This week’s Artificial Intelligence and Digital Health update highlights clinical AI advances, real-world workflow integration, preventive-health expansion, device innovation, and growing investment across healthcare and pharma.
In Today’s Newsletter
Dive deeper
🏥 Medtronic reshapes around higher-growth medtech areas [1] [01 Apr 2026]
Context: The source describes Medtronic’s restructuring, diabetes IPO path, and focus on cardiovascular, neuroscience, and surgical platforms.
Key point: Medtronic is presented as shifting from a diversified conglomerate model toward a more focused medtech growth story, with Renal Denervation and Pulsed Field Ablation highlighted as commercial drivers [1].
Implication: Signals portfolio focus and operating-model change, but the source is a company-focused market overview rather than a peer-reviewed clinical report.
🇹🇭 Thailand pushes predictive analytics in healthcare [2] [03 Apr 2026]
https://www.chiangraitimes.com/health/thailands-predictive-analytics-shift-in-healthcare/
Context: The article describes wearables, Mor Prom+, Health Link, and public-private initiatives tied to preventive care and aging-population needs in Thailand.
Key point: Thailand is promoting predictive analytics, AI-enabled screening, and unified health-data platforms to shift care toward earlier intervention for chronic disease [2].
Implication: May expand screening, initiation, and follow-up at scale.
🛏️ DeRUCCI launches AI mattress with sleep-tracking pitch [3] [Malaysia • 03 Apr 2026]
Context: DeRUCCI Malaysia launched the T11 Pro AI Mattress with dual-zone support, sleep monitoring, and a retail expansion plan.
Key point: DeRUCCI is positioning its mattress as a connected wellness product with real-time body mapping, sleep monitoring, and localized product development through a Getha partnership [3].
Implication: Signals consumer-health and sleep-tech expansion, though clinical utility is not established in the source.
🧓 LIVE4WELL debuts AI preventive-health zone in Hong Kong [4] [Hong Kong • 02 Apr 2026]
Context: LIVE4WELL said it will exhibit AI-enabled health assessments, lifestyle reporting, and app-based follow-up at the Smart Retirement Expo.
Key point: The platform combines multiple tests, AI interpretation, and behavior-change guidance as a preventive-medicine offering for older adults and health consumers [4].
Implication: May expand screening, initiation, and follow-up at scale.
⚖️ OMA launches OMAr for obesity-care conversations [5] [US • 02 Apr 2026]
Context: The Obesity Medicine Association launched Treating Obesity First with an AI-powered simulator, conversation guide, and CME resources.
Key point: OMA said OMAr is designed to help clinicians practice realistic, patient-centered obesity discussions using the 5As framework [5].
Implication: Could inform clinician communication and training, pending uptake and real-world evaluation.
💬 Insight Health raises Series A for clinical AI agents [6] [US • 03 Apr 2026]
https://www.mobihealthnews.com/news/insight-health-raises-11m-scale-clinical-ai-agents
Context: Insight Health raised $11 million in Series A funding led by Standard Capital.
Key point: The company said it will use the funding to expand Lumi, its patient-facing virtual care assistant, and broader AI agents for clinical workflows such as pre-visit intake and follow-up [6].
Implication: Signals pipeline investment and workflow automation expansion.
🗣️ Voice biomarkers may help flag vocal fold lesions [7] [05 Apr 2026]
https://scitechdaily.com/ai-may-soon-detect-cancer-just-by-listening-to-you-speak/
Context: The source summarizes a Frontiers in Digital Health study using the Bridge2AI-Voice dataset.
Key point: Researchers reported that voice features distinguished some men with vocal fold lesions from men without lesions, suggesting a possible future screening role for vocal biomarkers [7].
Implication: Could inform practice and payer discussions; interpretation depends on study design and confounding control.
❤️ ACC 2026 highlights mix device trials and AI workflow data [8] [US • 02 Apr 2026]
https://www.massdevice.com/the-biggest-cardiology-tech-stories-out-of-acc-2026/
Context: ACC 2026 coverage included updates from Boston Scientific, Medtronic, Abbott, Edwards, J&J MedTech, and others.
