything: ↳ Village doctors with specialist-level diagnostics ↳ Nurses spotting rare diseases ↳ Treatment starting years earlier ↳ Actual conversations replacing forms The Multiplication Effect: 1 AI diagnosis = catching disease while it's still treatable 10 hospitals equipped = entire regions healthier 100 systems deployed = specialist care everywhere At scale = no more "if only we'd caught it sooner" A doctor in rural Kenya sees what Johns Hopkins sees. A nurse in Bangladesh recognises patterns that take specialists decades to learn. Your local clinic finds answers that stumped university hospitals. We spent decades accepting that some cancers hide until they kill. Now AI shows us they were there all along. When doctors can see what was always invisible, they don't lose their purpose. They finally get to use it. Follow me, Dr. Martha Boeckenfeld for conversations about tech that makes humans better at being human. ♻️ Share if you believe AI should give doctors superpowers, not pink slips. Watch the Edge of Tomorrow with Dr. Cara Antoine to understand how AI becomes our superpower.

The headline that caught my eye this week was "AI Trial to Spot Heart Condition Before Symptoms." Here's my take: Artificial intelligence holds substantial promise to improve quality and reduce costs in healthcare. One example from Leeds involves an algorithm that scours medical records for early warning signs of atrial fibrillation (AF) before symptoms appear — potentially preventing thousands of strokes. The results suggest that by analyzing existing medical records for patterns that human physicians might miss, AI can flag high-risk patients for early intervention. The trial has already identified cases like a 74-year-old former Army captain who had no symptoms but can now manage his condition effectively. This is particularly significant given that AF contributes to around 20,000 strokes annually in the UK alone. As Professor Chris Gale notes, too often the first sign of undiagnosed AF is a stroke — an outcome this technology could help prevent. The broader implication here is about AI's role in healthcare: not replacing physicians but augmenting their ability to identify risks earlier and intervene before conditions become critical.  

AI Is Rewriting Cancer & Critical Care Faster Than Most Medical Students Realize What happens when AI can diagnose brain tumors, map cancer genes, and detect sepsis before clinicians even suspect it? Medicine is entering a new era where algorithms may become as essential as the stethoscope. Brain Tumor Breakthrough: The GMAP deep-learning model predicts key glioma genetic markers from routine pathology slides with >87% accuracy, potentially eliminating weeks of molecular testing in underserved regions. Personalized Cancer Care at Scale: Cedars-Sinai’s Path2Space can estimate spatial gene expression from standard biopsy images in minutes, dramatically lowering the cost and delay of precision oncology. Earlier Sepsis Detection Saves Lives: Bayesian Health’s FDA-cleared AI system identifies sepsis up to 48 hours earlier using EHR data alone, with studies showing reduced in-hospital mortality when clinicians respond to alerts. As I reflect on these advances, I see a future where physicians who understand AI will lead the transformation of patient care, diagnostics, and healthcare equity. The next generation of medical professionals must prepare not only to practice medicine, but to collaborate with intelligent systems that accelerate decision-making and expand access worldwide. How should medical schools evolve to prepare students for AI-integrated healthcare? Which breakthrough will have the biggest impact on patients first? Follow Harvey Castro, MD, MBA. #AI #ArtificialIntelligence #HealthcareInnovation #DigitalHealth #PrecisionMedicine #MedicalEducation #HealthTech #CancerResearch #Sepsis #FutureOfMedicine #MachineLearning #MedTech #PhysicianLeadership #DRGPT

