NEJM AI

On the latest episode of the NEJM AI Grand Rounds podcast, a mother shares how after 17 doctors missed her son’s diagnosis, it took combining AI analysis with hands-on medical expertise to finally get answers. Listen to the full episode hosted by NEJM AI Deputy Editors Arjun Manrai, PhD, and Andrew Beam, PhD: https://nejm.ai/ep24 #ArtificialIntelligence #AIinMedicine

Use of #ArtificialIntelligence (AI) in cardiovascular imaging may potentially augment clinical decision-making in disease management, but no prospective randomized controlled trials have assessed the impact on cardiovascular outcomes.? ? The PROTEUS trial aimed to evaluate whether AI-augmented decision-making in stress echocardiography is non-inferior to standard clinical decision-making for selecting patients for invasive coronary angiography.?? ? In the multicenter, parallel-group randomized controlled trial, conducted at 20 centers in the United Kingdom with 2341 participants, the primary endpoint was the appropriateness of referral for angiography.?? ? The results showed no significant difference in accuracy between AI-augmented and standard decision-making, with similar sensitivity and specificity between the groups.?? ? Although the AI augmentation did not meet the non-inferiority margin, subgroup analysis suggested potential benefits in low-volume stress echocardiography centers.? ? Read the Original Article “PROTEUS: A Prospective RCT Evaluating Use of AI in Stress Echocardiography” by R. Upton et al.: https://nejm.ai/3BRD9Cl? ? ?????????????? ??????????????? Editorial by David Ouyang, MD, and Joseph Hogan, ScD: We Need More Randomized Clinical Trials of AI https://nejm.ai/4hfETpq? ? #ClinicalTrials?

What role can?#ArtificialIntelligence?play in transforming health care and empowering patients?? ?? Find out in the latest episode of the NEJM AI Grand Rounds podcast, where Courtney Hofmann, a mother whose use of ChatGPT led to her son’s diagnosis of tethered cord syndrome after seeing 17 doctors over three years, and Dr. Holly Gilmer, the pediatric neurosurgeon who confirmed and treated the condition, discuss how?AI helped bridge diagnostic gaps, systemic health care challenges that led to missed diagnoses, and the evolving role of AI in patient advocacy and medical practice. ? ? The episode highlights the importance of combining AI insights with human medical expertise, while discussing both the potential and limitations of AI in health care.? ?? Listen to the full episode hosted by NEJM AI Deputy Editors Arjun Manrai, PhD, and Andrew Beam, PhD: https://nejm.ai/ep24? ? #AIinMedicine?

Ambient clinical visit recording will soon power emerging artificial intelligence–enabled technologies to support patients, clinicians, and health systems.?? ? To realize the potential of such routine encounter recording, we must prevent and ameliorate unintended yet foreseeable harms.?? ? Paul Barr, PhD, Robert Gramling, MD, DSc, and Soroush Vosoughi, PhD, focus on three areas of concern and offer strategies to mitigate them: burden to clinician–patient relationships, fairness in access and performance quality, and commoditization.?? ? The authors are optimistic about ambient recording technology and its potential to have a positive impact on health care communication, relationships, and decision-making, all with the patient voice at its center, as long as it is applied carefully and equitably.? ? Read the Perspective “Preparing for the Widespread Adoption of Clinic Visit Recording” by Paul Barr, PhD, Robert Gramling, MD, DSc, and Soroush Vosoughi, PhD: https://nejm.ai/3NDcYSo? ? #ArtificialIntelligence #AIinMedicine?

David Ouyang, MD, emphasizes the enduring value of deep expertise in medicine and computational skills. He advocates for future cardiologists to maintain critical thinking skills, especially in evaluating #ArtificialIntelligence algorithms and handling imperfect data. Listen to the full episode hosted by NEJM AI Deputy Editors Arjun Manrai, PhD, and Andrew Beam, PhD: https://nejm.ai/ep23 #AIinMedicine

Although large language models (LLMs), such as OpenAI GPT-4 or Google PaLM 2, are proposed as viable diagnostic support tools or even spoken of as replacements for “curbside consults,” past studies show that they may lack sufficient diagnostic accuracy for real-life applications.?? ? In an effort to improve their accuracy and reduce the risk of misdiagnoses, Barabucci et al. applied methods from the field of collective intelligence to produce synthetic differential diagnoses that aggregate answers from individual commercial LLMs.?? ? Using 200 clinical vignettes of real-life cases from the Human Diagnosis Project platform, the authors assessed and compared the accuracy of differential diagnoses from individual LLMs with those from aggregated LLM responses.?? ? They aggregated the LLM responses into synthetic differential diagnoses using a simple frequency-based, 1/r-weighted method, in which more weight is given to diagnoses appearing near the top of the LLM responses and appearing in the responses of multiple LLMs.?? ? They evaluated all possible combinations of LLMs by calculating various TOP-n accuracy metrics: that is, how frequently the correct diagnosis matches any of the first n diagnoses.?? ? The authors found that aggregating responses from multiple LLMs leads to more accurate differential diagnoses compared with the differential diagnoses produced by single LLMs. They also found that aggregating smaller and less capable models can rival the accuracy of the top-performing model.?? ? The use of collective intelligence methods to synthesize differential diagnoses, combining the responses of different LLMs, achieves three of the necessary steps toward advancing LLMs as a diagnostic support tool: demonstrating sufficiently high diagnostic accuracy, reducing the risk of misdiagnoses, and eliminating the dependence on a single commercial vendor.? ? Read the Case Study “Combining Multiple Large Language Models Improves Diagnostic Accuracy” by G. Barabucci et al.: https://nejm.ai/4eX6g5V? ? #ArtificialIntelligence #AIinMedicine?