Key point: The conference coverage emphasized new findings in LAAC, thrombolysis, TAVR, RDN, and EHR-driven clinician notifications, with several companies reporting new trial or outcome updates [8].
Implication: May influence prescriber choice and payer reviews pending full data.
🫀 Tempus AI and Medtronic report gains from cardiac EHR alerts [9] [US • 03 Apr 2026]
https://www.chicagobusiness.com/health-care/ccb-tempus-ai-notification-cardiac-04022026/
Context: The article highlights new trial data around automated notifications for aortic stenosis and mitral regurgitation.
Key point: Tempus AI and Medtronic said integrating automated EHR alerts improved the timeliness of evaluation and treatment for some cardiac valve patients [9].
Implication: May influence care-pathway design and referral workflows pending broader implementation data.
🧠 Epia Neuro launches BCI platform for stroke recovery [10] [US • 05 Apr 2026]
https://www.digitalhealthnews.com/epia-neuro-launches-with-a-bci-device-for-stroke-recovery
Context: Epia Neuro said its platform combines a minimally invasive BCI, assistive devices, and AI-driven support.
Key point: The company is initially targeting stroke rehabilitation and plans first-in-human demonstrations later in 2026 at Lenox Hill Hospital [10].
Implication: Signals pipeline investment and modality expansion.
🩻 AI chest X-rays could widen TB case finding in LMICs [11] [05 Apr 2026]
https://www.emjreviews.com/radiology/news/ai-assisted-chest-x-rays-in-tb-detection-across-lmics/
Context: EMJ Reviews summarized implementation evidence on AI-assisted chest X-ray use in low- and middle-income countries.
Key point: The report says AI-supported chest X-rays may improve sensitivity and speed in TB screening workflows, while portability may help reach remote communities [11].
Implication: Could inform practice and payer discussions; interpretation depends on study design and confounding control.
👁️ Portable AI eye scanner aims to broaden vision screening [12] [Japan • 02 Apr 2026]
Context: Tohoku University reported a portable slit-light device for anterior-segment assessment, with findings in Scientific Reports.
Key point: The team said the device showed strong agreement with AS-OCT for screening-oriented assessment and can run a lightweight AI model on-device [12].
Implication: May expand screening, initiation, and follow-up at scale.
🤖 Nurse-led AI platform improves RA follow-up metrics [13] [China • 02 Apr 2026]
Context: The report summarizes a Tongji Hospital single-center, multicampus real-world study over 6 months.
Key point: The AI-assisted platform group reportedly showed larger DAS28 improvement, better HAQ-II scores, higher adherence, and higher satisfaction than routine follow-up [13].
Implication: Could inform practice and payer discussions; interpretation depends on study design and confounding control.
🧬 NIH renews support for USC-led AI4AD2 Alzheimer’s program [14] [US • 01 Apr 2026]
https://www.eurekalert.org/news-releases/1122502
Context: NIH renewed funding for the AI4AD program’s next phase, AI4AD2, led by USC’s Stevens INI.
Key point: The renewal brings total cited support for AI4AD to $30.7 million, with work spanning imaging, genomics, disease subtyping, progression prediction, and genome-guided drug discovery [14].
Implication: Signals pipeline investment and modality expansion.
🫁 Smartphone photos support intraoperative LUAD assessment [15] [01 Apr 2026]
https://www.thelancet.com/journals/landig/article/PIIS2589-7500(25)00147-5/fulltext
Context: The Lancet Digital Health published a prospective multicenter diagnostic study of SuRImage in clinical stage IA lung adenocarcinoma.
Key point: SuRImage used smartphone surgical resection images for intraoperative identification, diagnosis, and grading, and the paper reports stronger performance than frozen section on several tasks [15].
Implication: May influence surgical workflow and decision support pending external implementation data.