🦓 Rare Disease Day is Tomorrow. Over the last year, 8 AI signals have surfaced that hint at how rare disease detection and care are beginning to shift 🔘 A study in npj Digital Medicine showed AI analysing unstructured EHR notes could detect rare diseases like AADC deficiency earlier by spotting subtle symptom patterns across time, potentially shortening diagnostic odysseys 🔘 UCB partnered with Citizen Health to apply AI to opt-in patient communities, using longitudinal real-world data to accelerate research in epilepsy and rare diseases 🔘 AstraZeneca partnered with Pangaea Data to deploy multimodal AI inside EHR systems to surface hidden rare and misdiagnosed patients and connect them to treatments and trials 🔘 Alexion Pharmaceuticals, Inc. also partnered with Pangaea Data to use AI within EHRs to identify undiagnosed hypophosphatasia patients, directly targeting rare-disease diagnostic delays 🔘 Novartis partnered with Atropos Health to use real-world data models to flag potential PNH cases earlier, aiming to reduce years-long diagnostic delays and associated complications 🔘 MENARINI Group teamed up with VisualDx to apply AI-powered image analysis to improve early detection of BPDCN, linking rare cancer identification more directly to targeted therapy 🔘 Novartis and Dawn Health launched Nelia, a digital companion app for rare kidney diseases, extending rare-disease strategy beyond diagnosis into ongoing digital support 🔘 Penn researchers used AI-driven drug repurposing to identify adalimumab as a life-saving treatment for idiopathic multicentric Castleman’s disease, demonstrating AI’s potential in ultra-rare rescue scenarios 💬Diagnosis remains the central challenge in rare disease, and understandably so. But we are now seeing movement beyond detection into digital patient support and AI-informed treatment optimisation. The next phase will likely shift from isolated case-finding tools toward longitudinal, integrated models that support patients across diagnosis, treatment, monitoring, and care navigation #digitalhealth #pharma #ai #raredisease

An AI paired with a single radiologist now detects breast cancer more accurately than two radiologists working together. A machine learning model predicts Alzheimer's disease with 92.87% accuracy from MRI scans — years before symptoms appear. The FDA has cleared the first AI tool to predict your five-year breast cancer risk from a standard mammogram alone. And a new class of "biological age" trackers can tell you not how old you are, but how fast you are aging — and what to do about it. In Blog #33 of the AI Frontier series, The Great Health, we go deep into AI's transformation of medicine: the workforce crisis it must help solve, the diagnostic breakthroughs already in clinical use, the new wave of AI-powered preventative health platforms — from Fountain Life and Human Longevity to continuous glucose monitors, epigenetic clocks, and AI-guided longevity protocols — and the serious risks of bias, misdiagnosis, and unequal access that could make AI in healthcare a force for widening the very gaps it promises to close. This is the most consequential edition of the AI Frontier yet. Because no technology matters more than the one that keeps you alive.

This paper discusses the application and potential of AI in cancer screening and surveillance, focusing on primary, secondary, and tertiary prevention strategies. 1️⃣ AI improves the cost-effectiveness of cancer prevention by enhancing the accuracy and efficiency of risk assessments and early diagnosis. 2️⃣ Predictive models powered by AI facilitate less invasive and more frequent tests, which improve the accuracy of individual risk profiles over time. 3️⃣ AI-based screening increases the probability of early cancer diagnosis, enabling proactive and personalized preventive treatments. 4️⃣ Liquid biopsy tests, which detect cancer biomarkers in blood samples, have advanced through AI integration, playing a significant role in primary and secondary cancer prevention. 5️⃣ Key challenges include long validation times for biomarkers, underrepresentation of subclinical populations in trials, and communication difficulties between doctors and patients regarding risk estimates. 6️⃣ AI aids in different screening programs—general population, targeted, and stratified screening—by improving the identification and management of high-risk individuals. 7️⃣ Ensuring the reliability of AI models through rigorous validation with external datasets is crucial for effective clinical application. AI models have been validated through multicentre studies across various cancer types, demonstrating their utility in improving early detection and monitoring.. 8️⃣ AI improves communication and data sharing among healthcare professionals, facilitating better-informed decision-making and treatment planning. 9️⃣ Continuous improvement and validation of AI models, particularly with real-time data, are essential to fully realize the benefits of AI in cancer prevention. ✍🏻 Gentile F , Malara N. Artificial intelligence for cancer screening and surveillance. European Society for Medical Oncology. 2024. DOI: 10.1016/j.esmorw.2024.100046