The European Union’s #ArtificialIntelligence Act (AIA), which took effect on August 1, 2024, establishes the world’s first comprehensive legal framework for AI, with significant implications for medical AI development and deployment worldwide.?? ?? A Perspective by Sebastian Porsdam Mann, PhD, I. Glenn Cohen, JD, and Prof. Timo Minssen, analyzes the AIA’s key implications for physicians and medical innovators in the United States.?? ?? The authors examine the Act’s risk-based approach, extraterritorial reach, and potential to influence global medical AI policy. They outline the key provisions of the AIA for medical AI systems and discuss its measures to support innovation.?? ?? As implementation begins, U.S. health care stakeholders must engage proactively with this new regulatory landscape to remain competitive and gain access to the EU market.? ?? Read the Perspective “The EU AI Act: Implications for U.S. Health Care” by Sebastian Porsdam Mann, PhD, I. Glenn Cohen, JD, and Prof. Timo Minssen: https://nejm.ai/3A5head? ?? #AIinMedicine?

#ArtificialIntelligence (AI) applications in medical imaging continue to evolve rapidly, with models now capable of interpreting medical images without being trained on explicit labels.?? ? A new Perspective, based on a conversation with Dr. Pranav Rajpurkar on NEJM AI Grand Rounds, discusses the progression of AI imaging models, starting from early successes in radiology, such as CheXNet, to more sophisticated recent models such as CheXzero.?? ? Dr. Rajpurkar emphasizes the importance of understanding the “data generation process,” including the artifacts and biases baked into data, which is illustrated by a specific example where an AI model exploited metadata rather than clinically relevant features.?? ? He addresses the urgent need for more open and accessible medical data with his initiative on Medical AI Data for All (MAIDA).?? ? Dr. Rajpurkar and AI Grand Rounds cohosts Andrew Beam, PhD, and Arjun Manrai, PhD (also deputy editors of the NEJM AI journal), also examine the changing role of clinicians in an AI-augmented health care system, and discuss a collaborative approach where human expertise guides AI development and implementation.?? ? Looking ahead, Drs. Rajpurkar, Beam, and Manrai envision a future where AI systems generate comprehensive medical reports and engage in natural language interactions, while emphasizing the need for ongoing focus on safety, efficacy, and equitable access.? ? Read the Perspective “Pixels and Pitfalls: Building Robust Artificial Intelligence for Medical Imaging” by Pranav Rajpurkar, PhD, Andrew Beam, PhD, and Arjun Manrai, PhD: https://nejm.ai/3TLbxVx? ? Listen to the full episode of the podcast: https://nejm.ai/3HFb2p5? ? #AIinMedicine?

Early detection of autism is important for timely access to diagnostic evaluation and early intervention services, which improve children’s outcomes. Despite the ability of clinicians to reliably diagnose autism in toddlers, diagnosis is often delayed.?? ? SenseToKnow is a mobile autism screening application (app) delivered on a smartphone or tablet that provides an objective and quantitative assessment of early behavioral signs of autism based on computer vision (CV) and machine learning (ML).?? ? A new study examined the accuracy of SenseToKnow for autism detection when the app was downloaded and administered remotely at home by caregivers using their own devices.?? ? The SenseToKnow app was administered by caregivers of 620 toddlers between 16 and 40 months of age, 188 of whom were subsequently diagnosed with autism by expert clinicians. The app displayed strategically designed movies and a bubble-popping game on an iPhone or iPad while recording the child’s behavioral responses through the device’s front-facing camera and touch/inertial sensors.?? ? Recordings of the child’s behavior were then automatically analyzed using CV. Multiple behavioral phenotypes were quantified and combined using ML in an algorithm for autism prediction.?? ? SenseToKnow demonstrated a high level of diagnostic accuracy with area under the receiver operating characteristic curve of 0.92, sensitivity of 83.0%, specificity of 93.3%, positive predictive value of 84.3%, and negative predictive value of 92.6%. Accuracy of the app for detecting autism was similar when administered on either a caregiver’s iPhone or iPad.?? ? These results demonstrate that a mobile autism screening app based on CV can be delivered remotely by caregivers at home on their own devices and can provide a high level of accuracy for autism detection.?? ? Remote screening for autism potentially lowers barriers to autism screening, which could reduce disparities in early access to services and support and improve children’s outcomes.? ? Read the full study results by P.R. Krishnappa Babu et al.: https://nejm.ai/3TIHFcw? ? ?????????????? ??????????????? Editorial by NEJM AI Editor-in-Chief Isaac Kohane, MD, PhD: Advancing Autism Detection: A Digital Step Forward https://nejm.ai/4eBCFyr? ? #ArtificialIntelligence #AIinMedicine?

While artificial general intelligence (AGI) and artificial superintelligence (ASI) remain speculative, the possibility of capabilities that meet or surpass expert levels in areas such as treatment planning, clinical reasoning, and cognitive empathy merits consideration and debate in anticipation of the day when these capabilities become viable, write the authors of a new Perspective.?? ? Current ethical debates about AI often center on immediate concerns like data bias, but the progression toward AGI and ASI presents a profound and novel challenge: these systems might develop ethical frameworks that fundamentally differ from human-derived ethics.?? ? Planning for them must anticipate these changes, ensuring that their ethical paradigms uphold human values. Given the transformative potential of AGI and ASI, a multidisciplinary dialogue among medical professionals, policy makers, and technology experts is essential to prepare for these advancements.? ? Continue reading the Perspective “If Machines Exceed Us: Health Care at an Inflection Point” by E. Klang et al.: https://nejm.ai/3XXlOjW? ? #ArtificialIntelligence #AIinMedicine?