🏥 Hopkins-LLM pushes EHR prediction into multi-task workflow support [16] [US • 02 Apr 2026]
https://www.nature.com/articles/s41746-026-02572-y
Context: The npj Digital Medicine paper describes a LLaMA-based framework trained on 42,160 patients within Johns Hopkins Health System.
Key point: The Hopkins-LLM framework reported mean ROC-AUC of 0.84 across tasks including 30-day readmission, 90-day mortality, ICU admission, and treatment recommendations [16].
Implication: Could inform practice and payer discussions; interpretation depends on study design and confounding control.
⌚ Sensor-based DHT endpoints are growing, unevenly [17] [03 Apr 2026]
https://www.nature.com/articles/s41746-026-02512-w
Context: A scoping review in npj Digital Medicine assessed 48 studies using sensor-based digital health technologies as outcome-measurement tools.
Key point: Most studies collected physiological data, especially continuous glucose monitoring, while fewer focused on sensor-based functional outcomes like activity or sleep [17].
Implication: May influence endpoint strategy and regulatory planning for digital measures.
🧾 Neural-symbolic agent improves biomedical concept mapping [18] [04 Apr 2026]
https://www.nature.com/articles/s41746-026-02594-6
Context: Columbia and collaborators described Medical Concept Mapping, which uses LLM reformulation before concept linking.
Key point: The workflow outperformed listed baselines on MedMentions, ST21pv, and MCN, with particular gains for underrepresented and abbreviated concepts [18].
Implication: May improve EHR standardization, retrieval, and downstream clinical AI workflows.
🖼️ Custom radiology-impression model nears human performance [19] [04 Apr 2026]
https://www.nature.com/articles/s41746-026-02586-6
Context: A blinded evaluation compared original radiologist impressions, a custom domain-specific model, and a general-purpose model on 200 oncologic CT reports.
Key point: The custom model reached near parity with human impressions, while generic-model impressions were longer and less concise; patient-harm ratings were low across groups [19].
Implication: Could streamline reporting workflows, though stakeholder preference and reliability remain important.
🫀 MedGuide-14B flags likely underdiagnosed HFpEF in MAFLD [20] [China • 31 Mar 2026]
https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001317
Context: A multicenter retrospective PLOS Digital Health study applied a domain-tuned LLM to structured EHR data and free-text notes in 24,011 patients with MAFLD.
Key point: The model identified 4,226 additional likely HFpEF cases beyond routine diagnosis, and the study reports 90.4% confirmation in a blinded validation subset [20].
Implication: Could inform case finding and referral strategies, though prospective validation across settings remains needed.
🧑🔬 LLMs match humans on some thematic-coding tasks [21] [03 Apr 2026]
https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001189
Context: The study compared ChatGPT-5, Claude 4 Sonnet, QualiGPT, and blinded human analysts on a healthcare focus-group transcript.
Key point: LLMs were reported non-inferior to human analysts for deductive coding, but inductive performance varied and quote verification remained necessary [21].
Implication: Could support research operations, but human review remains necessary for nuance and hallucination control.
🇯🇵 Japan digital health market piece highlights telehealth and wearables [22] [Japan • 02 Apr 2026]
https://vocal.media/01/how-digital-health-is-transforming-healthcare-delivery-in-japan
Context: The article is a market-style overview of Japan’s digital health ecosystem, covering telemedicine, wearables, EHRs, and analytics.
Key point: It frames aging demographics, chronic disease burden, and policy modernization as key demand drivers for telehealth, remote monitoring, and healthcare analytics in Japan [22].
Implication: Signals market expansion, but this is not a primary clinical evidence source.
🚨 MEDVi faces scrutiny over AI-generated personas and GLP-1 marketing [23] [US • 04 Apr 2026]
Context: The article reviews FDA warning history, public records, ad-library findings, and litigation connected to telehealth company MEDVi.
Key point: The report says MEDVi used AI-generated or fabricated promotional personas and that regulators had already warned the company over misbranding claims around compounded GLP-1 products [23].