Bonus AI ∩ Bio: Early Disease Diagnosis Following on my holiday edition of AI ∩ Bio on multimodal AI in oncology, I wanted to dig into a few related topics. Today: When Earlier Isn’t Always Better: Cost, Value & Overdiagnosis in Early Disease Detection Early detection is often treated as an unqualified good, or, at the other extreme, as one of the four horsemen of the apocalypse, destined to make an already strained healthcare system even more dysfunctional. In practice, its value is conditional. ➡️ Earlier detection improves outcomes only when it changes a decision that changes an outcome ⬅️ Simply finding disease sooner does not, by itself, improve survival or quality of life. For earlier detection to create real value, all three of the following must exist: 1️⃣ A preclinical detectable phase There must be a window in which disease is present and measurable before symptoms appear, but not yet so advanced that outcomes are already fixed. 2️⃣ A causal intervention window Acting during that earlier window must actually change the disease’s course, not merely shift the date of diagnosis (lead-time bias). 3️⃣ Detecting disease earlier does not count if treatment at that point is no more effective than treatment later. Technology sometimes becomes capable of finding abnormalities before medicine can reliably: ▪️ distinguish aggressive disease from indolent disease, ▪️ distinguish disease that requires treatment from disease that can safely be observed, or ▪️ offer effective treatment paths that can meaningfully alter the disease trajectory. The result is a trigger for additional tests, procedures, and treatments that may not improve survival or quality of life. Because these conditions vary by disease, early detection is not uniformly beneficial... ➖ Some heterogeneous cancers: In diseases like breast cancer, early detection can be life-saving for biologically aggressive subtypes, while simultaneously exposing patients with slow-growing disease to unnecessary treatment. The value depends on how well biology can be distinguished and acted upon. ➖ Rapidly progressive diseases with limited treatment leverage: In diseases like pancreatic cancer, even when detected earlier, outcomes may not meaningfully change because biology and treatment options, not timing of detection, dominate. ➖ Genetically determinative conditions with limited actionability: In conditions like Huntington’s disease, detection can be certain decades before symptoms, but without effective disease-modifying treatments, early knowledge needs to be handled carefully. The Right Question to Ask: Does earlier detection create decision leverage, given the biology of the disease, the effectiveness of available interventions, and how care is actually delivered? When early detection changes decisions in ways that improve outcomes, it is valuable. When it does not, earlier is not better. It is simply earlier.

The age of AI has truly begun, and its impact will surpass the internet, mobile, and cloud revolutions combined.   In 2022, DeepMind released AlphaFold. Using AI, they predicted the structure of nearly 200 million proteins. To put that in perspective until recently, discovering the structure of one protein could take the equivalent of a PhD’s lifetime of work. Why does this matter? Because most diseases are driven by proteins. Understanding how proteins fold and how other molecules interact with them fundamentally transforms how we discover drugs and treat disease. I genuinely believe that with the help of AI, we will find solutions to challenges like cancer and Alzheimer’s that have haunted humanity for decades. We have seen this transformation firsthand. Tuberculosis remains one of the biggest health challenges in South Asia, accounting for nearly 25 percent of global TB cases. In 2016 to 17, we spun out Qure.ai with a simple but ambitious goal to use AI to make early diagnosis accessible at scale. Today, Qure.ai can analyze X rays for tuberculosis and lung cancer in seconds, at one dollar per scan. For context, even with the best tools available, radiologists can miss findings nearly 25 percent of the time. In the Philippines, mobile X ray vans were sent to remote villages. Earlier, scans had to be sent back to cities for diagnosis often taking six weeks. By then, patients might infect others or never return for results. Now, with AI embedded directly into the machines, diagnosis happens instantly. Patients are told immediately whether they need specialist care, medication, or simply a follow up later. To date, Qure.ai has impacted over 25 million lives diagnosing TB, lung cancer, and paving the way for many more conditions in the future. This is what AI makes possible when applied with purpose. Not efficiency alone but scale, equity, and human impact. We are only at the beginning.