Implication: Highlights governance, compliance, and trust risks in AI-enabled telehealth marketing.
🏭 Lilly and Novo use AI across factories and trials [24] [31 Mar 2026]
Context: PYMNTS described Lilly’s use of digital twins and computer vision in GLP-1 manufacturing and Novo Nordisk’s use of AI agents in active trials.
Key point: The article frames AI as becoming part of the operating process itself in pharma, spanning plant optimization, clinical trial execution, and broader development workflows [24].
Implication: Signals pipeline investment and modality expansion.
🧪 Lilly signs $2.75B Insilico AI drug-discovery deal [25] [US • 30 Mar 2026]
Context: Eli Lilly entered a collaboration with Insilico Medicine for AI-based therapeutic discovery and development.
Key point: The agreement gives Lilly exclusive rights tied to the collaboration, while Insilico said its AI platform has generated at least 28 drug candidates, with nearly half at clinical stage [25].
Implication: Signals pipeline investment and modality expansion.
🧬 Mantis Biotech pitches digital twins for biomedical edge cases [26] [30 Mar 2026]
Context: The startup says it combines heterogeneous data sources, LLM routing, and physics engines to create predictive digital twins.
Key point: Mantis Biotech is positioning its platform as a way to generate synthetic data and human digital twins for rare conditions, procedure testing, and predictive modeling [26].
Implication: Signals pipeline investment and modality expansion.
🔍 XAI review maps where explainability is used in healthcare [27] [01 Apr 2026]
https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2026.1749527/full
Context: A systematic review in Frontiers in Artificial Intelligence examined 36 studies across imaging, diagnosis, and rehabilitation.
Key point: The review found a method-modality pattern, with imaging favoring saliency methods such as Grad-CAM and diagnosis or rehabilitation relying more on SHAP and LIME [27].
Implication: Could inform clinical AI governance and model-reporting practice.
🧠 Nature paper frames LLMs as predictive engines in healthcare [28] [02 Apr 2026]
https://www.nature.com/articles/s41746-026-02572-y
Context: The same Hopkins-LLM paper also positions clinical LLMs as a lower-friction way to deploy predictive models from structured EHR data.
Key point: The authors argue that a unified LLM framework may reduce implementation barriers for multiple constrained decision-support tasks in health systems [28].
Implication: Could streamline deployment of predictive analytics if external validation and workflow fit hold up.
📡 Nature review tracks sensor-based endpoints across recent trials [29] [03 Apr 2026]
https://www.nature.com/articles/s41746-026-02512-w
Context: The review spans studies published from 2021 through 2023 across therapeutic areas.
Key point: It highlights both the practical value of sensor-derived endpoints and recurring operational challenges in trial implementation, evidencing, and interpretation [29].
Implication: May shape sponsor decisions on trial instrumentation and endpoint design.
🧠 Neural-symbolic concept mapping strengthens long-tail normalization [30] [04 Apr 2026]
https://www.nature.com/articles/s41746-026-02594-6
Context: The Medical Concept Mapping paper combines LLM reformulation with symbolic linking.
Key point: Human evaluation in the paper found most LLM-generated expansions were judged reasonable and useful, supporting explainability claims alongside benchmark gains [30].
Implication: Could improve data harmonization and ontology-linked biomedical applications.
Why it matters
- AI is moving into practical healthcare use across diagnostics, screening, triage, care pathways, report generation, and device-enabled monitoring.
- The most immediate value is in workflow support and earlier detection, especially where health systems need faster evaluation, better routing, and more scalable follow-up.
- Healthcare AI is also advancing in the underlying infrastructure, including prediction from EHR data, terminology mapping, explainability, digital endpoints, and documentation workflows.
- Pharma adoption is expanding beyond discovery into manufacturing and clinical trial execution, showing that AI is becoming part of operational strategy as well as R&D.
- Consumer and preventive-health tools are widening the reach of digital health, from remote monitoring and sleep products to engagement platforms and proactive screening models.
- Evidence quality varies, with peer-reviewed studies alongside conference findings, company announcements, and market or media reports, so maturity and validation are uneven.
- Several use cases remain early, especially in brain-computer interfaces, voice biomarkers, and AI-enabled wellness products, where future potential is clear but real-world proof is still developing.
- Governance and compliance matter as much as model capability, particularly in telehealth, medical advertising, and any setting where AI-generated claims can influence care decisions or patient trust.
- The central question is shifting from whether these systems can perform in controlled settings to whether they can be deployed reliably, safely, and efficiently in real-world healthcare environments.
🎬 Watch on YouTube the Top AI in Healthcare and Digital health news from the past two weeks, curated by the LucidQuest team.
📚 Find your one-stop page for the full AI and Digital Health archive.
🚀 Accelerate your success. Contact us now
📂 Explore our case studies. See examples of our work.
💡 Read our insights. Learn from our latest reports and analysis
🎬 Watch on YouTube. Subscribe and never miss a video.
🧰 See our full range of services. Discover how we can help you.
FAQ
How is Medtronic’s 2026 strategy changing after the diabetes separation, renal denervation rollout, and pulsed field ablation push?
Medtronic is being framed as moving toward a more focused medtech model, with attention on newer growth areas such as renal denervation and pulsed field ablation, alongside the separation of its diabetes business [1].
What is Thailand doing with predictive analytics, Mor Prom, Health Link, and wearable-driven preventive healthcare?
Thailand is promoting predictive analytics, wearable monitoring, and shared health-data infrastructure to support earlier detection and preventive care, especially for aging and chronic-disease populations [2].
What is DeRUCCI’s T11 Pro AI Mattress, and how is it being positioned in health-tech sleep?
DeRUCCI is positioning the mattress as a connected wellness product with real-time body mapping, sleep monitoring, and localized product development through its Getha partnership, but the source does not establish clinical utility [3].
What is LIVE4WELL offering at the Hong Kong Smart Retirement Expo, and how does its AI preventive-health model work?
LIVE4WELL is presenting a preventive-health platform built around multi-source testing, AI interpretation, lifestyle recommendations, and app-based follow-up, aimed at older adults and health consumers [4].
What is OMAr, and how is the Obesity Medicine Association using it for obesity-care communication training?
OMAr is an AI-powered obesity conversation simulator from the Obesity Medicine Association, designed to help clinicians practice patient-centered discussions about obesity care [5].
What is Insight Health building with Lumi and clinical AI agents, and what will the Series A funding support?
Insight Health said it will use the new capital to expand Lumi, its patient-facing virtual care assistant, and broader AI agents for workflows such as pre-visit intake and follow-up [6].
Can voice biomarkers help detect benign or malignant vocal fold lesions before standard workup?
The cited study suggests that some acoustic features may distinguish men with vocal fold lesions from those without lesions, indicating a possible future screening role, but the work remains exploratory [7].
What were the Tempus Next ALERT trial results for EHR alerts in aortic stenosis and mitral regurgitation?
The reported finding is that automated EHR-based clinician alerts improved the timeliness of evaluation and treatment for some patients with aortic stenosis and mitral regurgitation [8][9].
What is Epia Neuro’s BCI platform for stroke recovery, and when are first-in-human demonstrations expected?
Epia Neuro is developing a minimally invasive brain-computer interface platform that combines neural signal interpretation with assistive devices and AI-driven support for stroke rehabilitation [10].
Can AI-assisted chest X-rays improve TB case finding in LMICs with limited imaging and radiology capacity?
The cited review says AI-assisted chest X-rays may improve sensitivity and reading speed in TB screening programs, especially where radiology access and workforce capacity are limited [11].
How did the Tohoku portable AI slit-light device perform versus AS-OCT for anterior-segment and angle-closure screening?
It is a low-cost, portable anterior-segment screening device that uses slit-light imaging and on-device AI, with the source saying it showed strong agreement with AS-OCT for screening-oriented use [12].
What did the Tongji Hospital AI-assisted rheumatoid arthritis management study report on DAS28, HAQ-II, adherence, and satisfaction?
The study summary says a nurse-led AI-assisted management platform was associated with better disease activity, adherence, and patient satisfaction than routine post-discharge care in one real-world setting [13].
What is AI4AD2, and what does the NIH renewal mean for dementia subtyping, progression modeling, and genome-guided drug discovery?
AI4AD2 is the next phase of a USC-led Alzheimer’s consortium using AI with imaging, genomics, cognition, and related data to improve disease subtyping, progression prediction, and treatment discovery [14].
How did SuRImage perform for stage IA lung adenocarcinoma intraoperative identification, diagnosis, and grading versus frozen section?
The SuRImage study used smartphone photos of surgical resection images to support intraoperative identification, diagnosis, and grading of early-stage lung adenocarcinoma, with the paper reporting stronger performance than frozen section on several tasks [15].
What is Hopkins-LLM, and how did it perform on readmission, mortality, ICU prediction, treatment recommendation, and external validation?
Hopkins-LLM is an LLM-based framework trained on structured EHR data and evaluated on tasks such as readmission prediction, mortality prediction, ICU admission, and treatment recommendation support [16].
How are sensor-based digital health technologies being used for clinical trial endpoints in CGM, sleep, activity, and physiological monitoring?
The scoping review says they are being used mainly to capture physiological measures, especially through continuous glucose monitoring, with less frequent use for functional outcomes like sleep and physical activity [17].
What is Medical Concept Mapping, and how did it perform on MedMentions, ST21pv, and MCN benchmark tasks?
Medical Concept Mapping is a neural-symbolic workflow that rewrites ambiguous biomedical text into clearer descriptions before linking it to standardized concepts and ontologies [18].
What did the AI radiology impression study show for custom domain models versus general-purpose LLMs and human radiologists?
It showed that a custom domain-specific model came close to human-written radiology impressions in a blinded evaluation, while general-purpose model outputs were less concise and not consistently preferred [19].
How did MedGuide-14B perform for HFpEF detection in MAFLD, and what did blinded validation show?
The cited MAFLD study suggests a domain-tuned LLM identified additional likely HFpEF cases beyond routine diagnosis, with blinded validation reported in a sampled subset [20].
Can LLMs match human analysts in healthcare thematic analysis for deductive coding and inductive focus-group interpretation?
The healthcare focus-group study found LLMs were non-inferior to human analysts for deductive coding, but performance was more variable in inductive thematic work and still required verification [21].
How is digital health adoption in Japan being linked to telemedicine, wearables, EHRs, and aging-population demand?
The market-style overview frames aging demographics, chronic disease burden, and policy modernization as key demand drivers for telehealth, remote monitoring, and healthcare analytics in Japan [22].
Why did MEDVi receive FDA scrutiny over compounded GLP-1 marketing and AI-generated doctor ads?
The article argues that MEDVi illustrates governance and compliance risks around AI-generated marketing personas, compounded-drug promotion, and potentially misleading medical advertising claims [23].
How are Eli Lilly and Novo Nordisk using AI in drug discovery, manufacturing, and live clinical trial operations?
The reporting says Lilly is using AI in manufacturing and discovery, while Novo Nordisk is applying AI agents inside live clinical trials to identify delays, gaps, and workflow inefficiencies [24].
What is Eli Lilly’s Insilico Medicine deal, and what does it mean for AI-generated therapeutic candidates?
Insilico Medicine entered a major discovery collaboration with Lilly in which AI is used to generate and develop novel therapeutic candidates, with Lilly holding exclusive rights under the deal structure described in the source [25].
What is Mantis Biotech building with digital twins, synthetic data, and edge-case biomedical modeling?
Mantis Biotech says it is building human digital twins from mixed data sources and synthetic data generation to support prediction, simulation, and edge-case biomedical modeling [26].
What did the Frontiers systematic review find about explainable AI use in imaging, diagnosis, and rehabilitation?
The review found a method-modality pattern, with imaging favoring saliency methods such as Grad-CAM and diagnosis or rehabilitation relying more on SHAP and LIME [27].
Entities / Keywords
Medtronic, MiniMed, Geoff Martha, Symplicity Spyral, renal denervation, PulseSelect, Affera, pulsed field ablation, Hugo RAS, Tempus AI, Tempus Next, ALERT trial, ACC 2026, aortic stenosis, mitral regurgitation, TAVR, Boston Scientific, Watchman FLX, Ekos, Abbott TriClip, Edwards Evoque, Impella CP.
Thailand predictive analytics, Mor Prom, Health Link, Ministry of Public Health Thailand, True Corporation, Digital Health Act, preventive medicine, wearables, aging population, chronic disease management.
DeRUCCI, T11 Pro AI Mattress, Getha, sleep tech, Sleepcare monitoring, wellness technology.
LIVE4WELL, Smart Retirement Expo, AI health assessment, preventive health, fundus analysis, posture assessment, health capital management.
Obesity Medicine Association, OMA, Treating Obesity First, OMAr, obesity care conversations, 5As framework, clinician training.
Insight Health, Lumi, clinical AI agents, virtual care assistant, pre-visit intake, automated clinical documentation, athenahealth Marketplace, Office Practicum.
Bridge2AI-Voice, vocal biomarkers, laryngeal cancer, vocal fold lesions, voice analysis, acoustic signatures.
Epia Neuro, brain-computer interface, BCI, stroke recovery, Lenox Hill Hospital, neurorehabilitation.
AI-assisted chest X-ray, tuberculosis screening, TB detection, LMICs, ultra-portable X-ray, active case finding.
Tohoku University, portable AI eye scanner, slit-light imaging, anterior segment screening, AS-OCT, angle-closure glaucoma, on-device AI.
Tongji Hospital, rheumatoid arthritis, AI-assisted management platform, DAS28, HAQ-II, medication adherence, remote monitoring.
USC Stevens INI, AI4AD2, NIH, Alzheimer’s disease, dementia subtyping, neuroimaging, genomics, genome-guided drug discovery, PreSiBO.
SuRImage, lung adenocarcinoma, LUAD, intraoperative diagnosis, frozen section comparison, smartphone surgical images.
Hopkins-LLM, Johns Hopkins Health System, EHR prediction, readmission, mortality, ICU admission, treatment recommendation, clinical LLMs.
Sensor-based digital health technologies, DHTs, digital endpoints, continuous glucose monitoring, CGM, sleep monitoring, physical activity endpoints, clinical trials.
Medical Concept Mapping, MCM, biomedical concept normalization, MedMentions, ST21pv, MCN, ontology mapping, entity linking, neural-symbolic AI.
AI radiology impressions, oncologic CT reports, report summarization, radiology workflow, clinical text generation.
MedGuide-14B, HFpEF, MAFLD, EHR case finding, blinded validation.
LLMs in qualitative research, thematic analysis, ChatGPT-5, Claude 4 Sonnet, QualiGPT, deductive coding, inductive coding.
Japan digital health, telemedicine, medical wearables, EHR, healthcare analytics, aging population.
MEDVi, FDA warning letter, compounded GLP-1s, semaglutide, tirzepatide, AI-generated doctor ads, telehealth marketing compliance, CareValidate, OpenLoop Health.
Eli Lilly, Novo Nordisk, Insilico Medicine, AI drug discovery, Pharma.ai, digital twins in pharma, manufacturing AI, AI agents in clinical trials, GLP-1 manufacturing.
Mantis Biotech, digital twins, synthetic data, rare disease modeling, predictive human modeling.
Explainable AI, XAI, SHAP, LIME, Grad-CAM, medical imaging explainability, trustworthy AI in healthcare.
